ADMINISTRATION AND CONTROL OF COMPUTER-BASED INFORMATION PROCESSING IN LOCAL GOVERNMENT A COLORADO STUDY
Wi111am Marshal 1 Leavitt
B. A., New York University, 1970 M. P. A., The University of Colorado, 1976
A thesis submitted to the
Faculty of the Graduate School of Public Affairs of the University of Colorado in partial fulfillment of the requirements for the degree of Doctor of Public Administration,
Graduate School of Public Affairs
This Thesis for the Doctor of Public Administration
William Marshall Leavitt has been approved for the Graduate School of Public Affairs by
Kathleen A. Archibald
Leavitt, William Marshall (D.P.A., Public Administration)
Administration and Control of Computer-Based Information Processing in Local Government: A Colorado Study
Thesis directed by Associate Professor Frank J. Cesario
The use of computer-based information processing in the public sector at all levels of government is now widespread. It is virtyally impossible for citizens to interact with their government at the federal, state, or local level without information processing technology playing a role in the interaction. This study examines the administration and control of computer-based information processing in Colorado local governments and describes the impact of the technology on the management of local governments.
The empirical portion of this study focuses on Colorado municipalities, specifically medium-sized (population 50,000 -250,000) municipalities. The research methodology is based primarily on semi-structured interviews with management-level personnel in the municipal governments of the targeted Colorado municipalities. Seven of the nine targeted municipalities participated in the study conducted in cooperation with the Colorado Municipal League. The semi-structured interviews focused on issues involving the management of computer-based information-processing, such as organizational arrangements for computing; the determination of information processing priorities; the role of computer
professionals and users; and the impact of computers on the management of local government.
Key findings indicate that all of the study cities have modern, near state-of-the-art computer operations based on on-line processing of data. In addition, information processing technology is considered an essential tool in the day-to-day operations of all the study cities. A profile of information processing use in the study cities indicates that: computing services are basically centralized in one department while the use of information processing is spread among several user departments; the users of information processing in the study cities are at least somewhat knowledgeable about the technology; and the greatest current values of information processing technology are clerical and information retrieval types of tasks.
One conclusion of this study is that microcomputing technology will play an increasingly important role in the operations of municipal government, especially the use of microcomputers as personal productivity tools.
Another important finding is that the introduction and use of computer-based information processing in municipal government has had little significant impact on the overall management of government operations.
It is a difficult task to identify and thank every individual who has contributed to this research effort. There are some individuals, however, whose important contributions to this research effort and to the development of the manuscript must be recognized.
My special thanks goes to my Dissertation Committee for their important contribution to the direction of this research and their valuable input and comments on the successive drafts of the manuscript. The Dissertation Committee: F. William Heiss, Frank Cesario, Kathleen Archibald and Robert Jensen all contributed their time and effort to make the dissertation a viable study.
It is also important to recognize the invaluable contribution of all the respondents who spent hours of their time answering questions during the interview process. Also, the efforts of all the study city liaison people greatly aided in setting up interviews and making necessary arrangements for conducting this research. In addition, I would like to thank Kenneth Bueche, Director of the Colorado Municipal League, for the League's cooperation in making this research possible.
A very special thanks to Suzan Knapp and Marbella Trujillo who put the dissertation manuscript in order and did an excellent job of typing it and making all the .necessary revisions. Although too many to name individually, I would like to thank all the manuscript reviewers for their many editing suggestions and other valuable input.
Last and perhaps most importantly, I would like to recognize my wife Nancy for her support throughout the entire process of writing this dissertation.
Definition of Computer-Based
Information Processing .................................... 1
The Development of the Computer 3
The Use of Computers in the
Government Sector .... ................................... 14
Notes ...................................................... 21
SURVEY OF THE LITERATURE OF COMPUTER-BASED INFORMATION PROCESSING IN LOCAL GOVERNMENT .............. 22
Information Technology and the
Early Theorists........................................... 23
The Later Computer Studies ................................. 32
The Litany to EDP in Local Government....................... 40
The Impact of Computer-Based Information Processing on the Management of Local
The Impact of Computer-Based Information Processing on Organization Structure ....................... 58
Departmentation .......................................... 60
Levels of Hierarchy....................................... 65
Span of Control ..................................... 69
The Impact on Certain Tasks and People...................... 71
Impacts on the Job of Top Management ................ 72
Impacts on the Job of Middle Management............... 77
vi i i
Impacts on the Job of Clerks and Supervisors .............. 79
Impacts on the Work Environment.............................. 81
The Impact on Authority and Control ......................... 85
Issues Involving the Management and Control of Information Technology
in Local Government ....................................... 90
The Information Processing Organization and Staff ... 98
Centralization Versus Decentralization of Computer Operations............................' . 101
Location of Computing Facilities in the Organization......................................... 107
The Priority-Setting Process ............................... 112
THE STUDY APPROACH AND METHODOLOGY ........................... 121
Study Purpose............................................... 121
Conceptual Framework ...................................... 122
Research Design ............................................ 123
The Experience Survey ...................................... 125
The Analysis of "Insight-Stimulating" Examples ............. 126
Data Collection............................................. 127
Choice of Population ....................................... 128
Development of the Survey Instrument ....................... 129
Validity and Reliability.................................... 133
Mechanics of Administering the Survey ...................... 135
Description of the Survey Instrument .............. ... 137
Characteristics of the Study Cities ....................... 143
Characteristics of Information
Processing in the Study Cities . !................... 145
Automated Information Processing Tasks
in the Study Cities....................................... 152
Intensity of Computer Usage
in the Study Cities....................................... 157
Personnel Variables in the Study Cities .................. 159
Assessment of User Knowledge of Computer-
Based Information Processing ............................. 161
Pol icy Variables........................................... 163
Organizational Arrangements for Computing .................. 163
The Administration and Control of the Priority-Setting Process ................................... 168
Management of Data Processing Capability ................... 172
Sources for the Development of New Applications ............................................... 176
Potentials and Constraints Regarding Improved Management of Information
Budgetary Support for Information Processing.................................................. 182
The Value of Information Processing Technology to Local Government ...................... 184
The Impact of Computer-Based Information Processing on the Management of Local
Changes in Management ...................................... 188
Impact on Organizational.Structure ........................ 197
Changes in Information Handling ......................... 199
The Usefulness of Information ........................... 200
Information Processing Problem Areas .................... 202
Availability of Data for Analysis of Specific Questions ................................... 203
Communication Between Information Processing Managers and Department Managers ........................ 204
Priority in Using Computer Resources ................. 206
Getting and Keeping Qualified
Information Processing Personnel ...................... 208
Inadequate Space for Information Processing Facilities, .................................. 210
Late Delivery of Information
Processing Products ................................... 211
Department Location of the
Information Processing Unit ........................... 212
Problems in Day-to-Day Operations of the Information Processing Unit ........................ 214
Proliferation of Computers ............................... 215
CHAPTER V .
CONCLUSIONS AND RECOMMENDATIONS FOR
FURTHER STUDY............................................. 219
Conclusions from Analysis of the Variables Involved with Information Processing in
Local Government......................................... 219
Personnel Variables .................................... 223
Policy Variables ......................................... 225
Potentials and Constraints Regarding Improved Management of Information
Values of Computing....................................... 235
The Goals of Information Management
In Local Government ...................................... 237
The Role of Top Management in
Information Processing................ ................ 239
Objectives of the Information
Processing Department .................................... 240
The Development of Microcomputer Policy .................... 242
Conclusions Concerning the Impact of Information Processing on the
Management of Local Government . . '.................. 246
Changes in Information Handling ............................ 251
Information Processing Problem Areas ....................... 252
Summary of Major Conclusions ............................... 253
Suggestions for Future Research ............................ 254
BIBLIOGRAPHY ................................................... 258
A. Dissertation Interview Questions ...................... 276
B. Number of Respondents by City and Job Category .... 293
2.1 A Typology of Information Processing
Tasks Characterization ............................... 42
2.2 Number and Kinds of Changes in the
Organization Structure of 18 Insurance
Companies as a Result of Information
Processing Technology .............................. 62
2.3 Effects on Work Environment Attributed
to Computing by Municipal Employees,
by Role............................................. 83
2.4 U.S. City and County Chief Executives'
Perceptions of Problems with Computing .............. 94
2.5 Impact of Centralization on Computing
Control, Operations and Staff ....................... 106
3.1 Major Issues Addressed in Each
Section of the Survey Instrument .................. 139
4.1 Characteristics of the Study Cities .................... 144
4.2 Information Processing Characteristics ................. 148
4.3 Hardware and Operating System
Configurations ...................................... 150
4.4 Automated and On-Line Applications ..................... 152
4.5 Automated Information Processing Tasks
and the Intensity of Computer Usage
in the Study Cities.................................. 156
4.6 Organizational Arrangements for Computing .............. 164
4.7 Sources of Application Programming and Design . 173
XI 1 1
4.8 The Greatest Value of Information
Processing Technology to Local
Government ........................................ 185
4.9 Managers' Beliefs about the Impact of
Computers on Policy-Making and
Flexibility in Resource Allocation ................. 189
4.10 Managers' Beliefs about the Impact of
Computers on Ability to Control and
Oversee Activities ................................... 191
4.11 Managers' Beliefs about the Impact of
Computers on the Complexity of Local
Government Operations ................................ 193
4.12 Managers' Beliefs about the Impact of
Computers on Costs and Performance
4.13 Managers' Beliefs about the Impact of
Computers on Staff ............................. 197
4.14 Managers' Beliefs about the Impact of
Computers on Organizational Structure ................ 198
4.15 Managers' Beliefs about the Impact of
Computers on Information Sources ..................... 200
4.16 Managers' Beliefs about the Impact of
Computers on the Usefulness of Information .... 201
4.17 Problems Associated with Data Availability ............. 203
4.18 Problems Associated with Communication
Between Information Processing
Departments and Operating Departments ................ 205
4.19 Problems Associated with the Determination
of Priorities for the Use of Computer
Resources ............................................ 207
4.20 Problems Associated with Getting and Keeping
Qualified Information Processing Personnel ....
Tables (conti nued)
4.21 Problems Associated with Space for Information Processing Facilities . 210
4.22 Problems Associated with the Late Delivery of Information Processing Products 211
4.23 Problems Associated with the Departmental Location of the Information Processing Unit . . 213
4.24 , Problems Associated with the Day-to-Day Operations of the Information Processing Unit . 214
4.25 Problems Associated with the Proliferation of Computers . 216
5.1 Characteristics of Computer-Based Information Processing Shared by the Majority of the Study Cities . 221
5.2 Interview Responses with which the Majority of the Managers in the Study Cities Agree Concerning the Impact of Information Processing on the Management of Local Government . 249
2.1 Leavitt and Whisler 1958 Model of
Structural Change from Computing ................... 66
2.2 Alternative Locations of Responsibility
for Providing Computer Services .................... 109
Definition of Computer-Based Information Processing
For the purposes of this study information technology is defined as the technology of sensing, coding, transmitting, translating, and transforming information. Traditionally, the technology has been divided into two general categories hardware and software. The hardware generally consists of a computer mainframe or mini or micro computer, a wide variety of peripheral input and output equipment, and a wide variety of data transmission equipment. The software consists of computer programs, which are generally hierarchical in character, with "executive" programs that "use" other application programs. Information technology does exactly what a great number of people in. modern organizations have traditionally done, that is, store, transmit, and transform information so that it can be applied to the solution of problems. The primary advantage of information technology seems to lie in its great capacity, its high speed, and the reliability of the computer as compared with human beings.
The most useful conceptualization of information technology is the view of the technology as a "package." The technology of computing can be viewed as an interdependent set of four basic
elements. The first element is the technology itself, which is composed of the hardware, systems software and applications software. This first element was discussed in the preceding paragraph. The second element of the "package" is policy. This element consists of the procedures, practices, and organizational arrangements which constitute the managerial approach to the development and use of information technology. The third element is personnel. These are the organizational actors, such as users, computer specialists, and managers, who play an active role in the development of information technology policy and in the use of the technology within the organization. The fourth element is the infrastructure of information technology. The infrastructure is made up of organizations and networks that comprise the broader "system world" of computing. Specifically, the infrastructure includes equipment manufacturers, vendors, software suppliers, professional associations, educational and training institutions, and other organizations that promote or regulate the use of the technology.
Frielich and Scheepmaker argue that there is a difference in kind between modern information technology and all previous technologies (Frielich and Scheepmaker, 1966, pp. 13-16). Earlier technologies were energy converters while information technology is strictly an information converter (ibid). The older technologies were thus extensions of human hands and muscles and were their tools and servants, while modern information technology could be. considered as an extension of the human brain and a partner.
