From AUTOMATING THE OFFICE:
TECHNOLOGY AND SKILL IN WOMEN'S CLERICAL WORK, 1940-1970
Margaret Lucille Hedstrom
Under the supervision of Professor Thomas J. McCormick
Ph.D. dissertation, University of Wisconsin
NB: Because this file became infected with a Word macro virus, I was unable to save
formatting for the text. This means that some embedded quotes now appear without
offset, so you can’t tell they’re quotes. It also means that the footnote numbering was
lost. I’ve included the notes themselves, at the end, but you won’t be able to tell what
goes with what. Sorry about this, and I’ll keep working on fixing the original file.
INSTALLING COMPUTERS: FROM BRUTE FORCE TO BLUE SKY, 1955-1965
In May 1954, General Electric's Small Appliance Division in Louisville became
the first commercial enterprise to install a large-scale electronic computer for business
data processing. G.E. based its decision to install a computer at Appliance Park on the
assumption that automating the payroll, materials handling, inventory control, order
service and billing, and general and cost accounting would justify the cost of the
equipment. G.E.'s initial use of computers was typical of early data processing
applications which followed the "brute force approach" to automation. Brute force
applications were intended to cut overall office costs by finding more efficient ways to
perform work already being done and by eliminating the need for large numbers of
clerical workers. Firms that applied the brute force approach to automate simple
accounting and bookkeeping operations in the mid-1950s began experimenting with
"blue sky" applications in the mid-1960s, which aimed to provide new types of
information to managers or new services to customers.
Improvements in computer hardware and programming methods during the
first decade of computer use contributed to a shift in the perception of potential uses for
new information technology. William Aspray and Donald Beaver examined some ten
thousand advertisements for computers and identified three major changes in the
popular understanding of and perceived uses for computers. During the mid- to
late-1950s, computers were viewed as faster and more versatile electronic calculators.
By the mid-1960s, popular perceptions of the computer had shifted from the "computer
as calculator" to the computer as an "information processing machine." This new
perception stressed the capability of computers to make "logical decisions" or to supply
vital information to decision makers. Although the term office automation has been
used in conjunction with computers since the early 1950s, Aspray and Beaver concluded
that the perception of computers as small-scale and multi-purpose office machines,
which could automate a wide variety of office tasks, emerged only recently with the
microcomputer revolution in the late 1970s.
The first two stages in this schema correspond closely to the brute force and blue
sky approaches of the 1950s and 1960s. Prior to any practical experience with
computers, many businessmen believed that these machines would transform office
work from a highly specialized and labor-intensive activity to an automated,
continuous flow process. Early experimentation with computers demonstrated,
however, that redesigning office work was considerably more complex than originally
anticipated. Technical limitations of the first generation of computers, inexperience
with systems analysis and programming, and a disregard for the complexity of office
work forced businessmen to scale down their original expectations for drastic
reductions in the size of the clerical work force. Businessmen, who had anticipated
almost instantaneous results from computers and a rapid shift from manual to
automated clerical work, recognized during the first decade of computer use that
automating office work would be a gradual and ongoing process.
During the first decade of computer use an important shift also occurred in
perceptions of the social consequences of office automation. Computers failed to
reduce the size of the overall office work force or to create interesting and more
challenging jobs for office workers. Accumulated evidence from the early computer
installations showed increases in routine data handling jobs and few opportunities for
advancement into the relatively small number of data processing jobs that required new
skills. Moreover, the use of computers reinforced, rather than diminished, the
distinctions between men's managerial and technical positions and women's jobs as
clerical workers. By the mid-1960s, concern over the potential displacement of large
numbers of clerical workers was superseded by concern over the increasingly limited
opportunities for advancement and rewarding work.
The successful installation of a Univac computer by the United States
government at the Bureau of the Census in April 1951 demonstrated to some private
firms that it was technically feasible to use computers for business data processing. In
1946, the National Bureau of Standards awarded a contract to John Eckert and John
Mauchley to prepare original design proposals for a computer that could be used for
the statistical tabulations performed by the Census Bureau. In June 1948, the Eckert
and Mauchley Computer Company was awarded a contract to build the first Univac
computer. Lacking sufficient capital to complete the project, Eckert and Mauchley sold
their company to Remington Rand in 1950, and in April 1951, Rand delivered the
Univac I to the Census Bureau. By 1954, the capacity of the first Univac was used
completely, and in 1955 the Census Bureau and the Internal Revenue Service jointly
purchased a second Univac.
Computers became commonplace in insurance companies by the end of the
1950s. In June 1954, Metropolitan Life was the first insurance company to install a
computer, followed by Allstate which installed a medium-sized computer at its
headquarters in Skokie, Illinois in December of that year. At Metropolitan Life, the
successful operation of the Univac I by the Census Bureau was a key factor in the firm's
final decision to purchase a similar machine. The pace of introducing computers in
insurance companies picked up in 1955 and by the end of 1957 at least fifty-six
insurance companies had installed computers. By 1963, almost three hundred
insurance companies, accounting for 90 percent of the total employment in the industry,
were using computers or had a computer scheduled for installation. The total value
of installed computer equipment in the public and private sectors in the United States
rose rapidly from $269 million in 1956, to $1 billion by 1959, and to $6 billion by 1965.
Large firms, especially insurance companies, banks, public utilities, and major
manufacturers, were the first businesses to experiment with electronic data processing.
But the increased availability of equipment, the introduction of less expensive models of
computers, and evidence of successful applications contributed to the rapid rate at
which smaller firms installed computers during the late 1950s and early 1960s. The
insurance companies that installed computers in 1954 and 1955 had an average of ten
thousand employees each. By 1960, more than two-thirds of the insurance companies
with at least five hundred employees and two-fifths of the smaller companies used at
least one computer. The rate of installation increased rapidly after 1960 as many small
firms installed their first computer and as large and medium-sized companies installed
Early discussions of the possible impact of automation assumed a framework in
which a set of clerical routines would be analyzed and transferred to the computer,
resulting in an "automated office." This assumption viewed the introduction of
computers as a significant, but discrete, change in office technology. Within a few
years of the first business uses of computers, however, it became apparent that office
automation was a more dynamic process. Companies that installed computers early in
the period exhausted the capacity of their first machines and rapidly replaced
equipment or installed more computers to handle new applications. Metropolitan Life
had three Univacs by 1956 which were operated for two shifts each day. By 1963,
insurance companies had an average of three computers each.
The drop in price of computer hardware coupled with technological
improvements in computers and peripheral devices made computers affordable for a
wider range of businesses during this period. Between 1958 and 1961, all computer
manufacturers converted from vacuum tube to transistor-based machines which
increased the reliability and speed of computers while simultaneously lowering the
costs of production and use. Memories were expanded and made more reliable,
and improvements in magnetic tape technology and high speed printers facilitated the
transfer of data in and out of the computer. Beginning in 1954 when IBM introduced
its 650 RAMAC computer, manufacturers started to offer smaller computers for
customers who did not need the capacity of a large system.
Changes in computer programming techniques also made computers easier to
use and more flexible for a variety of applications. The Univac I was the first
commercially available computer that provided the capability for stored programming.
With stored programming, the instructions needed for each computing job were stored
in the computer's high speed memory along with the data for each problem. Unlike
tabulating equipment, which had to be rewired in order to change instructions, the
capability to store instructions in the computer's memory made it much easier to change
from one data processing job to another. New programs were transferred to the
computer's memory from magnetic tape or punched cards, and once a job was
completed the program was saved on cards or tapes for reuse.
