Citation
Postdeployment family functioning : service member PTSD symptoms, couple and parenting alliance, and child behavioral difficulties

Material Information

Title:
Postdeployment family functioning : service member PTSD symptoms, couple and parenting alliance, and child behavioral difficulties
Creator:
Kenny, Jessica J.
Place of Publication:
Denver, CO
Publisher:
University of Colorado Denver
Publication Date:
Language:
English

Thesis/Dissertation Information

Degree:
Doctorate ( Doctor of philosophy)
Degree Grantor:
University of Colorado Denver
Degree Divisions:
Department of Psychology, CU Denver
Degree Disciplines:
Clinical health psychology
Committee Chair:
Allen, Elizabeth S.
Committee Members:
Ranby, Krista
Everhart, Kevin
Renshaw, Keith

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
Copyright Jessica J. Kenny. Permission granted to University of Colorado Denver to digitize and display this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.

Downloads

This item has the following downloads:


Full Text
POSTDEPLOYMENT FAMILY FUNCTIONING: SERVICE MEMBER PTSD
SYMPTOMS, COUPLE AND PARENTING ALLIANCE, AND CHILD BEHAVIORAL
DIFFICULTIES
by
JESSICA J. KENNY B.A., Pepperdine University, 2010 M.A., University of Denver, 2013 M.A., University of Colorado Denver, 2016
A dissertation submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Doctor of Philosophy Clinical Health Psychology Program
2019


ii
This dissertation for the Doctor of Philosophy degree by Jessica J. Kenny has been approved for the Clinical Health Psychology Program by
Elizabeth S. Allen, Chair Krista Ranby Kevin Everhart Keith Renshaw
Date: August 3, 2019


Ill
Kenny, Jessica J. (PhD, Clinical Health Psychology Program)
Postdeployment Family Functioning: Service Member PTSD Symptoms, Couple and Parenting Alliance, and Child Behavioral Difficulties Dissertation directed by Associate Professor Elizabeth Allen
ABSTRACT
As service members (SMs) and veterans are returning home, increasing research efforts are being devoted to assessing postdeployment family functioning. The association between PTSD symptoms and family functioning has been one of the most studied issues for returning SMs (e.g., Creech, Hadley, & Borsari, 2014; Sayers, 2011), and yet several questions related to these connections remain unanswered. This study explored associations between levels of postdeployment PTSD symptoms for male SMs and various aspects of family functioning including couple alliance around SM PTSD symptoms, parenting alliance, and child behavioral difficulties. Participants in the current study (N= 165 couples) are part of the Relationships Among Military Personnel (RAMP) study, a longitudinal study aimed at investigating the interpersonal and mental health of Army couples following deployment. Using a Structural Equation Modeling (SEM) framework, the study evaluated the degree to which (1) male SM PTSD symptom severity is associated with child behavioral difficulties (as reported by the partner) and (2) couple and/or parenting alliance mediate that relationship. In addition, these same associations were evaluated using only the emotional numbing symptoms of PTSD and using child gender as a moderator. Results showed that SM PTSD symptoms and child behavioral difficulties were associated, but only when using the partner report of the SM PTSD symptoms and not when using the SM self-report of PTSD symptoms. In a larger model


combining both partner’s report of SM PTSD symptoms, there was no association between SM PTSD and child behavioral difficulties and therefore there was no direct effect to be mediated. Couple alliance, as compared to parenting alliance, was most associated with both PTSD symptoms and child behavioral difficulties. The presence of PTSD emotional numbing symptoms (as compared to overall PTSD symptoms) was negatively associated with some couple variables but not child behavioral difficulties. Child gender did not moderate the relationship between SM PTSD and child behavioral difficulties. Outside of these hypothesized aims, several other noteworthy results emerged. Limitations of the study are explored and further implications for both clinical work and future research is discussed.
The form and content of this abstract are approved. I recommend its publication.
Approved: Elizabeth Allen


V
ACKNOWLEDGEMENTS
This research is made possible by a research grant that was awarded and administered by the U.S. Army Medical Research & Materiel Command (USAMRMC) and the Telemedicine & Advanced Technology Research Center (TATRC), at Fort Detrick, MD, under Contract Number W81XWH-12-1-0090. Opinions, interpretations, conclusions, and recommendations are those of the authors and are not necessarily endorsed by the Department of Defense.


VI
TABLE OF CONTENTS
CHAPTER
I. INTRODUCTION.....................................................1
Aim 1............................................................9
Aim 2...........................................................10
Aim 3...........................................................10
II. METHODS........................................................12
Procedure.......................................................12
Sample..........................................................13
Instruments.....................................................13
III. RESULTS.......................................................20
Aim 1...........................................................24
Aim 2...........................................................29
Aim 3...........................................................30
IV. DISCUSSION.....................................................37
Limitations.....................................................44
Implications....................................................45
Future Directions...............................................46
REFERENCES............................................................49
APPENDIX..............................................................55


1
CHAPTER I INTRODUCTION
The number of United States service members (SMs) and veterans who have been deployed to war zones has increased dramatically over the past decade (Bonds, Baiocchi, & McDonald, 2010). Deployment to war zones involves a number of significant stressors for United States SMs. When considered from a family systems perspective, the effect of deployment on a military family can also be significant, as more than half (52.4%) of active duty SMs are married and approximately 42% have children (Department of Defense, 2015). Even the most resilient of families can experience significant stress in the postdeployment reintegration period (Sayers, 2011).
Although reintegration after deployment presents challenges for any military family, certain vulnerabilities can contribute to an especially difficult and distressing time after deployment for families. In particular, the association of PTSD symptoms and family functioning for returning SMs has been a focus of empirical attention (e.g., Creech, Hadley, & Borsari, 2014; Sayers, 2011). The percentage of SMs returning from Iraq and Afghanistan who show elevated levels of PTSD symptoms is estimated at 10-18% (Hoge et al., 2006). SMs and veterans with high levels of PTSD symptoms report more numerous and severe relationship problems, and divorce at a higher rate, as compared to their veteran counterparts without PTSD symptoms (Monson et al., 2009). Additionally, high levels of SM and veteran PTSD symptoms are often accompanied by parenting stress and child behavioral difficulties (Creech et al., 2014; Flake, Davis, Johnson, & Middleton, 2009). For example, children of veterans with PTSD exhibit greater behavior problems than children of non-veteran parents without PTSD (Ahmadzadeh & Malekian, 2004). These issues can also be compounded; for example, non-service member


2
(partner) parenting stress at postdeployment, when accompanied by high levels of SM PTSD, predicts poorer psychological functioning among military children (Flake et al., 2009). Thus, PTSD symptoms can make an already difficult time after deployment even more difficult for military families. This highlights the need to increase understanding of how PTSD symptoms and various aspects of family functioning interact for military families.
To this end, this study will address three dynamics within the family system that may be associated with increased levels of SM PTSD symptoms. Dynamics related to the couple include (1) lower levels of couple alliance regarding the SM PTSD symptoms, as reported by both the SM and partner, including (la) lower levels of partner expressed positive emotional responses to the SM regarding SM PTSD symptoms and (lb) lower levels of couple connection around the SM PTSD symptoms. Dynamics related to parenting, as reported by both the SM and partner, include (2) lower levels of parenting alliance. Dynamics related to the children, as reported by the partner, include (3) child behavioral difficulties. How these three family dynamics and SM PTSD symptoms are associated has not previously been examined in combination.
In addition to associations between overall PTSD symptoms and family issues, some researchers have focused on how the specific symptoms of PTSD relate to family issues. Various factor analyses and conceptual categorizations have grouped specific symptoms into different clusters. King, Leskin, King, & Weathers (1998), using factor analysis, grouped self-report PTSD symptoms into four clusters: re-experiencing, avoidance, emotional numbing, hyperarousal. These clusters map onto the different diagnostic criteria groups of PTSD within the DSM-IV, with both avoidance and emotional numbing falling under Criterion C of the DSM-IV (American Psychiatric Association, 2013). Diagnostic criteria for PTSD within the DSM-5 has changed somewhat, but each of these types of symptoms are still included in the new criteria.


3
These King et al. clusters have subsequently been evaluated in relation to various family issues (Samper, Taft, King, & King, 2004; Ruscio, Weathers, King, & King, 2002), with studies demonstrating that the emotional numbing symptoms of PTSD in SMs are the most strongly associated with family issues such as relationship, parenting, and family distress (e.g., Galovski & Lyons, 2004; Wilson & Kurtz, 1997).
The specific symptoms that make up the emotional numbing symptoms are: a loss of interest in important, once positive activities; feeling distant or estranged from others; and difficulties experiencing positive feelings, such as happiness or love (American Psychiatric Association, 2013). Symptoms of emotional numbing often inhibit emotional expression, which is integral to healthy and well-functioning relationships (Riggs, Byrne, Weathers, & Litz, 1998). Thus, instead of emotionally expressing themselves to their family, the SM with emotional numbing symptoms is more likely to withdraw, which can inhibit healing and preclude successful reintegration with their family after deployment (Ray & Vanstone, 2009). Consequently, emotional numbing symptoms may be even more strongly associated with the three family system dynamics, listed above, than overall PTSD symptoms.
The purpose of this study is to look at how SM PTSD symptoms after deployment may be associated with three dynamics within the family system, as listed above. Specifically, this study will examine the relationship between SM PTSD and lower levels of couple alliance regarding SM PTSD symptoms, lower levels of parenting alliance, and child behavioral difficulties. This study will also examine if these relationships differ by child gender. Lastly, this study will investigate whether emotional numbing, as compared to overall PTSD symptoms, is more strongly associated with these family dynamic variables. The three family dynamics included in this study will be discussed in more detail below.


4
SMPTSD and Parenting Alliance
In military couples, parental roles and practices is one of the basic family dynamics that needs to be reestablished or renegotiated after deployment (Sayers, 2011). Parental roles and practices is also one of the family dynamics that may be affected by PTSD symptoms. Increases in PTSD symptoms post-deployment have been found to be significantly associated with impairments in SM parenting (Gewirtz et al., 2010). Therefore, when a SM experiences PTSD symptoms, one corresponding family systems dynamic may be decreased parenting alliance. Parenting alliance refers to the ways in which spouses provide support and show respect for each other in parental roles, and how they work together as a co-parenting team (Abidin & Brunner, 1995). Not surprisingly, low parenting alliance is often associated with marital distress (Abidin & Brunner, 1995). In fact, Allen et al. (2010) found that SM PTSD symptoms were negatively associated with parenting alliance among military couples.
As discussed above, the numbing symptoms of PTSD are associated with parenting dynamics in military families, including a decrease in perceived quality of father-child relationship (Ruscio, Weathers, King, & King, 2002) and decreased parenting satisfaction (Samper, Taft, King, & King, 2004). One explanation for these associations may be that when a SM has emotional numbing PTSD symptoms, the partners may feel particularly emotionally distant, finding it hard to trust one another, and not in agreement on several areas of their relationship, including their parenting alliance. Thus, the SMs’ emotional numbing symptoms may be more strongly negatively associated with parenting alliance, as compared to overall PTSD symptoms.


5
SMPTSD and PTSD Couple Alliance
Another possible family systems dynamic is a low couple alliance regarding the SM PTSD symptoms, or a low level of the couple coming together with regard to the SM PTSD symptoms. In the current study, couple alliance around the PTSD symptoms is posited to be comprised of two related constructs: (1) the partner’s emotional response to the SM as a result of his PTSD symptoms, including both positive/supportive emotions such as caring, compassionate, and connected, and negative/non-supportive emotions such as angry, resentful, and detached, and (2) the couple’s connection regarding the PTSD symptoms, including the effects of the PTSD symptoms on the couple’s relationship in the areas of communication, cohesion, and understanding/support.
There is a consistent link between SM PTSD symptoms and more numerous and severe romantic relationship problems (Galovski & Lyons, 2004; Monson, Taft, & Fredman, 2009). For example, when a SM or veteran has clinical levels of PTSD symptoms, as compared to SMs or veterans without clinical levels of PTSD, they are more likely to withdraw more from their partner during important conversations (Galovski & Lyons, 2004) and have more emotional intimacy difficulties, including being less self-disclosing and expressive with their partners (Riggs et al., 1998). Therefore, greater levels of SM PTSD symptoms are likely associated with lower couple alliance regarding the PTSD symptoms.
Of the specific types of PTSD symptoms, emotional numbing PTSD symptoms may be particularly associated with lower levels of PTSD couple alliance. By definition, emotional numbing PTSD symptoms particularly affect the SMs’ ability to connect and feel love towards others, which therefore often creates distance, confusion, and hurt feelings. This supports Riggs and colleagues (1998), who state that symptoms of emotional numbing decrease emotional


6
expression and intimate exchanges in a relationship. Thus, it is likely that emotional numbing symptoms, as compared to overall PTSD symptoms, may be even more strongly associated with a low couple alliance regarding the PTSD symptoms.
SM PTSD and Child Behavioral Difficulties
When a SM returns home with PTSD symptoms, their children are also affected. As noted above, SM PTSD and child behavioral difficulties are associated (Ahmadzadeh & Malekian, 2004; Jordan et al., 1992). Ruscio et al. (2002) suggest that the SMs’ PTSD symptoms of emotional numbing and withdrawal may directly impact a SMs’ ability to parent by decreasing his ability to engage in normal interactions with the child that are necessary for developing a meaningful parent-child relationship, thus increasing the likelihood for child behavioral difficulties. Therefore, emotional numbing symptoms may be more strongly associated with child behavioral difficulties, as compared to overall PTSD symptoms.
SMPTSD, Parenting Alliance, PTSD Couple Alliance, Child Behavioral Difficulties
It is likely that the above constructs interact and occur together within the family system. There are established connections between SM PTSD and marital distress (Riggs et al., 1998), marital distress and parenting alliance (Abidin & Brunner, 1995), and child behavioral difficulties and caregiver (non-service member parent) distress during the postdeployment period (Chandra et al., 2010; Flake et al., 2009; Lester et al., 2010). Drawing from general (i.e., nonmilitary specific) research, impaired parenting has long been recognized as a crucial detriment to children’s health, adaptation, resilience, and development, particularly in stressful situations (e.g., Masten, 2001; Gewirtz et al., 2014). As reviewed in Gewirtz et al. (2014), family stress models (e.g., Conger, Ge, Elder, Lorenz, & Simons, 1994) would suggest that the most stable and united parental units play a large role in buffering a child’s development of behavioral


7
and/or emotional difficulties in response to stress. In fact, low parenting alliance predicts more child behavioral difficulties (Bearss & Eyberg, 1998). Thus, it is likely that a low reported parenting alliance may help to explain some of the relationship between SM PTSD symptoms and child behavioral difficulties. Moreover, Snyder et al. (1988) found that mothers’ reports of disagreements with their husbands about childrearing correlate with children’s behavior problems. Thus, in the current study, it is likely that there is an association between low parenting alliance and child behavioral difficulties. Moreover, one can imagine how lower PTSD couple alliance may be related to lower parenting alliance due to the negative emotions, lack of connection, and a likely lack of trust and “we-ness.” Finally, it is possible that the negative association between PTSD and parenting alliance will be stronger when couple alliance regarding the PTSD symptoms is lower. That is, when the couple has lower levels of connection, support, and communication regarding the PTSD symptoms, it may further undermine a sense of teamwork and exacerbate the negative association of SM PTSD symptoms with parenting alliance.
Child Behavioral Difficulties by Child Gender
There is a host of research which details gender differences in emotional expression and the proclivity for children to develop internalizing versus externalizing symptoms when stressed. There is also a rich theoretical literature base which offers social, psychological, and biological explanations for such gender differences (e.g. Brody, 1999; Kring & Gordon, 1998; Brody & Hall, 2008; Chaplin & Aldao, 2013). In particular, gender role theory (Brody & Hall, 2008) proposes that, consistent with “gender-related display rules” in the United States and other Western cultures, girls are expected to display greater levels of emotions than boys (in particular, internalizing emotions such as sadness, fear, anxiety, shame and guilt). In contrast, boys are


8
expected to show less emotion and instead are expected or allowed to express themselves via externalizing behaviors (such as anger, contempt, aggression, disgust). Therefore, Brody et al. would argue that the gender differences between internalizing and externalizing behaviors in children are a result of traditional societal expectations and roles for each gender, and the internalization of such roles. Chaplin and Aldao (2013) conducted a recent comprehensive meta-analytic review of gender differences in emotional expression in children and did indeed find small but significant gender differences effect sizes; girls overall show more internalizing emotions (g = -. 10) than boys, and boys overall show more externalizing emotions (g = .09) than girls. However, these effect sizes are small, and several other theoretical and research findings note that such differences are varied, nuanced, and not necessarily static. For example, such differences could be strengthened or attenuated based on age, social and physiological development, and societal norms and influences (Brody & Hall, 2008; Chaplin & Aldao, 2013; Brown, 1999). Moreover, this is not a topic that has been extensively researched in the literature on military children. In fact, the unique culture, values, and expectations of the military may also affect gender differences in emotional expression for children. Thus, in the current study, I will explore the relationships of SM PTSD symptoms and both internalizing and externalizing child behavior for boys and girls separately.
Aims of Current Study
As reviewed above, PTSD symptoms may be associated with several different family systems dynamics, including a negative association with couple alliance around the SM PTSD symptoms and parenting alliance, and a positive association with child behavioral difficulties. Although prior research has studied how some of these dynamics are associated with PTSD symptoms, none have examined them in combination. By examining them in combination, it


9
may reinforce the idea that these processes are interrelated, and that intervening on one dynamic of the family system could alleviate distress in other parts of the family system (see the Clinical Implications section, below).
Aim 1: To evaluate basic associations among all variables, including understanding which aspects of child behavioral difficulties (Hyperactivity, Emotional Symptoms, Conduct Problems, Peer Problems, lower Prosocial Behaviors) are associated with which SM PTSD symptom clusters (Intrusions, Avoidance, Numbing, Hyperarousal). With all these aims, I will be evaluating the child behavioral difficulties of the youngest or only child.
Hypothesis 1:
(la) Service member PTSD symptomology will be significantly and positively correlated with child behavioral difficulties and negatively correlated with partner expressed positive emotions toward the SM related to his PTSD symptoms, couple connection regarding SM PTSD symptoms, and parenting alliance.
(lb) Emotional numbing symptoms will show a stronger association with the above variables compared to overall PTSD symptoms.
(lc) The PTSD clusters of Avoidance and Numbing will correlate with internalizing child behavioral difficulties (Emotional Symptoms and Peer Problems), and the PTSD clusters of Hyperarousal and Intrusions will correlate with externalizing child behavioral difficulties (Conduct, Hyperactivity, Prosocial).
Aim 2: To understand if the relationships between PTSD and child behavioral difficulties
are moderated by gender.


10
Hypothesis 2:
Gender will moderate the relationship between PTSD symptoms and child behavioral difficulties, such that the association between PTSD symptoms and externalizing behaviors will be stronger for boys than girls, and the association between PTSD symptoms and internalizing behaviors will be stronger for girls than boys.
The final aim of the study is to test a model where (1) higher SM PTSD symptoms are associated with lower parenting alliance, (2) this association will be moderated by PTSD couple alliance, such that the negative association between PTSD symptoms and parenting alliance is stronger when PTSD couple alliance is low, and (3) lower parenting alliance will be related to greater child behavioral difficulties.
Aim 3: To evaluate a moderated mediation model wherein parenting alliance (M) mediates the association between PTSD symptoms (X) and child behavioral difficulties (Y), and PTSD couple alliance (W) is a moderator of the association between PTSD and parenting alliance (see Model 1, below). Final implementation of the above model will be dependent upon issues such as the nature of the variables (e.g., distributions), tests of the measurement model (e.g., latent factors), and iterative tests of the structural model. Model fit will be compared for boys and girls. The model will also be estimated for both overall PTSD symptoms and only the emotional numbing PTSD symptoms.
Each construct will be measured at timepoint four of the overall study in order for the model to represent a snapshot of what is happening for a family at a particular time during postdeployment. As these are measured cross-sectionally and are not experimentally manipulated, it is acknowledged that the design does not provide any evidence of causation.


Figure 1. Hypothesized Moderated Mediation Model


12
CHAPTER II METHODS
Procedure
Participants in the current study (N= 165 couples) were participants who were part of the Relationships Among Military Personnel Study (RAMP), a longitudinal study aimed at investigating the interpersonal and mental health of Army couples following deployment. Recruitment for the study occurred between June 2013 and June 2014. RAMP utilized a variety of online recruitment methods to reach out to military couples. All recruitment materials directed interested individuals to the study website, which included information about the study, an online screening survey, and the informed consent document. Both members of the couple had to complete the screening to be considered for participation. Completed screening surveys were examined for the following eligibility criteria: (1) individuals had to be at least 18, English proficient, and able to pass informed consent comprehension questions, (2) individuals had to be a member of a dyad consisting of a male Army soldier (current or recent active duty) and civilian female with no prior military experience, (3) the male had to have returned from a deployment within the last 2 years, and (4) the couple had to be in a serious relationship for at least 1 year and beginning before the soldier’s past deployment. Additional individual evaluations of screening materials were conducted to rule out duplicate or questionable cases.
After this screening process, there were a total of 715 couples (1430 individuals) who were invited into the RAMP study. A total of 1,242 individuals participated in timepoint one, which served as a further screen for completeness, quality of responding, and level of PTSD symptoms. Based on this screen, couples were invited into a longitudinal sample, where they were given four additional surveys (i.e., timepoints two - five), spaced approximately six months


13
apart. Data for the current study were drawn from timepoint four. Five hundred and fifty-three individual timepoint four surveys were sent and 510 individuals (241 matched couples) completed timepoint four. Couples for the current study were selected if they met the following criteria: (1) couples who both have complete data at timepoint four and (2) couples who have a youngest or only child between the ages of 3 and 18, per maternal report, at timepoint four. This screening resulted in a final sample of 165 couples.
Sample
At timepoint four, participants (N= 165 couples) were aged 20-54, with an average age of 31.5 (SD = 5.42) for females and an average age of 32.68 (SD = 4.88) for males. Regarding marital status, 97% of the sample is married. In terms of race and ethnicity, 81.7% of individuals identify as White, 9.0% as Hispanic, 4.2% as Black or African American, and 5.1% as other or multiracial. Regarding education, 56% of males and 46% of females have some college but not a 4-year degree. After selecting for couples who have a youngest or only child between the ages of 3 and 18 at timepoint four, the range of ages of youngest children was 3-18, with a mean of 6.78 (SD = 3.47). When splitting the sample by child gender, boys’ and girls’ ages do not deviate significantly (boys: range = 3-17; mean = 6.46, SD = 3.00); girls: range = 3-17; mean = 6.78, SD = 3.47).
Instruments (See Appendix B for complete measures)
SMPTSD Symptoms.
To assess for SM PTSD symptoms, the PTSD Checklist - Military Version (PCL-M; Weathers et al., 1993) was given to male SMs. Consistent with the other variable names within this study, the PCL-M will be referred to hereafter as the PCL-SM, as it is a measure that was completed by the SM. The PCL-SM is a self-report scale comprised of 17 items that correspond


14
to the key symptoms of PTSD in the DSM-IV. The PCL-SM is specific to PTSD caused by military experiences. Respondents indicate how much they have been bothered by a symptom over the past month using a 5-point scale from 1 (not at all) to 5 (extremely). The PCL-SM is scored by summing the items for a total severity score. Allen et al. (2010) found this measure to be reliable in a similar sample (a = .94) and Weathers et al. (1996) found this to be a valid measure to assess PTSD symptoms in a military population.
To assess for female partners perceptions of their SM’s PTSD symptoms, the military version of the Posttraumatic Checklist was modified into the female partner perception of the PCL-SM (called the PCL-partner or PCL-p), which instructs partners to “indicate how much you think your spouse has had that problem..on the various PCL-SM items. A study by Renshaw, Rodrigues, and Jones (2008) similarly adapted the PCL-SM for spouses and the scale demonstrated strong internal consistency (a = .91).
Both the SM and partner versions of the PCL-SM were found to be reliable in the current RAMP sample (a = .96 and a = .94, respectively). As noted in Table 1, the average PCL-SM score was 40.48 (SD = 17.50) and the average PCL-p score was 35.07 (SD = 14.46). Bliese et al. (2008) suggests a cut-off of 34 for estimating a clinical diagnosis of PTSD. Thus, both the SM and partner report of average PCL scores represent risk for clinical levels of PTSD for a significant proportion of the sample.
Couple Alliance (regarding SM PTSD symptoms).
To assess the couples’ perception of the extent to which they have a positive alliance regarding the SM PTSD symptoms, two measures each were given to both the SM and the
partner.