The Development of the Computer
The use of machines to perform arithmetic operations (i.e.,
addition and subtraction) dates back many hundreds of years to the
Chinese abacus which came into use around 450 B.C. It wasn't until
1642 when the French philosopher-mathematician B1aise Pascal
invented an adding machine consisting of ten numbered wheels
connected through an assortment of gears that the next major advance
in computing machinery occurred. In 1671 Gottfried Wilhelm Leibnitz
designed a modified version of the Pascal Machine which could
multiply and divide as well as add and subtract. In 1694 the
Leibnitz designed machine was finally built. With some
modifications, the Leibnitz designed machine was essentially the
same as the modern electro-mechanical desk calculator. The device
was, however, primarily an experimental device until 1820 when
Thomas de Golmar produced it commercially. In 1891 W.T. Ohdner
developed the device to its present form.
At the same time that Thomas de Colmar was working on the production of his calculator, Charles Babbage was designing the first of his machines to calculate with numbers. Babbage's first design was called the "difference engine." This design was for a machine which would evaluate single polynomials, but, as with his later design, the intricate parts needed for the machine could not be produced with the technology then available. In 1833, ten years after his first design, Babbage developed a more complex version which he called the "analytical engine." This design was a
mechanical giant composed of many gears and cranks and would use steam as its source of power. Although the "analytical engine" was -beyond the technological capabilities of the time, the "analytical engine" did supply many of the basic concepts for present-day computers.
The Babbage computer design consisted of a "store" in which the numbers to be used in calculating were stored, and a "mill" in which actual operations were performed. The contemporary analogues of the "store" and the "mill" in modern computers are the "memory unit" and the "arithmetic unit," respectively. The "analytical engine" also theorized the use of punched cards, which would be adapted from the punched cards used in 1806 by the Frenchman Joseph Marie Jacquard who invented an automatic weaving machine controlled by punched cards. Jacquard was not attempting to invent a computer or a mechanical aid to computation, but his idea of using punched cards to hold information and insert instructions and data into the machine would be borrowed by a number of true computer pioneers.
The use of prepunched cards in combination with a "store" allowed a calculating machine to run itself for the first time, at least in theory. Babbage's machine, if it could have been produced to his specifications, would have been able to take numbers from the "store" or produce output.
Charles Babbage's work did not produce any sustained interest in computers or in the use of punched cards. It was not until 1885 that punched cards were used for anything other than creating Jacquard weaves. The work of Charles Babbage was largely
forgotten, not to be rediscovered until after his ideas had been implemented by other men almost a century later.
The next step toward the development of the modern computer occurred in the 1890s when the United States Census Bureau hired Herman Hollerith as a consultant. The United States Constitution requires that a census be conducted every ten years. The census of 1880 took, seven and one-half years to complete. Growth trends made it obvious that the 1890 census Would take even longer to tabulate -eleven years was considered to be a reasonable guess. The future for the Census Bureau appeared bleak with the prospect of falling further and further behind with each succeeding census.
Hollerith's inspiration in providing a solution to the census problem was Jacquard's automatic weaving machine and its punched cards. Hollerith believed that if a pattern of punched holes in a card could represent weaving instructions then similar holes could stand for numbers. As the data from the 1890 census began to come in, Hollerith had it punched into cards. To read and tabulate this data, Hollerith borrowed from the rapidly developing field of electricity. Since paper is a good insulator, cards do not conduct electricity. Hollerith's device simply counted the number of times that electric current flowed through a given wire as the cards were fed one at a time by hand. Each wire corresponded to one possible hole position on the card;' each hole position represented one value of one census statistic. Where there was a hole on the card a metal pin would make contact with the metal plate underneath allowing the current to flow and a count to register. The counts,
when added up, gave the results of the census. In effect, Hollerith's device introduced the binary number concept that has served as a foundation for modern computers. The electric current is either on or off which represents a value of one or zero in a particular field.
Due to Hollerith's invention the 1890 Census was completed within two years. Herman Hollerith had succeeded where others had failed primarily because electronics and manufacturing technology had advanced to the point where Hollerith's ideas could be implemented.
By the late 1920s, card-tabulating equipment had become quite common throughout the United States. In 1935 IBM was successful in obtaining the largest (to that day) card-tabulating equipment contract when the social security system came into being. Significant improvements had, of course, been made, but the equipment was based essentially on Hollerith's original idea.
In 1925, the modern era of mechanical computers was ushered in by Dr. Vannevar Bush of the Massachusetts Institute of Technology. Dr. Bush built a large-scale analogue computer which calculated mechanically with gears and levers like Babbage's machine, although it was powered by electricity. Unlike Babbage's digital design, the analogue computer represented numbers with the number of degrees through which certain gears rotated. The accuracy of the machine was limited by the precision with which the gears could measure an angle.
In 1935, Dr. Bush and his colleagues at MIT built another computer using electricity instead of gears to measure angles.
These machines were the first operating computers with a "mill" and "store," but they still had neither the flexibility nor the accuracy envisioned for the "analytical engine." Two copies of this machine were built under the Work Progress Administration (WPA), one for the U.S. Army's Aberdeen Proving Ground and one for the Moore School of Electrical Engineering at the University of Pennsylvania.
In the late 1930s Howard Aiker, while working on a Ph.D. thesis in physics at Harvard University, began to do a preliminary design on a general-purpose machine which could solve all kinds of problems. In 1939 the IBM Corporation agreed to support Aiker in his pursuit to develop his machine. Aiker, along with four IBM engineers began work on the "Automatic Sequence Controlled Calculator," or MARK I. This machine, when put into operation in 1944, was the world's first fully automatic computer, and it closely resembled Babbage's "analytical engine." The MARK I was a mechanical machine that was powered by electricity and took in instructions and data on a paper tape.
Later versions of the MARK I incorporated many technological innovations. The MARK III used vacuum tubes instead of mechanical relays for computing, and the MARK IV used transistors. The MARK series, however, was not compatible with the advances being made at the Moore School of the University of Pennsylvania, and so the series was not continued.
In the early 1940s, while development continued on the early computers, Dr. Norbert Weiner, of the Massachusetts Institute of Technology, developed the theory of cybernetics. Cybernetics is the study of human control functions and of the mechanical and electrical systems designed to replace them. Dr. Weiner was interested in using computers to emulate human communication. His work established several of the basic principles of the modern computer. Dr. Weiner sought the means to shift the computing function away from the human to the machine. It required the ability to perform multiple operations, make logical decisions on the basis of empirical evidence, and accomplish this without human intervention.
Dr. Weiner's architectural design for the computer involves five concepts or principles. He believed that computers should:
1. Have a numerical central adding and multiplying apparatus using registers for execution.
2. Use electronic circuitry rather than gears or mechanical
3. Use binary, or base 2, arithmetic rather than decimal arithmetic.
4. Not involve human intervention programs should be loaded in the machine.
5. Contain a machine-loadable apparatus to store, read, and erase data, and be removed for storage of new data.
The first fully electronic computer was developed in the early 1940s by Dr. John Mauchley and Dr. J. Presper Eckert at the
Moore School of the University of Pennsylvania. The computer they developed was called the "Electronic Numerical Integrator and Computer" or ENIAC. Unlike the MARK I this was not to be a general-purpose computer, but one designed solely to work out ballistics tables for the U.S. Army in its war efforts. The ENIAC computer was, in fact, financed by the Army. The computer itself weighed more than 30 tons, contained nearly 19,000 vacuum tubes, and literally took up rooms. The ENIAC was considered an advanced computer because it incorporated capacity to store program calculations, was completely electronic, and had no moving parts.
Its major limitation, other than its sheer size, was that it could process only one program or problem at a time and the switch to another program or problem was a major effort.
While the ENIAC computer was being developed in the United States, the British were also working on computers for their war effort. Professor M.H.A. Newman of the University of Cambridge developed the Colossus series of computers that first became operational in December of 1943. Although the control mechanism worked on the principle of an exterior, plug-in systems design, data was entered from punched paper tape, and stored programs were used to convey instructions through the value concept.
In 1945, Dr. John Von Neumann of the Institute of Advanced Study at Princeton University was retained by the Moore School to advance work on the ENIAC. Dr. Von Neumann, building on the work of Dr. Weiner and others, published a paper entitled "Preliminary Discussion of the Logical Design of an Electronic Computing
Instrument." This work formed the basis for the development of a stored-program computer using the binary number system, the Electronic Discrete Variable Automatic Computer (EDVAC). The EDVAC computer was completed in 1950, but the first stored-program computer to be put into operation was the Electronic Delay Storage Automatic Calculator (EDSAC), which began operating at the University of Cambridge, England in 1949.
With the publication of the Von Neumann paper, the era of major innovations in computer design came temporarily to an end. Modern computers use both the stored program and the binary number system. The construction of computers was, however, becoming industrialized. In the early computer days each machine had its own name and its own unique features. With the UNIVAC and the later IBM 701 and 650, the production of computers became standardized and big business.
Dr. John W. Mauchley and Dr. J. Presper Eckert resigned their posts at the University of Pennsylvania and founded the first computer company devoted exclusively to the production of computers. The Eckert and Mauchly Computing Company, formed in 1948, was later sold to the Remington Rand Corporation and in 1951 sold a computer named the UNIVAC I to the Census Bureau. The machine was retired in 1963 and now rests in the Smithsonian Institute.
Prior to 1950, companies, such as IBM, Honeywell, General Electric, and RCA, were not convinced of the commercial applicability and feasibility of computers. It was not until 1951
that IBM entered the computer market when it produced the IBM 701, the first of a long series. The IBM 701 computer was sold to the federal government primarily for large mathematical computations.
By 1952 UNIVAC still had the upper hand with its more efficient computer.
While UNIVAC and IBM progressed through the early 1950s with the first generation of computers, IBM remained dominant in the business calculator field. IBM had more than 5,000 installations for its business calculators by 1957. This provided a base for dominance in computers. IBM's strong marketing strategy moved customers from calculators to their computer line. By 1956 IBM had taken a marketing lead that it has yet to relinquish, selling the IBM 650 and 700 lines of computers. Other key competitors included Monroe, NCR, Burroughs, RCA, Underwood, and, of course, UNIVAC.
The 1950s marked the computer's first generation. The precise dates of this generation are difficult to pinpoint.
However, the beginning is set somewhere between 1950 and 1954 and the end generally falls somewhere around 1958 or 1959. What really distinguishes the first generation from subsequent generations is technology. The key electronic component of a first-generation computer was the electronic tube. By today's standards, these machines were very slow, being capable of executing approximately
1,000 instructions per second. These computers were also quite small in terms of capacity, holding perhaps 10,000 to 20,000 characters of data in their main memories.
The second generation of computers began with the invention
of the transistor, a device which won the 1956 Nobel Prize for three
American scientists. By 1957, Burroughs had developed a fully transistorized computer for the Air Force. The higher speed and greater reliability of transistors marked the end of the electronic tube-based first generation computers, which were virtually obsolete by 1960. Second generation computers were physically smaller, much faster (1,000,000 as opposed to 1,000 instructions per second) and much more reliable than the first generation models. The improved reliability of the second generation computers, which derived from solid-state electronics, made these machines much more attractive to the business market. By 1965, the value of installed computers had risen to an estimated three to four billions dollars.
Beyond the.field of electronics, the second generation saw a tremendous improvement in techniques for using computers, such as new operating systems and time-sharing. New computer languages were also introduced at this time which made the programmer's job easier. COBOL and RPG were the popular commercial languages supported by most manufacturers in an attempt to standardize. These languages were powerful tools for the second generation programmer and analyst, however, the trade-off was that large computer memories were required to take advantage of these "high-level" languages.
The third generation of computers, which began in approximately 1965, was marked by major advances in the area of printed and integrated circuits. As a result of these advances, computers became even more compact, faster, more reliable, and less
expensive. The IBM System/360 and IBM System'/370 computers became the dominant machines of the third generation. By the mid-1970s, the value of installed computers had risen above 24 billion dollars.
Perhaps one of the most important trends in third-generation computers was the rapid development of minicomputers and microprocessors. These small, relatively inexpensive machines were often dedicated to a specific task or application, such as police record keeping applications, telephone system management, or remittance processing for billings. With the minicomputer's ability to handle complex applications reliably at reasonable cost, new markets opened up and produced soaring sales for the manufacturers.
Many observers believe that we are now well into the fourth generation of computers with major advances in applications and techniques which include multiprogramming, multiprocessing, microcoding, and virtual memory. The growth in the number of computers has been almost exponential at the same time that -computers have become more powerful in terms of computing capacity. Over the past several years, the speed and reliability of computers has increased tremendously, while the size and, most importantly, the cost of these machines have tended to drop just as dramatically (Davis, 1978, p 15). Very few products in the history of manufacturing have ever followed this trend.