In spite of the advantages of stored programming, the first generation computers
were tedious to program and cumbersome to operate. All programs were written in
machine language which consisted of codes made up of various combinations of the
binary digits one and zero. During the late 1950s, assembly languages and general
programming languages, such as FORTRAN, were developed which enabled
programmers to use abbreviations and mnemonic devices to write programs. Most
second generation computers came with an operating system, which was a set of
multi-purpose programs for translating programming languages into machine language
and for controlling the input and output devices. By the mid-1960s, some operating
systems had provisions for time-sharing which allowed users to run several programs
simultaneously on the same computer.
During the first decade of computer use the number and variety of applications
grew steadily. Most companies launched their automation program with a few simple
applications which would allow them to gain experience in programming, assess the
difficulty of conversion, and reduce risks if complications arose with installation or
operation of the computer. Metropolitan Life, for example, first used its computer to
calculate actuarial statistics. This area was chosen for the first application in part
because the information was used internally and it had no direct bearing on contact
with policyholders. After the successful use of computers for this purpose,
Metropolitan Life rapidly added premium billing and accounting, policyholder
dividends, and agent commission accounting as four separate applications. By the
early 1960s, these applications were integrated into one single automated process.
Premium billing and accounting were the most common early computer
applications in insurance companies. Prior to the installation of computers, most
companies billed policyholders on a monthly basis from the central home office. In
spite of an extremely specialized division of labor and a high degree of mechanization,
these tasks required large amounts of clerical labor. Large numbers of clerks retrieved
billing information from policy files, determined of the amount of the premium due,
typed or produced policy notices on duplicating machines, and manually checked the
notices for errors.
Premium accounting was equally labor intensive because the history records on
each policy were updated to reflect payments, payments were distributed to several
different accounts, and general accounting and statistical data was revised. Agents'
commissions often were individually calculated as a portion of each premium. The
installation of computers eliminated some of the manual clerical jobs involved in
premium billing and accounting. Many firms transferred part of the history data on
each policy to magnetic tape or punch card files and developed computer programs to
determine the premiums and print notices as part of the regular billing cycle. When
premiums were paid, accounting data was added to the history record for each policy,
commissions were calculated automatically, and the premium payments were
distributed among the various accounts. Automation eliminated the need for much of
the retrieval of information from manual files, the calculation of premiums and
commissions, and the detailed checking of data in various files. Many firms also
consolidated their record keeping operations by making all changes to policy records,
such as changes in names and addresses, during one pass of the master policy file
through the computer, usually on a weekly or monthly cycle.
During the first decade of computer use, insurance companies rapidly added
more policy and accounting data to their computerized files and increased the number
of computer applications. By 1963, half of the insurance companies were using
computers for nine or more applications including premium billing and accounting,
commission accounting, valuation of reserves, agency and actuarial statistics, dividend,
mortgage and loan accounting, and payroll. Some companies had twenty or more
computer applications including general accounting, claims processing and
The experience that managers gained with adapting office work to the demands
of the new technology is another essential element of the first decade of
experimentation with computer technology which is often overlooked in
machine-oriented histories of technology. The detailed knowledge that they acquired
about the clerical labor process was necessary before they could profitably exploit the
potential of computers. Although early feasibility studies pointed out the need to
standardize and reorganize office work in order to make efficient use of computers,
only practical experience revealed the complexity of that task.
Discovering Skill in "Unskilled" Work
As computer use expanded rapidly during the late 1950s and early 1960s,
unexpected problems surfaced during the conversion of manual clerical work to
automated processes. Translating what appeared to be very simple clerical routines to
computers proved considerably more difficult than many managers had anticipated
because they often overlooked subtle complexities in even the most routine clerical jobs.
Personnel disruptions often accompanied the installation of computers as manual jobs
were eliminated and new positions were created in data handling, machine operation,
programming, and systems analysis. These unanticipated problems during the
conversion process created considerable turmoil for office workers and managers.
The initial enthusiasm for computers as a cure-all for economic and social
problems in offices started to wane as experience demonstrated that redesigning an
entire clerical process was considerably more difficult than most original planners had
anticipated. In preparation for the installation of a computer, most firms completed
detailed procedures analyses in order to redesign clerical routines for machine
processing. Yet the early studies greatly underestimated the time and cost required to
reorganize office work and to reach efficient operating capacity with the new machines.
At G.E., for example, feasibility studies estimated that the payroll could be processed in
two hours and that three other applications would require ten hours of computer time.
More than a year after installation, however, the payroll alone took twenty hours of
computer time and the three other applications were still being designed. A
vice-president of Prudential reported in 1957 that the target dates for its pilot operations
"bear no resemblance to original target dates which we had in our minds when our first
large-scale computer -- the 702 -- was delivered" because conversion from manual to
automated premium billing, estimated originally to take six months, actually took
thirteen months. These were not isolated incidents. As John Diebold concluded in
1959, fiascos in data processing were so common that "good applications seem to be the
Efforts to analyze clerical work processes and develop computer programs to
simulate the recording, checking and decision-making processes of clerical workers
revealed the hidden knowledge and skills of clerks and their supervisors. Ken
Kusterer, a sociologist who has studied a variety of so-called unskilled jobs, believes
that managers frequently overlook a wide range of skills that are essential to perform
most jobs. He argues that the successful performance of "unskilled" work requires a
basic knowledge of routine procedures as well as supplementary working knowledge
about the materials or documents handled, the machinery used, and patterns of
behavior among customers, clients, and management. Generally, managers recognize
only the basic knowledge necessary to master routine procedures, and they frequently
overlook the supplementary knowledge that workers develop in order to cope with
variations in the materials they process, the machines they operate, or the customers
The tendency of managers to ignore hidden skills, and not compensate workers
for them, is also a characteristic of jobs that are sex-typed as women's work. Recent
studies of comparable worth have demonstrated that the skills necessary to perform
typical women's jobs are often undervalued compared to those required for comparable
men's jobs. The historian, Susan Porter Benson, has also observed that department
store managers misunderstood saleswomen's skills of gauging customers desires and
meeting their demands. Similarly, Ann Machung, a political scientist, contends that
secretarial work requires social skills which go unnoticed and are uncompensated by
employers. In office work, the strict division of labor between women's clerical jobs
and men's professional and managerial positions exacerbated the tendency to overlook
or undervalue the skills required of many clerical workers.
By the late 1950s insurance companies subscribed to a well-established tautology
between women's work and routine work. Reflecting broader societal views of women
workers as uncommitted to a career and hence undeserving of jobs that required
initiative, judgment and skill, the notions of routine work and women's work were
almost inseparable. According to J. Howard Ditman, a manager at New York Life,
insurance companies had long recognized that many functions could be broken down
into routine operations that could be processed by clerical workers with relatively little
experience, allowing insurance companies to adopt the "theories of interchangeable
Efforts to automate clerical routines revealed that the information needed to
issue and monitor insurance policies was not as standardized or interchangeable as
parts on an assembly line. Even in paperwork industries such as insurance and
banking, where manual clerical procedures were rationalized and codified, lower-level
managers, supervisors and clerical workers still made many case-by-case decisions. In
insurance, for example, a portion of the policies failed to conform to established
guidelines and procedures. With manual routines, clerical workers and their
supervisors relied on experience and judgment to handle such cases.