15
Partner Emotional Response to SMRegarding SM PTSD Symptoms.
To assess the partners’ report of their own emotional responses to their SM as a result of his PTSD symptoms, the partner was given the Emotional Responses to PTSD Symptoms scale (called the ERPS-p). This measure was given after the PCL-p measure of PTSD symptoms, and the instructions note that such “problems” (the language used in the PCL-p) can have various effects on relationships. The instructions for the ERPS-p ask partners to rate 14 different feelings, on a scale of 1 (not at all) to 7 (very much), on the extent that they have each emotion towards their SMs as a result of the SM’s PTSD symptoms.
To assess for the SM’s perception of their partner’s emotional responses to them (SM) as a result of their (SM) PTSD symptoms, the SM was given the perception of the Emotional Responses to PTSD Symptoms scale (called the ERPS-SM). This ERPS-SM was given after the PCL-SM measure of PTSD symptoms. The instructions for the ERPS-SM ask the SMs to rate 14 different feelings, on a scale of 1 (not at all) to 7 (very much) on the extent that their partner has each emotion towards them (SMs) as a result of their PTSD symptoms. Thus, the scales given to the SMs and partners are both asking about the partner’s emotional responses.
Emotions include both positive and supportive emotions (e.g., caring, compassionate, connected) and negative or non-supportive emotions (e.g., angry, resentful, detached). This measure is scored by reverse scoring the negative or non-supportive emotions and then taking the average of the fourteen items, with higher scores reflecting more positive emotions and alliance around the SM’s PTSD symptoms. This


16
scale showed internal consistency in this sample of both SMs (ERPS-SM; a = .90) and partners (ERPS-p; a = .90).
Couple Connection Regarding SM PTSD Symptoms.
To assess SMs’ and partners’ perception of how well the couple communicates about and connects around PTSD symptoms, the PTSD Couple Connection Scale (called the PCC-SM for service members and PCC-p for partners) was given to both partners (Allen et al., 2012). The items in this scale are based on the themes noted by Nelson-Goff et al. (2006) on the impact of trauma on intimate relationships, such as increased or decreased communication, cohesion, and understanding/support. This scale is also given after the PCL-SM and PCL-p measures of PTSD symptoms, and the instructions note that such “problems” (the language used in the PCL-SM) can have various effects on relationships. The instructions then direct respondents to endorse 11 items from 1 (not at all true) to 7 (very true). The 11 items used for the current study are items that are exactly symmetrical between the male and female versions (the full scale includes 13 items for SMs and 15 items for partners). This measure is scored by reverse scoring items that reflect poor connection and then averaged for a total score, with higher scores indicating greater connection and couple alliance around the PTSD symptoms. As noted above, the items are based on Nelson-Goff et al., and focus on whether or not the couple communicates about symptoms (e.g., “When these problems come up, I tell my partner what is going on with me then”), cohesion (e.g., “These problems have brought us closer together”), and understanding/support (e.g., “My partner understands what I am going through).


17
Allen et al. (2012) tested a version of this measure in a sample of Army couples and found good internal consistency for both patients and partners, significant agreement between spouses, and strong evidence for criterion validity for the scale (i.e., significant positive associations with marital satisfaction and negative associations with reintegration adjustment after deployment). This measure also showed strong internal consistency for both SMs (PCC-SM; a = .91) and partners (PCC-p; a = .90) in the current sample. Parenting A lliance.
To assess for parenting alliance, five items from the original twenty item Parenting Alliance Inventory (PAI; Abidin & Brunner, 1995) were given to both the SM and the partner (called the PAI-SM and PAI-p, respectively). As discussed in Allen et al. (2010), the PAI assesses the degree to which parents perceive themselves to be in a cooperative, communicative, and mutually respectful alliance for the care of their children on a scale of 1 (never) to 5 (always). Abidin and Brunner found high convergence between the PAI and measures of marital distress and parenting style. The five items used in this study were chosen on the basis of representativeness of teamwork between the parents (e.g., “My spouse and I are a good parenting team”), rather than appraisals of the other parent's skill or enjoyment of parenting.
Allen et al. (2010) used this same five item PAI measure to measure parenting alliance in a separate military couple sample. The five chosen items all had factor loadings of at least .50 for both men and women in the original validation sample (Abidin & Brunner, 1995; Allen et al., 2010). This measure also showed strong internal consistency in the current RAMP sample for
both SMs (PAI-SM; a = .93) and partners (PAI-p; a = .96).


18
Child Behavioral Difficulties.
To assess for child behavioral difficulties, the Strengths and Difficulties Questionnaire (SDQ; Goodman, 1997) was given to female partners with instructions to answer the questions about the couples’ youngest child. The SDQ is a 25-item measure of a range of children's attributes. The SDQ has 5 scales, each containing 5-items: (1) Hyperactivity Scale, (2) Emotional Symptoms Scale, (3) Conduct Problems Scale, (4) Peer Problems Scale, and (5) Prosocial Scale. The sum of the 25 items included in the 5 scales generate a total difficulties score (Goodman, 1997) when the Prosocial Scale items are reverse scored to reflect child behavioral difficulties. Three different versions of the SDQ exist depending upon the age of the child. A slightly different version exists for children 3 years old, with 2 items reflecting oppositionality rather than antisocial behavior. Versions for 4-10 year-olds and 11-17 year-olds are virtually identical, aside from some items being worded differently for the 11-17 year-old version (Goodman,
1997). Responses require parents to report how true the item describing a behavior is to their child on a 3-point scale, from 1 (Not true) to 3 (Certainly true).
The SDQ has demonstrated concurrent validity with similar measures of child behavior. Additionally, the SDQ has established predictive validity by being able to discriminate between psychiatric and non-psychiatric samples (Goodman, 1997). In the current sample, reliability was adequate for the three age group versions (age 3 version a = .77; age 4-10 version a = .85; age 11-17 version a = .89). Of note, the reliability of the age 3 version excludes one item due to zero variance, but this item was included in scoring. Due to the three versions of the measure given to partners based on the age of their youngest or only child, an overall reliability for the sample is
not possible.


19
There is theoretical and preliminary empirical support for combining the SDQ's hypothesized emotional and peer subscales into an 'internalizing' subscale and the hypothesized behavioral, hyperactivity, and reverse scored prosocial subscales into an 'externalizing' subscale (Goodman, Lamping, Ploubidis, 2010). Within Aim 2, the internalizing and externalizing subscales are used.


20
CHAPTER III RESULTS
Prior to hypothesis testing, I first looked at descriptive and reliability statistics for all study variables. I also ran paired /-tests and correlations for all constructs that had both SM and partner reports (e.g., PCL-SM is SM’s report and PCL-p is partner’s report of SM PTSD symptoms). Please see Table 1, below. As noted above, all scales showed excellent internal consistency, with all alphas of .85 or above. The one exception is the SDQ items for children aged 3, which produced an alpha of .77. There was less variance in this age group (with scores ranging from 1-19 as compared to 0-33 and 2-31 on the SDQ scales for 4-10 year olds and 11-17 year olds, respectively), which may have contributed to the relatively lower alpha. Further, each correlation of couple reports was between r = .40 and .47, which offers evidence for criterion validity of the reports of these constructs. Although each construct with both SM and partner report was significantly correlated, the t-test results indicate that they are also all (except for the PAI-SM and PAI-p; p = .051) significantly different from one another in terms of absolute levels. That is, the positive correlation suggests that the couple reports generally move in the same direction, just with different levels.
Table 1.
N Mean (SD) Range a Paired t-test (effect size) Correlation
PCL-SM (17 items/sum) 164 40.17 (17.42) 17-85 .96 t = 3.94** (d=3l) 47** (JV=164)
PCL-p (17 items/sum) 164 35.07 (14.46) 17-77 .94
PCL-SM Re-experiencing subscale (5 items/sum) 164 10.85 (5.76) 5-25 .95 t = 3.49** (,d= .27) 47** (JV=164)
PCL-p Re-experiencing subscale (5 items/sum) 164 9.35 (4.74) 5-25 .92


21
Table 1. cont’d
PCL-SM Hyperarousal subscale (5 items/sum) 164 13.50 (5.77) 5-25 .88 t = 3.21** ( PCL-p Hyperarousal subscale (5 items/sum) 164 12.08 (5.20) 5-24 .86
PCL-SM Situational avoidance subscale (2 items/sum) 164 4.57 (2.60) 2-10 .89 t = 2.76** (,d= .22) 40** (N= 164)
PCL-p Situational avoidance subscale (2 items/sum) 164 4.00 (2.20) 2-10 .82
PCL-SM Emotional numbing subscale (5 items/sum) 164 11.25 (5.36) 5-25 .87 t = 3.82** (d= .30) 40** (N= 164)
PCL-p Emotional Numbing subscale (5 items/sum) 164 9.64 (4.41) 5-23 .84
ERPS-SM (14 items/mean, higher scores reflect more alliance) 148 4.64 (1.22) 1.36-7 .90 t = 2.55* (d= .21) 42** (/V=148)
ERPS-p (14 items/mean, higher scores reflect more alliance) 148 4.91 (1.24) 2-7 .91
PCC-SM (11 items/mean, higher scores reflect more alliance) 148 4.06 (1.37) 1-7 .91 t = 2.89** (<*=.24) 42** (/V=148)
PCC-p (11 items/mean, higher scores reflect more alliance) 148 4.40 (1.28) 1.91- 6.73 .90
PAI-SM (5 items/mean, higher scores reflect more alliance) 160 5.80 (1.28) 1-7 .93 t = 1.96 (P = .051) (d= .16) .45** (/V=160)
PAI-p (5 items/mean, higher scores reflect more alliance) 160 5.57 (1.52) 1-7 .96
SDQ Total (25 items/sum) 176 10.82 (6.99) 0-33 — — —
SDQ child age 3 (Reliability excludes one item due to zero variance, but this item is included in scoring) 19 10.47 (4.90) 1-19 .IT
SDQ child age 4-10 136 10.38 (6.81) 0-33 .85 — —
SDQ child age 11-17 21 14.00 (9.01) 2-31 .89 — —
SDQ Emotional Symptoms subscale (5 items) 176 2.12 (2.14) 0-9 — — —
SDQ Peer Problems subscale (5 items) 176 1.65 Q70) 0-8 — — —


22
Table 1. cont’d
SDQ Conduct Problems subscale (5 items) 176 2.14 (2.15) 0-9 — — —
SDQ Hyperactivity subscale (5 items) 176 4.18 (2.83) 0-10 — — —
SDQ Prosocial subscale (5 items) 176 1.47 (170) 0-8 — — —
Note:
PCL-SM = PTSD Checklist-Military. PCL-p = partner report of PTSD Checklist-Military.
ERPS-SM = SMperception of partner Emotional Responses to PTSD Symptoms scale.
ERPS-p = partner self-report of Emotional Responses to PTSD Symptoms scale.
PCC-SM = male report of PTSD couple connection scale.
PCC-p = partner report of PTSD couple connection scale.
SDQ =(partner report of Strengths and Difficulties Questionnaire
There are several other main themes of Table 1 to note. Overall, all of the PCL total and subscales from both SM and partner report are significantly correlated with one another, corresponding to what we would expect of such reports. One general theme that emerged was that males, on average, self-reported higher symptom severity and relationship distress than did their partners. This theme first emerged with SMs self-reporting a higher severity of PTSD symptoms on the PCL-SM (M= 40.17, SD = 17.42) as compared to their partners’ report of their SM’s PTSD symptoms on the PCL-p (M= 35.07, SD = 14.46) (*[163] = 3.94;p >_.05). This theme emerged on each PCL subscale as well (see Appendix A for the table that includes all of the PCL subscales). Further, the SM reported that the partner responds with less positive emotions towards him on the Emotional Responses to PTSD scale (ERPS-SM; M= 4.64, SD = 1.22) than she herself reported (ERPS-p; M= 4.91, SD = 1.24) (*[163] = 2.55,/? >_.05), with higher scores reflecting more positive emotions. The SM also reported that he perceived less couple connection related to his PTSD symptoms on the PTSD Couple Connection Scale (PCC-SM; M= 4.06, SD = 1.28) as compared to her perception of their connection (PCC-p; M= 4.40, SD = 1.28) (*[163] = 2.89,/) >_.01), with higher scores reflecting more connection. Thus, overall, he is reporting more individual and relationship distress. However, parenting alliance did not evidence significant differences, with both the SM (PAI-SM; M= 5.80, SD = 1.28) and the


23
partner (PAI-p; M= 5.57, SD = 1.52) reporting relatively equal levels of alliance as it relates to their parenting (/[ 163] = 1.96 ,p> .05).
Child Behavioral Difficulties Compared with National Norms
Although not an initial hypothesis, an important theme that emerged from these data was that the children in the RAMP sample, as reported by the partner, are generally more distressed as compared to the norm SDQ sample in the United States. The SDQ website (http://www.sdqinfo.com/norms/USNorml.pdf) provides norms for the SDQ within the United States for 4-17 year olds. I used only 4-17 year olds in the RAMP sample (N= 156) to match the age range for the United States norms (note, excluding 3 year olds did not change the RAMP sample means, varying by no more than .1) As seen in Table 2, the RAMP sample was statistically higher on overall total difficulties than the national norms. This trend remained true for every subscale except for the Prosocial Behavior subscale. Although the RAMP sample appears to be exhibiting a higher level of child behavioral difficulties overall, as compared to the national norms, it is helpful to ground the averages in the response scales. The SDQ is summed and responses are on a scale of 0 (not true), 1 (somewhat true) or 2 (certainly true). Thus, partners within my sample are rating their children, on average, as experiencing the stated difficulties over the past six months or school year on average as somewhere between not true and somewhat true.
Table 2.
SDQ score USA norms (A=9878 4-17 year olds) M (SD) RAMP sample (A=156 4-17 year olds) M (SD) One-sample /-test
Total difficulties (25 items summed) 7.1 (5.7) 10.86 (7.22) t= 6.54**
Emotional symptoms (5 items summed) 1.6 (1.8) 2.22 (2.20) t = 3.55**


24
Table 2. cont’d
Conduct problems (5 items summed) 1.3 (1.6) 2.22 (2.20) t= 5.26**
Hyperactivity (5 items summed) 2.8 (2.5) 4.11 (2.90) t= 5.69**
Peer problems (5 items summed) 1.4 (1.5) 1.69 (1.7) t = 2.09*
Prosocial behavior (reverse scored; 5 items summed) 1.4 (1.8) 1.45 (1.71) t= .34
Aim 1
My first aim was to evaluate basic associations among all variables, including understanding which aspects of child behavioral difficulties (Hyperactivity, Emotional Symptoms, Conduct Problems, Peer Problems, lower Prosocial Behaviors) are associated with which SM PTSD symptoms clusters (Intrusions, Avoidance, Numbing, Hyperarousal). With all these aims, I evaluated the child behavioral difficulties of the youngest or only child. Please see Appendix A for a bivariate correlation table of all study measures and subscales. I will go into detail on data analyses and specific hypotheses and provide pertinent bivariate correlation tables for Aim 1, below.
Hypothesis la
The first part of my hypothesis for Aim 1 was that SM PTSD symptomology will be significantly and positively correlated with child behavioral difficulties and negatively correlated with partner expressed positive emotions toward the SM related to his PTSD symptoms, couple connection regarding SM PTSD symptoms, and parenting alliance. I addressed this by running several Pearson’s correlations using SM and partner report of the PTSD and couple variables, and partner report of the child behavioral difficulties variable. Please see Table 3, below.
Overall, this hypothesis was partially confirmed. For the most part, PTSD was significantly


25
associated with the above couple variables (partner expressed positive emotions, couple connection, parenting alliance) in the expected direction, using both SM and partner report. The few exceptions to this finding were found in associations of SM report of various constructs with partner report of various constructs (i.e., constructs reported by the same reporter were all significant). For example, the only non-significant association of SM report of PTSD symptoms (PCL-SM) with a couples measure was with the partner report of PTSD Couple Connection Scale (PCC-p; r = -.14, ns). Likewise, non-significant associations with partner report of SM PTSD include SM report of the PTSD Couple Connection Scale (PCC-SM; r = -.08, ns) and parenting alliance (PAI-SM; r = -.10, ns). Lastly, SM PTSD symptoms were associated with overall child behavioral difficulties, but only when looking exclusively at partner report of SM PTSD symptoms (PCL-p; r =19,/? <05). This association disappeared when looking exclusively at SM report of his own PTSD symptoms (PCL-SM; r = .00, ns). Likewise, partner report of
various subscales on the PCL-p were correlated with various subscales of child behavioral difficulties on the SDQ, but SM self-report on the PCL-SM scales were not associated with any child difficulty subscales as reported by the partner on the SDQ.
Table 3, Bivariate correlations with all variables: Boys and girls
2 3 4 5 6 7 8 9 10 11
1. PCL-SM .47 .87 .39 -.50 -.26 -.40 -.14 -.22 -.25 .00
2. PCL-p — .39 .87 -.28 -.46 -.08 -.31 -.10 -.24 .19
3. Emotion Numb-SM — .40 -.59 -.31 -.46 -.11 -.28 -.28 .04
4. Emotion Numb-p — -.35 -.59 -.12 -.38 -.15 -.32 .21
5. ERPS-SM — .43 .66 .31 .39 .29 .03
6. ERPS-p — .34 .72 .18 .44 -.36
7. PCC-SM — .42 .41 .21 .03
8. PCC-p — .16 .34 -.33
9. PAI-SM — .45 -.03
10 PAI-p — -.21
11 SDQ Total —


26
Note: Bolded Pearson’s correlation signifies a statistically significant association (p< .05). Gray shaded numbers signify the association between partner report of all constructs with partner report of child behavioral difficulties (SDQ Total).
PCL-SM = PTSD Checklist-Military. PCL-p = partner report of PTSD Checklist-Military.
ERPS-SM = SMperception of partner Emotional Responses to PTSD Symptoms scale.
ERPS-p = partner self-report of Emotional Responses to PTSD Symptoms scale.
PCC-SM = male report of PTSD Couple Connection scale.
PCC-p = partner report of PTSD Couple Connection scale.
SDQ = (partner report of Strengths and Difficulties Questionnaire
Hypothesis lb
Hypothesis lb of Aim 1 was that emotional numbing PTSD symptoms would emerge as more strongly associated with the above variables as compared to overall PTSD symptoms. I addressed this by running Pearson’s correlations and then using an r-to-z Fisher’s calculation for the test of the differences between dependent correlations (Lee & Preacher, 2013). See Table 4 for a visual display of these associations. Although simply looking at the associations between overall PTSD symptoms versus emotional numbing symptoms suggests that these associations do increase when looking at only emotional numbing symptoms, the overall hypothesis is only partly confirmed when using Fisher’s r to z to statistically test the difference in magnitudes between the two dependent PTSD variables. Using partner report, there were two associations that did statistically increase in magnitude when going from overall PTSD symptoms to emotional numbing symptoms: her report of her emotional response to him regarding his PTSD symptoms (from -.46 to -.59; ERPS-p: z = 3.92 ,p < .00) and her report of parenting alliance (from -.24 to -.32; PAI-p: z = 2.09 ,p< .05). The z test using her report of PTSD Couple Connection approached significance (from -.31 to -.38; PCC-p: z = 1.87,p = .06). Using SM report, there was only one association that statistically increased in magnitude when going from overall PTSD symptoms to emotional numbing symptoms: his report of her emotional response to him regarding his PTSD symptoms (from -.50 to -.59; ERPS-SM: z = 2.74; p < .01). Finally,


27
the associations with child behavioral difficulties did not statistically increase in magnitude with
either SM or partner report of overall PTSD symptoms versus emotional numbing symptoms.
Table 4. Bivariate correlations: Comparing overall PTSD and emotional numbing PTSD symptoms with couples’ constructs_____________________________________________________________________
ERPS-SM ERPS-p PCC- SM PCC-p PAI-SM PAI-p SDQ Total
PCL-SM -.50 -.26 -.40 -.14 -.22 -.25 .00
Emotion Numb-SM -.59 -.31 -.46 -.11 -.28 -.28 .04
PCL-p -.28 -.46 -.08 -.31 -.10 .19
Emotion Numb-p -.35 -.59 -.12 -.38 -.15 -.32 .21
Note: Bolded Pearson’s correlation signifies a statistically significant association (p< .05). Blue shading signifies associations that significantly increased when going from SM report of overall PTSD to emotional numbing symptoms. Pink shading signifies associations that significantly increased when going from partner report of SM overall PTSD to SM emotional numbing symptoms.
Hypothesis lc
The third part of Aim 1 was related to understanding which SDQ scales or areas of child difficulty are correlated with which SM PTSD symptoms clusters (Please see Table 5). Hypothesis lc of Aim 1 was that the PTSD symptoms of Avoidance and Numbing will correlate with internalizing child behavioral difficulties (Emotional Symptoms and Peer Problems), and PTSD clusters of Hyperarousal and Intrusions will correlate with externalizing child behavioral difficulties (Conduct, Hyperactivity, Prosocial). I addressed this by running Pearson correlations using SM and female report of PTSD total and sub scale scores, and partner report of child behavioral difficulties total and subscale scores. See Table 5 for expected hypothesized associations.
There were times when these hypothesized correlations emerged (Emotion Numb-p with SDQ Emotional: r = .20, p < .05, and SDQ Peer Relations: r = .25, p < .05; Avoidance-p with


28
SDQ Emotional: r=.l6,p< .05; and Hyperarousal-p with SDQ Conduct: r = ,2\,p< .05). However, there were also significant correlations that did not follow the hypothesized pattern (EmotionNumb-p with SDQ conduct: r = 20,p< .05; Avoidance-p with SDQ Conduct: r = .17, p < .05; Hyperarousal-p with SDQ Emotional: r = .21 ,P< .05). Thus, there was not a clear differentiation between internalizing and externalizing symptoms in both SM and children symptoms.
Consistent with overall patterns of female’s report of overall PTSD symptoms (PCL-p) being more associated with child behavioral difficulties, the same pattern emerged when looking at subscales as well. None of the PCL-SM subscales emerged as associated with any child behavioral difficulties. The remainder of this paragraph discusses findings related to only partner report of SM PTSD symptoms. Partner report of SM reexperiencing symptoms was not associated with any child behavioral difficulties. However, the remainder of the PCL-p subscales did associate with at least one child behavioral difficulty subscale. Emotional numbing SM PTSD symptoms were most associated with child behavioral difficulties overall. Hyperarousal and Avoidance SM PTSD symptoms were also associated with two out of five of the child behavioral difficulty subscales. Emotional and Conduct behaviors were most associated with PTSD total and some subscale variables. Overall, child Prosocial and Hyperactivity behavior was not associated with any PTSD total or subscale variable.
Table 5. Bivariate correlations with PCL anc SDQ: Boys and Girls
SDQ Total SDQ Emotional SDQ Conduct SDQ Hyperactivity SDQ Peer Relations SDQ Pro Social
PCL-SM .00 .04 .03 -.01 -.00 .00
PCL-p .19 .20 .20 .10 .15 .09
Reexper-SM -.01 .04 .03 .02 -.04 -.02
Reexper-p .09 .12 .13 .05 .01 .06


29
Table 5. cont’d
Emotion Numb-SM .04 .06 .06 -.03 .10 -.03
Emotion Numb-p .21 .20 .20 .06 .25 .11
Avoidance-SM -.08 -.06 -.07 -.03 -.10 .01
Avoidance-p .15 .16 .17 .11 .11 .07
Hyperarousal-SM .01 .04 .04 -.02 -.03 .04
Hyperarousal-p .21 .21 .21 .13 .14 .07
Note. Bolded Pearson’s correlation signifies a statistically significant association (p< .05). Blue shaded numbers signify hypothesized associations among PCL and SDQ subscales.
Aim 2
The second aim was to understand if the relationships between PTSD symptoms and child behavioral difficulties are moderated by gender. My hypothesis was that gender will moderate the relationship between PTSD symptoms and child behavioral difficulties, such that the association between PTSD symptoms and externalizing behaviors will be stronger for boys than girls, and the association between PTSD symptoms and internalizing behaviors will be stronger for girls than boys. To address this hypothesis, I used the PROCESS macro of SPSS (Hayes, 2018). In these analyses, I combined SM and partner reports. That is, I averaged them together to form a combined latent score. I ran 10 moderation analyses to understand if the relationship between PTSD symptoms (combined SM and partner report of overall PTSD symptoms, emotional numbing symptoms, situational avoidance symptoms, hyperarousal symptoms, and re-experiencing symptoms) and child behavioral difficulties (internalizing and externalizing behaviors subscales of the SDQ) depended on gender. No interactions were significant (all ps > .05), indicating that none of the relationships depend on gender. However, for interest, the correlations are disaggregated for boys and girls in Table 6, below.