In the future there is little doubt that computers and related software will attain greater speed, processing power, and a tremendous profusion of new applications. Time Magazine, in its January 3, 1983 issue, declared the computer as the "Machine of the
Year" for 1982. Time decided that 1982 was the year of the computer because it had become an integral part of every American's life.
With the advent of the "personal computer" it is now possible for average citizens to afford their own computer for their own uses. Industry estimates the number of personal computers in use by the end of the century will run as high as 80 million (Friedrich, 1983, p. 16). A knowledge of the computer has now become essential to the modern educated man or woman.
The Use of Computers in the Government Sector
Computer-based information technology was first introduced in the Federal government in the late 1940s, primarily for military applications. The adoption of general use computers by the Federal government began in earnest in the 1950s. Today the United States Federal Government is the largest user of computer-based information technology in the United States and in the world (Sanders and Berkin, 1980). State governments have not lagged far behind the Federal government. By the mid-1960s every state government was utilizing computer-based technologies (ibid).
Computers were first introduced into U.S. local governments in the 1950s. Since that time their use and application has grown continuously. In 1975, more than half of all U.S. cities and counties with a population of over 10,000 used computers in one way or another (Kraemer, Dutton, and Matthews, 1975, p. 5). In a study conducted in 1976, Kraemer and King noted that the extent of computer use among local governments was directly related to the
size of the local governments as measured by population (Kraemer and King, 1977, Vol. 1, p. 23).
By 1975, ninety-nine percent of all large municipalities (cities over 100,000 population) were using electronic data processing. Ninety-two percent of all medium-sized cities (50,000 -
100.000 population) used computers and data processing. Only in small municipalities (those between 10,000 and 49,000 population) was there a significant proportion of governments not utilizing data processing in municipal operations (Kraemer, Dutton, and Matthews, 1975, p. 5). According to Kraemer, Dutton, and Matthews, computers were being used by fifty-eight percent of municipalities between
25.000 and 50,000 population and thirty-six percent of municipalities between 10,000 and 25,000 population in 1975 (Kraemer, Dutton, and Matthews, 1975, p. 5). The authors further noted that there was a continuously high rate of EDP adoption from the 1950s to 1975, with almost half (48 percent) of all adoptions occurring since 1970 (ibid).
The advent of computing in local government also coincided with a push for financial reforms, including performance budgeting, the uniform chart of accounts, cost accounting, and the establishment of centralized support services (Danziger, Dutton, Kling, and Kraemer, 1982, p. 114). Computing directly supported both the budgeting/accounting reforms and the move toward centralization. The accounting reforms were physically easier to accomplish because of the computer's capacity for sorting,, calculating, and printing information in a large variety of
formats. The centralization reforms were supported by the computer because it allowed the administrative reformers to claim that centralized, computerized accounting/finance would provide rapid turn-around for department reports, and instantaneous access to computerized information (Danziger, Dutton, Kling, and Kraemer,
1982, p. 115).
Computers were first adopted in government operations, in almost all cases, for handling routine operations, particularly in the finance function. The one major exception to this general trend was in the military functions of the federal government. Since the finance unit usually was the greatest user of computer-based technology in state and local governments, computer operations were most often located in that department. As computer use expanded into other government functions, the computing operation has tended to move into an independent department, and sometimes into operating departments (Kraemer, Danziger, and King, 1976, p. 3). The mainstream of local government computer usage still remains in relatively straightforward information processing activities, but there have been many experiments with more sophisticated uses such as data banks, urban models, and integrated municipal information systems.
In terms of technology, large computer mainframes have predominated in computing use by local government since the first computers were introduced in the early 1950s. This tendency toward the use of large computer mainframes for most computer-based tasks has been due to the propensity of local governments, at least the
larger ones, to consolidate computing in a single, centralized computing installation (King, 1982, p. 27).
Since the mid-1970s there has been a trend toward adoption of minicomputers, spurred largely by federal support for adoption of computers in law enforcement. There is no standard definition of what a minicomputer is, but generally minicomputers have less core memory capacity than large mainframes. They are also generally smaller in terms of physical size and usually cost less than mainframes. According to King, it is expected that the adoption of minicomputers will increase throughout the 1980s, both in larger local governments where they will be used for specialized applications, and in smaller local governments that may be adopting computer applications for the first time in finance and other functional areas (King, 1982, p. 27).
The impact of minicomputers in most small cities and towns will probably parallel the impact that larger computer systems have had in the larger cities. Even though the greatest impacts should be in the smaller cities, minicomputers, as noted above, will certainly impact large and medium-sized cities, which have already adopted some type of computer system. The low initial cost of minicomputers has appeal to user departments within the larger governments. Such user departments are currently adopting their own minicomputers and have begun building their own information processing installations, either independent of the existing central installation, or in coordination with the central information processing unit. It is not uncommon in the larger cities to have a
central information processing installation serving general municipal functions, with minicomputers located in satellite installations serving only one department, often the police. In some cities the police installations are totally independent of the central information processing installation, while others are physically separate, but centrally managed. The development of departmental computing, which is now more possible because of minicomputers and microcomputers, could signal the start of a major decentralization trend within local government computing.
Despite the generally optimistic predictions of many theorists, some users have discovered that computer technology, like other technologies, is a two-edged sword. The benefits of automation have been fewer than anticipated and the costs have been far greater (Kraemer, Dutton, and Northrop, 1981, p. 11). John Leslie King has noted that,
Most forecasts of the near future of information technology in governments and in other organizations rest on the assumption of continued phenomenal progress in computing hardware technology similar to that experienced over the past twenty years. (King, 1982, p. 25).
King goes on to state that,
Such forecasts are plausible, but there is not an a priori reason to expect that they will come to pass'in the near future. Among other things, there is no evidence that such visions represent real advancements over the way local governments presently do things. (King, 1982, p. 25)
Although the costs of computing may be greater than anticipated by local governments and the benefits fewer, local governments have made a substantial financial investment in computing. Investment in computing among local governments is about
one percent of annual operating budgets per year. This figure varies somewhat by size of government, with smaller governments spending approximately .5 percent of their operating budgets and larger governments slightly over one percent (Kraemer and King,
1977, vol. 1, p. 23). It should be noted that these figures reflect only direct expenditures on computing, including hardware, software, maintenance and information processing staff. They do not reflect the information processing-related costs of user departments. If these costs were taken into account, the total average information processing expenditures could be as high as two percent of operating budgets.
Personnel devoted to information processing operations average less than one percent of total local government personnel, regardless of size. Thus, the size of an information processing staff varies proportionately with the size of the government. In all cases, according to Kraemer and King, about 25 percent of information processing personnel are analysts and programmers. Information processing staff size in user departments is usually smaller than staff size in the information processing department itself (Kraemer and King, 1977, Vol. 1, p. 24).
This brief history of the adoption of computer-based information technology in government illustrates the rapid infusion of the technology into the government sector and the current reexamination of its role. Without question computers have long been recognized as a major tool for improving the management and use of information in local governments. Most local governments now use
computers, and those that don't, primarily the very smallest of local governments, will almost certainly adopt the technology (Kraemer, Dutton, and Matthews, 1975, p. 7). The adoption of computer technology is a matter of productivity improvement. Unlike many businesses, local governments generally cannot hope to improve productivity through automation of production processes and increased use of special tools alone. Local governments must generally rely on efforts to improve planning and decision making, which will in turn result in better scheduling, staffing, purchasing, and management of local programs.
The common denominator in planning and operational decisions is accurate, timely and relevant information. Both the need for operational information and mandated information-keeping activities demonstrate the importance of information to local governments. The use of computer-based information technology has improved the ability of local governments to deal effectively with information and, therefore, computers have earned an indispensable position as a major tool in the management of local governments.
NOTES CHAPTER I
1. There are numerous sources available concerning the history of computer-based information processing. The primary sources used in this study, all cited in the bibliography, are: Whisler, 1970a; Kraemer and King, vol. 2, 1977; Davis, 1978; Sanders and Birkin, 1980; and Clowes, 1982.
2. The three American scientists who won the 1956 Nobel Prize for physics for the invention of the transistor were: John Bardeen, Walter H. Brattain, and William Shockley.
SURVEY OF THE LITERATURE OF COMPUTER-BASED INFORMATION PROCESSING IN LOCAL GOVERNMENT
As noted in the introductory chapter, the use of information processing technology in the public sector at all levels of government is now widespread.. It is almost impossible for citizens to interact with government agencies at the federal, state, or local levels without information processing technology playing a role in the interaction. There seems to be little doubt that computerized systems are here to stay in the public sector and that these systems will play an increasingly important role in the administration of government. The primary purpose of this study is to explore, describe and document how managers at the local level of government administer and control the use of computer-based information processing in the agencies or units of government under their jurisdiction.
In order to understand better the nature of administering computer-based information processing systems, this chapter will review some of the previous work that has been done concerning management and computer-based information processing. Specifically, this chapter will review the work of the "early theorists" who studied.the impact of the new technology of computer-based information processing on the administration of organizations that
adopted the technology. Next, the impacts of computer-based, information processing on the administration of local governments will be discussed. Finally, issues involving the administration and control of computer-based information processing in local government will be explored.
Information Technology and the Early Theorists
Management theorists and practitioners first began to examine the interrelationship between management and computer-based information processing in the late 1950s just a few years after the introduction of the first commercial computers. In the late 1950s and early 1960s, the number of commercial computer installations was very limited. There was very little known about the impact of the technology on the organizations that employed this new technology. The early writers in the field of computer-based information processing consequently confined themselves, to a large extent, to speculations and predictions concerning the impact of the new technology on organizations. The rapid development and diffusion of computer-based information processing in its early years is not very surprising in light of the predictions of the early theorists concerning the beneficial impacts of computers on organizations (Leavitt and Whisler, 1958; Hoos, 1960; Simon, 1965).
In their now famous article, "Management in the 1980s," Leavitt and Whisler noted that the "new technology" was so new that it did not yet have a single established name. They suggested the name "information technology" and defined it as being composed of
several related parts, including techniques for processing large amounts of information rapidly and epitomized by the high-speed computer (Leavitt and Whisler, 1958, p. 41). The term "information technology" is still in use today to describe the practice of computer-based information processing.
One of the earliest articles that speculated on the impact of information technology on organizations was written by Robert Slater, a manager who had just completed one of the first large-scale applications of computer systems to insurance office operations (Slater, 1958). Slater made several significant predictions in his article. They were as follows:
1. A shift of operations to the EDP (electronic data processing) department with other departments becoming, in effect, planning departments.
2. Fewer departments in the firm.
3. Fewer management levels in the firm.
4. More power and influence for the data processing executives.
5. Systems specialists moving from firm to firm, interested in effective procedures but not involved in goal setting.
6. More emphasis upon goal setting as the primary executive function.
7. A change in traditional career patterns, with lower level experience by itself no longer adequate preparation for upper level executive activities.
8. University training in procedures for systems specialists and on-the-job training in goal setting for executives (Slater, 1958).
Slater's article did not make any mention of short-run worker displacement or longer-run hidden displacement, except as the latter was implied by the prospect of fewer levels of management and a reduced number of departments, and thus a fewer number of managers needed to oversee the operations of an organization.
Unlike Slater, who was a practitioner, Leavitt and Whisler were academicians without Slater's managerial experience. However, their predictions, made in their article which appeared shortly after Slater's, were substantially in agreement with Slater's predictions. Leavitt and Whisler's major predictions were as follows:
1. Greater organizational centralization.
2. A reduction in the number of middle managers and more programming of the jobs of the remaining middle managers.
3. The creation of a "block" to the movement of middle managers within the organization.
4. More emphasis in top management on innovation and the search for new goals.
5. A change in the managerial career pattern from the intrafirm ladder to the interfirm movement.
6. University training for top managers and on-the-job training through high level apprenticeships (Leavitt and Whisler, 1958, pp. 41-48).
The computer was viewed as a technological breakthrough by many of the early theorists (Slater, 1958; Leavitt and Whisler;
1958; Hoos, 1960; Simon, 1960 and 1965; Meyer, 1968) that would revolutionize the management and performance of organizations. A number of these early theorists also predicted that computers would offer improved information for decision making (Leavitt and Whisler, 1958; Simon, 1965; Meyer, 1968). Computers would provide more accurate and timely information, a broader range of information, and more easily obtainable information, according to the theorists.
In 1960, Herbert Simon, a highly respected academician, forecast in detail the computer's impact not only on managers, but also on other members of the organization (Simon, 1960). He predicted:
1. Fewer manual and clerical tasks concerned with in-line production, except for those requiring flexible eye-and-brain coordination.