The use of data processing techniques required that clerical routines be reduced
to simple logic, stripped of inessentials and ambiguities, and made perfectly uniform
with standardization of decision making at all levels. As one study pointed out, office
automation implied a strict rational system of work organization, without which the
machines were useless. One management expert concluded that no motivation had
previously existed for the strict and formal policies and procedures which were
absolutely necessities for data processing.
In a manual system . . . policy develops on a day-to-day basis, some even at lower
management levels. People who are designing highly integrated control systems are
only now beginning to discover the amount of subtle decision making which takes
place at the clerical level.
Early attempts to write computer programs that accounted for all of the possible
variations in clerical procedures demonstrated that even routine paperwork processes
were more complex than many managers assumed. During the conversion process,
when planners had to confront direct comparisons between the capabilities of
computers and the hidden skills and ingenuity of the workers they were to replace, it
became apparent that clerical work required a type of intelligence that the new
machines lacked. Melvin E. Davis, a vice-president of Metropolitan Life, admitted that
initial studies had lost sight of some of the practical limitations to the idea of completely
We do not always have the right information in the right place. A numeral can be
transposed in a policy number. John Smith, who bought a policy ten years ago,
became John Smythe five years ago, and may now be Jonathan Smythe. An
experienced clerk dealing with many practical situations can detect and correct small
discrepancies of this kind and find the records. Automatic machinery, however,
slavishly follows a given routine . . ."
This type of hidden skill needed for even the most routine clerical jobs, such as filing,
did not become apparent to managers until systems analysts attempted to design
automated systems for processing clerical work.
The conversion of manual routines to computers usually began with a concerted
effort to analyze existing procedures and to extract from supervisors and clerical
workers their detailed knowledge of the work process. In spite of well-developed and
codified procedures for handling the normal flow of paperwork, managers' detailed
knowledge of the labor process was somewhat limited. Written procedures, where
they did exist, were often out of date and they did not include the informal methods
that clerical workers and supervisors used to handle exceptions to the rules. In
preparation for automation, procedures analysts often required workers and
supervisors to prepare detailed reports on their activities or they observed the various
work units. At the Pacific Mutual Life Insurance Company, for example, a time study
was conducted for thirty days during which each employee was required to keep a full
record of her daily activities. At the end of each week, supervisors reviewed and
summarized the data for their work units and provided additional details if the
information was insufficient. At the end of the time study, all of the forms were turned
over to the computer committee for summary and analysis. Systems analysts and
programmers used this information to write computer programs that replicated the
implicit logic of manual paperwork processes.
The one to three year conversion period, intended to allow a smooth transition
from manual to automated systems, was unexpectedly chaotic because the process of
converting information from written records or punched cards into a form acceptable to
computers was also more complicated than originally projected. Conversion was labor
intensive even where advanced punched card systems had been used, because
numerous card files had to be consolidated into a single master file and errors and
discrepancies had to be analyzed and reconciled. During the conversion process at
Prudential, for example, thirteen million punched cards and manual files were
converted into a single master file on magnetic tapes. Quality control checks were
performed on all of the major files and three hundred thousand additional cards were
punched. Pre-conversion checks revealed 40 percent of the errors that had to be
corrected, but during the conversion process more than one hundred thousand
discrepancies were investigated and resolved.
In most firms, the workload increased dramatically during the conversion period
because clerks, coders, and key punch operators were needed to convert records for use
with computers. Meanwhile, manual systems could not be phased out until the
automated process was fully operational. At one medium-sized insurance company,
conversion added between fifty- and sixty thousand transactions per week to the
existing workload, requiring the services of thirty additional full-time clerks and several
part-time key punch operators. At Prudential, two hundred fifty people worked on
conversion at the height of the project, chosen from the operating department, the
planning department, and from outside departments on a "borrowed" basis. At
United Benefit Life (Mutual of Omaha), estimates indicated that it would take forty
people working for nineteen months to transfer name, address and identifying
information from addressograph plates to punch cards. In order to find enough
operators to key punch four million cards, the firm brought back former employees,
housewives, and other workers who could be hired on a temporary basis without any
promises of employment after the conversion was completed.
A combination of demands for speed, accuracy and precision, the increased
workload, and frequently missed deadlines made the conversion period extremely
stressful for office workers and managers. To shorten the conversion period and to
cope with shortages of data entry workers, some firms added excessive overtime to the
schedules of their permanent staff. A supervisor at a telephone company, where
billing was converted from a manual to an automated function, described the
conversion period as one of confusion. "When we first started, it was a hell of a mess.
There was a lot of pressure. My desk was piled high. There I was, trying to break in
clerks, checking their work. Being new on the job myself, it was quite a job." A
supervisor at a utility company also remarked that the conversion period was an awful
mess. "There were plenty of bottlenecks, very slow. You had to work harder. The
bills would take longer. There was plenty of overtime." A clerk at the same
company believed that the problems resulted from being half on the old system and
half on the new. "We'd feed the Univac, and we got in such a mess. They got behind,
lost the work, and then we had to do it all over again." At least initially, systems
designed to streamline clerical procedures and impose order on office work created
As long as the complexity of office work and the subtlety of office skills remained
unrecognized, managers assumed naively that machines could easily replace office
workers. Problems with early conversion projects forced managers to reexamine this
premise. Management expert began to recognize that workers' consent was necessary
to secure cooperation during the design phase when methods analysts had to gather
complete and accurate information about existing procedures from the office staff.
Furthermore, many managers believed that positive attitudes toward automation
would help to maintain workers' morale during the chaotic and often stressful
conversion period. Managers used the experience gained from early conversion
projects to establish a set of guidelines which were intended to ease the transition from
manual to automated clerical work.
Building Consent and Encountering Resistance
Management experts placed considerable emphasis on building support for
automation. As we saw in Chapter Three, even before any computers were installed,
unions and social critics expressed concern over the displacement potential of this new
technology. Furthermore, data processing systems were especially vulnerable to
passive resistance and direct sabotage. According to Lowell Hattery, a professor of
management at American University, one of the factors that made automation different
from simply introducing faster machines was that a data processing system could
"easily be slowed or rendered ineffective by organizational units or personnel who may
resist the necessary change-over." Similarly, an editorial in The Automatic Office
pointed out that sabotage could take many forms, and even a lack of interest on the part
of workers in making the system work could cause unintentional damage.
Many managers believed that workers' resistance could be curbed through
well-planned personnel programs, and most firms introduced broad ranging programs
to ease the transition to automation. Assumptions about women's life cycle patterns
and their behavior as workers were essential elements of plans for a smooth transition
to automated office work. Because routine clerical work, performed by young women
who were in scare supply, was most suitable for automated processing, most firms were
able to assure workers that no layoffs would result from automation. Metropolitan
Life, for example, had 1,200 unfilled vacancies for entry level clerical jobs when its first
computer was installed. Likewise, a manager at Aetna, explained that one of the
objectives of mechanized policy writing and accounting was to "give relief in areas of
critical personnel shortages by substituting machines that can be obtained for personnel
that cannot be hired (emphasis in the original)."
Managers also counted on high turnover rates among young women to reduce
the size of the office staff. In firms that were not facing an immediate shortage of
clerical workers, managers assumed that "natural" attrition during the conversion
period would eliminate surplus workers without the need to resort to layoffs. If firms
simply stopped filling vacancies once a decision was made to automate, "cupid and the
stork" would cushion the effects of displacement. At Metropolitan Life, workers were
assured that no one would be laid off or downgraded and that each employee's job
preference would be taken into account when reassignments were made.