30
Note. Associations using only boy children are top left of the cells. Associations using only girl children are bottom left of the cells. Bolded Pearson’s correlation signifies a statistically significant association {p< .05). Blue shaded numbers signify hypothesized associations among PCL and SDQ subscales.
Aim 3
The third aim was to evaluate a moderated mediation model wherein parenting alliance (M) is a mediator between PTSD symptoms (X) and child behavioral difficulties (Y), and PTSD couple alliance (W) is a moderator between PTSD symptoms and parenting alliance (see Figure 1, above). All analyses within Aim 3 were also run using combined (averaged) SM and partner report on all variables except for child behavioral difficulties (which I only have partner report on). The model testing the interaction effect of PTSD symptoms and PTSD couple alliance predicting parenting alliance was run first to assess initial model fit in MPlus. Using the scaled scores for male and female reports of SM PTSD symptoms, PTSD couple alliance, parenting alliance, and interaction terms of PTSD symptoms and PTSD couple alliance, the interaction


31
effect was not significant. In other words, the effect of PTSD on parenting alliance does not depend on level of PTSD couple alliance ((3 = .06, SE = .15, p= .72). When looking at a plot of this relationship, it was clear that couples who were high in couple alliance were generally reporting low levels of SM PTSD symptoms, and this contributed to the inability to detect a moderation effect. Because the interaction was not significant, the iterative process led to a different proposed structural model (see below). Of note, however, was the fact that all loadings of scaled scores on their respective latent factors (e.g., PCL-SM and PCL-p on overall PTSD latent factor; ERPS-SM, ERPS-p, PCC-SM, PCC-p on overall PTSD couple alliance latent factor; PAI-SM and PAI-p on overall parenting alliance latent factor) were adequate, with estimates of .52 or above. Therefore, the measurement model of the latent constructs was sufficient to use in future, iterative models.
Multiple Mediation Model
My proposal stated that final implementation of the above model would be dependent upon issues such as the nature of the variables (e.g., distributions), tests of the measurement model (e.g., latent factors), and iterative tests of the structural model. Therefore, I shifted to a different structural model. I tested a multiple mediation model in MPlus (see Figure 2, below), wherein both parenting alliance and couple alliance were tested as mediators of the association between SM PTSD symptoms and child behavioral difficulties. This model adheres to the broad conceptual themes developed in the background, wherein higher PTSD symptoms may show negative associations with child behavior via lower levels of parenting and relationship functioning. Whereas I knew from Aim 1 that SM self-report of PTSD symptoms did not correlate with child behavioral difficulties, I wanted to adhere to the original conceptualization of the predictors all being latent, combined factors. This model was run six times, based on


variations of overall PTSD versus numbing symptoms only, and for all children versus boys versus girls. See Table 7 below, for a list of all path estimates among six different iterations of the final multiple mediation model.
32
Figure 2. Multiple Mediation Model
Notes: PCL-SM = SM self-report of PTSD Checklist-Military. PCL-p = partner perception of PTSD Checklist-Military.
ERPS-SM = SM perception ofpartner Emotional Responses to PTSD Symptoms scale.
ERPS-p = partner self-report of Emotional Responses to PTSD Symptoms scale.
PCC-SM = SM self-report of PTSD Couple Connection scale. PCC-p = partner self-report of PTSD Couple Connection scale.
Interpreting Model Fit. There are several model fit statistics to consider when deciding the fit of a model (Geiser, 2013; Hooper et al., 2008). The chi-square tests whether misfit of the data is significant (Hu & Bentler, 1999). A significant chi-square rejects the null hypothesis that the model fits the data; thus, a non-significant chi-square is ideal. However, chi-square is dependent on sample size and a large sample size is considered over-powered and will almost


33
always have a non-significant chi-square test. The CFI is the comparative fit index and indicates by how much the model fits better than an independence model in which variables are not related (Bentler, 1990). Values for this statistic range from 0-1, with values closer to 1 indicating good fit. A CFI greater than .95 is preferred. The RMSEA is a measure of approximate model fit and has a penalty for non-parsimonious models (Steiger, 1990). A RMSEA less than .05 is preferred. The SRMR is a standardized measure of the model residuals. A value less than .05 is preferred.
Model 1: Overall model with all PTSD symptoms. Fit indices show that the overall model did not fit perfectly (x2 = 106.20 [df= 21]; p = .00; RMSEA = .16; CFI = .82; SRMR = .10). To improve model fit, I allowed correlation between the errors of the two member reported measures within the couple alliance latent factor (SM reported PTSD couple connection and emotional responses to PTSD). Nonetheless, there was still a lot of misfit, likely due to misfit in the measurement model. Although the loadings of scaled scores on the respective latent factors was considered adequate, there was some misfit as the indicators were not highly correlated within the latent factor, particularly among female and male reports. When looking at path estimates of this overall multiple mediation model, PTSD symptoms did not have an overall effect on child behavioral difficulties ((3 = -.05; p = .68). Within the model, the only significant predictor of child behavioral difficulties was couple alliance ((3 = -.36; p < 01).
Models 2 & 3: To assess whether child gender affected the relationships within the model, I ran the same overall model using first only the couples who reported child behavioral difficulties on a boy child (N=79) and then using only the couples who reported child behavioral difficulties on a girl child (A=86). Fit indices again indicated that the model using couples of a boy child did not fit perfectly (x2= 64.64 \df= 21];p = .00; RMSEA = .16; CFI = .81; SRMR =
. 11). Fit estimates were very similar when compared to the entire sample. When looking at path


34
estimates of model, PTSD symptoms also did not have an overall effect on boy child behavioral difficulties (P = -.13; p = .50). The only significant predictor of boy child behavioral difficulties was couple alliance (P = -.48; p < 01). When using only couples who reported on girl child behavioral difficulties, fit indices again indicated that the model did not fit perfectly (x2 = 66.26 \df= 21]; p = .00; RMSEA = .16; CFI = .82; SRMR = .10). When looking at path estimates of model, PTSD symptoms also did not have an overall effect on girl child behavioral difficulties (P = -.04; p = .84). The only significant predictor of girl child behavioral difficulties was couple alliance (P = -.29; p < 01).
Model 4: Model with only emotional numbing PTSD symptoms. The emotional numbing symptoms of PTSD are the PTSD symptoms that have been most strongly associated with various family issues, such as relationship, parenting, and family distress (e.g., Galovski & Lyons, 2004; Wilson & Kurtz, 1997). Therefore, to assess whether isolating the X variable to just emotional numbing symptoms instead of overall PTSD symptoms would affect the relationships with the model, I ran the overall model using only SM and partner reported emotional numbing PTSD symptoms. Fit indices show that this model also did not fit perfectly (X2= 131.24 \df= 21];p = .00; RMSEA = .18; CFI = .79; SRMR = .11). Emotional numbing symptoms did not significantly predict child behavioral difficulties (P = -.09; p = .56). Again, the only significant predictor of child behavioral difficulties was couple alliance (P = -.38; p < 01). Both couple and parenting alliance had larger associations with emotional numbing as compared to overall PTSD symptoms (couple alliance with overall PTSD symptoms: P = -.54; p <.01 as compared to emotional numbing symptoms: P = -.71 \P< .01; parenting alliance with overall PTSD symptoms: P= -.34;p <.01 as compared to emotional numbing symptoms: P = -.46;p
<01).


35
Models 5 & 6: To assess whether child gender affected the relationships within the model using only emotional numbing symptoms, I ran this same model using first only the couples who reported child behavioral difficulties on a boy child (N=79) and then using only the couples who reported child behavioral difficulties on a girl child (N= 86). The model using only couples with a boy child (N=79) similarly showed not a perfect fit (x2 = 84.71 \df= 21]; p = .00; RMSEA = .19; CFI = .75; SRMR = .12). There was no relationship between emotional numbing symptoms and boy child behavioral difficulties ((3 = .02; p = .93). Couple alliance was no longer a significant predictor of boy child behavioral difficulties in this model, likely due to smaller sample size ((3 = -.39; p= .07). Finally, the model using only couples who report on girl child behavioral difficulties (N= 86) not surprisingly also did not fit perfectly fit (x2 = 74.99 \df= 21]; p = .00; RMSEA = .17; CFI = .80; SRMR= .12). The effect of emotional numbing symptoms on child behavior was not significant for girls (J3 = -.21; /? = .44). One difference between boys and girls in this model is found in the relationship between couple and parent alliance, in the model with girls this relationship is non-existent ((3 = .02; p = .94), whereas there is a significant association in the model with boys ((3 = .39; p < .02).
Table 7, Path Estimates for Multiple Mediation Models
Model PTSD - Couple Alliance PTSD - Parenting Alliance Couple Alliance -Parenting Alliance PTSD -Child behavioral difficulties Couple Alliance -Child behavioral difficulties Parenting Alliance -Child behavioral difficulties
1. Overall PTSD -.54** _ 33** .37 -.05 . 36** -.05
2. Boys _ -.25 (p = .08) .55** -.13 _ 4g** -.00
3. Girls _ 57** . 51** .14 -.04 _ 29** -.09
4.Emotional Numbing _ 7t** _ 45** .21 (P = °75) -.09 _ 38** -.07


36
Table 7. cont’c
5. Boys _ 72** . 45** 20** .02 -.39 (p = .065) -.02
6. Girls 72** -.53** .02 -.21 _ 4Q** -.13
Note. ** = p < .05


37
CHAPTER IV DISCUSSION
Taking a systems perspective, service member (SM) PTSD symptoms after deployment can take a toll on family functioning. This study tested relationships among the variables of SM PTSD, couple alliance (including partner’s emotional responses to the SM as a result of his PTSD symptoms and overall couple communication and connection around PTSD symptoms), parenting alliance, and child behavioral difficulties. Results showed that, in general, higher levels of SM PTSD related to worse couple and parenting alliance. SM PTSD symptoms and child behavioral difficulties were associated on a bivariate level, but only when using the partner report of the SM PTSD symptoms. Thus, SM self-report of PTSD symptoms was not correlated with child behavioral difficulties, nor was SM PTSD related to child behavioral difficulties when combining these reports into a latent variable in a larger model. Similarly, partner reports of aspects of couple alliance and parenting alliance were both significantly correlated with child behavioral difficulties, but these associations were generally not found when using SM reports.
In a larger model combining both parents’ perceptions into a latent variable of couple alliance, there was a link between couple alliance and child behavioral difficulties; however, this association is carried by the partner report (based on the bivariate correlations). When evaluating these associations for PTSD numbing symptoms specifically (compared to overall PTSD symptoms), some of these associations increased in magnitude, but the overall pattern of results was generally the same. There are several important themes to discuss within these overall
results.


38
Basic Levels of Variables
Before discussing patterns of association, there were interesting differences in the basic levels of the variables as reported by the SM and the partner. When looking at general frequencies of the PTSD symptoms and couple variables, the SM reported more individual (PTSD symptoms) and relationship (less positive feelings from his partner towards him as a result of the PTSD symptoms, less connection related to the PTSD) distress as compared to his partner’s report on the same constructs. There are several potential explanations for this finding. One way to understand this is in the context of the SM’s PTSD symptoms. By nature of the internally experienced symptoms that make up many of the diagnostic criteria of the DSM-5 (American Psychiatric Association, 2013), including PTSD, it makes sense that even loved ones cannot fully grasp the extent of their SM’s PTSD symptoms. Moreover, the SM may be more sensitive to his partner’s lapses in connection and expression of positive emotions due to his higher individual distress; such distress may “spill over” and lead to a more negative perception of the relationship.
There were also interesting differences in the level of child behavioral difficulties in this sample as compared to national norms. Specifically, the children in the current sample, as reported by the partner, had higher levels of behavioral difficulties compared to the norm SDQ sample in the United States. Whereas we know that military children exhibit a great deal of resiliency (Lincoln & Sweeten, 2011), we also know that children of veterans with PTSD on average exhibit greater behavior problems than children of non-veteran parents without PTSD (Ahmadzadeh & Malekian, 2004), and that parental deployment is associated with an increase in child emotional and behavioral symptoms (White, de Burgh, Fear & Iversen, 2011). This is


39
consistent with the higher level of child distress in this current sample of families who have experienced at least one deployment and were selected based on elevated PTSD symptoms. Salience of Couple Alliance
As noted above, PTSD symptoms were generally associated with the couple variables (parenting alliance, partner’s emotional responses to the SM as a result of his PTSD symptoms, couple communication and connection around PTSD symptoms) in the expected direction, using both SM and partner report. Yet, the connections between PTSD symptoms and these couple variables with child behavioral difficulties were inconsistent and depended largely on whether it was the SM or partner report. However, couple alliance had a significant association with PTSD symptoms (whereas parenting alliance did not), and couple alliance also emerged as the only significant predictor of child behavioral difficulties within the final models. This is interesting because couple alliance does not directly assess interactions with children or parenting, but rather it is related to the couple’s responses to the SM PTSD symptoms. Thus, symptoms of PTSD in of itself may not be directly related to child behavioral difficulties, nor is the couple’s ability to be on the same page for parenting. Instead, the most important variable related to child behavioral difficulties may be the couples’ negative emotional reactions and lack of connection related to the PTSD symptoms. If this possibility, based on these cross-sectional findings, held in studies which better evaluate effects on children of changing the couples’ connection regarding PTSD, then focusing on such dynamics in assessment and treatment of family functioning at post deployment may be useful when addressing child behavioral difficulties.
Comparisons of Overall PTSD Symptoms with Emotional Numbing Symptoms
There were several analyses that focused on subscales of the various study variables. I was especially interested in the emotional numbing symptoms subscale of the PCL, as this


40
constellation of symptoms in SMs has been shown to be most strongly associated with various family issues, such as relationship, parenting, and family distress (e.g., Galovski & Lyons, 2004; Wilson & Kurtz, 1997). When comparing associations of overall PTSD symptoms versus only emotional numbing symptoms, the magnitude of association increased with both partner and SM report of partner emotional response to SM PTSD symptoms, and partner report of parenting alliance. The remainder of associations did not statistically increase in magnitude, including child behavioral difficulties. Isolating to emotional numbing symptoms also did not statistically change the association of PTSD symptoms with child behavioral difficulties within the final multiple mediation models. Emotional numbing symptoms were, however, more strongly associated with the couple variables (couple alliance, parenting alliance) within the final models than were overall PTSD symptoms, similar to the pattern of associations of the basic correlations. Thus, if similar results emerged in longitudinal and interventional studies, this may suggest that emotional numbing symptoms negatively impact the couple. Symptoms of emotional numbing decrease emotional expression and intimate exchanges in a relationship (Riggs et al., 1998), and we see that the couples in our sample do in fact report lower levels in partner’s positive emotional responses (using both SM and partner report) and parenting alliance (using only partner report) in the context of elevated symptoms of feeling distant or estranged from others and having difficulty experiencing or having positive feelings (American Psychiatric Association, 2013).
Emotional numbing also emerged as a salient factor in a closer evaluation of which SDQ subscales were correlated with which SM PCL subscales. In general, there was not a clear differentiation between internalizing and externalizing patterns among SM and child symptoms as hypothesized. Instead, there were certain subscales of both the PCL and SDQ that emerged as


41
most frequently associated. Emotional and Conduct behaviors were the SDQ subscales that were most frequently associated with PTSD total and subscale variables. Emotional numbing SM PTSD symptoms, as reported by the partner, were associated with the most child difficulty subscales (Emotional, Conduct, Peer Relations). Ruscio et al. (2002) suggests that emotional numbing symptoms may impact a SM’s ability to engage in normal interactions with a child that are necessary for developing a meaningful parent-child relationship. If this cross-sectional finding held in studies designed to focus and intervene on such constructs with both the SM and child, this may suggest that this same concept may extend to the child’s ability to develop meaningful peer relations as well. If this is true, then when assessing a family post deployment it may be especially important to focus on SM emotional avoidance symptoms and child emotional symptoms, conduct behaviors, and peer relations.
Reporter Method Variance
A noteworthy theme that emerged across results of this study is reporter method variance. Most notably, SM PTSD symptoms predicted child behavioral difficulties (as reported by the partner), but only when using partner perception, and not SM self-report, of the PTSD symptoms. This pattern was also noted when looking at associations among subscales of PTSD and child behavioral difficulties. There are several potential explanations for this finding. First, it is possible that the partner’s perception of both her SM’s PTSD symptoms and her child’s behavioral difficulties may actually be partly a reflection of the partner’s own mental or emotional distress. There are several studies which show that the more emotionally impaired the mother, the greater the degree to which she perceives her child to have behavior problems (e.g., Najman et al., 2001; Youngstrom et al., 2000). However, not all studies are in agreement that maternal distress prevents mothers from providing un-biased reports of child behavior (e.g.,


42
Querido, Eyberg, & Boggs, 2001). This first explanation, that of distress “spillover” into reports on multiple constructs, would suggest that if the study had also included the SM reports of child difficulties, then there might have been significant associations between his report of his own PTSD symptoms and his report of child behavioral difficulties. Thus, although parental reports of child behavior/emotional problems have been shown to have a high level of concordance (Luoma, Koivisto & Tamminen, 2003), these issues of method variance attributable to reporter could have strongly influenced the pattern of associations among constructs.
Another possible explanation is that partners are more distressed/behaviorally affected to the degree that they perceive their SM having symptoms of PTSD, and that child behavioral difficulties are in fact generally associated with maternal distress. That is, the more that mothers are perceiving problems in their partner’s functioning, the more this may affect the child’s behavior through mechanisms such as changes in maternal parenting/responsiveness, increased maternal distress and transmission of such distress, or similar processes. The association between maternal mental illness/distress and child behavioral difficulties is consistently found across a variety of samples and contexts and reporters of child behavior (e.g., O'Connor, Monk, & Burke, 2016), including wives of SMs with PTSD in the post deployment context (Flake et al., 2009). Moreover, the link between mother-child emotional distress may also have a level of reciprocity, as child behavioral difficulties can also lead to maternal mental health impairment (Kingsbury, Clavarino, Mamun, Saiepour & Najman, 2017).
The Role of Child Gender
I had hypothesized that child gender would moderate the relationship between PTSD symptoms and child behavioral difficulties, such that the association between PTSD symptoms and externalizing behaviors would be stronger for boys than girls, and the association between


43
PTSD symptoms and internalizing behaviors would be stronger for girls than boys. However, this was not found. Although gender differences in emotional expression in children has been shown (Chaplin & Aldao, 2013), several studies note that such differences are varied, nuanced, not necessarily static, and could be strengthened or attenuated on various other variables, including age (Brody & Hall, 2008; Chaplin & Aldao, 2013; Brown, 1999). It is possible that if I had power to focus on certain subgroups of children by age and gender (e.g., adolescent girls; elementary aged boys) or could look at changes in behavior by gender over time, I may have found a significant interaction.
One interesting finding related to child gender within the final multiple mediation models was that in both main models (overall PTSD symptoms and emotional numbing PTSD symptoms), the association between couple and parenting alliance was significant for boys but not for girls. There are no obvious differences (e.g., age) between boys and girls that could help to explain this finding and further exploration into this area would entail additional analyses looking at potential differences among boys and girls in this sample.
Limitations
It is important to address limitations of the study. Perhaps the most important to highlight is the not perfect fit of the final model. Whereas fit indices are a useful guide, Hooper, Couglhan and Mullen (2008) argue that a structural equation model should also be examined in the context of substantive theory, and the rationale for the proposed model was strong in the current study. Nonetheless, the fit indices of the model deviating from published recommended thresholds is a clear limitation of my study and indicates that results of the model cannot be interpreted with full
confidence.