2. Relatively more maintenance tasks.
3. Fewer stressful interpersonal relations, such as supervising and expediting the actions of others.
4. Certain managerial tasks taken over by computers: (a) solving well-structured problems (rapid and early takeover), (b) solving ill-structured problems (less rapid takeover), (c) supervising (very slow substitution).
5. Activities of middle management more completely
automated than others.
6. Management attention shifting to the design and modification of large systems (Simon, 1960, pp. 17-55).
Simon's predictions both support some of the earlier ones and add some new ideas.. Perhaps the most significant departure in Simon's predictions from those of Slater and Leavitt and Whisler concern his opinion about the primary activity of post-computer top level management. Simon predicts it will be systems design instead of the earlier forecasts of goal setting and the search for new goals that dominates the time of these managers in organizations.
Ida Hoos, also writing in 1960, reported on a study of computer impacts in the office (Hoos, 1960). Her article was based on a two-year study which began in 1957 of 19 organizations in the San Francisco Bay Area that had introduced electronic data processing. The Hoos study was perhaps the first study of electronic data processing in the office environment to include large and small private business concerns and a few government agencies. Banks, insurance companies, public utilities, manufacturers, distributors, and processors were among the firms studied. The Hoos study identified a few major organizational changes due to the introduction of electronic data processing that she believed were trends that would persist. These trends were:
1. Worker displacement.
2. Recentralization of authority.
3. Some middle management jobs downgraded.
4. Departments and jobs consolidated (Hoos, 1960,
The trends reported by Hoos were very similar to the predictions of those who had written earlier on the subject.
Although most of the early theorists appeared to treat the impacts of computing as largely apolitical, a few, including Hoos, predicted that computer-based technology might have power payoffs, in addition to more technical benefits (Leavitt and Whisler, 1958; Hoos, 1960; Downs, 1967; Meyers, 1967). Basically, power shifts were predicted to occur as a result of the information obtained from computerized systems.
John Burlingame, writing in 1961, was one of the few early theorists who disagreed with the argument that information technology induces organizational centralization (Burlingame,
1961). Burlingame believed that the new technology would make organizational decentralization much more effective by providing lower-level managers with the appropriate information when they needed it to make decisions. Burlingame's central thesis was that, since there were sound reasons for decentralization which were entirely unrelated to computers, decentralization should be an expected consequence of computer application. Burlingame did note, however, that computers might make an already centralized organization more effective; and, as a consequence, such an organization might decide to centralize even further (Burlingame, 1961). In effect, Burlingame was arguing that information technology would leave the control structure of an organization as a matter of choice for the organization's executives.
Another "dissenter" among the early theorists was Melvin Anshen. Anshen was in basic agreement with Burlingame concerning the issue of centralization versus decentralization; however, he did believe that in large organizations centralization, or at least a slowdown in the trend toward decentralization, would occur as a result of the introduction of information technology (Anshen,
1962). Anshen also disagreed with the majority of the early theorists concerning the fate of middle-level managers. Anshen did believe that the biggest changes in job structure due to the introduction of information technology would occur at the middle management levels; however, he believed that middle level management jobs would be enlarged and upgraded rather than the reverse (Anshen, 1962). Anshen's other significant predictions concerning the impact of information technology on organizations were more in line with those of the other early theorists. They were:
1. Management time would be freed for "creative" thinking, identifying problems that previously were not even known to exist.
2. The scheduling function would be located at the top of the organization.
3. It would be necessary for all managers to understand the capabilities and uses of computer systems (Anshen, 1962, pp. 56-83).
There were two other important articles written in 1962 which influenced future perceptions concerning information technology. One of these articles, written by Allen Kraut, a social psychologist, reviewed the impact of electronic data processing on workers in organizations (Kraut, 1962). The second article, written
by L. R. Fiock, Jr., concerned the dangers inherent in the use of electronic data processing (Fiock, 1962). Kraut's work was important because it was directly concerned with the effect of introducing electronic data processing in an organization on workers and their jobs. Kraut noted that the question that had generated the greatest amount of concern and speculation was the question of how many clerical workers would lose their jobs due to automation. His conclusion was that although the introduction of computers makes many employees redundant, especially clerical employees, there were very few layoffs (Kraut, 1962, p. 39). Kraut explained that layoffs were infrequent due to normal attrition and the absorption of excess personnel in ways that made layoffs unnecessary (ibid).
In conclusion, Kraut found that although there were not a great many layoffs due to automation, it did permit a reduction in the size of the work force. He also found that as a result of automation there was a substantial amount of upgrading of employees' jobs throughout the organization, but not very much among the employees whose work was directly affected by the change to electronic data processing (Kraut, 1962, p. 43). Kraut's article served as the first basic primer for managers in order to help them deal with any conflict or discontent that might arise from the transition from manual systems to electronic data processing.
The article by L. R. Fiock, Jr., which appeared in 1962, entitled "Seven Deadly Dangers in EDP," was one of the first to point out the perils of the new technology of electronic data processing. Fiock pointed out that the new technology could produce
tremendous savings in time and money; however, as he further noted, the new technology required skill, better planning, and more precise control to administer with the price of failure being dramatic (Fiock, 1962, p. 129). The basic thesis of Fiock's article was simply that if businessmen were more knowledgeable about some of the dangers involved in setting up and evaluating an EDP installation, then maybe they would be better able to protect their organizations from a waste of time and money (Fiock, 1962, p. 130).
Fiock identified the "seven deadly dangers" as: poor procurement; ignorance of procedures; "service" over "control"; middle-management resistance; inadequate EDP staff; poor staff location; and poor evaluation practices (Fiock, 1962). Fiock concluded that managers must first of all establish some definite goal for their EDP group and set up a competent organization to direct all the facilities toward the achievement of this goal. He further noted that the entire organization must be aware of the goals for EDP and what responsibilities are borne by the EDP staff and which responsibilities are to be borne by other units of the organization (Fiock, 1962, p. 136). A number of the issues raised by the Fiock article, such as centralization versus decentralization of computer resources, the location of the EDP'staff, and evaluation criteria, are still major concerns of managers and theorists alike
The Later Computer Studies
In the 1960s and the early 1970s, the study of computer systems and computer system impacts intensified dramatically. The use of information technology in organizations had become so widespread that researchers had little difficulty in selecting organizations which employed the technology for their intensive studies. Researchers and theorists in the study of computer-based information processing were moving out of the early era of speculation and predictions of computer use and impacts. The widespread use of the emerging technology made it possible and practical for researchers to study actual cases of computer use and computer impacts in subject organizations.
For the most part researchers in the 1960s and early 1970s studied private business firms rather than public sector organizations. The large majority of these studies were conducted through the use of interviews with top managers, middle managers, and other professional level employees. A small minority of these studies employed questionnaires or other survey methodologies, and some used these research techniques in conjunction with interviews. The common thread in almost all these research efforts is that they were based on the perceptions of the respondents and therefore relied heavily on the premise that the perceptions of the respondents were "in line" with the reality of computer use and computer impacts in the subject organizations. The research studies themselves ranged in size from small, intense studies of two or more
firms lasting a few months to studies of more than 170 different firms using questionnaires and interviews and lasting more than two years.
As might be expected, the computer studies of the 1960s and early 1970s tended to confirm some of the predictions of the early theorists and to discount others of these predictions. A number of the studies confirmed the prediction that information technology would first be used by firms employing the technology to computerize routine clerical operations (Garrity, 1963; Knight and Miller, 1968; and Churchill, Kempster, and Uretsky, 1969). The Whisler study also notes that clerical jobs tended to be routinized and supervisory jobs enlarged in areas where information technology was employed (Whisler, 1970). Many of these studies also confirmed that the accounting function of the subject organizations was by far the most likely function to be computerized first (Knight and Miller, 1968; Baum and Burack, 1969; Churchill, Kempster, and Uretsky, 1969; and Ruskin, 1973). The tendency to computerize the accounting function first led to the placement of the electronic data processing group under the supervision of the finance department in many organizations.
Many of the first computer installations experienced significant "problems" with the use of the new resource. In effect, the computer was not living up to the expectations of many managers and users. Garrity, in his 1963 study, noted that most major computer-related problems were managerial and organizational in scope (e.g., the organizatonal location of the data processing
department) rather than technical problems with the hardware or software development (Garrity, 1963). Another study determined that one of the significant problems associated with the management of the computer resource was the tendency for the management of the resource to be characterized by isolation (Churchill, Kempster, and Uretsky, 1969). The authors of this study concluded that there was
very little attempt made by most organizations to apply broadly applicable management knowledge and techniques to the new computer resource. The lack of top management and appropriate user management involvement in the management of information processing technology has been a continuing problem into the decade of the 1980s. The Garrity study and later studies have shown that the level of computer systems development is positively related to:
- Leadership provided by top management
- The use of planning and control tools, including the use of regular management audits of computer activities
- The caliber of the computer systems staff
- The involvement of operations management in the design of computer systems (Garrity, 1963; Taylor and Dean, 1966).
A number of the early studies focused on management decision making and its relationship to the introduction of computer-based information processing in the subject organizations. In 1970, Whisler used a questionnaire to survey managers in twenty different firms concerning computer impact on decision making. Whisler found that decision-making activities tended to move to higher levels in the subject organizations and that control was increasingly
centralized (Whisler, 1970). W. E. Reif, in his 1968 in-depth study of three business enterprises, also reached the conclusion that decision making becomes more centralized as^a result of computerization (Reif, 1968).
Two other important conclusions of the Whisler study were that decisions within the subject organizations tended to be made on a more "rational" basis and that some "inflexibility" was introduced into decision-making activities as a result of increasing computerization (Whisler, 1970). Basically, computerization made it possible to structure many decisions by providing needed data in an organized fashion. This resulted in more "rational" decisions since decisions on a given subject were based on organized sets of data, while at the same time causing a certain amount of "inflexibility" in the decision-making process since pertinent data that was not computerized tended to be ignored in the process. Although the Whisler and Reif studies indicated that decision-making activities tended to move to higher organization levels and become more centralized, the 1969 Churchill, Kempster, and Uretsky study found that there was little use of computer resources at high management levels to assist in making strategic decisions (Churchill, Kempster, and Uretsky, 1969). In fact, top managers in both the public and private sector have tended to make important strategic decisions without the direct aid of their organization's computer resources. This fact appears to be due to the unstructured nature of strategic decisions in terms of the number and variety of the different variables that are considered in the making of strategic decisions.
There has been some progress made in recent years in the development of various computer simulation models to aid top level managers in strategic decision making, but the models are generally not yet sophisticated enough for managers to place a high level of confidence in them.
In 1968, Sollenberger completed a study of fifteen different business firms and found that the use of computer resources did provide improved decision-making potential. The improved decision-making potential was found to be the result of several factors. First, the number of factors to be included in decision making could be increased; secondly, the computer increased the availability of data; thirdly, increased time would be available for decision making due to the speed of the computer; and fourthly, there was an improvement in the reaction time to problems (Sollenberger, 1968).
One of the major predictions of the early Leavitt and Whisler study in 1958 concerned the reduction of middle management positions. This prediction was contradicted by some of the later computer impact studies. In 1970, R. S. Jackson studied the impact of computerization on middle level managers in five firms. He found that as a result of computerization, middle managers' jobs are expanded in scope and have a broader perspective (Jackson, 1970).
He also found that middle managers' decision-making capabilities are expanded and that their jobs tended to require increased specialization with a substantial increase in skill levels needed to perform their jobs (Jackson, 1970). Jackson's findings tended to be
supported by two earlier studies, one by Lee in 1968 that found that jobs require a higher level of "mental application" and "job knowledge" (Lee, 1968) and a second study by Shaul in 1964 that found that there was an expansion in scope of managers' jobs (Shaul, 1964). In fact, Shaul also concluded that there was no evidence of a reduction in the number of managers in his subject organizations (Shaul, 1964). The only study that tended to support the Leavitt and Whisler prediction concerning middle managers was the 1968 Reif study which found a reduction in the number of middle managers in the three organizations he studied (Reif, 1968).
One other important area of focus of the early research into computer system impacts concerned the impact on management jobs themselves. In a 1968 study involving intensive interviews with 49 managers in two business firms, Lee found that managers had a generally favorable attitude toward the new work conditions brought on through the advent of computer technology (Lee, 1968). However, the managers in Lee's study also noted that due to the introduction of computer technology there were more "deadlines" to be met, there was less control over the "work place," and there was a large increase in inter-departmental contacts (Lee, 1968).
In the 1972 study by Guthrie, the findings were similar to those of Lee. Guthrie found that managers in the subject organizations generally perceived that computer systems development would have a positive effect on job satisfaction (Guthrie, 1972). However, Guthrie also found that managers in his study were concerned with the potential problem of "information overload" as a
direct result of the computerization of increasing numbers of activities within their organizations (Guthrie, 1972).