Companies generally informed office workers of anticipated changes in advance
to minimize the damage of rumors and gossip. One government study found that all
but one of twenty firms told the office staff about the impending computer installation
through house organs, bulletin boards, meetings with supervisors and similar means.
At Pacific Mutual Life, which had a "model" personnel relations program, the entire
staff was informed in 1953 that the firm was considering automation. When the
computer was delivered in 1955, the staff was invited to watch to it being unloaded.
More than half of the staff volunteered to attend ten hours of lectures on computers,
mostly on their own time, and much to the surprise of management. Companies that
tried to keep plans for automation secret from employees risked heightened fear,
resentment, and hostility because information about the proposed installation often
leaked out and spread informally among office workers rather than through channels
that management controlled. A clerical worker in a telephone company reported that
she got word of the change two years before the official announcement. "The man
from IBM used to visit our boss, and then everybody figured it out. There was no
secret after that."
Employee education programs tried to build consent on the benefits of
computers by emphasizing what the American Management Association called the
"public relations value" of data processing and automation.
They spell progress: modern approaches to problems, upgrading of workers, less
routine work, better customer service, better quality products, better p & e ratios for
stockholders' and directors' attention. There is an appeal for every segment of the
Insurance companies consistently assured office workers that the jobs subject to
automation were repetitive recording, filing and calculating tasks and that workers
could look forward to more rewarding and demanding work after automation. The
program at the Pacific Mutual, for example, stressed the viewpoint that "freedom from
these highly repetitive and monotonous operations would be a boon to mankind and
specifically to employees of Pacific Mutual."
In spite of managerial efforts to enlist workers' support for automation, indirect
and passive resistance occurred and there were occasional cases of direct sabotage. In
some cases, efforts to undermine the system began during the initial procedures
analysis when workers recognized that the systems analyst was the modern equivalent
of the efficiency expert. In one large corporation where a time and motion study was
conducted, the management informed workers that their jobs would not be changed in
any way. But one of the clerks told Ida Hoos, a sociologist who studied office
automation in the early 1960s, that there was much conjecture among the clerks. "We
figured that all this measuring had something to do with the million-dollar baby they
were talking about."
Workers resented their lack of involvement in the planning process. As early as
1954, Paul R. Laurence had warned in the Harvard Business Review that blind spots
among staff specialists could contribute to workers' resistance because specialists did
not have the intimate contact with day-to-day operations that was necessary to develop
"a natural respect for the knowledge and skill that these people have." Nevertheless,
one 1960 study found that clerical workers and supervisors in four of six firms were
offended by management's failure to consult them during the changeover. Methods
analysts and planning experts, who assumed that workers did not understand the
relationship of their jobs to company-wide procedures, ignored the potential for advice
and assistance from workers. A methods analyst at a Southern Bank, for example,
claimed that consultation with employees was unnecessary because the system was
very simple and only a few revisions were made to the old methods. "There was no
useful purpose in the supervisors' or employees' participating. Anyway, they had only
limited knowledge of their jobs; they couldn't be helpful." Office workers, on the
other hand, often attributed problems with the new automated systems to the failure of
managers or systems analysts to consult them. A supervisor in a telephone company
claimed that if the methods analysts had discussed the new systems with the
supervisors "they would have known of the loopholes they would encounter in the
everyday workings of the job." A clerk in the same company believed that the analysts
should have also consulted the clerks because "[a] supervisor may know the work but
not the details."
Workers' resistance during the conversion phase ranged from sabotage to a
subtle lack of cooperation. Some managers complained that even when workers were
consulted in the design phase, it was difficult to obtain complete and accurate
information about the existing procedures. In one firm, male tabulating equipment
operators, whose jobs were about to be replaced by a large computer, turned out
incorrect data, failed to meet deadlines, and placed numerous obstacles in the path of
the new system. According to Hoos, managers did not know whether to interpret
these tactics as unconscious resistance to change or as concerted opposition. Indirect
and subtle forms of resistance appear to have been most common. A clerk at one
public utility believed that workers' reactions to automation caused some of the
difficulties during the conversion process. "Most of the people didn't want
automation. It scared them. The attitude of the people slowed it down quite a bit."
While it is difficult to gauge the full extent of resistance to automation, it is
apparent that resistance was not limited to clerical workers. Supervisors often were
the most vociferous opponents of automation because they had the most to lose from
the transition to automated processing. As numerous observers noted, the
introduction of data processing diminished the importance of experience in favor of
technical knowledge. As a consequence, supervisors felt threatened by computer
programs which included routines for handling the complex and exceptional cases that
had required their experience, judgment, and special skills in the past. Supervisors
whose departments were abolished or trimmed substantially suffered from a loss of
status even if their grade level remained unchanged. One study of responses to
automation among managers, methods analysts, supervisors and workers found that
supervisors had the most negative response in two of six companies examined. In a
southern bank, a supervisor reported that his job was ripped from under his feet, while
others had the impression that they were being treated like "old shoes, no longer useful,
about to be discarded."
The resistance of supervisors posed special problems for managers of automation
projects. Postwar management experts and sociologists believed that supervisors
played a crucial role in communicating managerial objectives to workers and in
maintaining the morale of the work force. During the transition to automation,
supervisory support was considered essential because morale often dropped once the
announcement of a computer installation was made, and the transition period was
fraught with uncertainty. In many cases, however, managers were not able to gain the
support of supervisors for the proposed change because the supervisors felt threatened
by the new methods. Moreover, supervisors usually posed the most difficult transfer
and reassignment problems. Whereas managers reported little difficulty in
transferring displaced clerical workers either to routine data handling jobs or to
vacancies in other departments, supervisors' skills usually were unique to one specific
Some middle managers were also a source of resistance to automation. While
younger, technically trained men in middle management positions generally promoted
automation vigorously, older men who had ascended to their management positions
through longevity and experience felt threatened by the changes that automation
brought about. One management expert concluded that middle management
resistance was common enough to be considered one of the "seven deadly dangers" of
electronic data processing. Systems design required precise definitions of instructions
which were previously verbal, and many middle managers felt vulnerable to criticism
of their methods when systems analysts scrutinized their procedures and developed a
"better way" to do the work. Middle managers who believed that hard work or
cunning formed the basis of managerial success were also reluctant to accept the idea of
Studies of office automation in the late 1950s and early 1960s provide little
evidence to suggest that workers' resistance was universal among computer conversion
projects. Hoos found a variety of responses in addition to occasional cases of passive
resistance. Some workers were resigned to let fate take its course while others took
advantage of the extra overtime, pressure for upgrading, and special attention from
management that accompanied the transition period. Still others subscribed to the
"exclusion hypothesis," assuming that even though jobs all around them were being
automated, their particular position could not be transferred to a machine. One team
of social scientists who studied workers' attitudes toward automation in a
medium-sized insurance company in the late 1950s also found that the exclusion
hypothesis was quite common. Three-quarters of the employees believed that
machines had replaced workers in insurance companies, yet 80 percent felt that their
own chances of being displaced were less than for most other jobs.
The ways that office workers and managers responded to office automation
depended on the conditions surrounding the project and on individual characteristics
such as personality, age, training, position and length of service. Studies often showed
that older workers and workers with long job tenure had a more difficult time adapting
to the new automated routines. There is also some evidence that informing
employees of the impending changes and involving office workers in the changeover
elicited greater cooperation. Management's concern over possible resistance to
automation and the accompanying employee relations programs added a human
relations element to many computer installation projects. This effort to build workers'
consent was the first attempt to synthesize human relations management with a broad
effort to rationalize office work. While the human relations programs were not
completely successful, managers subsequently applied the lessons of the early
automation projects to expanded data processing activities.