44
Moreover, constructs are gathered cross-sectionally, are not experimentally manipulated, are all self-report, and use only one reporter on child behavioral difficulties. This has limitations related to response bias, reporter method variance, and the inability to make any claims about causation or temporal relationships.
Further, it is likely that there are other measured or unmeasured variables that are not a part of the study that would help to better explain the relationship between a SMs’ PTSD symptoms, couple and parenting alliance, and child behavioral difficulties, including maternal distress or depression.
Lastly, I only studied American, English-speaking, male Army SMs and female civilian partners in this project. This does not allow me to generalize to female service members/male partners, gay/lesbian/transgender couples, dual military couples, non-English speaking military couples, or couples not in the military. Couple, family, and parent-child dynamics, as well as the understanding or expression of PTSD symptoms may be different in such samples, which may change the associations among the various study variables.
Clinical Implications
Despite the study’s limitations, there are several potential clinical implications to the degree that these cross sectional associations do in fact suggest the type of reciprocal causal relationships suggested by family systems theory. Within family systems theory, it is impossible to separate an individual’s functioning from the family’s functioning (Bowen, 1978). For families who are overwhelmed by the combination of the processes within the family system discussed throughout this study, such a narrative may help to identify potential intervention points within the family system. As mentioned above, by putting these processes together in a larger model, it may reinforce the idea that they are all interrelated and that modifying one part


45
of the family system could influence other parts of the family system. For example, if a mother is concerned about her child’s behavior, results from this study would suggest that helping the couple to find ways to connect more supportively or positively around the SM’s PTSD symptoms, as opposed to dynamics specifically related to parenting or child, may be the most effective in improving child outcomes.
This idea, that to improve the child’s behavior and well-being, interventions should focus on reducing parental conflict and disruption is not new (e.g., Carlson & Corcoran, 2001; Demo & Fine, 2010; Brown, 2010). However, the usual go-to empirically supported treatments targeting negative child behavioral difficulties typically include parent-child management or cognitive-behavioral therapy focused on the child (Society of Clinical Child & Adolescent Psychology, 2016) and often do not focus on the parental relationship itself. Results of this study are a good reminder that in such situations as described above, targeting the relationship and, more specifically, how the couple is coming together around the SM PTSD, may be a part of how to help the family system address the child’s distress. Bernet, Wamboldt, and Narrow (2016) discuss that a new condition in the DSM-5, “Child Affected by Parental Relationship Distress,” was intended to capture such a situation, “when the focus on clinical attention is the negative effects of parental relationship discord on a child in the family, including effects on the child’s mental or other medical disorders” (American Psychiatric Association, 2013). Whereas this diagnosis is often an after-thought, the authors consider how it covers a broad and common set of family systems’ issues often seen in treatment. This current study may provide insight into one complex situation, specific to military members but potentially applicable to any family system, by which this DSM-5 condition may be applicable to young patients and their parents and may help to inform treatment.


46
It is encouraging that there are interventions already developed for couples experiencing both PTSD and relationship distress, including targeting decreasing PTSD symptoms and increasing couple communication skills (Erbes et al., 2008; Monson et al., 2012; Blow et al., 2015). Results of this study would suggest that screening for and addressing SM emotional numbing symptoms may be particularly important within such couples’ interventions, given that within this study, this constellation of symptoms was most associated with couple difficulties. Future Directions
There are several next steps I could take from here. Using my current dataset, it may be helpful to further investigate the nature of problems that could differentiate between boys and girls. This could include looking at any differences between boys and girls on the subscale or item level of the SDQ. Another future direction could be including more maternal mental health and well-being variables within bivariate and model analyses to better understand how the partner’s distress is associated with the study variables. It may also be interesting to explore if any SM self-report individual functioning variable, including other mental or physical health variables or specific variables related to the military or deployment, may be associated with child outcomes as reported by the partner. Within the current dataset, I could also look at altering the models in an attempt to improve model fit and/or better explain the relationships among the variables. Improving model fit could involve using a simpler construct for the couple alliance variable with less indicator items or dropping parenting alliance from the model as a mediator because it was overall less associated with both PTSD and child outcomes as compared to couple alliance.
There are several limitations of the dataset that could only be addressed in a future study. Likely because the SDQ instructions asked respondents to report on their youngest or only child,


47
the child age range was skewed right; less than 10% of children in the study were aged 11-17. If I had a larger sample or had a sample with a more even spread of child age ranges, it is possible that I could have analyzed associations for boys and girls in different age groups and possibly seen significant results when comparing results for boys versus girls (Brody & Hall, 2008; Chaplin & Aldao, 2013). As discussed previously, it would be very helpful to have the service member’s report on child behavioral difficulties in order to better understand relationships between SM PTSD and child behavioral difficulties. Although the SDQ is a valid and well-known measure of child behavioral difficulties, it would also be helpful to have more than one measure of child behavioral difficulties to use in analyses. Specifically, more objective (i.e., observational) or third party (i.e., teacher self-report) data may provide a more comprehensive assessment of child functioning. Lastly, a future study should include a longitudinal assessment of family functioning, including data from both pre and post deployment to measure overall changes or stability in family functioning dynamics that would paint a fuller picture of how these constructs work within each family structure over time.


48
REFERENCES
Abidin, R. R. & Brunner, J. F. (1995) Development of a parenting alliance inventory.
Journal of Clinical Child & Adolescent Psychology, 24: 1, 31 - 40.
Ahmadzadeh G. & Malekian A. (2004) Aggression, anxiety, and social development in
adolescent children of war veterans with PTSD versus those of non-veterans. Journal of Research in Medical Sciences, 9(5) :231-234.
Allen, E. S., Rhoades, G. K., Stanley, S. M., & Markman, H. J. (2010). Hitting Home:
Relationships between Recent Deployment, Post Traumatic Stress Symptoms, and Marital Functioning for Army Couples. Journal of Family Psychology, 24(3), 280-88.
Allen, E., Stanley, S., Rhoades, G., Markman, H., and Loew, B. (November, 2012).
Reintegration with the spouse after deployment: Does talking about PTSD and combat help ? Talk presented at the meeting of the Association for Behavioral and Cognitive Therapies Convention, National Harbor, MD.
Allen, E. S., Renshaw, K. D., & Fredman, S. J (2016, October). Mothers’ Parental Buffering in the Context of Husband PTSD. In S. J. Fredman (Chair), Partnering and parenting in the presence of PTSD. Symposium presented at the annual meeting of the Association for Behavioral and Cognitive Therapies, New York, NY.
American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Arlington, VA: American Psychiatric Publishing.
Bearss, K. E. & Eyberg, S. (1998) A Test of the Parenting Alliance Theory. Early Education and Development, 9(2): 179-185, DOI: 10.1207/sl5566935eed0902_5
Bernet, W., Wamboldt, M. Z., & Narrow, W. E. (2016). Child Affected by Parental
Relationship Distress. Journal of the American Academy of Child & Adolescent Psychiatry, 55(7): 571-579.
Bentler, P. M. (1990). Comparative fit indexes in structural models. Psychological Bulletin, 107, 238 -246.
Bliese, P. D., Wright, K. M., Adler, A. B., Cabrera, O., Castrol, C. A., & Hoge, C. W.
(2008). Validating the Primary Care Posttraumatic Stress Disorder screen and the Posttraumatic Stress Disorder Checklist with soldiers returning from combat. Journal of Consulting and Clinical Psychology, 76, 272-281. doi: 10.1037/0022-006X.76.2.272
Blow, A. J., Curtis, A. F., Wittenborn, A. K., & Gorman, L. (2015). Relationship
Problems and Military Related PTSD: The Case for Using Emotionally Focused Therapy for Couples. Contemporary Family Therapy, 37(3), 261-270.
DOI: 10.1007/s 10591-015-9345-7


49
Bonds T. M., Baiocchi D., & McDonald L. L. (2010) Documented briefing: Army
Deployments to OEF and OIF. Santa Monica, CA: Arroyo RAND Corporation. Retrieved from:
http ://www.rand. org/content/dam/rand/pubs/documented_briefings/2010/RAND_DB 587.pdf.
Bowen, M. (1978) Family therapy in clinical practice. New York: Aronson
Brody, L.R. (1999). Gender, emotion, and the family. Cambridge, MA: Harvard University Press.
Brody, L.R., Hall, J.A. (2008). Gender and emotion in context. In: Lewis, M., Haviland-Jones, J.M., Barrett, L.F. (eds). Handbook of emotions. 3rd The Guilford Press;
New York, NY: 2008. pp. 395-408.
Brown, L.M. (1999). Raising their voices: The politics of girls’ anger. Cambridge, MA: Harvard University Press. Retrieved from EBSCOhost.
Brown, S. L. (2010). Marriage and Child Well-Being: Research and Policy Perspectives. Journal of Marriage and Family Therapy, 72(5): 1059-1077.
Carlson, M.J. & Corcoran, M.E. (2001). Family structure and children's behavioral and cognitive outcomes. Journal of Marriage and Family, 63:779-792.
Chaplin, T. M. & Aldao, A. (2013). Gender Differences in Emotion Expression in Children:
A Meta-Analytic Review. Psychological Bulletin, 759(4): 735-765.
Chandra, A., Lara-Cinisomo, S., Jaycox, L. H., Tanielian, T., Burns, R. M., Ruder, T., &
Han, B. (2010). Children on the homefront: The experience of children from military families. Pediatrics, 725(1), 16-25.
Conger, R. D., Ge, X., Elder, G. H., Jr., Lorenz, F. O., & Simons, R. L. (1994). Economic stress, coercive family process, and developmental problems of adolescents. Child Development, 65, 541-561.
Creech, S. K., Hadley, W., & Borsari, B. (2014). The Impact of Military Deployment and Reintegration on Children and Parenting: A Systematic Review. Professional Psychology, Research and Practice, 45(6), 452-464. http://doi.org/10.1037/a0035Q55
Demo, D. M. & Fine, M,A. (2010). Beyond the average divorce. Sage; Thousand Oaks, CA.
Department ofDefense (2015). 2015 Demographics: Profile of the Military Community.
Washington, DC: Department ofDefense, Office of the Deputy Assistant Secretary of Defense for Military Community and Family Policy.


50
Erbes, C., Polusny, M., MacDermid, S., & Compton, J. (2008). Couple therapy with combat veterans and their partners. Journal of Clinical Psychology, 64(8), 972-983.
Flake, E. M., Davis, B.E., Johnson, P. L., & Middleton, L. S. (2009). The psychosocial
effects of deployment on military children. Journal of Developmental and Behavioral Pediatrics, 30, 271-278. doi: 10.1097/DBP.0b013e3181aac6e4
Fredman, S. J., Monson, C. M., & Adair, K. C. (2011). Implementing cognitive-behavioral conjoint therapy for PTSD with the newest generation of veterans and their partners. Cognitive and Behavioral Practice, 75(1), 120-130.
Galovski, T & Lyons, J. A. (2004). Psychological sequelae of combat violence: A review of the impact of PTSD on the veteran's family and possible interventions. Aggression and Violent Behavior, 9(5): 477-501.
Geiser C. (2013) Data Analysis with Mplus. NewYork: Guilford Press.
Gewirtz, A. EL, Polusny, M. A., DeGarmo, D., Khaylis, A., & Erbes, C. (2010).
Posttraumatic stress symptoms among National Guard soldiers deployed to Iraq: Associations with parenting behaviors and couple adjustment. Journal of Consulting and Clinical Psychology, 78, 599- 610. doi:10.1037/a0020571
Gewirtz, A. H., Pinna, K. L. M., Hanson, S. K. & Brockberg, D. (2014). Promoting Parenting to Support Reintegrating Military Families: After Deployment, Adaptive Parenting Tools. Psychological Services, 77(1): 31-40.
Goodman, R. (1997) The Strengths and Difficulties Questionnaire: A Research Note. Journal of Child Psychology and Psychiatry, 38: 581-586.
Goodman, A., Lamping, D.L. & Ploubidis, G.B. (2010). When to use broader internalising and externalising subscales instead of the hypothesised five subscales on the Strengths and Difficulties Questionnaire (SDQ): data from British parents, teachers and children. Journal of Abnormal Child Psychology, 35(8): 1179-91.
Hanley, K. E., Leifker, F. R., Blandon, A. Y., & Marshall, A. D. (2013). Gender
differences in the impact of posttraumatic stress disorder symptoms on community couples’ intimacy behaviors. Journal of Family Psychology, 27, 525-530. doi:10.1037/a0032890
Hayes, A. F. (2018). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach (2nd Ed.) New York: The Guilford Press.
Hoge, C. W., Auchterlonie, J. L., & Milliken, C. S. (2006). Mental health problems, use of mental health services, and attrition from military service after returning from deployment to Iraq or Afghanistan. Journal of the American Medical Association,
295, 1023-32.


51
Hooper, D., Coughlan, J. and Mullen, M. R. (2008). Structural Equation Modelling:
Guidelines for Determining Model Fit. The Electronic Journal of Business Research Methods, 6(1): 53 - 60. available online at www.ejbrm.com
Hu, L.T. and Bentler, P.M. (1999). Cutoff Criteria for Fit Indexes in Covariance Structure Analysis: Conventional Criteria Versus New Alternatives. Structural Equation Modeling, 6( 1), 1-55.
Jordan, B.K., Marmar, C.R., Fairbank, J.A., Schlenger, W.E., Kulka, R.A., Hough, R.L., &Weiss, D.S. (1992). Problems in families of male Vietnam veterans with post-traumatic stress disorder. Journal of Consulting and Clinical'Psychology, 60.916-926.
King, D. W., Leskin. G. A., King, L. A. and Weathers, F. W. (1998). Confirmatory factor
analysis of the Clinician-Administered PTSD Scale: Evidence for the dimensionality of posttraumatic stress disorder. Psychological Assessment. 10.90-96.
Kingsbury, A. M., Clavariono, A., Mamun, A., Saiepour, N. & Najman, J.M. (2017). Does having a difficult child lead to poor mental health? Public Health, 146: 46-55.
Kring, A.M. & Gordon, A.H. (1998). Sex differences in emotion: Expression, experience, and physiology. Journal of Personality and Social Psychology, 74:686-703.
[PubMed: 9523412],
Lee, I. A., & Preacher, K. J. (2013, September). Calculation for the test of the difference between two dependent correlations with one variable in common [Computer software]. Available from http://quantpsv.org.
Lester, P., Peterson, K., Reeves, J., Knauss, L., Glover, D., Mogil, C., Duan, N., Saltzman, W., Pynoos, R., Wilt, K. & Beardslee, W. (2010). The Long War and Parental Combat Deployment: Effects on Military Children and At-Home Spouses. Journal of the American Academy of Child and Adolescent Psychiatry, 49(4), 310-320.
Lincoln, A.J. & Sweeten, K. (2011). Considerations for the Effects of Military Deployment on Children and Families. Social Work in Health Care, 50(1 ):73-84, DOI: 10.1080/00981389.2010.513921
Luoma, I., Koivisto, A.M. & Tamminen, T. (2003). Fathers’ and mothers’ perceptions of their child and maternal depressive symptoms. Nordic Journal of Psychiatry, 55(3): 205-211. DOI: 10.1080/08039480410006299
Masten, A. S. (2001). Ordinary magic: Resilience processes in development.
American Psychologist, 56, 227-238. doi:10.1037/0003-066X.56.3.227
Monson, C., Taft, C. T., & Fredman, S. J. (2009). Military-related PTSD and Intimate Relationships: From Description to Theory-Driven Research and Intervention Development. Clinical Psychology Review, 29(8): 707-714.


52
Monson, C.M., Fredman, S.J., Adair, K.C., Stevens, S.P., Resick, P.A., Schnurr, P.P.,
Macdonald, A. (2011). Cognitive-behavioral conjoint therapy for PTSD: Pilot results from a community sample. Journal of Traumatic Stress, 24.91 101. doi: 10.1002/jts.20604.
Monson, C., Fredman, S., MacDonald, A., Pukay-Martin, N., Resick, P., & Schnurr, P. (2012). Effects of cognitive-behavioral couple therapy for PTSD: A randomized controlled trial. Journal of the American Medical Association, 308(7), 700-709.
Najman, J. M., Williams, G. M., Nikles, J., Spence, W., Bor, W., O’Callaghan, M., Le
Brocque, R., Anderson, M. J. & Shuttlewood, G. J. (2001). Bias Influencing Maternal Reports of Child Behaviour and Emotional State. Social Psychiatry and Psychiatric Epidemiology, 36 (4):186-194. DOI: 10.1007/s001270170062
Nelson-Goff, B. S, Reisbig, A. M., Bole, A., Scheer, T., Hayes, E., Archuleta, K. L., ... & Sanders-Hahs, E. (2006). The effects of trauma on intimate relationships: A qualitative study with clinical couples. American Journal of Orthopsychiatry, 76(4), 451.
O'Connor, T. G., Monk, C., & Burke, A. S. (2016). Maternal Affective Illness in the Perinatal Period and Child Development: Findings on Developmental Timing, Mechanisms, and Intervention. Current Psychiatry Reports, 18(3), 5. doi: 10.1007/s 11920-016-0660-y
Querido, J. G., Eyberg, S. M., & Boggs, S. R. (2001). Revisiting the accuracy hypothesis in families of young children with conduct problems. Journal of Clinical Child Psychology, 30(1), 253-261.
Ray, S. L., & Vanstone, M. (2009). The impact of PTSD on veterans' family relationships: an interpretative phenomenological inquiry. International Journal of Nursing Studies, 46(6), 838-847.
Renshaw, K.D., Rodrigues, C.S. & Jones, D. H. (2008). Psychological symptoms and marital satisfaction in spouses of Operation Iraqi Freedom veterans: Relationships with spouses' perceptions of veterans' experiences and symptoms. Journal of Family Psychology, 22(4): 586-594. doi.org/10.1037/0893-3200.22.3.586
Riggs, D.S., Byrne, C.A, Weathers, F.W, & Litz, B.T (1998). The quality of the intimate relationships of male Vietnam veterans: Problems associated with posttraumatic stress disorder. Journal of Traumatic Stress, 11: 87-101
Ruscio, A. M., Weathers, F. W., King, L. A. & King, D. W. (2002), Male war-zone veterans' perceived relationships with their children: The importance of emotional numbing. Journal of Traumatic Stress, 15: 351-357. doi:10.1023/A:1020125006371


53
Sayers, S.L. (2011). Family Reintegration Difficulties and Couples Therapy for Military Veterans and Their Spouses. Cognitive and Behavioral Practice, 18: 108-119.
Samper, R. E., Taft, C. T., King, D. W. & King, L. A. (2004). Posttraumatic stress disorder symptoms and parenting satisfaction among a national sample of male Vietnam veterans. Journal of Traumatic Stress, 17: 311-315.
Snyder, D. K., Klein, M. A., Gdowski, C. L., Faulstich, C., & LaCombe, J. (1988).
Generalized dysfunction in clinic and nonclinic families: A comparative analysis. Journal of Abnormal Child Psychology, 16, 97-109.
Society of Clinical Child & Adolescent Psychology (2016). Effective child therapy:
Evidence-based mental health treatment for children and adolescents. Retrieved January 8, 2017 from http://effectivechildtherapv.org/content/ebp-options-specific-dis orders.
Steiger, J. H. (1990). Structural model evaluation and modification. Multivariate Behavioral Research, 25, 214-12.
Weathers, F. W., Litz, B. T., Herman, D. S., Huska, J. A. & Keane, T. M. (1993). The PTSD checklist: reliability, validity, & diagnostic utility. Paper presented at the Annual Meeting of the International Society for Traumatic Stress Studies, San Antonio, TX.
Weathers, F. W., & Ford, J. (1996). Psychometric properties of the PTSD Checklist (PCL-C, PCL-S, PCL-M, PCL-PR). InB. H. Stamm (Ed.), Measurement of stress, trauma, and adaptation (pp. 250-252). Lutherville, MD: Sidran Foundation & Press.
White, C.J., de Burgh, T., Fear, N. T. & Inersen, A. C. (2011). The impact of deployment to Iraq or Afghanistan on military children: A review of the literature. International Review of Psychiatry,2 3(2): 210-217, DOI: 10,3109/09540261.2011.560143
Wilson, J. P., & Kurtz, R. R. (1997). Assessing posttraumatic stress disorder in couples and families. In J. P. Wilson & T. M. Keane (Eds.), Assessing psychological trauma and PTSD (pp. 349 - 372). New York: Guilford Press
Youngstrom, E., Loeber, R., & Stouthamer-Loeber, M. (2000). Patterns and correlates of agreement between parent, teacher, and male adolescent ratings of externalizing and internalizing problems. Journal of Consulting and Clinical Psychology, 68(6), 1038-1050.