In terms of management activities, the 1964 study by Shaul of 67 top and middle level managers in eight firms, conducted by interview methodology, found that managers spend more time on planning activities, supervision, and the training of subordinates after the introduction of computer technology (Shaul, 1964).
Shaul's study also found that managers, as a group, spend less time on control activities after the introduction of computer technology (Shaul, 1964). Shaul's findings were supported by Jackson's 1970 study in which he determined that middle managers spent more time on planning functions and less time on control activities after the introduction of computer technology (Jackson, 1970).
Shaul also found no evidence that managers' jobs were becoming highly structured as a result of computerization. In fact, Shaul found that there was an enhancement in managers' job status as a result of computerization (Shaul, 1964). This apparent enhancement in job status was explained by Whisler in terms of a raise in the perceived level of skills needed by managers to perform their jobs in organizations subject to computerization (Whisler, 1970).
In terms of organizational power shifts as a result of computerization, only the 1968 study by Reif addressed this issue directly. Reif found that staff groups tended to become more powerful as a result of computerization at the expense" of line groups (Reif, 1968). This finding is due, in large part, to the
fact that staff groups, particularly in the areas of accounting and finance, tended to be the first groups in the subject organizations to use computers. Line groups in many organizations were the last to employ computer technology, but that situation has changed markedly in the interviewing years since the early studies were conducted.
In the early years of computerization, there was a tendency on the part of many organizations to centralize the data processing function in one department, often under the direction of the finance department (Sollenberger, 1968; Whisler, 1970). However, the 1966 study by Taylor and Dean found no relationship between the effectiveness of computer usage and the centralization or decentralization of computer activities (Taylor and Dean, 1966).
This issue has been hotly debated by many later theorists and empirical researchers and will be explored in more depth later in this study.
The first two sections of this chapter have reviewed the work of the early theorists and later researchers in the field of computer-based information processing. The focus of the large majority of these studies was on the impact of information processing on the management of organizations or on the interrelationship between management and information processing technology.
Perhaps the single most influential study of information processing during this period was conducted by Leavitt and Whisler and published in their 1958 article entitled "Management in the
1980's" (sic). This article was cited by many of the later researchers in their own studies. However, many of Leavitt and Whisler's predictions about the impact of the technology on organizations, such as their prediction that the ranks of middle management would shrink dramatically as a result of the introduction of information processing technology, have been largely discredited by the findings of later researchers.
Two issues concerning the impact of information technology on management, organizational issues and control issues, were addressed in many of these studies (Slater, 1958; Leavitt and Whisler, 1958; Hoos, 1960; Burlingame, 1961; Fiock, 1962;
Garrity, 1963; Taylor and Dean, 1966; Lee, 1968; Reif, 1968; Whisler, 1970). In terms of organizational issues, there was little agreement on the organizatonal location of the information processing department or on whether information processing activities should be centralized or decentralized. However, in terms of control, a number of researchers stressed the need for better planning and supervision in order to receive maximum value from the investment in information processing technology.
The following sections of this chapter review the literature concerning the administration and control of information processing technology specifically as it relates to local government.
The Litany to EDP in Local Government
There is no question that during the 1960s and 1970s a variety of "information processing tasks" in local government
emerged where the use of computer-based information processing technology was particularly valuable. The computer was capable of providing data that was:
- Stored efficiently and economically
- Easily retrieved and transferred.
Moreover, the computer provided the means to manipulate and analyze data at a level of sophistication, speed, and magnitude that was not possible with the manual systems used prior to.the computer.
Table 2.1 presents several types of information processing tasks where EDP has made fundamental and valuable contributions to the activities of city and county governments.
A TYPOLOGY OF INFORMATION PROCESSING TASKS CHARACTERIZATION
1. Record Keeping Activities which primarily
involve the entry, updating, and storage of data, with a secondary need for access; .the computer facilitates manageable storage and easy updating for nearly uniimited amounts of i nformation.
Inventories, such as registration files and land use files; stati sti cs-keepers, such as Uniform Crime Reports data; throughput systems, such as accounting, ledgers.
2. Calculating/ Printi ng
3. Record Searching
Activities which primarily involve sorting, calculating, and printing of stored, data to produce specific operational outputs; utilizes the computer's capabilities as a high-speed data processor.
Activities where access to and search of data files is of primary importance; by defining parameters, relevant cases can be retrieved from a file with speed and comprehensiveness; on-line capability of computer is particularly useful.
Payroll processing, uti1ity billing, preparation of mai1-ing 1i sts, simple budget preparation.
Regional, state, and national wanted warrant files among police agencies; parking ticket "scofflaw" systems; jury selection.
Activities which involve reorganization, reaggregation, and/or analysis of data; the computer is used to link data from diverse sources or to summarize large volumes of data as management and planning i nformation.
Social services information and referral systems; program budgeting systems; geoprocessing systems; such as ACG/DIME.
TABLE 2.1 (continued)
5. Sophisticated Analytics
Activities which utilize sophisticated visual, mathematical, simulation, or other analytical methods to examine data; the special capabilities of computers make possible the manipulation of data about complex, interdependent phenomena.
Computer mapping and graphics systems such as SYMAP, regression models to estimate, the appraised value of real property, planning simulation models, revenue and expenditure forecasting.
6. Process Control Activities which approximate a cybernetic system; data about the state of a system is continually monitored and fed back to a human or automatic controller which steers the system towards a performance standard; the computer's capability for real-time monitoring and direction of activities is utilized.
Police, fire and ambulance dispatch; budget monitoring and control; traffic signal control; water and power distribution control.
Source: Danziger. "Computers, Local Governments, and the Litany to EDP." Public Administration Review, Jan./Feb. (1977).
It seems quite clear that, under appropriate conditions, almost all local governments should be able to make effective and beneficial use of EDP for various information processing tasks. However, as James Danziger pointed out in 1977, there had been little effort by practitioners and researchers to specify either the full range of impacts of computer-based information processing
technology or the conditions for the effective use of the technology (Danziger, 1977, p. 29).
The early literature on the use of computers tended to be
characterized by a promoter's bias which basically gave little
coverage to the impacts of the technology. Danziger observed that the obvious potential of information processing technology, as outlined above, combined with the promotion-oriented flow of evaluative information in the early literature to create a favorable climate of opinion among top-level management about information processing in local government (ibid).
It was Danziger's belief, expressed in his 1977 article, that a broadly circulated and generally accepted credo about the impacts of computers had risen in local government. Danziger labeled this credo the litany to EDP. Danziger based his views on research findings derived from exploratory case studies in 12 cities and counties across the United States which utilized numerous open-ended interviews and the analysis of written documents. It is important to note that Danziger suggested that the findings of his research, while supported by substantial evidence, were relatively subjective (ibid). Later research tended to confirm many of the findings which Danziger first presented in his 1977 article on the litany to EDP (King, 1982; Kraemer, Dutton, and Northrop, 1981; Kraemer and King, 1981; Danziger, Dutton, Kling, and Kraemer, 1982).
Obviously, there was some variation in the content of the
litany to EDP by place and by person ; however, as Danziger notes, there was a striking recurrence of certain themes when those
involved in local government computing discussed computer-based information processing technology (Danziger, 1977, p. 29). The following is a catalogue of some of the central canons of the EDP litany as presented by James Danziger:
1. EDP tends to be staff reducing and cost reducing.
2. EDP turns mountains of data into molehills.
3. EDP provides better information for decision makers.
4. EDP increases the supervisor's ability to manage subordinates.
5. Inadequate utilization of EDP is primarily a function of either the user's resistance to or the user's failure to understand computers.
6. Transfer of computer technology among local governments will prevent the continual reinvention of the wheel (Danziger, 1977, pp. 29-30).
Perhaps the most attractive consequence of automation for many local governments has been the anticipated reduction of staff. There is no doubt that the speed and efficiency of computers can replace the staff required for many of the routinized information processing tasks. In some cases, the number of staff has remained constant while the number of transactions has increased. In other cases, computers have enabled an operating unit to expand the scope of a task or undertake new tasks (ibid).
However, staff reductions are not the rule for the majority of the more complex tasks which have been computerized (Kraemer, Dutton, and Matthews, 1975; Clowes, 1982). In many cases, data
coding and entry involve more staff time per transaction than did the manual system that was replaced. In other cases, new or expanded data collection was undertaken which had questionable utility to the task at hand. It is also important to note that in all cases substantial technical staff time was required to service computer operations and software.
As Danziger points out, it has been extremely difficult to evaluate the claim that data processing has been "cost reducing" for local governments. He notes that the cost-saving, argument has often been linked closely to purported staff reductions; but if the reduction of operating staff has not been very substantial and if the data processing staff commands moderately high salaries, as they most often do, then this aspect of cost-reduction is at best problematic (Danziger, 1977, p. 31).
One other important aspect of the cost equation has been pointed out by several authors. Normally, a local government, like many other types of organizations, expands its programmer and analyst staff to meet the work demands of a development phase in which automation of most tasks appears cost-effective. However, after these systems are completed, there may be underutilization of technical staff, since maintenance activities, in theory at least, are less time consuming. In addition, many organizations acquire hardware with capacity far in excess of current needs. All of the resulting "underutilization" leads to the obvious solution of initiating new projects. The momentum of the development/design cycle can, over time, take control with the result that there is
diminished attention paid to the intrinsic cost-effectiveness of automating a particular task (Danziger, 1977; Kraemer, Dutton, and Matthews, 1975). The central question relating to data processing development/design can easily become, What next? instead of the more important question, Anything next?
There is little doubt that one of the greatest values of EDP is its ability to store, manipulate, and retrieve large amounts of information. This is a particularly important benefit for many local governments because a number of information processing tasks are required by statute, for example, tax and property records, certain police records, and employee payroll records. In most instances, as Danziger points out, the marginal costs of expanding the information available for any information processing task are small (Danziger, 1977, p. 32). According to Kenneth Laudon, a risk for local government activities is that computer-based information processing will stimulate the over-accumulation of information (Laudon, 1974, p. 155). This problem is really the issue of information overload, which has been much discussed in' the literature but largely ignored by practitioners in local government.
An important expectation expressed in the literature on computer-based information processing and found among practitioners in local government has been that EDP would increase the usable information available to decision makers. In fact, the costs involved in the search and analysis of information usually do drop significantly with the use of information processing, particularly when large amounts of data must be handled. It is also important to
note that it is often assumed that the quality of the information will be enhanced, since it has been collected and analyzed more systematically than with most manual systems. In the cities and counties examined in Danziger's study, he found a substantial number of cases of "mystification via automation" (Danziger, 1977, p. 32). Computer-based data and decision making should not be viewed with either more or less trust than other forms of data and decisions.
The problem, according to Danziger, is that many local government administrators harbor some predisposition about the credibility of data and decisions where information processing technology has been employed (ibid).
The fourth canon of the EDP litany, according to Danziger,
is that EDP increases the supervisor's ability to manage subordinates. This canon has been an issue in the literature on computer-based information processing since Leavitt and Whisler's article in 1958. In a landmark article written in 1967, Anthony Downs argued that some of the most important organizational "power shifts" with information processing technology would involve an increase in the power and control of higher-level officials over intermediate and lower-level personnel. Downs' belief was that information processing facilitates the close monitoring of subordinates by capturing performance data from their routine reports (Downs, 1967, p. 208).
There are many examples of the work monitoring capacities of information processing technology. Recently at the Denver. Water Department, Denver, Colorado, a new meter reading system was
installed to replace the manual system which had been in use for many years. The new system involves the use of hand-held computers with very large memory capabilities. The new hand-held devices include a feature that was completely new to the meter reading system. The device's memory records the exact time that each meter reading was punched into the hand-held units by the meter readers. The meter reading supervisors are thus provided with new information which allows them to measure the productivity of each meter reader and to investigate any "suspicious" gaps of time when no meter readings are punched in. Obviously, this new system has led to a period of adjustment on the part of the meter readers who were used to operating quite independently of any close supervision.
Computer systems, like the one outlined in the above example, tend to rely on strictly quantitative and often unidimensional measurement. Those employees monitored are more or less consciously driven to "do well" as defined by the particular quantitative measures used by the system in question. In the above example, the new quantitative measure of productivity fails to take into account the time spent by meter readers in talking to customers at their homes and answering their questions concerning water service. In fact, if the quantitative measure is overemphasized, the meter readers would be reluctant to deal with customers, thus negating an important aspect of the Water Department's public relation effort.
The obvious solution is to develop work monitoring systems that utilize sensitive measures which take qualitative factors into
account. There is little doubt that information processing technology has provided many supervisors with easy access to performance data and enhanced their ability to manage their personnel. As Danziger points out, however, the unintended consequences of computer systems upon the behavior and morale of employees must be studied. These systems have a tendency to supplant rather than supplement more qualitative criteria of control and evaluation and therefore could actually reduce the quality of service provided (Danziger, 1977, p. 33).