Automation and Skill
The impact of early computer installations on skill merits special attention
because promises of upgrading were an important aspect of managers' efforts to secure
cooperation from office workers. Moreover, historians, economists, and sociologists
have begun to examine the relationship between technology and skill, especially in
response to issues raised by Harry Braverman in his pathbreaking work Labor and
Monopoly Capital. Braverman argued that capitalists install new technology in order
to lower labor costs by allowing a substitution of less skilled machine operators for
skilled craft workers. Several historians have offered a slightly different explanation
for managers' motivations in the selection and deployment of particular technologies.
David Noble, in particular, contends that managers' foremost consideration in
technological choice is their ability to limit the control that workers maintain over levels
of output, the pace of work, and other aspects of workplace relationships.
Computer Programming: Reproducing the Gender Division of Labor
Computer programming was the main area of data processing work that
required an entirely new skill. Therefore, the recruitment of workers for data
processing positions provides insights into the extent of upgrading and the
incorporation of gender-based assumptions into the requirements for many new
computer-related jobs. More than half of the workers displaced by automation were
women, yet by the mid-1960s only one woman had obtained a position as a systems
analyst or programmer for every six men. In life insurance companies, where women
made up 70 percent of all non-sales personnel, only 15 percent of the planners, systems
analysts and programmers were women.
At the outset, a number of factors suggest that women should not have been
excluded from the new trade of computer programming. At the Moore School, Adelle
Goldstine led a group of women college graduates in the development and detailed
coding of the first computer programs used for the ENIAC. When computers were
first introduced into offices, managers did not have any established criteria for
determining the attributes of good programmers and there was no tradition of
programming to suggest that either women or men made better programmers.
Moreover, there was no pool of trained applicants stacked in favor of one sex or the
other. Most firms recruited programmers from within, usually drawing between 70
and 80 percent of the data processing staff from employees on the payroll. Had
programmers been recruited only among workers whose
jobs were eliminated by automation, the pool of potential applicants would have
included more women than men.
In business data processing there is no evidence that the experiences of the first
women programmers established a precedent or influenced assumptions about the
characteristics of good programmers. In fact managers were unsure of the qualities
that a good programmer should possess. George E. Wallace, a manager at John
Hancock, asked how managers should classify the programmer.
Is he a professional man, a creative artist, a mathematician, a skilled technician? He
may be any of these things in different installations, and no one is yet quite sure what
qualities he needs.
During the early 1950s, managers tended to overstate the education and training
need for success in computer programming. Many believed that formal training in
mathematics at least at the college level was essential for programmers. Yet the lack of
candidates with such qualifications forced managers to substitute more modest criteria,
such as an aptitude for mathematical or logical thinking. As Edward A. Robie,
personnel director at the Equitable explained, "In our initial state of awe, I believe we
tended to overstate the difficulties of training people to master the complexity of EDPM
[electronic data processing machine] jobs."
Early experience convinced many managers that existing staff could be trained to
program computers and that knowledge of company operations was as important to
programming as technical qualifications. Most firms responded to the shortage of
trained programmers by sending candidates to schools operated by the computer
manufacturers and by establishing their own in-house training schools. Until the late
1960s, when technical schools and colleges began offering courses in computer
programming, programmers were not expected to have any prior training or
The willingness of firms to train their own programmers might have created
opportunities for women to enter the field. The recruitment criteria for programmer
candidates focussed on aptitude rather than demonstrated ability, and the indicators of
aptitude did not fall clearly into prevailing sex-linked characteristics of men or women.
Managers considered an ability for logical or abstract reasoning, patience, an obsession
for detail, extreme accuracy, and a creative imagination among the most important
attributes of workers who were likely to succeed as programmers. While men were
often assumed to be superior in abstract thinking and creative problem solving;
patience, attention to detail, and accuracy were often used to describe attributes that
made women especially adept at clerical work.
In spite of the absence of an established gender tradition or clearly sex-typed
criteria for programming work, men were recruited for most programming jobs. The
increasing emphasis on experience and knowledge of the business as criteria for
successful programming is one structural factor that worked against women.
Programmers were often recruited from accounting and related professional work
where women had made few inroads. According to one study, the nature of the work
in which programmers were previously involved was the major determining factor in
their choice of a programming career. At some point, their previous work brought
them into contact with the computing field and resulted in a decision to take up
Informal social contacts were very influential in shaping the gender composition
of the new programming jobs. As Joan Greenbaum has pointed out, during the 1950s
and 1960s, most programmers "literally stumbled into the field." Many of the young
men who became programmers either came into contact with someone who encouraged
them to apply for a programming job or they entered the field "by accident" without
any clear understanding of what the work involved. These informal practices,
however, seem to have excluded women from consideration for programming work.
Broader social assumptions about women as workers and the role of women in
the family also influenced the recruitment of workers into this new field. Because most
firms paid to train programmers, many managers were reluctant to pay for women's
training on the assumption that they might not remain active in the work force. While
managers used the human capital argument to limit the number of women in skilled
data processing jobs, they grudgingly accepted high turnover rates among
programmers generally because the overall scarcity of trained workers made job
mobility easy. According to John Diebold, the demand for experts of all types in the
computing field had created "a large group of job hoppers who would be considered
`floaters' in any other field." Firms that did hire women programmers found that
their tendency to remain in a programming job rather than seek advancement to
supervisory or managerial positions was a major advantage of women programmers.
Some firms excluded women from programming and computer operating jobs
because of the demand for night shift work. Prudential, for example, had three women
working in computer operations, but the company decided to restrict this work to men
because of the shift arrangement. The Equitable and Metropolitan Life also excluded
women from night shift work. Yet managers admitted that shift work was unpopular
among workers of either sex and that special inducements often were necessary to
encourage workers to accept shift work. Moreover, some firms operated evening
shifts for key punching and other data entry work which were made up entirely of
women. When women took evening key punching jobs, some experts viewed the
evening shifts as a special opportunity for housewives, who might otherwise be unable
to leave children at home, to accept an office job.
Advertising and recruitment literature discouraged women from applying for
programming jobs because most jobs were advertised in the "Help Wanted - Male"
section of newspapers, and the advertisements did not indicate that women applicants
would be considered. Only a few of the companies that prepared career recruitment
literature about programming developed special literature aimed at women, while one
government publication on automation and employment opportunities in office work
stated explicitly that men were preferred as programmer trainees in most areas.
Nevertheless, some women did apply and were accepted for programming positions.
At the Virginia National Bank, where the management used the same selection
procedures for men and women programmers, two hundred of the seven hundred
employees attended a meeting to learn about programming jobs, ninety-eight took the
screening test, and eight of the thirteen workers chosen to take the training course were
The widely-held assumption during the 1950s and early 1960s that programming
would become a high status occupation with potential for mobility into management
positions was another key factor in the limited recruitment of women into this field.
Although managers were unsure of exactly where programmers fit in the office
hierarchy, many believed that experience in analyzing business problems systematically
would eventually qualify programmers for management positions. Sam Cohn has
argued that managers are reluctant to hire women for entry level management jobs and
trainee positions because discrimination against women restricts the supply of potential
applicants and keeps the salaries of managers artificially high. This practice is most
common when only a small number of workers are involved and the higher wages paid
to male employees have little effect on overall operating costs.