54
APPENDIX A
Bivariate correlations with all variables: Boys and girls
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
1. PCL-SM .4 1 .9 2 .4 4 .8 7 .3 9 .8 3 .4 1 .9 2 .4 3 .5 0 .2 6 .4 0 .1 4 .2 2 .2 5 .0 0 .04 .0 3 .0 1 .0 0 .0 0
2. PCL-p - .4 5 .8 7 .3 9 .8 7 .3 7 .8 1 .4 5 .9 3 .2 8 .4 6 .0 8 .3 1 .1 0 .2 4 .1 9 .20 .2 1 .1 0 .1 5 .0 9
3. Reexper-VI - .4 7 .6 9 .3 2 .8 3 .4 3 .7 8 .3 8 .3 5 .1 8 .2 8 .1 1 .1 5 .1 9 .0 1 .04 .0 T .0 1 .0 4 .0 i
4. Reexper-p - .3 1 .5 9 .3 6 .7 3 .4 0 .7 1 .1 7 .2 2 .0 4 .1 2 .1 1 .0 7 .0 9 .12 9 â– i J 1 .0 1 | f
5. Emotion Numb-M - .4 0 .6 2 .2 9 .7 4 .3 4 .5 9 .3 1 .4 6 .1 1 .2 8 .2 8 .0 4 .0 6 .0 3 i .0 3
6. Emotion Numb-p - .2 7 .6 4 .3 5 .7 9 .3 5 .5 9 .1 2 .3 8 .1 5 .3 2 .2 1 s .2 0 .0 6 3 1 .1 1
7. Avoidance-M - .4 0 .6 7 .2 9 .3 8 .1 7 .3 7 .1 3 .1 7 .1 5 .0 8 â–  .0 7 .0 3 .1 I .0 1
8. Avoidance-P - .3 7 .6 6 .1 4 .4 0 .0 5 .2 9 .1 0 .1 5 .1 5 0 .1 7 .1 1 .0 7
9. Hyperarous al-M - .4 6 .4 3 .2 4 .3 3 .1 5 .2 0 .2 5 .0 1 .04 .0 f .0 1 .0 2 .0 |
10. Hyperarous al-p - .2 8 .4 1 .0 6 .3 0 .0 9 .2 5 .2 1 .21 3 | â– i 3 .1 4 .0 V
11. ERPS-SM - 3 3 .6 6 .3 1 .3 9 .2 9 .0 3 .03 .0 5 .0 8 .0 5 .0 7
12. ERPS-p - .3 4 .7 2 .1 8 .4 4 .3 6 .33 .3 4 .2 1 .2 8 .1 8
13. PCC-SM - | 2 .4 1 .2 1 .0 3 .08 .0 6 .0 4 .0 2 .0 3
14. PCC-p - .1 6 .3 4 .3 3 .29 .2 8 .2 2 .2 1 .2 5
15. PAI-SM - .4 I .0 3 .07 .0 6 .0 1 .0 3 .0 7
16. PAI-p - .2 1 .12 .1 1 .1 2 .1 4 .1 6
17. SDQ Total - .71 .7 1 .7 4 .6 8 .6 2
18. SDQ Emotional - .9 8 .3 4 .5 4 .2 1
19. SDQ Conduct - .3 4 .5 2 .2 1
20. SDQ Hyperactiv ity - .2 5 .3 1
21. SDQ Peer Relations - .3 0
22. SDQ Pro Social -
Note. Bolded Pearson’s correlation signifies a statistically significant association (p< .05). Gray shaded numbers signify the association between partner and SM report on the same construct. Blue shaded numbers signify hypothesized associations among PCL and SDQ subscales.
PCL-SM = PTSD Checklist-Military. PCL-p = partner report ofPTSD Checklist-Military.
ERPS-SM = SM perception of partner Emotional Responses to PTSD Symptoms scale.
ERPS-p = partner self-report of Emotional Responses to PTSD Symptoms scale.
PCC-SM = male report ofPTSD Couple Connection scale.
PCC-p = partner report ofPTSD Couple Connection scale.
SDQ = (partner report of) Strengths and Difficulties Questionnaire


Full Text

PAGE 1

POSTDEPLOYMENT FAMILY FUNCTIONING: SERVICE MEMBER PTSD SYMPTOMS , COUPLE AND PARENTING ALLIANCE, AND CHILD BEHAVIORAL DIFFICULTIES by JESSICA J. KENNY B.A., Pepperdine University, 2010 M.A., University of Denver, 2013 M.A., University of Colorado Denver, 2016 A dissertation submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Doctor of Philosophy Clinical Health Psychology Program 2019

PAGE 2

ii This dissertation for the Doctor of Philosophy degree by Jessica J. Kenny has been approved for the Clinical Health Psychology Program by Elizabeth S. Allen, Chair Krista Ranby Kevin Everhart Keith Renshaw Date: August 3 , 2019

PAGE 3

iii Kenny, Jessica J. (PhD, Clinical Health Psychology Program) Postdeployment Family Functioning: Service Member PTSD Symptoms, Couple and Parenting Alliance, and Child Behavioral Difficulties Dissertation directed by Associate Professor Elizabeth Allen ABSTRACT As service members (SM s ) and veterans are returning home, increasing research efforts are being devoted to assessing postdeployment family functioning. The association between PTSD symptoms and fam ily functioning h as been one of the most studied issues for returning SM s (e.g., Creech, Hadley, & Borsari, 2014; Sayers, 2011), and yet several questions related to these connections remain unanswered. This study explored associations between levels of postdeployment PTSD symptoms for male SM s and various aspects of family functio ning including couple alliance around SM PTSD symptoms , parenting alliance, and child behavioral difficulties . Participants in the current study ( N = 165 couples) are part of the Relationships Among Military Personnel (RAMP) study, a longitudinal study aimed at investigating the interpersonal and mental health of Army couples following deployment. Using a Structural Equation Model ing (SEM) framework, the study evaluated the degree to which (1) male SM PTSD symptom severity is associated with child behavioral difficulties (as reported by the partner) and (2) couple and/or parenting all iance mediate that relationship. In addition, these same associations were evaluated using only the emot ional numbing symptoms of PTSD and using child gender as a moderator. Results showed that SM PTSD symptoms and child behavioral difficulties were assoc iated, but only when using the partner report of the SM PTSD symptoms and not when using the SM self report of PTSD symptoms. In a larger model

PAGE 4

iv PTSD and child behavior al difficulties and therefore there was no direct effect to be mediated . Couple alliance, as compared to parenting alliance, was most associated with both PTSD symptoms and child behavioral difficulties . The presence of PTSD emotional numbing symptoms (as compared to overall PTSD symptoms) was negatively a ssociated with some couple variables but not child behavioral difficulties . Child gender did not moderate the relationship between SM PTSD and child behavioral difficulties. Outside of these hypothesized aims, several other noteworthy results emerged. Limitations of the study are explored and further implications for both clinical work and future research is discussed. The form and content of this abstract are approved. I recommend its publication. Approved: Elizabeth Allen

PAGE 5

v ACKNOWLEDGEMENTS This research is made possible by a research grant that was awarded and administered by the U.S. Army Medical Research & Materiel Command (USAMRMC) and the Telemedicine & Advanced Technology Research Center (TATRC), at Fort Detrick, MD, under C ontract Number W81XWH 12 1 0090 . Opinions, interpretations, conclusions, and recommendations are those of the authors and are not necessarily endorsed by the Department of Defense.

PAGE 6

vi TABLE OF CONTENTS CHAPTER I. Aim Aim 2. Aim 3 ... II. .1 2 2 3 Instruments. III. 20 Aim 2. Aim 3 IV. ..44 Impli . . Future .. . APPENDIX ... 55

PAGE 7

1 CHAPTER I INTRO DUCTION The number of United States service members (SM s ) and veterans who have been deployed to war zones has increased dramatically over the past decade (Bonds, Baiocchi, & McDonald, 2010). Deployment to war zones involves a number of significant stressors for United States SMs . When considered from a family s ystems perspective, the effect of deployment on a military family can also be significant, as more than half (52.4%) of active duty SM s are married and approximately 42% have children (Department of Defense, 2015). Even the most resilient of families can e xperience significant stress in the postdeployment reintegration period (Sayers, 2011). Although reintegration after deployment presents challenges for any military family, certain vulnerabilities can contribute to an especially difficult and distressing time after deployment for families . In particular, the association of PTSD symptoms and family functioning for returning SM s has been a focus of empirical attention (e.g., Creech, Hadley, & Borsari, 2014; Sayers, 2011). The percentage of SM s returning fro m Iraq and Afghanistan who show elevated levels of PTSD symptoms is estimated at 10 18% (Hoge et al., 2006). SM s and veterans with high levels of PTSD symptoms report more numerous and severe relationship problems, and divorce at a higher rate, as compared to their veteran counterparts without PTSD symptoms (Monson et al., 2009). Additionally, high levels of SM and veteran PTSD symptoms are often accompanied by parenting stress and child behavioral difficulties (Creech et al., 2014; Flake, Davis, Johnson, & Middleton, 2009). For example, children of veterans with PTSD exhibit greater behavior problems than children of non veteran parents without PTSD (Ahmadzadeh & Malekian, 2004). These issues can also be compounded; for example, non service member

PAGE 8

2 (partner) parenting stress at postdeployment, when accompanied by high levels of SM PTSD, predicts poorer psychological functioning among military children (Flake et al., 2009). Thus, PTSD symptoms can make an already difficult time after deployment even more diffi cult for military families. This highlights the need to increase understanding of how PTSD symptoms and various aspects of family functioning interact for military families . To this en d, this study will address three dynamics within the family system that may be associated with increased levels of SM PTSD symptoms. Dynamics related to the couple include (1) lower levels of couple alliance regarding the SM PTSD symptoms, as reported by both the SM and partner, includin g ( 1 a) lower levels of partner expressed positive emotional responses to the SM regarding SM PTSD symptoms and ( 1 b) lower levels of couple connection around the SM PTSD symptoms . Dynamics related to parenting , as reported by both the SM and partner, includ e (2 ) lower levels of parenting alliance. Dynamics re lated to the children , as reported by the partner, include (3 ) child behavioral difficulties . How these three family dynamics and SM PTSD symptoms are associated has not previously been examined in comb ination . In addition to associations between overall PTSD symptoms and family issues, some researchers have focused on how the specific symptom s of PTSD relate to family issues. Various factor analyses and conceptual categorizations have grouped specific symptoms into different clusters. King, Leskin, King, & Weathers (1998), using factor analysis, grouped self report PTSD symptoms into four clusters: re experiencing, avoidance, e motional numbing, hyperarousal . These clusters map onto the different diagnostic criteria groups of PTSD within the DSM IV, with both avoidance and emotional numbing falling under Criterion C of the DSM IV (American Psychiatric Association, 2013) . D iagnost ic criteria for PTSD within the DSM 5 has changed somewhat , b ut each of these types of symptoms are still included in the new criteria.

PAGE 9

3 These King et al. clusters have subsequently been evaluated in relation to various family issues ( Samper, Taft, King , & King, 2004 ; Ruscio, Weathers, King , & King, 2002), with studies demonstrating that the emotional numbing symptoms of PTSD in SM s are the most strongly associated with family issues such as relationship, parenting, and family distress (e.g., Galovski & Lyon s, 2004; Wilson & Kurtz, 1997). The specific symptoms that make up the emotional numbing symptoms are: a loss of interest in important, once positive activities; feeling distant or estranged from others; and difficulties experiencing positive feelings, su ch as happiness or love (American Psychiatric Association, 2013). Symptoms of emotional numbing often inhibit emotional expression, which is integral to healthy and well functioning relationships (Riggs, Byrne, Weathers, & Litz, 1998). Thus, instead of emo tionally expressing themselves to their family, the SM with emotional numbing symptoms is more likely to withdraw, which can inhibit healing and preclude successful reintegration with their family after deployment (Ray & Vanstone, 2009). Consequently, emotional numbing symptoms may be even more strongly associated with the three family system dynamics, listed above, than overall PTSD symptoms. The purpose of this study is to look at how SM PTSD symptoms after deployment may be associated with three dynamics within the family system, as listed above. Specifically, this study will examine the relationship betw een SM PTSD and lower levels of couple alliance regarding SM PTSD symptoms, lower levels of parenting alliance, and child behavioral difficulties . This study will also examine if these relationships differ by child gender. Lastly, this study will investiga te whether emotional numbing, as compared to overall PTSD symptoms, is more strongly associated with these family dynamic variables. The three family dynamics included in this study will be discussed in more detail below.

PAGE 10

4 SM PTSD and Parenting Alliance In military couples, parental roles and practices is one of the basic family dynamics that needs to be reestablished or renegotiated after deployment (Sayers, 2011). Parental roles and practices is also one of the family dynamics that may be aff ected by PTSD symptoms. Increases in PTSD symptoms post deployment have been found to be significantly associated with impairments in SM parenting (Gewirtz et al., 2010). Therefore, w hen a SM experiences PTSD symptoms, one corresponding family systems dynamic may be decreased parenting alliance. Parenting alliance refers to the ways in which spouses provide support and show respect for each other in parental roles, and how they work together as a co paren ting team (Abidin & Brunner, 1995). Not surprisingly, low parenting alliance is often associated with marital distress (Abidin & Brunner, 1995). In fact, Allen et al. (2010) found that SM PTSD symptoms were negatively associated with parenting a lliance amo ng military couples. As discussed above, the numbing symptoms of PTSD are associated with parenting dynamics in military families, including a decrease in perceived quality of father child relationship (Ruscio, Weathers, King , & King, 2002) and decreased parenting satisfaction (Samper, Taft, King , & King, 2004). O ne explanation for these associations may be that when a SM has emotional numbing PTSD symptoms, the partner s may feel particularly emotionally distant, finding it hard to trust one another , and not in agreement on several areas of their relationship, including their parenting alliance. Thus, the SM may be more strongly negatively associated with parenting alliance, as compared to overall PTSD symptoms.

PAGE 11

5 SM PTSD and P TSD Couple Alliance Another possible family systems dynamic is a low couple alliance regarding the SM PTSD symptoms , or a low level of the couple coming together with regard to the SM PTSD symptoms. In the current study, couple alli ance around the PTSD sy mptoms i s posited to be comprised of two related constructs: (1) as a result of his PTSD symptoms , including both positive/supportive emotions such as caring, compassionate, and connected, and negative/non supportive emotions such as angry, resentful, and detached , and (2) the connection regarding the PTSD symptoms , including the effects of the PTSD under standing/support. There is a consistent link between SM PTSD symptoms and more numerous and severe romantic relationship problems (Galovski & Lyons, 2004; Monson, Taft , & Fredman, 2009). For example, w hen a SM or veteran has clinical levels of PTSD sympt oms, as compared to SM s or veterans without clinical levels of PTSD, they are more likely to withdraw more from their partner during important conversations (Galovski & Lyons, 2004) and have more emotional intimacy difficulties, including being less self d isclosing and expressive with their partners (Riggs et al., 1998). Therefore , greater levels of SM PTSD symptoms are likely associated with lower couple alliance regarding the PTSD symptoms . Of the specific types of PTSD symptoms, e motional numbing PTSD symptoms may be particularly associated with lower levels of PTSD couple alliance. By definition, emotional numbing PTSD symptoms particularly affect the SM feel love towards others, which therefore often creates distance, confusio n, and hurt feelings . This supports Riggs and colleagues (1998), who state that symptoms of emotional numbing decrease emotional

PAGE 12

6 expression and intimate exchanges in a relationship. Thus, it is likely that emotional numbing symptoms, as compared to overall PTSD symptoms, may be even more strongly associated with a low couple alliance regarding the PTSD symptoms . SM PTSD and Child Behavioral D ifficulties When a SM returns home with PTSD symptoms, their children are also affected. As noted above, SM PTSD and child behavioral difficulties are associated (Ahmadzadeh & Malekian, 2004; Jordan et al., 1992). Ruscio et al. (2002) suggest that the SM of emo tional numbing and withdrawal may directly impact a SM decreasing his ability to engage in normal interactions with the child that are necessary for developing a meaningful parent child relationship, thus increasing the likelihood fo r child behavioral difficulties . Therefore, emotional numbing symptoms may be more strongly associated with child behavioral difficulties , as compared to overall PTSD symptoms. SM PTSD, Parenting Alliance, PTSD Couple Alliance, Child Behavioral D ifficulti es It is likely that the above constructs interact and occur toge ther within the family system. There are established connections between SM PTSD and marital distress (Riggs et al., 1998), marital distress and parenting alliance (Abidin & Brunner, 1995), and child behavioral difficulties and caregiver (non service member parent) distress during the postdeployment period (Chandra et al., 2010; Flake et al., 2009; Lester et al., 2010). Drawing from general (i.e., non military specific) research, i mpaired pa renting has long been recognized as a crucial detriment to (e.g., Masten, 2001; Gewirtz et al., 2014). As reviewed in Gewirtz et al. (2014), family stress mode ls (e.g., Conger, Ge, Elder, Lorenz , & Simons, 1994) would suggest that the most stable

PAGE 13

7 and/or emotional difficulties in response to stress. In fact, low parenting alliance predicts more child behavioral difficulties ( Bearss & Eyberg, 1998). Thus, it is likely that a low reported parenting alliance may help to explain some of the relationship between SM PTSD symptoms and child behavioral difficulties . Moreover, Snyd problems. Thus, in the current study, it is likely that there is an association between low parenting alliance and child behavioral difficulties . Moreover, o ne can imagine how lower PTSD couple alliance may be related to lower parenting alliance due to the negative emotions, Finally, i t is possible that the negative association between PTSD and parenting alliance will be stronger w hen coupl e alliance regarding the PTSD symptoms is lower. That is, when the couple has lower levels of connection, support, and communication regarding the PTSD symptoms, it may fur ther undermine a sense of teamwork and exacerbate the negative association of SM PTSD symptoms with parenting alliance . Child Behavioral Difficulties by Child Gender There is a host of research which details gender differences in emotional expression and the proclivity for children to develop internalizing versus externalizing symptoms when stressed. There is also a rich theoretical literature base which offers social, psychological, and biological explanation s for such gender differences (e.g. Brody, 199 9; Kring & Gordon, 1998; Brod y & Hall, 2008; Chaplin & Aldao, 2013). In particular, gender role theory (Brody & Hall, 2008) proposes that Western cultures, girls are expected t o display greater levels of emotions than boys (in particular, internalizing emotions such as sadness, fear, anxiety, shame and guilt). In contrast, boys are

PAGE 14

8 expected to show less emotion and instead are expected or allowed to express themselves via extern alizing behaviors (such as anger, contempt, aggression, disgust). Therefore, Brody et al. would argue that the gender differences between internalizing and externalizing behaviors in children are a result of traditional societal expectations and roles for each gender , and the internalization of such roles . Chaplin and Aldao (2013) conducted a recent comprehensive meta analytic review of gender differences in emotional expression in children and did indeed find small but significant gender differences effect sizes; girls overall show more internalizing emotions ( g = . 10) than boys, and boys overall show more e xternalizing emotions ( g = .09) than girls. However, these effect sizes are small, and several other theoretical and research findings note that such d ifferences are varied, nuanced, and not necessarily static. For example, such differences could be strengthened or attenuated based on age, social and physiological development, and societal norms and influences (Brody & Hall, 2008; Chaplin & Aldao, 2013; Brown, 1999). Moreover , this is not a topic that has been extensively researched in the literature on military children. In fact, t he unique culture, values, and expectations of the military may also affect gender differences in emotional expression for children. Thus, in the current study, I will explore the relationships of SM PTSD symptoms and both internalizing and externalizing c hild behavior for boys and girls separately. Aim s of Current Study As reviewed above, PTSD symptoms may be associated with several different family systems dynamics , including a negative association with couple alliance around the SM PTSD symptoms and p arenting alliance, and a positive association with child behavioral difficulties . Although prior research has studied how some of these dynamics are associated with PTSD symptoms, none have examined them in combination. By examining them in combination, it

PAGE 15

9 may reinforce the idea that these processes are interrelated, and that intervening on one dynamic of the family system could alleviate distress in other parts of the family system (see the Clinical Implications section, below). Aim 1: To evaluate basic associations among all variables, including understanding which aspects of child behavioral difficulties (Hyperactivity, Emotional Symptoms, Conduct Probl ems, Peer Problems, lower Prosocial Behaviors ) are associated with which SM PTSD symptom clusters (Int rus ions, Avoidance, Numbing, Hyperarousal). With all these aims, I will be evaluating the child behavioral difficulties of the youngest or only child. Hypothesis 1 : (1 a ) Service member PTSD symptomology will be significantly and positively correlated with child behavioral difficulties and negatively correlated with partner expressed positive emotions towa rd the SM related to his PTSD symptoms, couple connection regarding S M PTSD symptoms , and parenting alliance . (1b) Emotional numbing symptoms will show a stronger association with the above variables compared to overall PTSD symptoms. (1c) T he PTSD clu sters of Avoidance and Numbing will correlate with internalizing child behavioral difficultie s (Emotional Symptoms and Peer Problems), and the PTSD clusters of Hyperarousal and Intrusions will correlate with externalizing child behavioral difficulties (Conduct, Hyperactivity, Prosocial). Aim 2: To understand if the relation ships between PTSD and child behavioral difficulties are moderated by gender.

PAGE 16

10 Hypothesis 2: Gender will moderate the relationship between PTSD symptoms and child behavioral difficulties , such that the association between PTSD symptoms and externalizing behaviors will be stronger for boys than girls, and the association between PTSD symptoms and internalizing behaviors will be stronger for girls than boys. The final aim of the study i s to test a model where (1) higher SM PTSD symptoms are associated with low er parenting alliance , (2) t his association will be moderated by PTSD couple alliance, such that the negative association between PTSD symptoms and parenting alliance is stronger wh en PTSD couple alliance is low , and (3) lower parenting alliance will be related to greater child behavioral difficulties . Aim 3: To evaluate a moderated mediation model wherein parenting alliance (M) mediat es the association between PTSD symptoms (X) and child behavioral difficulties (Y), and PTSD couple alliance (W) is a moderator of the association between PTS D and parenting alliance (see Model 1, below). Final implementation of the above model will be dependent upon issues such as the nature of the variables (e.g., distributions), tests of the measurement model (e.g., latent factors), and iterative tests of the structural model. Model fit will be compared for boys and girls . The model will also be estimated for both overall PTSD symptoms and only the emotional numbing PTSD symptoms. Each construct will be measured at timepoint four of the overall study in order for the model to represent a snapshot of what is happening for a family at a particular time during postdeployment. As these are measured cr oss sectionally and are not experimentally manipulated, it is acknowledged that the design does not provide any evidence of causation.

PAGE 17

11 Figure 1. Hypothesized Moderated Mediation Model PTSD Parenting Alliance PTSD Couple Alliance Child Behavioral Difficulties

PAGE 18

12 CHAPTER II METHODS Procedure Participants in the current study ( N = 165 couples) were participants who were part of the Relationships Among Military Personnel Study (RAMP), a longitudinal study aimed at investigating the interpersonal and mental health of Army couples following deployment. Recruitment for the study occurred between Ju ne 2013 and June 2014. RAMP utilized a variety of online recruitment methods to reach out to military couples. All recruitment materials directed interested individuals to the study website, which included information about the study, an online screening s urvey, and the informed consent document. Both members of the couple had to complete the screening to be considered for participation. Completed screening surveys were examined for the following eligibility criteria: (1) individuals had to be at least 18, English proficient, and able to pass informed consent comprehension questions, (2) individuals had to be a member of a dyad consisting of a male Army soldier (current or recent active duty) and civilian female with no prior military experience, (3) the mal e had to have returned from a deployment within the last 2 years, and (4) the couple had to be in a serious relationship for at least 1 year screening materials were c onducted to rule out duplicate or questionable cases. After this screening process, there were a total of 715 couples (1430 individuals) who were invited into the RAMP study. A total of 1,242 individuals participated in timepoint one, which served as a f urther screen for completeness, quality of responding, and level of PTSD symptoms. Based on this screen, couples were invited into a longitudinal sample, where they were given four additional surveys (i.e., timepoints two five), spaced approximately six months

PAGE 19

13 apart. Data for the current study were drawn from timepoint four. Five hundred and fifty three individual timepoint four surveys were sent and 510 individuals (241 matched couples) completed timepoint four. Couples for the current study were select ed if they met the following criteria: (1) couples who both have complete data at timepoint four and (2) couples who have a youngest or only child between the ages of 3 and 18, per maternal report, at timepoint four. This screening resulted in a final samp le of 165 couples. Sample At timepoint four, participants ( N = 165 couples) were aged 20 54, with an average age of 31.5 (SD = 5.42) for females and an average age of 32.68 (SD = 4.88) for males . Regarding marital status, 97% of the sample is marri ed. In terms of race and ethnicity, 81.7% of individuals identify as White, 9.0% as Hispanic, 4.2% as Black or African American, and 5.1% as other or multiracial. Regarding education, 56 % of males and 46% of females have some college but not a 4 year degree. Af ter selecting for couples who have a youngest or only child between the ages of 3 and 18 at timepoint four , the range of ages of youngest children was 3 18, with a mean of 6.78 (SD = 3.47). When splitting the sample by child gen der, boys and girls ages d o not deviate significantly ( boys : range = 3 17; mean = 6.4 6, SD = 3.0 0 ); girls : range = 3 17; mean = 6.78, SD = 3.47). Instruments (See Appendix B for complete measures) SM PTSD S ymptoms . To assess for SM PTSD symptoms, the PTSD Checklist Military Version (PCL M; Weathers et al., 1993) was given to male SM s . Consistent with the other variable names within this study, the PCL M will be referred to hereafter as the PCL SM, as it is a measure that was compl eted by the SM. The PCL SM is a self report scale comprised of 17 items that correspond

PAGE 20

14 to the key symptoms of PTSD in the DSM IV. The PCL SM is specific to PTSD caused by military experiences. Respondents indicate how much they have been bothered by a sym ptom over the past month using a 5 point scale from 1 (not at all) to 5 (extremely). The PCL SM is scored by summing the items for a total severity score. Allen et al. (2010) found this measure to be reliable in a similar sample ( = .94) and Weathers et al. (1996) found this to be a valid measure to assess PTSD symptoms in a military population. To assess for female partners perceptions of their SM the military version of the Postt raumatic Checklist was modified into the female partner perception of the PCL SM (called the PCL partner or PCL p) , which instructs on the various PCL SM items. A study by Renshaw, Rodrigues, and Jones (2008) sim ilarly adapted the PCL SM for spouses and the scale Both the SM and partner versions of the PCL SM were found to be reliable in the current RAMP sample ( = .96 and = .94, respectively). As noted in T able 1, the average PCL SM score was 40.48 (SD = 17.50) and the average PCL p score was 35.07 (SD = 14.46). Bliese et al. (2008) suggests a cut off of 34 for estimating a clinical di agnosis of PTSD. Thus, both the SM and partner report of average PCL score s represent risk for clinical levels of PTSD for a significant proportion of the sample. Couple Alliance (regarding SM PTSD symptoms) . regarding the SM PTSD symptoms, two measures each were given to both the SM and the partner.