The fifth canon of the EDP litany, according to Danziger, is that the inadequate utilization of EDP is primarily a function of either the user's resistance to or the user's failure to understand computers. There is at least some support in the literature for this canon of the EDP litany (Garrity, 1963; Churchill, Kempster, and Uretsky, 1969; Sollenberger, 1968; and Whisler, 1970). Some managers and more staff and line personnel, who are perplexed or threatened by information processing technology, may exhibit classic bureaucratic behavior by resisting change, subverting innovations, and persisting with old operating procedures.
However, the case studies conducted by Danziger, in his research, suggest that inadequate utilization of information processing technology in local governments is not primarily a function of failures and shortcomings internal to user departments (Danziger, 1977, p. 33). The most common source of problems is the unit in charge of information processing. According to Danziger, a surprisingly large proportion of automated tasks are poorly
designed. In some cases at least, this situation is due to the technical inadequacies of the analyst and programmer staff.
Danziger states that,
In such settings, certain conditions are likely: (1) skill levels are inadequate to mount first-rate development projects; (2) projects are characterized by major delays, large cost overruns, or abandonment; (3) a large proportion of programmer and analyst activity concerns debugging and software maintenance rather than redesign or development;
(4) documentation is insufficient or nonexistent; and (5) there is talk of a "credibility gap" between EDP and users (Danziger, 1977, p. 34).
The level of user involvement in information processing systems design is another aspect of this EDP canon. Automated local government operations may be poorly designed, regardless of whether the information processing staff is technically competent.
Decisions about the selection and priority of information processing tasks to be automated may not be based on an understanding of data processing capabilities. The user may lack insight regarding the computers capabilities, and the information processing staff may lack insight regarding the requirements of user department tasks.
In addition, the automation of some tasks may either fail to fulfill the requirements of the task or fail to facilitate the task.
There appears to be no simple explanation for the inadequate development and use of automated tasks for various local governments. To a limited extent, such problems may relate to the fear or ignorance of computing among potential users. However, in many cases it appears that the problems can be attributed to the quality of the information processing staff, to their misunderstandings of the user's tasks or to overambitious systems
analysts who design systems that are not appropriate for the tasks to which they are applied.
The sixth canon of the EDP litany, according to Danziger, is
that the transfer of computer technology among local governments
would prevent the continual reinvention of the wheel. This
particular canon has been widely advocated by practitioners in local
government,.primarily due to the high in-house costs for developing
automated systems. In Danziger1s study, however, he found that when
particular local governments were examined, successful examples of
software transfer were rare. (Danziger, 1977, p. 34). For many
local governments, serious consideration of technology transfer
seldom occurs. There have been some computer software packages
developed by computer vendors, consultants, federal and state
agencies, and clearing houses that provide generalized software for 3
particular tasks. Even these generalized software packages, which are supposed to meet the requirements of large numbers of local governments, have been adopted sporadically and, in many cases, have been found to be unsatisfactory by users.
Perhaps the greatest single problem encountered by local governments wishing to pursue technology transfer has been technical problems with the proposed transfer. Many information processing personnel would argue that differences in hardware and/or languages between the two information processing units result in adoption costs which are prohibitively expensive. In addition, the documentation on many of the systems proposed for transfer has been
woefully inadequate, thus resulting in major technical difficulties in accomplishing a successful transfer.
Other problems with technology transfer have resulted from a lack of adequate information about what other local governments have developed in terms of automated systems. The greater the distance between local governments with similar needs and situations, the less chance that a workable system will be discovered and the more prohibitive the costs of studying such a system will be. The attitudes of information processing personnel and users may constitute another obstacle to successful technology transfer. Both users and information processing personnel may be very concerned that, due to the "unique" requirements of their particular systems, any system transferred from another local government will fail to meet important design criteria for their proposed system.
The problems illustrated above have led to a very low level of successful technology transfer between local governments. Apparently, most local governments would prefer to "go it alone" in the design, development, and implementation of new systems despite the fact that most local government administrators "support" the concept of technology transfer.
The evidence seems to suggest that the impacts of information processing technology upon many local government operations are complex and might be, at least in part, negative (Danziger, 1977; Kraemer, Dutton and Northrop, 1981). For many information processing tasks, automation does not seem to result in cost reduction or staff reduction. The technology can stimulate the
collection, coding, and storage of large amounts of data for which there may be no compelling purpose. There may also be unintended or unsatisfactory consequences from the use of information processing technology to justify decisions. And there are organizational arrangements and attitudes and behaviors on the part of users and information processing personnel that can lead to poor design of automated tasks and the minimal use of technology transfer.
None of the observations outlined above and discussed by Danziger in his 1977 article are particularly startling to those who have studied computer-based information technology in local governments. What has been interesting, as Danziger pointed out, was the apparent absence of objective observers of local government computing. Most of those publicly reflecting upon information processing in local government have been committed to its success, and their attention has been drawn to successful uses of the technology (Danziger, 1977, p. 35). The result has been the confusion of potential with actual performance.
The Impact of Computer-Based Information Processing on the Management of Local Government
The early literature on the impacts of computing in organizations is quite extensive and suggests that computing is likely to have many dysfunctional side effects as well as positive benefits for organizations (Kraemer and King, 1977; Westin and Baker, 1972; Kraemer, Dutton, and Northrop, 1981). However, it should be pointed out that the literature dealing specifically with
computer impacts in local governments is very limited in comparison to that dealing with business. A few works have expressed concern over potential and actual dysfunctional side effects (Westin and Baker, 1972), but on the whole, the prominence of speculative works which lack empirical descriptions of actual dysfunctional impacts have tended to lessen concern over such impacts (Kraemer, Dutton, and Northrop,. 1981, p. 51). It appears that the anticipated benefits of computer-based information processing have had a correspondingly greater effect on the decisions of organizations seeking ways to improve the quality of their information systems than concern for potential dysfunctional impacts.
Two of the earlier in-depth studies of the impact of computer-based information processing on management capabilities and management organization in private industry were made by Booz, Allen and Hamilton in 1967 and by the McKinsey Company in 1968. Both studies determined that the direct repercussion of the introduction of computers on top management organization had not materialized with the intensity and speed initially forecast by the early theorists (Ream, 1968, p. 495). However, both studies emphasized that, in the proper environment, all signs pointed to an acceleration in the impact of computers on organizations.
In 1965, Gilbert Burck and the editors of Fortune magazine made a sweeping assessment of the impact of computers on
management. Their opinion was that, as as result of the impact,
management might never be the same again. Burck commented:
Certainly no other single item of capital goods has changed the basic terms of so many human activities in so short a time. Within a few years, as the engine of modern information technology, it has profoundly altered the techniques of science, has begun to make government efficient, and has provided a new basis for the strategies of national defense. Above all, it is radically changing business' production methods and the art and science of management. Although the machine is the bete noire of critics who fear it will accelerate unemployment and compound the worst problems of modern society, it seems destine to shine as a powerful instrument for making business more creative and efficient and hence for raising the nation's real income per person, for eliminating a vast amount of drudgery, and for increasing leisure. In short, for measurably expanding free man's range of choices, which by definition is the goal of any good society.
One characteristic of the computer that makes it unique among technical achievements is that it has forced men to think about what they are doing with clarity and precision. A man cannot instruct the computer to perform usefully until he has arduously thought through what he is up to in the first place, and where he wants to go from here. Even scientists, once they have wrestled with a computer's demands on knowledge and logic, are astonished to discover how much of their mental activity travels in ruts. The rethinking process gets more difficult as the computer gets better. Wherever the machine is used, it is improving enormously the quantity and quality of human cogitation; and it is rapidly becoming a kind of Universal Disciplinarian. (Burke, 1965, pp. 2-3)
In a 1968 article, entitled "The Computer and its Impact on Public Organization," by Norman J. Ream, then the Special Assistant to the Secretary of the Navy, he stated that while progress in the use of computers in the federal government and in state and local governments had been acceptable to a degree, the gap between the computer's technical capability and its practical application was growing wider. Ream further noted that computers had more than paid
their own way when used in clerical operations within given
government functional organizations, but that the government had
failed to use computers effectively in adapting to new and
increasing demands (Ream, 1968, p. 501). Ream cited a June 28, 1966
memorandum from President Lyndon B. Johnson to all heads of
departments and agencies which stated:
I want the head of every Federal agency to explore and apply all possible means to
- use the electronic computer to do a better job
- manage computer activity at the lowest possible cost I want my administration to give priority emphasis to both of these objectivesnothing less will suffice. (Ream,
1968, p. 496)
Ream concluded that the effective use of computers would require a change in government management strategy, in management's attentiveness to their full exploitation, and ultimately to necessary changes in public organization structures (Ream, 1968, p. 501).
From today's perspective it appears that Ream's assessment was correct. The effective use of computer-based information processing has required changes in management and management strategy. The desire to effectively utilize computer resources has, without doubt, been the leading "cause" of impacts resulting from the use of computer resources. One important point that should be noted concerning the large majority of the research that has been conducted on impacts is that these studies look mainly at the changes brought about in the work environment by the introduction or modification of computer-based information systems. In general, this research assumes that the politics and community attributes of
local governments do not change, and therefore most of the research does not address the possible dynamic impacts of computerized information systems. It appears from the research that computer-based information processing does not lead to changes in the goals and purposes of organizations, but rather to a shift in the structures and processes for achieving the goals and purposes.
The Impact of Computer-Based Information Processing on Organization Structure
Some theorists have predicted that organizations will undergo massive reorganization as a result of computer-based information systems (Slater, 1958; Leavitt and Whisler, 1958).
These authors suggested that firms would recentral i'ze as a result of new computer technology; the availability of more information than previously possible would allow management to centralize. The trend until the development of computer systems had been toward decentralization because centralized management could not cope with the amount of information and the number of decisions required in a large organization. Computers offered the power to make centralized management possible so the organization could be tightly controlled by a group of top managers. Other researchers have contended that information is independent of the organization and would have little affect on structural characteristics (Simon, 1965; Lucas, 1975).
The available empirical research does little to support or refute these differing contentions (Kraemer and King, Volume 2, 1977). Rather, it presents a complex and confusing pattern of structural
impacts that raises new questions for each answer provided. The issue of centralization versus decentralization is further changed and complicated by the introduction of minicomputers and microcomputers in organizations.
Hoos1 1960 study of employees at all levels in 19 public and private organizations found that computing had brought reorganizations within particular divisions, branches, and offices of the organizations ("internal" reorganization), as well as reorganizations in divisions, branches, and offices at dispersed geographical locations ("external" reorganization) (Hoos, 1960, p. 110). Unfortunately, the nature of these changes is unclear, and the evidence is not compelling.
Similarly, a 1966 Organization for Economic Cooperation and Development (OECD) study reported reorganization findings in case studies of an insurance company in Canada, the Public Pension Department in Germany, and the Internal Revenue Service in the United States. These studies indicated that the number of jobs increased in some locations, while decreasing in other locations, and that the size and importance of department subunits changed as a result of computerization. Again, the precise nature of these changes is unclear; the research only reports their occurrence (Kraemer and King, 1977, p. 147). This is characteristic of the early research on computer impacts.
The available research does indicate, at least to some extent, that computer systems do bring changes to the processes and structure of local governments and other organizations (ibid).
Sometimes these changes are intended as part of the design for the new computer-based systems, but more frequently they are indirect, unintended, and unanticipated. It appears these changes are contingent both upon the initial organizational conditions and the processes by which change is introduced (Mumford and Banks, 1967). However, these contingent features are rarely addressed explicitly
in the research literature (Kraemer and King, 1977). Most of the
research treats only the outcomes of computerization and not the antecedents of change.
One area of research in organization theory that has not received much attention is a systematic explanation of the grouping of activities, jobs, and people in organizations. Some research suggests, however, that the costs of communication are a fundamental determinant of the pattern of departmentation within any organization (Whisler, 1970, p. 46). It is Whisler's contention that organizations, by trial and error, seek that structure which will minimize their costs, given the number and kinds of problems that must be solved (ibid). These problems are dependent on the character and volume of an organization's outputs.
No matter how the departments of any organization are laid out, each one must gather and transmit information, analyze and transform it. With the growth of computer technology, pervasive changes in the techniques and economics of information processing could be expected, as well as continuous changes in the arrangement
of departments in any given organization. In a study of 18 life insurance companies using computers for periods ranging from 18 months to 10 years, Whisler found that 177 departmental changes were ascribed to the use of computer systems (Whisler, 1970, pp. 46-47). These changes tended to be concentrated in larger companies with longer experience and with more extensive computer applications.