Beth Parkhurst and Joan Richards have studied the women in the early decades
of computer programming and have found that women were better represented as
programmers in scientific and engineering applications than in business data
processing. They explain this difference by pointing out that in scientific fields
programming was relatively low in status compared to the work of engineers or
scientists, but in business programming offered upward mobility. The small number
of programmers provided little incentive to search for workers who would accept low
wages while the assumption that programmers could rise to managerial positions
reinforced informal preferences for male workers in this field.
Changes in skill requirements for office work and the rapid sex-typing of skilled
data processing jobs as men's work heightened the differences in jobs opportunities for
young men and women who entered the office work force in the early 1960s. A
reduced demand for clerks in firms that had installed data processing equipment
coupled with an increase in the number of high school graduates made it more difficult
for young women to find office jobs. While female office employment in the
traditional areas of typing, secretarial and stenographic work continued to grow, high
school graduates who lacked training in these areas faced a reduced demand for clerks.
One study of automation in a large Milwaukee insurance company found that hiring of
new employees was 50 percent lower in 1962 than in 1956. The study concluded that
girls graduating from high school in 1964, who did not have specific typing or
shorthand skills, would have to look elsewhere for employment.
The employment situation for young men was somewhat brighter, especially for
those with college degrees or some college education. Some analysts predicted that the
use of computers might actually improve a boy's chances of finding a job because many
of the routine jobs were open only to women, while men were preferred for skilled data
processing positions. In insurance companies, the increase in demand for agents and
salesmen, 90 percent of whom were male, also enhanced the employment opportunities
for young men. The Milwaukee study found that the increase in sales positions
created an entry area for a significant number of men with some college education.
Demand for women with college degrees or some college education, however, did not
increase. The study concluded that "typing and shorthand ability, rather than college
education, would enhance the employability of female applicants at this firm."
The limited opportunities for women in the new fields of systems analysis and
programming widened the gap between men's and women's office jobs. Although
women had filled the most routine jobs in the office hierarchy prior to automation, there
had been some possibility of mobility for women with long job tenure. Women who
became familiar with complex clerical routines were often promoted to positions where
they supervised small groups of clerical workers. When supervisory positions were
eliminated during the transition to automation, the only source of mobility from many
clerical jobs was abolished. The separate career tracks for clerical workers and for
technicians or managers corresponded more closely than ever to the positions held by
women and men in offices.
Toward Blue Sky: Top Management Responses to Automation
After a decade of computer use in businesses, many managers were
disappointed with their limited effect on cutting office costs and curbing the overall
growth of the office work force. Managers often admitted that beneath the surface of
remarkable progress in data processing systems, there was considerable concern over
the lack of evidence of an adequate return on their investment. Studies of computer
installations rarely provided evidence that automation resulted in significant cost
savings. One study of 120 medium-sized firms found that computers had not paid off
in 72.5 percent of the cases. Another assessment of both return on investment and less
tangible long-term benefits from computers determined that the results in nine
companies were unmistakably successful, but in eighteen companies they were
marginal at best.
Several factor contributed to the unspectacular results of many early computer
installations. First, early projections had underestimated both the costs of equipment
and the difficulty of conversion. The conversion period often stretched over several
years, and by the time the system was fully operational revisions or new enhanced were
being envisioned or planned. Second, early feasibility studies did not anticipate the
extent to which automation would transform information handling beyond basic
clerical routines. Cost benefit analyses rarely accounted for the myriad of new reports
and detailed management information that computers would make available and
managers would request. Finally, managers did not recognize that data processing
would become an expensive and growing operation which would not enhance a firm's
profits unless it was subjected to managerial controls.
Few managers were willing to abandon computer technology. Rather, they
used new arguments to justify further investment and experimentation with computers
and they explored new ways to control data processing costs. In an important shift
from the brute force to the blue sky approach, managers turned to the intangible
benefits of automation as the primary justification for investments in data processing
systems. As evidence of meager returns from computer installations mounted, some
experts and equipment salesmen argued that the new information made available by
advanced data processing technology, rather than direct reductions in office costs, was
the main benefit of computers.
The argument that intangible benefits could justify the expense of large data
processing operations was not completely convincing. Studies often demonstrated that
managers were incapable of handling larger volumes of information, yet the ability of
the data processing system to deliver information rapidly increased their desire to be
informed. According to one expert, there was no evidence that managers needed more
A man who asks for a monthly statement of specific purchases by 35,000 customers, as
one divisional manager of a chemical company does, may read scarcely any of it, but he
continues to want it because he "feels that this helps him do a better selling job."
Some managers even complained that computers added to their problems of
information overload by producing too many useless reports and providing too many
undigested facts. According to Herbert Cherry, Director and Data Processing and
Planning at the Minnesota Mutual Life Insurance Company, planners were unable to
define management information or to describe the types of information they needed to
plan for the future, evaluate the company's progress, or control its operations.
Furthermore, intangible benefits were a difficult selling point for hard-nosed
businessmen who were most concerned with profitability. According to N. J. Dean of
the consulting firm Booz, Allen and Hamilton, one of the difficulties of selling the idea
of benefits from new information was that the concept had the word "intangible"
attached to it. "Intangible results do not interest the board of directors."
Managers who were disappointed with the results of office automation and
disillusioned with high costs and frequent project delays usually were willing to write
off the initial experimentation as a learning experience and a significant step in the
long-term process of rationalizing office work. Robert W. Harvey, a second
vice-president at Prudential, told the Life Office Management Association, that his firm
had reason to be proud of the progress they had made in data processing, yet
I will be quite frank and tell you that if I thought our first set of computer systems was
to be the final result by which the value of our investment in electronics was to be
measured, I would be disappointed in that result. But, in actuality, pricing the
economies and the improvements resulting from this first set of systems gives only a
partial accounting of what we have received for our money, for it does not measure the
value of a staff of trained and experienced people who will eventually -- or, in some
cases, are even today -- planning the more extensive and more effective systems which
will be installed tomorrow.
Harvey went on to insist that people in charge of computer projects could not fulfill
their responsibilities unless "they also have a clear appreciation of the power which the
machine puts at their command. No manager can afford to be an uninitiated slave to
an inscrutable machine attended by a secret priesthood."
Rather than abandon computer technology, managers looked toward the future
and redoubled their efforts to apply profit-oriented criteria to data processing activities.
Numerous studies of computer installations pointed to areas where management had
failed through ignorance, lack of planning, and delegation of managerial
responsibilities to middle-level technical experts. A McKinsey Corporation study, for
example, concluded that involvement and leadership from top management was the
single most important factor in determining the success of a computer installation.
These early studies convinced many managers that the unimpressive benefits of
automation were not caused by technical or engineering problems, but resulted from
managerial decisions which could be reversed.
In an effort to enhance the profitability of computers, managers extended their
control over data processing operations by establishing more realistic completion dates
and then insisting that systems analysts and programmers meet the deadlines. The
autonomy that programmers enjoyed during the early years of computer use was
curbed significantly when the process of programming was subdivided into planning,
analysis, writing code and coding. This fragmentation of programming made it easier
for managers to develop and insist upon tighter schedules. Some firms taught basic
computer concepts to middle managers so that they could have more control over data
processing operations, while other companies taught management principles to their
Much of the growing managerial acceptance of computers in the mid-1960s was
based on projections of future benefits from more sophisticated uses of information
technology. With expanded computer memory capacities, improvements in
communications technology, and more complex programming methods, many firms
extended the organizational and geographic coverage of their data processing systems.