PAGE 21

15 Partner Emotional Response to SM Regarding SM PTSD Symptoms. To assess th e partner s report of their own emotional responses to their SM as a result of his PTSD symptoms, the partner was given the Emotional Responses to PTSD Symptoms scale (called the ERPS p ) . This measure was given after the PCL p measure of PTSD symptoms, and the PCL p ) can have various effects on relationships. The instructions for the ERPS p ask partners to rate 14 different feelings, on a scale of 1 (not at all) to 7 (very much), on the extent that they have each emotion towards their SM PTSD symptoms. To assess for the SM perception them (SM ) as a result of their (SM) PTSD symptoms, the SM was given the perception of the Emotional Responses to PTSD Symptoms scale (called the ERPS SM ) . This ERPS SM was given after the PCL SM measure of PTSD symptoms. The instructions for the ERPS SM ask the SM s to rate 14 different feelings, on a scale of 1 (not at all ) to 7 (very much) on the extent that their partner has each emotion towards them (SMs) as a result of their PTSD symptoms. Thus, the scales given to the SMs and partners are both asking about Emotions include both positive and supportive emo tions (e.g., caring, compassionate, connected) and negative or non supportive emotions (e.g., angry, resentful, detached). This measure is scored by reverse scoring the negative or non supportive emotions and then taking the average of the fourteen items, with higher scores reflecting more positive emotions and alliance around the SM This

PAGE 22

16 scale showed internal consistency in this sample of both SM s ( ERPS SM ; = .90) and partners (ERPS p ; = .90). Couple Connection Regarding SM PTSD Symp toms . To assess perception of how well t he couple communicate s about and connect s around PTSD symptoms, the PTSD Couple Connection Scale (called the P CC SM for service members and P CC p for partners) was given to both partners ( Allen et al., 2012 ). The items in this scale are based on the the mes noted by Nelson Goff et al. (2006) on the impact of trauma on intimate relationships, such as increased or decreased communication, cohesion, and understanding/support. This scale is also given after the PCL SM and PCL p measure s of PTSD symptoms, and the instructions note that PCL SM ) can have various effects on relationships. The instructions then direct respondents t o endorse 11 items from 1 ( not at all true ) to 7 ( very true ) . The 11 items used for the current study are items that are exactly symmetrical between the male and female versions (the full scale includes 13 items for SMs and 15 items for partners). This measure is s cored by reverse scoring items that reflect poor connection and then averaged for a total score, with higher scores indicating greater connection and couple alliance around the PTSD symptoms. As noted above, the items are based on Nelson Goff et al., and f ocus on whether or not the couple r have brought us closer together and understanding/support (e.g., rtner understands what I am going through).

PAGE 23

17 Allen et al. (2012 ) tested a version of this measure in a sample of Army couples and found good internal consistency for both patients and partners, significant agreement between spouses, and strong evidence fo r criterion validity for the scale (i.e., significant positive associations with marital satisfaction and negative associations with reintegration adjustment after deployment). This measure also showed strong internal consistency for both SMs (P CC SM; = .91) and partners (P CC p; = .90) in the current sample. Parenting A lliance . To assess for parenting alliance, five items from the original twenty item Parenting Alliance Inventory (PAI; A bidin & Brunner, 1995) were given to both the SM and the part ner (called the PAI SM and PAI p, respectively) . As discussed in Allen et al. (2010), the PAI assesses the degree to which parents perceive themselves to be in a cooperative, communicative, and mutually respectful alliance for the care of their children on a scale of 1 (never) to 5 (always). Abidin and Brunner found high convergence between the PAI and measures of marital distress and parenting style. The five items used in this study were chosen on the basis of representativeness of teamwork between the p Allen et al. (2010) used this same five item PAI measure to measure parenting alliance in a separate military couple sample . The five chosen items all had factor loadings of at least .50 for both men and women in the original validation sample (Abidin & Brunner, 1995; Allen et al., 2010). This measure also showed strong internal consistency in the current RAMP sample f or both SMs (PAI SM; = .93) and partners (PAI p; = .96 ).

PAGE 24

18 Child B ehavior al Difficulties . To assess for child behavioral difficulties , the Strengths and Difficulties Questionnaire (SDQ; Goodman, 1997) was given to female partners with instructions to answer the questions is a 25 item measure of a range of children's attributes. The SDQ has 5 sc ales, each containing 5 items: (1) Hyperactivity Scale, (2) Emotional Symptoms Scale, (3) Conduct Problems Scale, (4) Peer Problems Scale, and (5) Prosocial Scale. The sum of the 25 items included in the 5 scales generate a total difficulties score (Goodma n, 1997) when the Prosocial Scale items are reverse scored to reflect child behavioral difficulties . Three different versions of the SDQ exist depending upon the age of the child. A slightly different version exists for children 3 years old, with 2 items reflecting oppositionality rather than antisocial behavior. Versions for 4 10 year olds and 11 17 ye ar olds are virtually identical, aside from some items being worded differently for the 11 17 year old version (Goodman, 1997). Responses require parents to report how true the item describing a behavior is to their child on a 3 point scale, from 1 (Not tr ue) to 3 (Certainly true). The SDQ has demonstrated concurrent validity with similar measures of child behavior. Additionally, the SDQ has established predictive validity by being able to discriminate between psychiatric and non psychiatric samples (Good man, 1997). In the current sample, reliability was adequate for the three age group versions (age 3 version = .77; age 4 10 version = . 85; age 11 17 version = . 89). Of note, the reliability of the age 3 version excludes one item due to zero variance, but this item was included in scoring. Due to the three versions of the measure given to partners based on the age of their youngest or only child , an overall reliability for the sample is not possible.

PAGE 25

19 There is theoretical and preliminary empirical support for combining the SDQ's hypothesized emotional and peer subscales into an 'internalizing' subscale and the hypothesized behavioral, hyperactivity, and reverse scored prosocial subscales into an 'extern alizing' subscale (Goodman, Lamping, Ploubidis, 2010). Within Aim 2, the internalizing and externalizing subscales are used.

PAGE 26

20 CHAPTER III RESULTS Prior to hypothesis testing, I first looked at descriptive and reliability statistics for all study variables . I also ran paired t tests and correlations for all construct s that had both SM and partner reports (e.g., PCL SM symptoms). Please see Table 1, below. As noted abo ve , all scales showed excellent internal consistency, with all alphas of .85 or above. The one exception is the SDQ items for children aged 3, which pr oduced an alpha of .77. T here was less variance in this age group (with scores ranging from 1 19 as comp ared to 0 33 and 2 31 on the SDQ scales for 4 10 year olds and 11 17 year olds, respectively), which may have contributed to the relatively lower alpha. Further, each correlation of couple reports was between r = .40 and .47, which offers evidence for crit erion validity of the reports of these constructs. Although each construct with both SM and partner report was significantly correlated, the t test results indicate that they are also all (except for the PAI SM and PAI p; p = .051) significantly different from one another in terms of absolute levels. That is, the positive correlation suggests that the couple reports generally move in the same direction, just with different levels. Table 1. N Mean (SD) Range Paired t test (effect size) Correlation PCL SM (17 items/sum) 164 40.17 (17.42) 17 85 .96 t = 3.94** ( d =.31) .47** ( N =164) PCL p (17 items/sum) 164 35.07 (14.46) 17 77 .94 PCL SM Re experiencing subscale (5 items/sum) 164 10.85 (5.76) 5 25 .95 t = 3.49** ( d = .27) .47** ( N =164) PCL p Re experiencing subscale (5 items/sum) 164 9.35 (4.74) 5 25 .92

PAGE 27

21 PCL SM Hyperarousal subscale (5 items/sum) 164 13.50 (5.77) 5 25 .88 t = 3.21** ( d = .25) .46** ( N =164) PCL p Hyperarousal subscale (5 items/sum) 164 12.08 (5.20) 5 24 .86 PCL SM Situational avoidance subscale (2 items/sum) 164 4.57 (2.60) 2 10 .89 t = 2.76** ( d = .22) .40** ( N =164) PCL p Situational avoidance subscale (2 items/sum) 164 4.00 (2.20) 2 10 .82 PCL SM Emotional numbing subscale (5 items/sum) 164 11.25 (5.36) 5 25 .87 t = 3.82** ( d = .30) .40** ( N =164) PCL p Emotional Numbing subscale (5 items/sum) 164 9.64 (4.41) 5 23 .84 ERPS SM (14 items/mean, higher scores reflect more alliance) 148 4.64 (1.22) 1.36 7 .90 t = 2.55* ( d = .21) .43** ( N =148) ERPS p (14 items/mean, higher scores reflect more alliance) 148 4.91 (1.24) 2 7 .91 PCC SM (11 items/mean, higher scores reflect more alliance) 148 4.06 (1.37) 1 7 .91 t = 2.89** ( d =.24) .42** ( N =148) PCC p (11 items/mean, higher scores reflect more alliance) 148 4.40 (1.28) 1.91 6.73 .90 PAI SM (5 items/mean, higher scores reflect more alliance) 160 5.80 (1.28) 1 7 .93 t = 1.96 (p = .051) ( d = .16) .45** ( N =160) PAI p (5 items/mean, higher scores reflect more alliance) 160 5.57 (1.52) 1 7 .96 SDQ Total (25 items/sum) 176 10.82 (6.99) 0 33 ---SDQ child age 3 ( a reliability excludes one item due to zero variance, but this item is included in scoring) 19 10.47 (4.90) 1 19 .77 a --SDQ child age 4 10 136 10.38 (6.81) 0 33 .85 --SDQ child age 11 17 21 14.00 (9.01) 2 31 .89 --SDQ Emotional Symptoms subscale (5 items) 176 2.12 (2.14) 0 9 ---SDQ Peer Problems subscale (5 items) 176 1.65 (1.70) 0 8 ---

PAGE 28

22 SDQ Conduct Problems subscale (5 items) 176 2.14 (2.15) 0 9 ---SDQ Hyperactivity subscale (5 items) 176 4.18 (2.83) 0 10 ---SDQ Prosocial subscale (5 items) 176 1.47 (1.70) 0 8 ---Note: PCL SM = PTSD Checklist Military. PCL p = partner report of PTSD Checklist Military. ERPS SM = SM perception of partner Emotional Responses to PTSD Symptoms scale. ERPS p = partner self report of Emotional Responses to PTSD Symptoms scale. PCC SM = male repo rt of PTSD couple connection scale. PCC p = partner report of PTSD couple connection scale. SDQ =(partner report of) Strengths and Difficulties Questionnaire There are several other main themes of Table 1 to note. Overall, all of the PCL total and subs cales from both SM and partner report are significantly correlated with one another, corresponding to what we would expect of such reports. One general theme that emerged was that males, on average, self reported higher symptom severity and relationship di stress than did their partners. This theme first emerged with SMs self reporting a higher severity of PTSD symptoms on the PCL SM ( M = 40.17, SD p ( M = 35.07, SD = 14.46) ( t [163] = 3.94; p > .05). This theme emerged on each PCL subscale as well (see Appendix A for the table that includes all of the PCL subscales). Further, the SM reported that the partner responds with less positive emotions towards him on the Emotional R esponses to PTSD scale (ERPS SM; M = 4.64, SD = 1.22) than she herself reported (ERPS p; M = 4.91, SD = 1.24) ( t [163] = 2.55, p > .05), with higher scores reflecting more positive emotions. The SM also reported that he perceived less couple connection related to his PTSD symptoms on the PTSD Couple Connection Scale ( PCC SM ; M = 4.06, SD = 1.28) as compared to her perception of their connection ( PCC p ; M = 4.40, SD = 1.28) ( t [163] = 2.89, p > .01), with higher scores reflecting more connection . Thus, overall, he is reporting more individual and relationship distress. However, parenting alliance did not evidence significant differences , with both the SM (PAI SM; M = 5.80, SD = 1.28) and the

PAGE 29

23 partner (PAI p; M = 5.57, SD = 1.52) reporting relatively equal levels of alliance as it relates to their parenting ( t [163] = 1.96, p > .05). Child Behavioral D ifficulties Compared with National Norms Although not an initial hypothesis, an important theme that emerged from these data was that the children in the RAMP sample, as reported by the partner, are generally more distressed as compared to the norm SDQ sample in the United States. The SDQ website ( http://www.sdqinfo.com/norms/USNorm1.pdf) provides norms for the SDQ within the United States for 4 17 year olds. I used only 4 17 year olds in the RAMP sample ( N = 156) to match the age range for the United States norms (note, excluding 3 year olds did not change the RAMP sample means, varying by no more than .1) As seen in Table 2, the RAMP sample was statistically higher on overall total difficulties than the nati onal norms. This trend remained true for every subscale except for the Prosocial Behavior subscale. Although the RAMP sample appears to be exhibiting a higher level of child behavioral difficulties overall, as compared to the national norms, it is helpful to ground the averages in the response scales. The SDQ is summed and responses are on a scale of 0 (not true), 1 (somewhat true) or 2 (certainly true). Thus, partners within my sample are rating their children, on average, as experiencing the stated diffic ulties over the past six months or school year on average as somewhere between not true and somewhat true. Table 2. SDQ score USA norms ( N =9878 4 17 year olds) M (SD) RAMP sample ( N =156 4 17 year olds) M (SD) One sample t test Total difficulties (25 items summed) 7.1 (5.7) 10.86 (7.22) t = 6.54** Emotional symptoms (5 items summed) 1.6 (1.8) 2.22 (2.20) t = 3.55**

PAGE 30

24 Conduct problems (5 items summed) 1.3 (1.6) 2.22 (2.20) t = 5.26** Hyperactivity (5 items summed) 2.8 (2.5) 4.11 (2.90) t = 5.69** Peer problems (5 items summed) 1.4 (1.5) 1.69 (1.7) t = 2.09* Prosocial behavior (reverse scored; 5 items summed) 1.4 (1.8) 1.45 (1.71) t = .34 Aim 1 My first aim was to evaluate basic associations among all v aria bles, including understanding which aspects of child behavioral difficulties (Hyperactivity, Emotional Symptoms, Conduct Problems, Peer Problems, lower Prosocial Behaviors) are associated with which SM PTSD symptoms clusters (Intrusions, Avoidance, Num bing, Hyperarousal). With all these aims, I evaluated the child behavioral difficulties of the youngest or only child. Please see Appendix A for a bivariate correlation table of all study measures and subscales. I will go into detail on data analyses and s peci fic hypotheses and provide pertinent bivariate correlation tables for Aim 1, below. Hypothesis 1a The first part of my hypothesis for Aim 1 was that SM PTSD symptomology will be significantly and positively correlated with child behavioral difficulties and negatively correlated with partner expressed positive emotions toward the SM related to his PTSD symptoms, couple connection regarding SM PTSD symptoms, and parenting alliance . I addressed this by running several using SM and partner report of the PTSD and couple variables, and partner report of the child behavioral difficulties variable. Please see Table 3 , below. Overall, this hypothesis was parti ally con firmed. For the most part, PTSD was significantly

PAGE 31

25 associated with the above couple variables (partner expressed positive emotions, couple connection , parenting alliance) in the expected direction, using both SM and partner report. The few exceptions to thi s finding were found in associations of SM report of various constructs with partner report of various constructs (i.e., constructs reported by the same reporter were all significant) . For example, the only non significant association of SM report of PTSD symptoms ( PCL SM ) with a couples measure was with the partner report of PTSD Couple Connection Scale ( PCC p ; r = .14 , ns ). Likewise, non significant associations with partner report of SM PTSD include SM report of the PTSD Couple Connection Scale ( PCC SM ; r = .08 , ns ) and parenting alliance (PAI SM; r = .10 , ns ). Lastly, SM PTSD symptoms were associated with overall child behavioral difficulties , but only when looking exclusively at partner report of SM PTSD symptoms (PCL p; r =.19, p < . 05 ). This association disappeared when looking exclusively at SM report of his own PTSD symptoms ( PCL SM ; r = .00, ns ). Likewise, partner report of various subsca les on the PCL p were correlated with various subscales of child behavioral difficulties on the SDQ, but SM self report on the PCL SM scales were not associated with any child difficulty subscales as reported by the partner on the SDQ. T able 3 . Bivariate correlations with all variables: Boys and girls 2 3 4 5 6 7 8 9 10 11 1. PCL SM .47 .87 .39 .50 .26 .40 .14 .22 .25 .00 2. PCL p -.39 .87 .28 .46 .08 .31 .10 .24 .19 3. Emotion Numb S M -.40 .59 .31 .46 .11 .28 .28 .04 4. Emotion Numb p -.35 .59 .12 .38 .15 .32 .21 5. ERPS SM -.43 .66 .31 .39 .29 .03 6. ERPS p -.34 .72 .18 .44 .36 7. PCC SM -.42 .41 .21 .03 8. PCC p -.16 .34 .33 9. PAI SM -.45 .03 10. PAI p -.21 11. SDQ Total -

PAGE 32

26 Note : p < .05). Gray shaded numbers signify the association between partner report of all constructs with partner report of child behavioral difficulties (SDQ Total) . PCL SM = PTSD Checklist Military. PCL p = partner report of PTSD Checklist Military. ERPS SM = SM perception of partner Emotional Responses to PTSD Symptoms scale. ERPS p = partner self report of Emotional Responses to PTSD Symptoms scale. PCC SM = male report of PTSD Couple Connection scale. PCC p = partner report of PTSD Couple Connection scale. SDQ = (partner report of) Strengths and Difficulties Questionnaire Hypothesis 1b Hypothesis 1b of Aim 1 was that emotional numbing PTSD sym ptoms would emerge as more strongly associated with the above variables as compared to overall PTSD symptoms. I r to z the test of the differences between dependent correlations (Lee & Preacher, 2013). See Table 4 for a visual display of these associations. Although simply looking at the associations between overall PTSD symptoms versus emotional numbing symptoms suggests that these associations do increase when looking at only emotional numbing symptoms, the overall hypothesis is only r to z to statistically test the difference in magnitudes between the two dep endent PTSD variables. Using partner report, there were two associations that did statistically increase in magnitude when going from overall PTSD symptoms to emotional numbing symptoms: her report of her emotional response to him regarding his PTSD sympto ms (from .46 to .59; ERPS p: z = 3.92, p < .00) and her report of parenting alliance (from .24 to .32; PAI p: z = 2.09, p < .05). The z test using her report of PTSD Couple Connection approached significance (from .31 to .38; PCC p : z = 1.87, p = .06 ). Using SM report, there was only one association that statistically increased in magnitude when going from overall PTSD symptoms to emotional numbing symptoms: his report of her emotional response to him regarding his PTSD symptoms (from . 50 to .59; ER PS SM: z = 2.74; p < .01). Finally,

PAGE 33

27 the associations with child behavioral difficulties did not statistically increase in magnitude with either SM or partner report of overall PTSD symptoms versus emotional numbing symptoms. T able 4 . Bivariate correlations : Comparing overall PTSD and emotional numb ing PTSD constructs ERPS SM ERPS p PCC SM PCC p PAI SM PAI p SDQ Total PCL SM .50 .26 .40 .14 .22 .25 .00 Emotion Numb S M .59 .31 .46 .11 .28 .28 .04 PCL p .28 .46 .08 .31 .10 .24 .19 Emotion Numb p .35 .59 .12 .38 .15 .32 .21 Note p < .05). Blue shading signifies associations that significantly increased when going from SM report of overall PTSD to emotional numbing symptoms. Pink shading signifies associations that significantly increased when going from partner report of SM overall PTSD to SM emotional numbing symptoms . Hypothesis 1c The third part of Aim 1 was related to understanding which SDQ scales or areas of child difficulty are correlated with which SM PTSD symptoms clusters (Please see Table 5). Hypothesis 1c of Aim 1 was that the PTSD symptoms of Avoidance and Numbing will cor relate with internalizing child behavioral difficulties (Emotional Symptoms and Peer Problems), and PTSD clusters of Hyperarousal and Intrusions will correlate with externalizing child behavioral difficulties (Conduct, Hyperactivity, Prosocial). I addresse d this by running Pearson correlations using SM and female report of PTSD total and subscale scores, and partner report of child behavioral difficulties total and subscale scores. See Table 5 for expected hypothesized associations. There were times when these hypothesized correlations emerged (Emotion Numb p with SDQ Emotional: r = .20, p < .05, and SDQ Peer Relations: r = .25, p < .05; Avoidance p with

PAGE 34

28 SDQ Emotional: r = .16, p < .05; and Hyperarousal p with SDQ Conduct: r = .21, p < .05). However, there were also significant correlations that did not follow the hypothesized pattern (Emotion Numb p with SDQ conduct: r = 20, p < .05; Avoidance p with SDQ Conduct: r = .17, p < .05; Hyperarousal p with SDQ Emotional: r = .21, p < .05). Thus, there was not a clear differentiation between internalizing and externalizing symptoms in both SM and children symptoms. p) being more associated with child behavioral difficulties , the sa me pattern emerged when looking at subscales as well. None of the PCL SM subscales emerged as associated with any child behavioral difficulties . The remainder of this paragraph discusses findings related to only partner report of SM PTSD symptoms. P artner report of SM reexperiencing symptoms was not associated with any child behavioral difficulties. However, the remainder of the PCL p subscales did associate with at least one child behavioral difficulty subscale. Emotional numbing SM PTSD symptoms were most associated with child behavioral difficulties overall. Hyperarousal and Avoidance SM PTSD symptoms were also associated with two out of five of the child behavioral difficulty subscales. Emotional and Conduct behaviors were most associated with PTSD total and some subscale variables. Overall, child Prosocial and Hyperactivity behavior was not associated with any PTSD total or subscale variable. Table 5. Bivariate correlations with PCL and SDQ: Boys and Girls SDQ Total SDQ Emotional SDQ Conduct SDQ Hyperactivity SDQ Peer Relations SDQ Pro Social PCL SM .00 .04 .03 .01 .00 .00 PCL p .19 .20 .20 .10 .15 .09 Reexper S M .01 .04 .03 .02 .04 .02 Reexper p .09 .12 .13 .05 .01 .06

PAGE 35

29 Emotion Numb S M .04 .06 .06 .03 .10 .03 Emotion Numb p .21 .20 .20 .06 .25 .11 Avoidance S M .08 .06 .07 .03 .10 .01 Avoidance p .15 .16 .17 .11 .11 .07 Hyperarousal S M .01 .04 .04 .02 .03 .04 Hyperarousal p .21 .21 .21 .13 .14 .07 Note signifies a statistically significant association ( p < .05). Blue shaded numbers signify hypothesized associations among PCL and SDQ subscales. Aim 2 The second aim was to understand if the relationships between PTSD symptoms and child behavioral difficulties are moderated by gender. My hypothesis was that gender will moderate the relationship between PTSD symptoms and child behavioral difficulties , such that the association between PTSD symptoms and externalizing behaviors will be stron ger for boys than girls, and the association between PTSD symptoms and internalizing behaviors will be stronger for girl s than boys. To address this hypothesis, I used the PROCESS macro of SPSS (Hayes, 2018 ). In these analyses, I combined SM and partner re ports. That is, I averaged them together to form a combined latent score. I ran 1 0 moderation analyses to understand if the relationship between PTSD symptoms (combined SM and partner report of overall PTSD symptoms, emotional numbing symptoms, situational avoidance symptoms, hyperarousal symptoms , and re experiencing symptoms) and child behavioral difficulties (internalizing and externalizing behaviors subscales of the SDQ ) depended on gender . No interactions w ere significant (all p s > .05), indicating tha t none of the relationships depend on gender. However, for interest, the correlat ions are disaggregated for boys and girls in Table 6, below.