NUMBER AND KINDS OF CHANGES IN THE ORGANIZATION STRUCTURE OF 18 INSURANCE COMPANIES AS A RESULT OF INFORMATION PROCESSING TECHNOLOGY
Type of Chanqe Number of Instances Reported Percent of Total
(1) Creation of new departments 52 29
(2) Elimination of old departments 43 24
(3) Transfer of activities (departments) to EDP 33 19
(4) Other transfers of activities (departments) 21 12
(5) Consolidations (mergers) of departments 4 2
(6) Splits of departments 4 2
(7) Reorganization of parallel into functional
departments 20 12
Source: Whisler. Information Technology and Organizational Change, p. 47 (1970).
Computing in an organization does not affect all departments equally. Those most affected are the initial user departments (Mumford and Banks, 1967). In local governments, the initial users have traditionally been finance departments (Kraemer and Howe, 1968; Dial, Kraemer, MitcheV, and Weiner, 1970). The initial primary user department tends to be responsible for operation and control of the computing unit. This often requires the creation of a new departmental subunit. The department most frequently responsible
for computing operations in local government has usually been the primary user (Kraemer and King, 1977, vol. 2).
In many organizations, computers have been used extensively by certain departments with little noticeable change resulting in the structure of these departments. It is certainly possible that lack of visible impact may be the result of factors such as growth or a lag in adaptation of organizational structure to the new systems. However, Whisler has found in his research that certain kinds of departments dominate the list of those showing little structural change. Departments typical of those showing little structural change in Whisler's research were: engineering design, research and development, industrial relations, and advertising (Whisler, 1970, p. 50). Whisler compares these departments with those in which major changes are most often visible, such as production, accounting, premium billing, and product distribution (ibid). The differences which Whisler found in the two groups are summarized as follows:
Degree of job specialization. In general, task specialization is more complete in departments which are most affected by computerization.
Degree of job interdependence. The degree to which the outcome of one member's effort depends upon the behavior of other members is largely a function of the degree of job specialization. Effort is substantially more interdependent in departments which are most affected by computerization.
Degree of management control and coordination required. In a highly specialized department, a high degree of management control is required to properly coordinate the efforts of the department's members. Departments most affected by computerization tend to require more control and coordination than departments less affected by computerization.
Character of departmental taskproblem solving vs. problem definition. In departments where organizational problems have been defined, and especially where they recur, it is feasible to develop a high degree of specialization, interdependence, and control and coordination. In departments where the task is primarily that of discovering which problems the organization must deal with, the ill-structured and one-time character of the task may make it difficult to extensively exploit specialization. Departments which are most affected by computerization tend to be those which are primarily concerned with problem-solving rather than with problem defini tion.
Time press. In some departments, deadlines are accurately and precisely defined, strongly enforced, and critical in nature.
In other departments, deadlines tend to be more vague and less stringent. The imposition of deadlines tends to correlate to a high degree with high levels of task specialization. Again, departments which are most affected by computerization tend to be those departments with many deadlines to be met (Whisler, 1970, pp. 50-51).
The introduction of computer-based information processing into departments, which by their nature require high degrees of task
specialization, tends to make these departments into man-machine operations, rather than strictly human systems. It could be reasonably expected that the structural characteristics of these departments would change markedly with the introduction of computerization.
In summary, the literature is in agreement that the initial impact of computing is felt most strongly by departments which are routine/schedule oriented (Kraemer and King, 1977, vol. 2., p. 151). The computing unit is likely to be controlled by one of these departmental users, particularly the accounting or finance department, and the impacts of computing are likely to be greatest there (ibid). Kraemer and King believe that the short-term effect of computing on departmentation is primarily the introduction of new department subunits, including the computing unit, and various department and interdepartment policy boards and user committees.
The long-run effect might be the introduction of a new department into the organization structure (ibid).
Levels of Hierarchy
The classic 1958 essay by Leavitt and Whisler predicted that the structure of organizations which adopted computer systems would change from the traditional pyramid shape to a shape that resembles the outline of a football balanced on top of a bell, as shown in Figure 2.1.
Supervi sory and Clerical
Tradi tional Leavitt and Whis1er's
Organizational Shape Predicted
LEAVITT AND WHISLER 1958 MODEL OF STRUCTURAL CHANGE FROM COMPUTING
Source: Leavitt and Whisler, p. 44 (1958).
The rationale for this model was that organizations compete in external environments that respond to innovation. This situation tends to move research, development, and information management personnel toward the top of the organization hierarchy. Leavitt and Whisler believed that these personnel shifts would mean that a horizontal slice of the pyramid-type organization (middle management) would break in two. This would leave fewer middle managers devoted to managing day-to-day operations since computer-based systems would take over many of their functions. It was Leavitt and Whisler1s contention that a small portion of middle
managers devoted to creative thinking would proliferate and rise to the top management level (Leavitt and Whisler, 1958).
In the years since Leavitt and Whisler made their predictions concerning the impact of computerization on organizational hierarchy, researchers have sought to assess their accuracy. The major works on this subject include C. A. Myer's work in 1966 of case studies in business firms and his summary of previous research; the Whisler and Meyer survey-oriented field study of 23 insurance companies in 1967; and the independent research conducted by Whisler in 1970. The results of these studies were generally helpful, although not conclusive. All of the researchers agreed that reorganization stemming from computerization was likely to result in the same or fewer levels in the hierarchy.
Myer suggested that managerial philosophy could be more important than the technology itself in affecting the structure of organizations that use computer-based information processing (Myer, 1966). Myer concluded that his evidence failed to support the shrinking size of the middle management group. In fact, Myer indicated that in some organizations middle management was expanding due to the addition of system design and programming staff in the computing unit (Myer, 1966).
In Whisler's 1970 review of the research, he noted that there was a modest amount of evidence that the number of levels tend to decline as a result of computerization. He further noted that the data had been collected in such a way that it was very difficult to determine if both the span of control and the number of levels
decline in the same organization or if one tends to substitute for the other (Whisler, 1970 a, p. 40). Only in the life insurance study was there any evidence presented that in the same set of companies both the span of control and the number of levels tended to decline (Whisler and Meyer, 1967).
There are documented cases where large? integrated computer systems imposed on organizations over a short period of time resulted in a reduction in the number of organizational levels. One example discussed by Whisler was the application of the SAGE system to the North American Defense Command (NORAD). This command has the function of detecting and intercepting all unidentified aircraft on the North American continent and destroying hostile aircraft. Prior to the installation of the SAGE computer system, NORAD relied on World War II technology to carry out this function, using a complex combination of radar, radio, and plotting boards to work out the critical problem of the optimal deployment of interceptor weapons whenever an unknown aircraft appeared. The SAGE system was designed to solve this problem accurately, quickly, and automatically. As a consequence of the SAGE installation, the number of levels of command in NORAD was reduced from five to four (Whisler 1970 a, p. 40).
In summary, although some internal and external restructuring of organizations occurs, the predictions of Leavitt and Whisler regarding the impacts of computerization on organizational shape and scalar levels generally are unsupported by later research (Kraemer and King, 1977, vol. 2). Based on the small
amount of evidence at hand, a reasonable conclusion is that if there is any change at all, the number of levels in the "chain of command" of organizations is more likely to decrease rather than increase as a consequence of computerization.
Span of Control
Theorists have long maintained that the span of control, the
number of people reporting to a supervisor or manager, should be
affected by the introduction of computer-based information
processing in a given organization (Leavitt and Whisler, 1958;
Whisler, 1970 b; and Neuman, 1978). The common prediction about the
computer's impact has been that it would alter the relationships
between staff personnel and their supervisors, primarily by
improving the ability of supervisors to monitor subordinates. In
1978, Neuman predicted that advances in computer-based information
processing and office technology would mean that managers in future
offices would have more people under their direct supervision
(Neuman, 1978, p. 30). Neuman went on to predict that the
administrative result of expanding the span of control is that it
reduces the need for middle managers. He stated that,
When an administrator can control 12 functions rather than six, he or she knocks out the person to whom the six new functions reported under the old structure The overhead eliminated by office automation is not low order clerical wages, but higher order supervisory and management salaries. (Neuman, 1978, p. 31)
Whisler was not as positive about the impacts of computer-based information processing on span of control as Neuman. In Whisler's study of insurance companies, he noted that an
examination of company organization charts showed that at the top level of the companies (the level immediately under the president), there was no relationship between the use of computers and the span of control (Whisler, 1970 b, p. 55). Whisler felt that it would be many years before a relationship between computers and span of control showed up in the life insurance industry.
In' 1969, Churchill, Kempster, and Uretsky conducted major research demonstrating the impacts of computers on top management in twelve firms. In regard to span of control, the authors found that top managers' supervision of immediate subordinates decreased. The managers were found to have fewer people to manage as a result of computerization, and those remaining were more skilled and self-sufficient, thus reducing the importance of supervision in the managers' role (Churchill, Kempster, and Uretsky, 1969). It appears that much of the research in regard to computerization and span of control is conflicting and inconclusive.
In terms of research directly linked to computerization and span of control in local government, very little has been done. In perhaps the only research to address the issue, Kraemer, Danziger, and King found that chief executives generally do not believe that the introduction of computer-based information processing in local government has had any impact on supervisor/staff relationships including span of control (Kraemer, Danziger, and King, 1976, p. 18). The authors caution, however, that subordinates would be more sensitive to this issue than top executives. The authors' research indicates that subordinates do feel they are being more
closely supervised where their superiors have access to computer-generated workload statistics (ibid).
The Impact on Certain Tasks and People
No elements in an organization are more acutely affected by computer-based information processing than certain tasks; some are eliminated, others are added to the organization's total pool of tasks, and many others are modified or adjusted. Each job in an organization is defined by one or more tasks which can be assigned to an individual. Thus, it should be obvious that changes in tasks result in changes to the jobs of individuals within the organization.
Many early studies of the impact of computer-based information processing were concerned with user reactions. Some of these studies dealt with the psychological reaction of workers, and others concentrated on overall changes in levels of employment (Lucas, 1978, p. 69). Although some changes in employment have occurred as a result of computerization, it is difficult to find an overall trend. Lucas believed that it was safe to say that the impact of computers on unemployment levels has been no greater than that of any other technological change. He felt that the lack of an adverse impact was particularly significant in view of the short period of time that has elapsed since computers were introduced and the rapid development of computer systems (ibid).
Impacts on the Job of Top Management
For the purpose of clarity, top and, to a lesser extent, middle management jobs are defined as jobs that have as their primary specification the making and/or the administration of an organization's policies, programs, and practices.
Generally, the research to date indicates that the jobs of top managers have been affected only slightly and usually indirectly by computer-based information processing systems, and few top managers use computers directly in their work (Kraemer and King, 1977, vol. 2, p. 155). Most top managers receive the majority of their information from staff and middle managers who may process computer data for top managers. In the great majority of cases, information systems in organizations cannot be economically justified solely for top management and in fact are not designed primarily to serve top managers, but rather to serve the operating departments.
Churchill, Kempster, and Uretsky, as discussed earlier in this chapter, have done major research concerning the impacts of computers on top management. They found the following impacts on the jobs of top managers in firms that had been using sophisticated computer systems for more than five years:
1. Top managers' skill requirements were extended to include analytic problem solving for effective use of the expanded information base available with the computerized system.
2. Top managers' decision making required a broader range of considerations as the increased integration of information in the new system provided a more complex decision base than previously.
3. Top managers' control function shifted to management by exception. This resulted from the fact that needed information was provided automatically by the system rather than derived by the manager.
4. Top managers' time delay between an action and its reported effect decreased dramatically. This shortening of the decision-action-feedback cycle required a conceptual broadening of the manager's role and decision-making ability. This was considered as the biggest area of impact from computer-based information processing systems (Churchill, Kempster, and Uretsky, 1969).
The impacts cited by Churchill, Kempster, and Uretsky above are all indirect rather than direct impacts, because most top managers do not use computers directly (Brady, 1967; Dunlop, 1971). Brady's study of the decision-making process of 100 top managers in 12 companies found that top managers did not use computers directly, but rather received computer-based reports which had been filtered for them by middle management. Brady found considerable use of computers by middle managers to support top management decision processes. However, Brady's study further showed that top managers believed that the future impacts of computers on their decision-making processes would be much greater. Impacts were listed as slight in 1967 but were predicted to be great by 1980 (Brady, 1967).
Dunlop also found that top executives did not use computers directly. In his case study of top managers at IBM (International Business Machines), Dunlop found that executives provided with direct access to corporate information systems via computer terminals in their offices usually moved the terminals into the office of a secretary or staff person within a short time. Dunlop explained this behavior in light of traditional management concepts and practices. He noted that managers traditionally use information in the form of reports and have little need for or understanding of the real-time information available from a computer-based system (Dunlop, 1971).