Insurance companies, for example, were developing the "total systems" concept where
all of the data needed for policy processing and internal financial controls was stored in
one centralized system and where data could be transferred between the home office
and agents. These systems, many still in the design phase, began to resemble the
integrated information systems that managers had assumed would be the principle
advantage of automation.
Top managers also became more supportive of new information technology
when they recognized its potential to enhance their control over middle-level decision
making. In 1958, Harold Leavitt and Thomas Whisler suggested that the new
information technology was beginning to move the boundary between planning and
execution upward in the management hierarchy. They predicted that by the 1980s,
decision making would be centralized among a few top managers as middle-level
decisions were either automated or eliminated. "One reason that we expect top
acceptance of information technology," they argued "is its implicit promise to allow the
top to control the middle just as Taylor allowed the middle to control the bottom."
With such projections, managers recognized that they were only beginning to exploit
the potential of office automation to transform a wide range of clerical, professional and
middle management jobs.
Byron F. Burch, "The Computer at Work on Payrolls," American Management
Association, Office Management Series No. 136 (New York: AMA, 1954), pp. 16-25; W.
W. Smith, "The Computer in Industry: Impact of Computer Methods," American
Management Association, Office Management Series No. 136, pp. 3-16; and Statement of
Ralph J. Cordiner, in U.S. Congress, Joint Committee on the Economic Report,
Automation and Technological Change (Washington, D.C.: U.S. Government Printing
Office, 1955), pp. 445-46.
Thomas Kenny, "Where Do You Go From Here? Start Now to Plan Your Studies,"
Dun's Review and Modern Industry (October 1955), p. 105, and A. B. Toan, Jr.,
"Business Strategy and Electronics," Proceedings of the 1956 Annual Conference of the
Life Office Management Association (New York: LOMA, 1956), pp. 56-62.
Kenny, "Where Do You Go From Here?" p. 105.
William Aspray and Donald Beaver, "Marketing the Monster: Advertising
Computer Technology," Annals of the History of Computing 28 (April 1986), pp. 127-43.
Morris H. Hansen and James McPherson, "Potentialities and Problems for EDP: the
Role of the Census and the Need for Data Processing Equipment," in Electronics in
Management, Lowell H. Hattery and George P. Bush, eds. (Washington, D.C.:
University Press of Washington, D.C., 1956), pp. 55-56, and Joan M. Greenbaum, In the
Name of Efficiency: Management Theory and Shopfloor Practice in Data-Processing
Work (Philadelphia: Temple University Press, 1979, p. 26.
Testimony of Robert W. Burgess, in U.S. Congress, Automation and Technological
Change, pp. 72-73, and Hansen and McPherson, "Potentialities and Problems for EDP,"
U.S. Department of Labor, Bureau of Labor Statistics, Studies of Automatic
Technology, "No. 2 - The Introduction of a Computer in a Large Insurance Company,"
reprinted in U.S. Congress, Automation and Technological Change, p. 294, and U.S.
Department of Labor, Bureau of Labor Statistics, Impact of Office Automation in the
Insurance Industry, Bulletin No. 1468 (Washington, D.C.: Bureau of Labor Statistics,
1966), pp. 4, 7-13.
Gerald W. Brock, The U.S. Computer Industry: A Study in Market Power
(Cambridge, Mass.: Ballinger Publishing Company, 1975), pp. 13-15.
U.S. Department of Labor, Impact of Office Automation in the Insurance Industry,
U.S. Department of Labor, Impact of Office Automation in the Insurance Industry,
p. 11, and R. Hunt Brown, Office Automation: Insurance (New York: Office Automation
Consultants, Inc., 1959), Part III, pp. D6 4.
Brock, The U.S. Computer Industry, p. 14; U.S. Department of Labor, Impact of
Office Automation in the Insurance Industry, p. 9, and Greenbaum, In the Name of
Efficiency, pp. 27-29.
With stored programming, computers were able to execute a long sequence of
instructions or to perform a series of simple instructions repeatedly on a large volume
of data without interruptions to change instructions or rewire plug boards. For a
discussion of the origins of the stored programming concept, see Herman H. Goldstine,
The Computer from Pascal to von Neumann (Princeton: Princeton University Press,
For discussions of the history of programming languages, see Jean E. Sammet,
Programming Languages: History and Fundamentals (Englewood Cliffs, N.J.: Prentice
Hall, 1969), and History of Programming Languages, R. L. Wexelblat, ed. (New York:
Academic Press, 1981).
U.S. Department of Labor, Bureau of Labor Statistics, "The Introduction of a
Computer in a Large Insurance Company," p. 294; U.S. Department of Labor, Impact
of Office Automation in the Insurance Industry, pp. 17-18, and Brown, Office
Automation: Insurance, Part III, pp. D6 1-4.
U.S. Department of Labor, Impact of Office Automation in the Insurance Industry,
p. 11. For descriptions of the manual billing process, see Brown, Office Automation:
Insurance, Part III, pp. D2 3-5 and D8 2-5.
For descriptions of automated billing and accounting procedures in life insurance
companies, see Brown, Office Automation: Insurance, Part II, pp. D2 5-6, D7 3-4, and D8
U.S. Department of Labor, Impact of Office Automation in the Insurance Industry,
"Effects of Mechanization and Automation in Offices: I," International Labor
Review (Feb. 1960), p. 162; Walter Bloomberg, Jr., The Age of Automation (New York:
League for Industrial Democracy, 1955), pp. 21-23; John Diebold, "False Starts in Office
Automation and how to avoid them," The Management Review (July 1957), p. 81;
Society of Actuaries, Committee on New Recording Means and Computing Devices,
Report (Chicago: The Society of Actuaries, 1957), p. 75, and Stewart Toy, "Bugs in
Automation," Computers and Automation 10 (May 1961), p. 10.
Ralph J. Cordiner, testimony in U.S. Congress, Automation and Technological
Change, pp. 445-46 and Richard B. Cole, "Brain Trouble," Wall Street Journal (Nov. 11,
1955), pp. 1 and 10.
Thomas Allsopp, "Getting the Electronics Show on the Road," Proceedings of the
1957 Annual Conference of the Life Office Management Association (New York:
LOMA, 1957), pp. 124-25.
John Diebold, Automation: Its Impact on Business and Labor (Washington, D.C.:
National Planning Association, 1959), p. 5.
Ken Kusterer, Know-How on the Job: The Important Working Knowledge of
"Unskilled" Workers (Boulder, Colo.: Westview Press, 1978).
For examples, see Comparable Worth and Wage Discrimination: Technical
Possibilities and Political Realities, Helen Remick, ed. (Philadelphia: Temple University
Press, 1984); Sex Discrimination in the Workplace, Barbara F. Restin, ed. (Washington,
D.C.: National Academy Press, 1984), and Mary Witt and Patricia K. Naherny, Women's
Work: Up From 878-Report on the DOT Research Project (Madison, Women's Education
Resources, University of Wisconsin-Extension, 1975).
Susan Porter Benson, Counter Cultures: Saleswomen, Managers and Customers in
the American Department Store (Urbana: University of Illinois Press, 1986), especially
Anne Machung, "From Psyche to Technic: the Politics of Office Work," (Ph D diss.,
University of Wisconsin-Madison, 1983).