PAGE 36

30 Table 6. Bivariate correlations with PCL and SDQ: Boys/Girls SDQ Total SDQ Emotional SDQ Conduct SDQ Hype ractiv e SDQ Peer Relations SDQ Pro Social PCL SM .01 .06 .03 .0 8 .02 .07 .04 .01 .0 6 .01 .12 .16 PCL p .23 .23 .23 .22 .2 4 .20 .17 .11 .23 .18 .00 .14 Reexper S M .0 4 .07 .02 .07 .0 2 .05 .01 .05 .05 .05 .16 .16 Reexper p .0 2 .15 .07 .17 .07 .15 .12 .01 .03 .11 .12 .16 Emotion Numb S M .1 2 .05 .09 .10 .09 .08 .06 .06 .20 .11 .0 9 .08 Emotion Numb p .32 .21 .24 .21 .24 .19 .12 .09 .39 .24 .1 5 .06 Avoidance S M .18 .03 .14 .03 .14 .05 .03 .04 .25 .07 .13 .16 Avoidance p .13 .16 .18 .10 .18 .09 .08 .13 .20 .07 .06 .16 Hyperarousal S M .03 .09 .0 4 .11 .03 .09 .0 7 .01 .06 .02 .06 .20 Hyperarousal p .2 9 .26 .2 9 .24 .29 .22 .23 .15 .24 .18 .01 .13 Note . Associations using only boy children are top left of the cells. Associations using only girl children are bottom left of the cells. p < .05). Blue shaded numbers signify hypothe sized associations among PCL and SDQ subscales. Aim 3 The third aim was t o evaluate a moderated mediation model wherein parenting alliance (M) is a mediator between PTSD symptoms (X) and child behavioral difficulties (Y), and PTSD couple alliance (W) is a moderator between PTS D symptoms and parenting alliance (see Figure 1, above ). All analyses within Aim 3 were also run using combined (averaged) SM and partner report on all variables except for child behavioral difficulties (which I only have partner report on). The model testing the interaction effect of PTSD symptoms and PTSD couple alliance predicting parenting alliance was run first to assess initial model fit in MPlus. Using the scaled scores for male and female reports of SM PTSD symptoms , PTSD c ouple alliance, parenting alliance, and interaction terms of PTSD symptoms and PTSD couple alliance, the interac tion

PAGE 37

31 effect was not significant. In other words, the effect of PTSD on parenting alliance does not depend on level of PTSD couple alliance ( = .06 , SE = .15, p = .72 ). When looking at a plot of this relationship, it was clear that couples who were high in couple alliance were generally reporting low levels of SM PTSD symptoms , and this contributed to the in ability to detect a moder ation effect. Because the interaction was not significant, the iterative process led to a different proposed structural model (see below) . Of note, h owever, was the fact that all loadings of scaled scores on their respective latent factors (e.g., PCL SM an d PCL p on overall PTSD latent factor; ERPS SM, ERPS p, PCC SM , PCC p on overall PTSD couple alliance latent factor; PAI SM and PAI p on overall parenting alliance latent factor) were adequate, with estimates of .52 or above. Therefore, the measurement mod el of the latent constructs was sufficient to use in future, iterative models. Multiple Mediation Model My proposal stated that final implem entation of the above model would be dependent upon issues such as the nature of the variables (e.g., distribution s), tests of the measurement model (e.g., latent factors), and iterative tests of the structural model. Therefore, I shifted to a different structural model . I tested a multiple mediation model in MPlus (see Figure 2 , below) , wherein both parenting alliance and couple alliance were tested as mediators of the association between SM PTSD symptoms and child behavioral difficulties . This model adheres to the broad conceptual themes developed in the background, wherein higher PTSD symptoms may show negative associations with child behavior via lower levels of parenting and relationship functioning. Whereas I knew from Aim 1 that SM self report of PTSD symptoms did not correlate with child behavioral difficulties, I wanted to adhere to the original conceptualization of the predictors all being latent, combined factors. This model was run six times, based on

PAGE 38

32 variations of overall PTSD versus numbing symptoms only, and for all children versus boys versus girls. See Table 7 below, for a list of all path estimates among six different iterations of the final multiple mediation model. Figure 2. Multiple Mediation Model Notes: PCL SM = SM self report of PTSD Checklist Military. PCL p = partner perception of PTSD Checklist Military. ERPS SM = SM perception of partner Emotional Responses to PTSD Symptoms scale. ERPS p = p artner self report of Emotional Responses to PTSD Symptoms scale. PCC SM = SM self report of PTSD Couple Connection scale. PCC p = part ner self report of PTSD Couple Connection scale. Interpreting Model Fit . There are several model fit statistics to consider when deciding the fit of a model (Geiser, 2013 ; Hooper et al., 2008 ). The chi square tests whether misfit of the data is significant (Hu & Bentler, 1999). A significant chi square rejects the null hypothesis that the model fits the data; thus, a non significant chi square is ideal. However, chi square is dependent on sample size and a lar ge sample size is considered over powered and will almost SM PTSD Parenting Alliance PTSD Couple Alliance Child Behavioral Difficulties PAI p PAI SM PCL p PCL SM ERPS SM ERPS p PCC SM PCC p

PAGE 39

33 always have a non significant chi square test. The CFI is the comparative fit index and indicates by how much the model fits better than an independence model in which variables are not related (Be ntler, 1990) . Values for this statistic range from 0 1, with values closer to 1 indicating good fit. A CFI greater than .95 is preferred. The RMSEA is a measure of approximate model fit and has a penalty for non parsimonious models (Steiger, 1990). A RMSEA less than .05 is preferred. The SRMR is a standardized measure of the model residuals. A value less than .05 is preferred. Model 1: Overall model with all PTSD symptoms . Fit indices show that the overall model did not fit perfectly ( 2 = 106.20 [ df = 21] ; p = .0 0 ; RMSEA = .16; CFI = .82; SRMR = .10 ). To improve model fit, I allowed correlation between the errors of the two member reported measures within the couple alliance latent factor ( SM reported PTSD couple connection and emotional responses to PT SD). Nonetheless, there was still a lot of misfit, likely due to misfit in the measurement model . Although the loadings of scaled scores on the respective latent factors was considered adequate, there was some misfit as the indicators were not highly corre lated within the latent factor, particularly among female and male reports. When look ing at path estimates of this overall multiple mediation model, PTSD symptoms did not have an overall effect on child behavioral difficulties ( = .05 ; p = .68). Within t he model, the only significant predictor of child behavioral difficulties was couple alliance ( = .36 ; p <.01). Models 2 & 3: To assess whether child gender affected the relationships within the model, I ran the same overall model using first only the c ouples who reported child behavioral difficulties on a boy child ( N =79) and then using only the couples who reported child behavioral difficulties on a girl child ( N =86). Fit indices again indicated that the model using couples of a boy child did not fit p erfectly ( 2 = 64.64 [ df = 21] ; p = .00; RMSEA = .16; CFI = .81; SRMR = .11 ). Fit estimates were very similar when compared to the entire sample. When looking at path

PAGE 40

34 estimates of model, PTSD symptoms also did not have an overall effect on boy child behavioral difficulties ( = .13 ; p = .50). The only significant predictor of bo y child behavioral difficulties was couple alliance ( = .48 ; p <.01). When using only couples who reported on girl child behavioral difficulties , f it indices again indicated that the model did not fit perfectly ( 2 = 66.26 [ df = 21] ; p = .00; RMSEA = .16 ; CFI = .82; SRMR = .10 ). When looking at path estimates of model, PTSD symptoms also did not have an overall effect on girl child behavioral difficulties ( = .04 ; p = .84). The only significant predictor of girl child behavioral difficulties was couple alliance ( = .29 ; p <.01). Model 4: Model with only e mo tional n umbing PTSD s ymptoms . The emotional numbing symptoms of PTSD are the PTSD symptoms that have been most strongly associated with various family issues, such as relationship, parenting, and family distress (e.g., Galovski & Lyons, 2004; Wilson & Kurtz, 1997). Therefore, to assess whether isolating the X variable to just emotional numbing symptoms instead of overall PTSD symptoms would affect the relationships with the model, I ran the overall model using only SM and partner reported emotional numbing PTSD symptoms. Fit indices sho w that this model also did not fit perfectly ( 2 = 131.24 [ df = 21] ; p = .00 ; RMSEA = .18; CFI = .79; SRMR = .11 ). Emotional numbing symptoms did not significantly predict child behavioral difficulties ( = .09; p = .56 ). Again, the only significant predi ctor of child behavioral difficulties was couple alliance ( = .38 ; p <.01). Both couple and parenting alliance had larger associat ions with emotional numbing as compared to overall PTSD symptoms (couple alliance with overall PTSD symptoms: = .54; p <. 01 as compared to emotional numbing symptoms: = .71; p <.01; parenting alliance with overall PTSD symptoms: = .34; p <.01 as compared to emotional numbing symptoms: = .46; p <.01).

PAGE 41

35 Models 5 & 6: To assess whether child gender affected the relationships within the model using only emotional numbing symptoms, I ran this same model using first only the couples who reported child behavioral difficulties on a boy child ( N =79) and then using only the co uples who reported child behavioral difficulties on a girl child ( N= 86). The model using only couples with a boy child ( N =79) similarly showed not a perfect fit ( 2 = 84.71 [ df = 21] ; p = .00; RMSEA = .19; CFI = .75; SRMR = .12 ). There was no relationship between emotional numbing symptoms and boy child behavioral difficulties ( = .02; p = .93 ). Couple alliance was no longer a significant predictor of boy child behavioral difficulties in this model, likely due to smaller sample size ( = .39; p = .07 ). Fi nally, the model using only couples who report on girl child behavioral difficulties ( N =86) not surprisingly also did not fit perfectly fit ( 2 = 74.99 [ df = 21] ; p = .00; RMSEA = .17; CFI = .80; SRMR = .12 ). The effect of emotional numbing symptoms on child behavior was not significant for girls ( = .21; p = .44 ). One differen ce between boys and girls in this model is found in the relationship between couple and parent alliance, in the model with girls this rel ationship is non existent ( = .02; p = .94 ), whereas there is a significant association in the model with boys ( = .39; p < .02 ). Table 7 . Path Estimates for Multiple Mediation Models Model PTSD Couple Alliance PTSD Parenting Alliance Couple Alliance Parenting Alliance PTSD Child behavioral difficulties Couple Alliance Child behavioral difficulties Parenting Alliance Child behavioral difficulties 1. Overall PTSD . 54 ** .33 ** .37 .05 .36 ** .05 2. Boys .48 ** .25 ( p = .08 ) .55 ** .13 .48 ** .00 3. Girls .57 ** .51 ** .14 .04 .29 ** .09 4. Emotional Numbing .71 ** .46 ** .21 ( p = .075) .09 .38 ** .07

PAGE 42

36 5. Boys .71 ** .45 ** . 39 ** .02 .39 ( p = .065) .02 6. Girls .73 ** .53 ** .02 .21 .40 ** .13 Note. ** = p < .05

PAGE 43

37 CHAPTER IV DISCUSSION Taking a systems perspective, service member (SM) PTSD symptoms after deployment can take a toll on family functioning. This study tested relationships among the variables of SM the SM as a result of his PTSD symptoms and overall couple communication and conne ction around PTSD symptoms), parenting alliance, and child behavioral difficulties. Results showed that, in general, higher levels of SM PTSD related to worse couple and parenting alliance. SM PTSD symptoms and child behavioral difficulties were associated on a bivariate level, but only when using the partner report of the SM PTSD symptoms. Thus, SM s elf report of PTSD symptoms was not correlated with child behavioral difficulties, nor was SM PTSD related to child behavioral difficulties when combining thes e reports into a latent variable in a larger model. Similarly, partner reports of aspects of couple alliance and parenting alliance were both significantly correlated with child behavioral difficulties, but these associations were generally not found when using SM reports. there was a link between couple alliance and child behavioral difficulties; however, this association is carried by the partner report (based on the bivariate correlations) . When evaluating these associations for PTSD numbing symptoms specifically (compared to overall PTSD symptoms), some of these associations increased in magnitude, but the overall pattern of results was generally the same. T here are several important themes to discuss within these overall results.

PAGE 44

38 Basic L evels of V ariables Before discussing patterns of association, there were interesting differences in the basic levels of the variables as reported by the SM and the partne r. When looking at general frequencies of the PTSD symptoms and couple variables, the SM reported more individual (PTSD symptoms) and relationship ( less positive feelings from his partner towards him as a result of the PTSD symptoms , less connection relate d to the PTSD) distress as compared to his One way to understand this is in the context of the SM internally experienced symptoms that make up many of the diagnostic criteria of the DSM 5 (American Psychiatric Association, 2013), including PTSD, it makes sense that even loved ones cannot fully grasp the extent of thei r SM connection and expression of positive emotions due to his more negative perception of the relat ionship. There were also interesting differences in the level of child behavioral difficulties in this sample as compared to national norms. Specifically, the children in the current sample, as reported by the partner, had higher levels of behavioral dif ficulties compared to the norm SDQ sample in the United States. Whereas we know that military children exhibit a great deal of resiliency (Lincoln & Sweeten, 2011), we also know that children of veterans with PTSD on average exhibit greater behavior proble ms than children of non veteran parents without PTSD (Ahmadzadeh & Malekian, 2004), and that parental deployment is associated with an increase in child emotional and behavioral symptoms (White, de Burgh, Fear & Iversen, 2011). This is

PAGE 45

39 consistent with the higher level of child distress in this current sample of families who have experienced at least one deployment and were selected based on elevated PTSD symptoms . Salience of C ouple A lliance As noted above, PTSD symptoms w ere generally associated with the couple variables al responses to the SM as a result of his PTSD symptoms, couple communication and connection around PTSD symptoms) in the expected direction, using both SM and partner report. Yet, the connections between PTSD symptoms and these couple variables with child behavioral difficulties were inconsistent and depended largely on whether it was the SM or partner report. However, c ouple alliance had a significant association with PTSD symptom s ( whereas parenting alliance did not ), and couple alliance also emerged as the only significant predictor of child behavioral difficulties within the final models. This is interesting because couple alliance does not directly assess interactions with chil dren or parenting, but rather symptoms of PTSD in of on the same page for parentin g. Instead, the most important variable related to child behavioral and lack of connection related to the PTSD symptoms. If this possibility, based on these cross sectional findings, held in stu dies then f ocusing on such dynamics in assessment and treatment of family functioning at post deployment may be useful when addressing child behavioral difficulti es. Comparisons of Overall PTSD S ymptoms with E motional Numbing S ymptoms There were several analyses that focused on subscales of the various study variables . I was especially interested in the emotional numbing symptom s subscale of the PCL , as this

PAGE 46

40 constellation of symptoms in SM s has been shown to be most strongly associated with various family issues, such as relationship, parenting, and family distress (e.g., Galovski & Lyons, 2004; Wilson & Kurtz, 1997). When comparing associations of o verall PTSD symptoms v ersus only emotional numbing symptoms, the magnitude of association increased with both partner and SM report of partner emotional response to SM PTSD symptoms, and partner report of parenting alliance. The remainder of associations d id not statistically increase in magnitude, including child behavioral difficulties. Isolating to emotional numbing symptoms also did not statistically change the association of PTSD symptoms with child behavioral difficulties wi thin the final multiple med iation models. Emotional numbing symptoms were, however, more strongly associated with the couple variables (couple alliance, parenting alliance) within the final models than were overall PTSD symptoms, similar to the pattern of associations of the basic c orrelations. Thus, if similar results emerged in longitudinal and interventional studies, this may suggest that emotional numbing symptoms negatively impact the couple. Symptoms of emotional numbing decrease emotional expression and intimate exchanges in a relationship (Riggs et al., 1998), and we see that the couples in our sample do in fact report lower levels in (using only partner report) in the context of e levated symptoms of feeling distant or estranged from others and having difficulty experiencing or having positive feelings (American Psychiatric Association, 2013). Emotional numbing also emerged as a salient factor in a closer evaluation of which SDQ s ubscales were correlat ed with which SM PCL subscales. In general, there was not a clear differentiation between internalizing and externalizing patterns among SM and child symptoms as hypothesized . Instead, there were certain subscales of both the PCL and SDQ that emerged as

PAGE 47

41 most frequently associated . Emotional and Conduct behaviors were the SDQ subscales that were most frequently associated with PTSD total and subscale variables. Emotional numbing SM PTSD symptoms, as reported by the partner, were associa ted with the most child difficulty subscales (Emotional, Conduct, Peer Relations). Ruscio et al. (2002) suggests that emotional are necessary for developing a mea ningful parent child relationship. I f this cross sectio nal finding held in studies designed to focus and intervene on such constructs with both the SM and child, this may suggest that meaningfu l peer relations as well. If this is true, then when assessing a family post deployment it may be especially important to focus on SM emotional avoidance symptoms and child emotional symptoms, conduct behaviors, and peer relations. Reporter Method V arianc e A noteworthy theme that emerged across results of this study is reporter method variance. Most notably , SM PTSD symptoms predicted child behavioral difficulties (as reported by the partner), but only when using partner perception, and not SM self report , of the PTSD symptoms. This pattern was also noted when looking at associations among subscales of PTSD and child behavioral difficulties. There are several potential explanations for this finding. First, i both her SM symptoms and her child behavioral difficulties may actually be partly emotional distress. There are several studies which show that the more emo tionally impaired the mother, the greater the degree to which she perceives her child to have behavior pro blems (e.g., Najman et al., 2001 ; Youngstrom et al., 2000). However, not all studies are in agreement that maternal distress prevents mothers from pro viding un biased reports of child behavior (e.g.,

PAGE 48

42 Qu erido, Eyberg , on multiple constructs, would suggest that if the study had also included the SM reports of child difficult ies, then there might have been significant associations between his report of his own PTSD symptoms and his report of child behavioral difficulties. Thus, although parental reports of child behavior/emotional problems have been shown to have a high level of concordance (Luoma, Koivisto & Tamminen, 2003) , these issues of method variance attributable to reporter could have strongly influenced the pattern of associations among constructs. Another possible explanation is that partners are more distressed/beh aviorally affected to the degree that they perceive their SM having symptoms of PTSD, and that child behavior al difficulties are in fact generally as sociated with maternal distress. That is, the more that mothers behavior through mechanisms such as changes in maternal parenting/responsiveness, increased maternal distress and transmission of such distress, or similar processes. The association between maternal mental illness/distress and child behavioral difficulties is consistently found across a variety of samples and contexts and reporters of child behavior (e.g., O'Connor, Monk , & Burke, 2016), including wives of SM s with PTSD in the post deployment context (Flake et al., 2009). Moreover, the link between mother child emotional distress may also have a level of reciprocity, as child behavioral difficulties can also lead to maternal mental health impairment (Kingsbury, Clavarino, Mamun, Saiepour & Najman, 2017). Th e R ole of Child G ender I had hypothesized that child gender would moderate the relationship between PTSD symptoms and child behavioral difficulties, such that the association between PTSD symptoms and externalizing behaviors would be stronger for boys tha n girls, and the association between

PAGE 49

43 PTSD symptoms and internalizing behaviors would be stronger for girls than boys . However, this was not found. Although gender differences in emotional expression in children has been shown (Chaplin & Aldao, 2013), sever al studies note that such differences are varied, nuanced, not necessarily static, and could be strengthened or attenuated on various other variables, including age (Brody & Hall, 2008; Chaplin & Aldao, 2013; Brown, 1999). It is possible that if I had powe r to focus on certain subgroups of children by age and gender (e.g., adolescent girls; elementary aged boys) or could look at changes in behavior by gender over time, I may have found a significant interaction. One interesting finding related to child ge nder wi thin the final multiple mediation models was that in both main models (overall PTSD symptoms and emotional numbing PTSD symptoms), the association between couple and parenting alliance was significant for boys but not for girls. There are no obvious differences (e.g., age) between boys and girls that could help to explain this finding and further exploration into this area would entail additional analyses looking at potential differences among boys and girls in this sample. Limitations It is important to address limitations of the study. Perhaps the most important to highlight is the not perfect fit of the final model. Whereas fit indices are a useful guide, Hooper, Couglhan and Mullen (2008) argue that a structural equation model should also be examined in the context of substantive theory, and the rationale for the proposed model was strong in the current study. Nonetheless, the fit indices of the model deviating from published recommended thresholds is a clear limitation of my study an d indicates that results of the model cannot be interpreted with full confidence.