The data which are contained in the computer-based information processing systems of organizations may also contribute to a lack of top management use of computing. For example, Lucas found that only 11 percent of the reports produced by the computer-based systems he examined were directed to top or middle management. The majority of the systems dealt with routine data about internal operations used by clerical and supervisory level personnel, while middle and top managers required different data for their decision making (Lucas, 1978).
Thomas Whisler argues in his classic work entitled Information Technology and Organizational Change (1970) that there was little empirical evidence concerning computer-induced changes in the content of managerial tasks (Whisler, 1970 a, p. 81). He further noted that the absence of evidence was not very surprising since computer applications had produced the most substantial
changes in organization and jobs at the lowest levels of the hierarchy (ibid). Whisler's most important contribution in this area was his recognition that organizational theorists had not agreed on a useful set of categories of managerial tasks. Whisler offered his own schema and divided managerial tasks into four categories. Briefly reviewed, these categories are as follows:
1. ComputationIf the word is broadly defined, much of what a manager does is computation, working out the best, quickest, or cheapest way to get something done. This computation must often be done over and over again as unforeseen events invalidate previous calculations.
2. CommunicationThe manager sends and receives information and also originates information. A number of studies have shown that a high percentage of a manager's time is devoted to communication with others by memorandum, letter, telephone, or face-to-face.
3. Goal settingThis important function is usually performed jointly with other managers, to decide what is desirable for the organization and in what direction the organization should move.
4. Pattern PerceptionThis task approximates the common sense notion of creativity. The manager basically scans his or her environment and perceives patterns among events that have not been discovered before. It is an activity critically important for the organization and also for the society in which the organization exists (Whisler, 1970 a, p. 8).
Whisler reviews the comparative advantages of the manager and the computer for each of the activities discussed above. In computation, he notes that the ultimate superiority of the computer seems undeniable. Not only is the individual manager an inferior computer, but whole groups of managers are unlikely to be able to perform computation tasks as effectively as computers themselves.
In terms of communication, Whisler believes that in large measure it will also be taken over by advanced telecommunications systems (Whisler, 1970 a, p. 82). Economic considerations may determine to what degree telecommunications will substitute, but more and more task-related communication problems are being transferred to technological systems. In terms of goal setting, people are the dominant partner in the human-machine system. Although machines may at times control people, this control has been delegated by humans to the machine in order that humans may achieve the goals they want. Pattern perception will, for the foreseeable future, be a uniquely human function. Computers can create patterns, but their sensory apparatus is not yet sophisticated enough to scan the environment to discover patterns.
In summary, Whisler speculates that, in the highly developed manager-machine system of the future, computer-based information processing systems will perform most of the computation function and much of the communication task. The manager's time will be increasingly dominated by goal setting and pattern perception tasks (Whisler, 1970 a, pp. 82-83).
The large number of organizations .using computers today suggests that the overall attitude of top managers toward computer-based information processing systems is positive. Despite this evidence of use and support, the research indicates that all is not completely well between computers and top management. Problems occur with the computer unit and with computer-generated problems of the line departments. The two main problems with the computer unit are involvement of users in the design and development of applications and control of system building projects (Kraemer and King, 1977, vol. 2, p,p. 163-164).
Impacts on the Job of Middle Management
The prediction that computer impacts on middle managers would be greater than those on top management appears to be true, although the impacts to date appear to be less than expected.
Leavitt and Whisler predicted that computerization would result in middle management jobs that required less conformity to social norms and more conformity to the routines set by the computer system (Leavitt and Whisler, 1958). They also predicted that many jobs now held by middle managers would become "programmed" by more precise descriptions of job specifications, while some, although a lesser number, would be deprogrammed through the use of generalized job specifications. This aspect of Leavitt and Whisler1s work seemed to be supported by Hoos, who found that middle management jobs in computerized organizations were becoming more programmable, with
middle managers supervising errors rather than training subordinates or making policies (Hoos, 1960).
However, in work six years after Hoos's article appeared, Myers found a decrease in the routine and repetitive aspects of middle management jobs and an expansion of the nonroutine aspects (Myers, 1966). Other researchers found that middle managers often needed improved capabilities for decision-making ability and leadership. They suggested that these skills must be accompanied by improvements in managers' abilities to plan, make judgmental decisions, and perform duties accurately (Mumford and Banks, 1967).
Middle managers more so than top managers are subjected to increased time pressures and information overload as a result of computerization. The time lag between taking an action and receiving feedback of results is shortened for middle managers, thus requiring faster decision making ability (Brady, 1967). Middle managers filter information for top management. The increase in volume of information for filtering brought by computerization has made middle managers overloaded with information and information processing tasks. This has "decreased" their job performance by traditional standards, causing a negative attitude about computer-based information processing systems (Guthrie, 1972).
There has been very little research on changes in the productivity of middle managers as a result of computer systems (Kraemer and King, 1977, vol. 2). Most researchers appear to agree with the statement by Myers that the middle manager, like the top executive, will be interacting more in the future with the computer
system and using it as an aid in decision making, rather than as a substitute (Myers, 1966).
Middle managers are actual users of computer systems, so their behavior toward computers could be much different than that of top managers. Leavitt and Whisler predicted major resistance to the expected routinization of relatively autonomous and unprogrammed middle management jobs (Leavitt and Whisler, 1958). Other researchers supported this finding, noting that middle managers make only the minimal required use of computer systems (Mumford and Banks, 1967; Guthrie, 1972). If these views were generally correct today, then Myers's beliefs concerning interactions between computers and managers would be drastically wrong. More recent research indicates a much higher level of acceptance from management personnel of computer-based information processing systems (Dutton and Kraemer, 1976; Dutton, Kraemer, and Pearson, 1975; McGowan and Stevens, 1983).
Impacts on the Job of Clerks and Supervisors
The majority of the research on computer impacts in organizations has focused on clerks and supervisors and indicates that computers have their greatest impact on these clerical and supervisory jobs (Kraemer and King, 1977, vol. 2). Not only do the jobs change with respect to content, process, and skills, but so does productivity.
In 1965, H. C. Lee found that many traditional clerical, accounting, bookkeeping, and filing jobs were eliminated by
automation in computerized offices, while the clerical and supervisory jobs that remained became more technical, rationalized, standardized, and increasingly subject to rigid rules and regulations (Lee, 1965). Lee and other researchers (Borodin and Gotlieb, 1972; Mann and Williams, 1960; Mann, 1962; and Whisler,
1970 b) found that computer systems often require supervisors and clerical personnel to:
1. Structure their work more rigidly due to system-imposed deadlines and standards.
2. Upgrade their work in terms of speed, accuracy, and precision.
3. Face greater responsibility for their work since it can usually be traced to the individual.
4. Perform shift work.
5. Accept measurement of work output and performance.
These changes result from the requirement for closer
coordination among people and departments brought about by the computer's technical characteristics and changes in work flow (Lee, 1965).
There does not appear to be any definitive research regarding impacts of computer-based information processing on the productivity of clerks and supervisors. Mumford and Banks found that work volume per person increased in some cases and decreased in other cases in a study of an English bank and a corporate office (Mumford and Banks, 1967). The decrease in productivity was attributed to the poor reliability of computer service. Other
researchers found at least some increase in the volume of work per person in a computerized environment (Kraemer and King, 1977, vol. 2).
In conclusion, the amount of research on the impact of computers on tasks and jobs in organizations has not been very substantial. From the research that has been conducted, the indication is that the impact of computer-based systems on tasks and jobs has been considerable at the lower levels of the organization and decreases as one moves up the organization levels. Clerical and supervisory jobs have become rationalized and generally require specific machine-related skills. Middle management positions have been burdened with more information processing tasks, in addition to the traditional tasks of middle managers. The jobs of top managers have been affected only slightly and indirectly, primarily through a speedup in the decision-action-feedback cycle.
Impacts on the Nork Environment
The work environment is a domain where major impacts from computer technology have been predicted. Computers were supposed to improve the work environment of government officials by making work more varied and interesting, by relieving day-to-day work pressures, by providing easy access to information they need, and by increasing
their influence over others (Leavitt and Whisler, 1958; Pfeffer,
1981). There have been a number of claims made about the impact of computers on the professional workplace. The effects that have been
empirically validated are more modest. The early empirical research suggested that computing tended to reduce the quality of working life, primarily by producing greater time pressure and reducing individual satisfaction with the job (Kraemer and King, 1977, vol. 2). The more recent empirical research indicates that generally computer technology has had limited rather than major impacts on the character of white collar working life. It also appears that the patterns and levels of computer impacts on work vary across roles, with more positive impacts attributed to the computer as the employee's role ascends the organizational hierarchy (Kraemer and Danziger, 1984).
The majority of the empirical research that has addressed the impact of computing on work environment and work life has examined a particular role type in a single organization or across a variety of organizations (Dutton, Kraemer, and Pearson, 1975). Perhaps one of the most careful and complete pieces of research on this subject was conducted by Kenneth L. Kraemer and James N. Danziger in 1984. Their research employed a systematic, comparative framework for a large sample of individuals within a single class of organizations. Presented below, in tabular form, are the results of the Kraemer and Danziger research.
EFFECTS ON WORK ENVIRONMENT ATTRIBUTED TO COMPUTING BY MUNICIPAL EMPLOYEES, BY ROLEa
Comoutina Effect UDon: Managers (N=498)
SuDervision of work
More Closely supervised 17%
No difference 78
Less closely supervised 5
Influence over others
No change 56
Not affected 48
Sense of accomolishment
Not affected 44
Staff Street level Desk Top
Professionals Bureaucrats Bureaucrats
6% 18% 9%
78 73 78
16 8 13
3 3 4
54 68 67
43 30 28
34 19 37
51 36 40
IS 45 24
3 5 6
46 40 42
52 55 51
a Respondents are 1,448 professional service workers in 42 American municipal governments who indicated that they use computers or receive computer-based reports and have had some contact with data processing personnel. Responses were from a self-administered questionnaire which included these questions:
As a result of computing is your work more or less closely supervised? (less closely supervised, no difference, more closely supervised)
Has computing given you more or less influence over the actions of others?
(less influence, no change, more influence)
Has computing increased or decreased time pressures in your job? (decreased, not affected, increased)
Has computing increased or lowered your sense of accomplishment in your work?
(lowered, not affected, raised)
All questions also had a "Don't Know" response.
Source: Kraemer and Danziger, p. 35 (1984).
Kraemer and Danziger note that two broad generalizations can be derived from these data. First, the changes in work life caused by computing are widespread, but are not pervasive. The large majority in every role (73-78 percent) find that computing has not altered the extent to which their work is supervised, and most (54-68 percent) indicate that computing has not affected their capacity to influence others. In contrast, the authors report that a majority within each of the four roles does report a notable impact of computing on the sense of accomplishment with their work, and only the staff professionals (at 49 percent) fall below a majority among all roles in attributing changes in time pressure to computing (Kraemer and Danziger, 1984, p. 34).
The second broad generalization that emerges from Kraemer and Danziger's research is that the effects of computing on work life are largely jobr-enhancing. According to the authors, about half of those in every role reported that computer technology had not altered their capacity to influence others; nearly all those who had experienced an impact reported that they had greater influence due to computing, ranging from 28 to 43 percent across all roles. Less than one in five employees in any role reports that computing has increased the level of supervision of their work. Even the impact of the computer on time pressure has been generally benign;
64 to 81 percent of the respondents reported that computing either had not affected or had actually decreased the pressure they felt on the job (Kraemer and Danziger, 1984, p. 35).
Overall, Kraemer and Danziger note that there is little support for the view that computer technology, at least in its current modes of implementation and use, has been a dehumanizing or demoralizing force in the work life of professional service workers (Kraemer and Danziger, 1984).
The Impact on Authority and Control
A number of researchers expected that the use of computer-based information processing would improve the ability of top managers to control subordinate departments and individuals through increased information about the quantity and quality of performance, through more frequent reporting of such information, and through the elimination of "filtering" information by lower-level staff (Whisier and Meyer, 1967; Simon, 1971). Within local governments, some improvement has been achieved in the area of fiscal control, but computing has not improved the general ability of top managers for monitoring and control (Kraemer, Dutton, and Northrop, 1981).
Most control-oriented uses of computing are a by-product of operational uses and consist of simple listings and comparisons of inputs to government activities rather than the outputs/outcomes of these activities. Therefore, the use of such information for control is likely to be limited. According to Kraemer, Dutton, and Northrop, top managers tend to feel that computers have had little effect on their ability to control units under their responsibility, to identify problems, abuses, or inefficiencies in these units, or