PAGE 50

44 Moreover, constructs are gathered cross sectionally , a re not experimentally manipulated, are all self report, and use only one reporter on child behavioral difficulties. T his has limitations related to response bias, reporter method variance , and the inability to make any claims about causation or temporal relationships. Further, it is likely that there are other measured or unmeasured variables that are not a part of the study that would help to better explain the relationship between a SM symptoms, couple and parenting alliance, and child behavioral difficulties, including maternal distress or depression. Lastly, I only studied American, English speaking, male Ar my SM s and female civilian partners in this project. This does not allow me to generalize to female service members/male partners, gay/lesbian /transgender couples, dual military couples, non English speaking military couples, or couples not in the military . Couple, family, and parent child dynamics, as well as the understanding or expression of PTSD symptoms may be different in such samples, which may change the associations among the various study variables. Clinical Implications tations, there are several potential clinical implications to the degree that these cross sectional associations do in fact suggest the type of reciprocal causal relationships suggested by family systems theory . Within family systems theory, it is impossib le families who are overwhelmed by the combination of the processes within the family system discussed throughout this study, such a narrative may help to identify pot ential intervention points within the family system. As mentioned above, by putting these processes together in a larger model, it may reinforce the idea that they a re all interrelated and that modifying one part

PAGE 51

45 of the family system could influence other parts of the family system. For example, if a mother is helping the couple to find ways to connect symptoms , as opposed to dynamics specifically related to parenting or child , may be the most effective in improving child outcomes. This idea , being, interventions should focus on reducing parental conflict and disruption is not ne w (e.g., Carlson & Corcoran, 2001; Demo & Fine, 2010; Brown, 2010). However, the usual go to empirically supported treatments targeting negative child behavioral difficulties typically include parent child management or cognitive behavioral therapy focused on the child (Society of Clinical Child & Adolescent Psychology, 2016) and often do not focus on the parental relationship itself . Results of this study are a good reminder that in such situations as described above, targeting the relationship and , more specifically, how the couple is coming together around the SM PTSD, may be a part of how to h elp the family system (2016) discuss that a new condition in the DSM al Relationship negative effects of parental relationship discord on a child in the family, including effects on the this diagnosis is often an after thought, the authors consider how it covers a broad and common one compl ex situation, specific to military members but potentially applicable to any family system , by which this DSM 5 condition may be applicable to young patients and their parents and may help to inform treatment.

PAGE 52

46 It is encouraging that there are interventi ons already developed for couples experiencing both PTSD and relationship distress, including targeting decreasing PTSD symptoms and increasing couple communication skills (Erbes et al., 2008; Monson et al., 2012; Blow et al., 2015). Results of this study would suggest that screening for and addressing SM emotional numbing symptoms may be particularly important wit given that within this study, this constellation of symptoms was most asso ciated with couple difficulties. Futur e Directions There are several next steps I could take from here. Using my current dataset, it may be helpful to further investigate the nature of problems that could differentiate between boys and girls. This could include looking at any differences betw een boys and girls on the subscale or item level of the SDQ. Another future direction could be including more maternal mental health and well being variables within bivariate and model analyses to better understand how the with the study variables. It may also be interesting to explore if any SM self report individual functioning variable, including other mental or physical health variables or specific variables related to the military or deployment, may be associated with c hild outcomes as reported by the partner. Within the current dataset, I could also look at altering the models in an attempt to improve model fit and/or better explain the relationships among the variables. Improving model fit could involve using a simpler construct for the couple alliance variable with less indicator items or dropping parenting alliance from the model as a mediator because it was overall less associated with both PTSD and child outcomes as compared to couple alliance. There are several l imitations of the dataset that could only be addressed in a future study. Likely because the SDQ instructions asked respondents to report on their youngest or only child,

PAGE 53

47 the child age range was skewed right; less than 10% of children in the study were age d 11 17. If I had a larger sample or had a sample with a more even spread of child age ranges, it is possible that I could have analyzed associations for boys and girls in different age groups and possibly seen significant results when comparing results fo r boys versus girls (Brody & Hall, 2008; Chaplin & Aldao, 2013). As discussed previously, it would be very helpful to have the service between SM PTSD and child be havioral difficulties. Although the SDQ is a valid and well known measure of child behavioral difficulties, it would also be helpful to have more than one measure of child behavioral difficulties to use in analyses. Specifically, more objective (i.e., obse rvational) or third party (i.e., teacher self report) data may provide a more comprehensive assessment of child functioning. Lastly, a future study should include a longitudinal assessment of family functioning, including data from both pre and post deplo yment to measure overall changes or stability in family functioning dynamics that would paint a fuller picture of how these constructs work within each family structure over time.

PAGE 54

48 REFERENCES Abidin, R. R. & Brunner, J. F. (1995) Development of a parenting alliance inventory. Journal of Clinical Child & Adolescent Psychology, 24 : 1, 31 40. Ahmadzadeh G. & Malekian A. (2004) Aggression, anxiety, and social development in adolescent children of war veterans with PTSD versus those of non veterans. Journal of Research in Medical Sciences, 9 (5):231 234. Allen, E. S., Rhoades, G. K., Stanley, S. M., & Markman, H. J. (2010). Hitting Home: Relationships between Recent Deployment, Post Traumatic Stress Symptoms, and Mar ital Functioning for Army Couples. Journal of Family Psychology, 24 (3), 280 88. Allen, E., S tanley, S., Rhoades, G., Markman, H., and Loew , B. (November, 2012). Reintegration with the spouse after deployment: Does talking about PTSD and combat help? Talk presented at the meeting of the Association for Behavioral and Cognitive Therapies Convention, National Harbor, MD. Allen, E. S. in the Context of Husband PTSD. I n S. J. Fredman (Chair), Partnering and parenting in the presence of PTSD. Symposium presented at the annual meeting of the Association for Behavioral and Cognitive Therapies, New York, NY. American Psychiatric Association. (2013). Diagnostic an d stat istical manual of mental disorders (5th ed.). Arlington, VA: American Psychiatric Publishing. Bearss, K. E. & Eyberg, S. (1998) A Test of the Parenting Alliance Theory. Early Education and Development , 9 (2): 179 185, DOI: 10.1207/s15566935eed0902_5 Bernet, W., Wamboldt, M. Z., & Narrow, W. E. (2016). Child Affected by Parental Relationship Distress. Journal of the American Academy of Child & Adolescent Psychiatry , 55 (7): 571 579. Bentler, P. M. (1990). Comparative f it indexes in structural models. Psychological Bulletin , 107, 238 246. Bliese, P. D., Wright, K. M., Adler, A. B., Cabrera, O., Castrol, C. A., & Hoge, C. W. (2008). Validating the Primary Care Posttraumatic Stress Disorder screen and the Posttrau matic Stress Disorder Checklist with soldiers returning from combat. Journal of Consulting and Clinical Psychology, 76 , 272 281. doi: 10.1037/0022 006X.76.2.272 Blow, A. J., Curtis, A. F., Wittenborn, A. K., & Gorman, L. (2015). Relationship Problem s and Military Related PTSD: The Case for Using Emotionally Focused Therapy for Couples. Contemporary Family Therapy , 37 (3), 261 270. DOI: 10.1007/s10591 015 9345 7

PAGE 55

49 Bonds T. M., Baiocchi D., & McDonald L. L. (2010) Documented briefing: Army Deployments to OEF and OIF . Santa Monica, CA: Arroyo RAND Corporation. Retrieved from: http://www.rand.org/content/dam/rand/pubs/documented_briefings/2010/RAND_DB 587 .pdf. Bowen, M. (1978) Family therapy in clinical practice. New York: Aronson Brody, L . R. (1999). Gender, emotion, and the family. Cambridge, MA: H arvard University Press. Brody , L . R . , Hall , J . A. (2008). Gender and emotion in context. In: Lewis , M ., Haviland Jones, J . M . , Barrett , L. F . (eds) . Handbo ok of emotions. 3rd The Guilford Press; New York, NY: 2008. pp. 395 408. Harvard University Press. Retrieved from EBSCOhost. Brown, S. L. (2010). Marriage and Chil d Well Being: Research and Policy Perspectives. Journal of Marriage and Family Therapy, 72 (5): 1059 1077. Carlson , M . J . & Corcoran , M . E. (2001). Family structure and children's behavioral and cognitive outcomes. Journal of Marriage and Family , 63:779 792. Chaplin, T. M. & Aldao, A. (2013). Gender Differences in Emotion Expression in Children: A Meta Analytic Review. Psychological Bulletin, 139 (4): 735 765. Chandra, A., Lara Cinisomo, S., Jaycox, L. H., Tanielian, T., Burns, R. M., Ruder, T., & Han, B. (2010). Children on the homefront: The experience of children from military families. Pediatrics , 125 (1), 16 25. Conger, R. D., Ge, X., Elder, G. H., Jr., Lorenz, F. O., & Simons, R. L. (1994). Economic stress, coercive family process, and developmental problems of adolescents. Child Development, 65, 541 561. Creech, S. K., Hadley, W., & Borsari, B. (2014). The Impact of Military Deployment and Reintegration on Children and Parenting: A Systematic Review. Professional Psy chology, Research and Practice , 45 (6), 452 464. http://doi.org/10.1037/a0035055 Demo, D . M. & Fine, M , A. (2010). Beyond the average di vorce. Sage; Thousand Oaks, CA. Department of Defense (2015). 2015 Demographics: Profile of the Military Community. Washington, DC: Department of Defense, Office of the Deputy Assistant Secretary of Defense for Military Community and Family Policy.

PAGE 56

50 Erbes, C., Polusny, M., MacDermi d, S., & Compton, J. (2008). Couple therapy with combat veterans and their partners. Journal of Clinical Psychology, 64 (8), 972 983. Flake, E. M., Davis, B.E., Johnson, P. L., & Middleton, L. S. (2009). The psychosocial effects of deployment on milita ry children. Journal of Developmental and Behavioral Pediatrics , 30 , 271 278. doi: 10.1097/DBP.0b013e3181aac6e4 Fredman, S. J., Monson, C. M., & Adair, K. C. (2011). Implementing cognitive behavioral conjoint therapy for PTSD with the newest generatio n of veterans and their partners. Cognitive and Behavioral Practice , 18 (1), 120 130. Galovski, T & Lyons, J. A. (2004). Psychological sequelae of combat violence: A review of the impact of PTSD on the veteran's family and possible interventions. Aggres sion and Violent Behavior , 9(5): 477 501. Geiser C. (2013) Data Analysis with Mplus . NewYork: Guilford Press. Gewirtz, A. H., Polusny, M. A., DeGarm o, D., Khaylis, A., & Erbes, C. (2010). Posttraumatic stress symptoms among National Guard soldiers deployed to Iraq: Associations with parenting behaviors and couple adjustment. Journal of Consulting and Clinical Psychology, 78, 599 610. doi:10.1037/a0020571 Gewirtz, A. H., Pinna, K. L. M., Hanson, S. K. & Brockberg, D. (2014). Promoting Parenting to Support Reintegrating Military Families: After Deployment, Adaptive Parenting Tools. Psychological Services, 11 (1): 31 40. Goodman, R. (1997) The Strengths and Difficulties Questionnaire: A Research Note. Journal of Child Psychology and Psychiatry , 38 : 581 586. Goodman, A., Lamping, D.L. & Ploubidis, G.B. (2010). When to use broader internalising and externalising subscales instead of the hypothesised five subscales on the Strengths and Difficulties Questionnaire (SDQ): data from Br itish parents, teachers and children. Journal of Abnormal Child Psychology, 38 (8): 1179 91. Hanley, K. E., Leifker, F. R., Blandon, A. Y., & Marshall, A. D. (2013). Gender differences in the impact of posttraumatic stress disorder symptoms on community Journal of Family Psychology, 27, 525 530. doi:10.1037/a0032890 Hayes, A. F. (2018). Introduction to mediation, moderation, and conditional p rocess analysis: A regression based approach (2nd Ed.) New York: The Guilford Press. Hoge, C. W., Auchterlonie, J. L., & Milliken, C. S. (2006). Mental health problems, use of mental health services, and attrition from military service after returning from deployment to Iraq or Afghanistan. Journal of the American Medical Association , 295, 1023 32.

PAGE 57

51 Hooper, D., Coughlan, J. and Mullen, M. R. (2008). Structural Equation Modell ing: Guidel ines for Determining Model Fit. The Electronic Journal of Business Research Methods , 6 (1): 53 60. available online at www.ejbrm.com Hu, L.T. and Bentler, P.M. (1999). Cutoff Criteria for Fit Indexes in Covariance Structure Analysis: Co nventional Criteria Versus New Altern atives. Structural Equation Modeling, 6 (1), 1 55. Jordan, B.K., Marmar, C.R., Fairbank, J.A., Schlenger, W.E., Kulka, R.A., Hough, R.L., &Weiss, D.S. (1992). Problems in families of male Vietnam veterans with po st traumatic stress disorder. Journal of Consulting and Clinical Psychology,60 :916 926. King, D. W., Leskin. G. A., King, L. A. and Weathers, F. W. (1998). Confirmatory factor analysis of the Clinician Administered PTSD Scale: Evidence for the dimensionality of posttraumatic stress disorder. Psychological Assessment. 10.90 96. Kingsbury, A. M., Clavariono, A., Mamun, A., Saiepour, N. & Najman, J.M. (2017). Does having a dif ficult child lead to poor mental health? Public Health, 146: 46 55. Kring , A . M . & Gordon , A . H. (1998). Sex differences in emotion: Expression, experience, and physiology. Journal of Personality and Social Psychology, 74 :686 703. [PubMed: 9523412] . L ee, I. A., & Preacher, K. J. (2013, September). Calculation for the test of the difference between two dependent correlations with one variable in common [Computer software]. Available from http://quantpsy.org . Lester, P., Peterson, K., Reeves, J., Knauss, L., Glover, D., Mogil, C., Duan, N., Saltzman, W., Pynoos, R., Wilt, K. & Beardslee, W. (2010). The Long War and Parental Combat Deployment: Effects on Military Children and At Home Spouses. Journal of th e American Academy of Child and Adolescent Psychiatry , 49 (4), 310 320. Lincoln, A.J. & Sweeten, K. (2011). Considerations for the Effects of Military Deployment on Children and Families. Social Work in Health Care, 50 (1): 73 84, DOI: 10.1080/00981389.2010.513921 their child and maternal depressive symptoms. Nordic Journal of Psychiatry, 58 ( 3): 205 211. DOI: 10.1080/08039480410006299 Masten, A. S. (2001). Ordinary magic: Resilience processes in development. American Psychologist, 56, 227 238. doi:10.1037/0003 066X.56.3.227 Monson, C., Taft, C. T., & Fredman , S. J. (2009). Military related PTSD and Intimate Relationships: From Description to Theory Driven Research and Intervention Development. Clinical Psychology Review, 29 (8): 707 714.

PAGE 58

52 Monson , C . M . , Fredman , S . J . , Adair , K . C . , Stevens , S . P . , Resick , P . A . , Schnurr , P . P . , Macdonald , A. (2011). Cognitive behavioral conjoint therapy for PTSD: Pilot results from a community sample. Journal of Traumatic Stress, 24 :97 101. doi:10.1002/jts.20604. Monson, C., Fredman, S., MacDonald, A., Pukay Ma rtin, N., Resick, P., & Schnurr, P. (2012). Effects of cognitive behavioral couple therapy for PTSD: A randomized controlled trial. Journal of the American Medical Association, 308 (7), 700 709. Najman, J. M., Williams, G. M., Nikles, J., Spence, W., B Brocque, R., Anderson, M. J. & Shuttlewood, G. J. (2001). Bias Influencing Maternal Reports of Child Behaviour and Emotional State. Social Psychiatry and Psychiatric Epidemiology, 36 (4):186 194. DOI: 10.1007/s001270170062 Nelson Goff, B. S , Reisbig, A. M., Bole, A., Scheer, T., Hayes, E., Archuleta, K. L., ... & Sanders Hahs, E. (2006). The effects of trauma on intimate relationships: A qualitative study with clinical couples. American Journal of Orthopsychiatry , 76 (4), 451. O'Connor, T. G., Monk, C., & Burke, A. S. (2016). Maternal Affective Illness in the Perinatal Period and Child Development: Findings on Developmental Timing, Mechanisms, and Intervention. Current Psy chiatry Reports, 18 (3), 5. doi:10.1007/s11920 016 0660 y Querido, J. G., Eyberg, S. M., & Boggs, S. R. (2001). Revisiting the accuracy hypothesis in families of young children with conduct problems. Journal of Clinical Child Psychology, 30(1), 253 2 61. Ray, S. L., & Vanstone, M. (2009). The impact of PTSD on veterans' family relationships: an interpretative phenomenological inquiry. International Journal of Nursing Studies, 46(6), 838 847. Renshaw, K.D., Rodrigues, C.S. & Jones, D. H. (2008). Psychological symptoms and marital satisfaction in spouses of Operation Iraqi Freedom veterans: Rel ationships with spouses' perceptions of veterans' experiences and symptoms. Journal of Family Psychology , 22(4): 586 594. doi.org/10.1037/0893 3200.22.3.586 Riggs, D.S., Byrne, C.A, Weathers, F.W, & Litz, B.T (1998). The quality of the intimate rela tionships of male Vietnam veterans: Problems associated with posttraumatic stress disorder . Journal of Traumatic Stress , 11: 87 101 Ruscio , A. M., Weathers, F. W., King, L. A. & King, D. W. (2002), Male war zone veterans' perceived relationships with their children: The importance of emotional numbing. Journal of Traumatic Stress, 15 : 351 357. doi:10.1023/A:1020125006371

PAGE 59

53 Sayers, S.L. ( 2011). Family Reintegration Difficulties and Couples Therapy for Military Veterans and Their Spouses. Cognitive and Behavioral Practice, 18: 108 119. Samper, R. E., Taft, C. T., King, D. W. & King, L. A. (2004). Posttraumatic stress disorder symptoms and parenting satisfaction among a national sample of male Vietnam veterans. Journal of Traumatic Stress, 17 : 311 315. Snyder, D. K., Klei n, M. A., Gdowski, C. L., Faul stich, C., & LaCombe, J. (1988). Generalized dysfunction in clinic and nonclinic families: A comparative analysis. Journal of Abnormal Child Psychology, 16 , 97 109. Society of Clinical Child & Adolescent Psychology (2016). Effective child therapy: Evidence based mental health treatment for child ren and adolescents. Retrieved J anuary 8, 2017 from http://effectivechildtherapy.org/content/ebp options specific disorders . Steiger , J. H. (1990). Structural model evaluation and modification. Multivariate Behavioral Research , 25, 214 12. Weathers, F. W., Litz, B. T., Herman, D. S., Huska, J. A. & Keane, T. M. (1993). The PTSD checklist: reliability, validity, & diagnostic utilit y. Paper presented at the Annual Meeting of the International Society for Traumatic Stress Studi es, San Antonio, TX. Weathers, F. W., & Ford, J. (1996). Psychometric properties of the PTSD Checklist (PCL C, PCL S, PCL M, PCL PR) . In B. H. Stamm (Ed.), Measurement of stress, trauma, and adaptation (pp. 250 252). Lutherville, MD: Sidran Foundation & Press. White, C.J., de Burgh, T., Fear, N. T. & Inersen, A. C. (2011). The impact of deployment to Iraq or Afghanistan on militar y children: A review of the literature. International Review of Psychiatry,23 (2): 210 217, DOI: 10.3109/09540261.2011.560143 Wilson, J. P., & Kurtz, R. R. (1997). Assessing posttraumatic stress disorder in couples and families. In J. P. Wilson & T. M. Keane (Eds.), Assessing psychological trauma and PTSD (pp. 349 372). New York: Guilford Press Youngstrom, E., Loeber, R., & Stouthamer Loeber, M. (2000). Patterns and correlates of agreement between parent, teacher, and male adolescent ratings of externalizing and internalizing problems. Journal of Consulting and Clinical Psychology, 68 (6), 1038 1050.

PAGE 60

54 APPENDIX A Bivariate correlations with all variables: Boys and girls 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 1. PCL SM . 4 7 .9 2 . 4 4 .8 7 .3 9 .8 3 .4 1 .9 2 .4 3 .5 0 .2 6 .4 0 .1 4 .2 2 .2 5 .0 0 .0 4 .0 3 .0 1 .0 0 .0 0 2. PCL p -.4 5 .8 7 .3 9 .8 7 .3 7 .8 1 .4 5 .9 3 .2 8 .4 6 .0 8 .3 1 . 1 0 .2 4 .1 9 . 20 .2 1 . 1 0 .1 5 .0 9 3. Reexper M -.4 7 .6 9 .3 2 .8 3 .4 3 .7 8 .3 8 .3 5 .1 8 .2 8 .1 1 .1 5 .1 9 .0 1 .04 .0 3 .0 2 .0 4 .0 2 4. Reexper p -.3 1 .5 9 .3 6 .7 3 .4 0 .7 1 .1 7 .2 2 .0 4 .1 2 .1 1 .0 7 .0 9 .12 9 .1 3 . 0 5 .0 1 . 0 6 5. Emotion Numb M -.4 0 .6 2 .2 9 .7 4 .3 4 .5 9 .3 1 .4 6 .1 1 .2 8 .2 8 .0 4 .06 .0 6 .0 3 .1 0 .0 3 6. Emotion Numb p -.2 7 .6 4 .3 5 .7 9 .3 5 .5 9 .1 2 .3 8 .1 5 .3 2 .2 1 .20 .2 0 .0 6 .2 5 .1 1 7. Avoidance M -.4 0 .6 7 .2 9 .3 8 .1 7 .3 7 .1 3 .1 7 .1 5 .0 8 .06 .0 7 .0 3 .1 0 .0 1 8. Avoidance p -.3 7 .6 6 .1 4 .4 0 .0 5 .2 9 .1 0 .1 5 .1 5 .16 .1 7 .1 1 .1 1 .0 7 9. Hyperarous al M -.4 6 .4 3 .2 4 .3 3 .1 5 .2 0 .2 5 .0 1 .04 .0 4 .0 1 .0 2 .0 4 10. Hyperarous al p -.2 8 .4 1 .0 6 .3 0 .0 9 .2 5 .2 1 .21 .2 1 .1 3 .1 4 .0 7 11. ERPS SM -.4 3 .6 6 .3 1 .3 9 .2 9 .0 3 .03 .0 5 .0 8 .0 5 .0 7 12. ERPS p -.3 4 .7 2 .1 8 .4 4 .3 6 .33 .3 4 .2 1 .2 8 .1 8 13. PCC SM -.4 2 .4 1 .2 1 .0 3 .08 .0 6 .0 4 .0 2 .0 3 14. PCC p -.1 6 .3 4 .3 3 .29 .2 8 .2 2 .2 1 .2 5 15. PAI SM -.4 5 .0 3 .07 .0 6 .0 1 .0 3 .0 7 16. PAI p -.2 1 .12 .1 1 .1 2 .1 4 .1 6 17. SDQ Total -.71 .7 1 .7 4 .6 8 .6 2 18. SDQ Emotional -.9 8 .3 4 .5 4 .2 1 19. SDQ Conduct -.3 4 .5 2 .2 1 20. SDQ Hyperactiv ity -.2 5 .3 1 21. SDQ Peer Relations -.3 0 22. SDQ Pro Social -Note . Bold ed fies a statistically significant association ( p < .05). Gray shaded numbers signify the association between partner and SM report on the same construct. Blue shaded numbers signify hypothesized associations among PCL and SDQ subscale s. PCL SM = PTSD Checklist Military. PCL p = partner report of PTSD Checklist Military. ERPS SM = SM perception of partner Emotional Responses to PTSD Symptoms scale. ERPS p = partner self report of Emotional Responses to PTSD Symptoms scale. PCC SM = male repo rt of PTSD Couple Connection scale. PCC p = partner report of PTSD Couple Connection scale. SDQ = (partner report of) Strengths and Difficulties Questionnaire