J. Howard Ditman, "The Application of Data Processing Equipment to Life Office
Operations," Proceedings of the 1955 Annual Conference of the Life Office Management
Association (New York: LOMA, 1955), p. 79.
For examples, see Edmund C. Berkeley, "Bugs in People," Computers and
Automation 10 (Sept. 1961), pp. 13-14, and Leonard Rico, The Advance Against
Paperwork (Ann Arbor: University of Michigan, Bureau of Industrial Relations, 1967),
"Effects of Mechanization and Automation in Offices: I," p. 161.
R. L. Fiock, Jr., "Seven Deadly Dangers in E.D.P," Harvard Business Review
(May/June 1962), p. 91.
Society of Actuaries, Committee on New Recording Means and Computing
Devices, Report (Chicago, Society of Actuaries, 1952), p. 16. For other examples, see
Berkeley, "Bugs in People," pp. 13-14; Fiock, "Seven Deadly Dangers of E.D.P.," pp.
91-92; and Ben Miller, Gaining Acceptance for Major Methods Changes (New York:
American Management Association, 1960), pp. 18-20, 31-32, and 45-46.
Berkeley, "Bugs in People," pp. 13-14, and Rico, The Advance Against Paperwork,
R. D. Dotts, "An Approach to Electronics by a Medium-Sized Company,"
Proceedings of the 1954 Annual Conference of the Life Office Management Association
(New York: LOMA, 1954), pp. 311-14.
Allsopp, "Getting the Electronics Show on the Road," p. 125.
Harold F. Craig, Administering a Conversion to Electronic Accounting (Boston:
Harvard University, Graduate School of Business Administration, 1955), p. 52.
Allsopp, "Getting the Electronics Show on the Road," p. 126.
George C. Boddiger, "Panel Discussion: Getting the Electronics Show on the Road,"
1957 LOMA Conference Proceedings, pp. 129-30. See also Albert Kushner, "People and
Computers," Personnel 40 (Jan./Feb. 1963), p. 29, and Floyd C. Mann and Lawrence
Williams, "Observations of the Dynamics of Change to Electronic Data Processing
Equipment," Administrative Science Quarterly 5 (Sept. 1960), pp. 234-39.
Miller, Gaining Acceptance for Major Methods Changes, p. 18.
Ibid., p. 38.
Lowell H. Hattery, "Management Impact of Electronic Systems," in Electronics and
Management, p. 5.
"How to Recognize and Overcome Resistance to Procedures Analysis," Automatic
Office 1 (March 1958), p. 2.
"Effects of Business Automation in the '60s," Management and Business
Automation 5 (Jan. 1961), p. 39, and Brown, Office Automation: Insurance, Part III, p.
Interoffice Communication, Sept. 4, 1956, Data Processing Division Records, Aetna
Resource Collection, Aetna Insurance Company Archives, Hartford, Conn.
Peter F. Drucker, America's Next Twenty Years (New York: Harper and Row,
1955), pp. 16-30; "Automation: Special Report," Business Week (October 1, 1955), pp.
88-92; Ida Hoos, Automation in the Office (Washington, D.C.: Public Affairs Press,
1961), p. 31, and U.S. Bureau of Labor Statistics, Impact of Automation, Bulletin No.
1287 (Washington, D.C.: U.S. Department of Labor, 1960), p. 53.
U.S. Department of Labor, "The Introduction of a Computer in a Large Insurance
Company," p. 293, and Richard F. Barry, "Placement and Retraining of Displaced
Personnel - Problem or Opportunity," 1957 LOMA Conference Proceedings, pp. 161-65.
U.S. Department of Labor, Bureau of Labor Statistics, Adjustments to the
Introduction of Office Automation, Bulletin No. 1276 (Washington, D.C.: U.S.
Department of Labor, Bureau of Labor Statistics, 1960), pp. 18-19. See also Raymond
Dreyfack, "How Seven Companies Prepared Their Employees for Automation,"
American Business 29 (August 1959), pp. 28-31.
Wesley S. Bagby, "The Human Side of Electronics," in American Management
Association, Pioneering in Electronic Data Processing: Company Experience with
Electronic Computers (New York: American Management Association, 1956), pp. 41-42.
For guidelines and advice on introducing computers, see U.S. Department of Labor,
Adjustments to the Introduction of Office Automation, Appendix E; Paul R. Lawrence,
"How to Deal with Resistance to Change," Harvard Business Review (May/June 1954),
pp. 49-57; American Management Association, Establishing an IDP Program: Blueprint
for a Company Program, Special Report No. 11 (New York: AMA, 1956), pp. 100-04;
Keith Davis, "Human Adjustments of Automation," Advanced Management 29 (Jan.
1961), pp. 23-27; Robert B. Curry, "Preparing Employees for Changeover," in Electronics
and Management, pp. 123-38, and Warren C. Stevens, "Seventeen Ways to Fight
Resistance to Change," Supervisory Management (1965), pp. 35-37.
Lowell H. Hattery, "Electronic Computers and Personnel Administration,"
Personnel Administration 19 (Mar./Apr. 1956), p. 12; Dreyfack, "How Seven Companies
Prepared Their Employees for Automation," pp. 28-30, and Miller, Gaining Acceptance
for Major Methods Changes, pp. 19-20.
Miller, Gaining Acceptance for Major Methods Changes, pp. 19-20.
American Management Association, Establishing an IDP System, p. 104.
Dotts, "An Approach to Electronics by a Medium Sized Company," p. 311.
Hoos, Automation in the Office, p. 60.
Laurence, "How to Deal with Resistance to Change," p. 53.
Miller, Gaining Acceptance for Major Methods Changes, p. 32.
Ibid., p. 19.
"How to Recognize and Overcome Resistance to Procedures Analysis," p. 2, and
Davis, "Human Adjustments to Automation," p. 27.
Hoos, Automation in the Office, p. 11.
Miller, Gaining Acceptance for Major Methods Changes, p. 38.
Ida Hoos, "When the Computer Takes Over the Office," Harvard Business Review
(July/Aug. 1960), p. 107; Harold J. Leavitt, "Dealing with Management Obsolescence,"
in American Management Association, Computer-Based Management (New York:
AMA, 1963), pp. 51-52; Leonard Rico, "The Staffing Process and the Computer,"
Management of Personnel Quarterly 1 (Autumn/Winter 1962), p. 35, and Fiock, Jr.,
"Seven Deadly Dangers of EDP," pp. 91-92.
Hoos, Automation in the Office, pp. 72-73, and Arlen Gray, "Problems of
Adjustment in the Automated Office," Personnel 41 (July/August 1964), pp. 43-48.
Miller, Gaining Acceptance for Major Methods Changes, p. 29.
American Management Association, Establishing an IDP Program, p. 102, and
Miller, Gaining Acceptance for Major Methods Changes, p. 29.
U.S. Department of Labor, Impact of Office Automation in the Insurance Industry,
p. 39; U.S. Department of Labor, Adjustments to the Introduction of Office
Automation, p. 28, and Wesley S. Bagby, "How to Win Active Employee Support for an
Electronics Program," Management Methods 10 (May 1956), p. 44.
Fiock, "Seven Deadly Dangers of E.D.P.," p. 91.
Milton M. Stone, "Management Attitudes Toward an Information System," in
Computer-Based Management, pp. 21-22.
Hoos, Automation in the Office, pp. 60-64.
Eugene J. Jacobsen, et. al., "Employee Attitudes Toward Technological Change in a
Medium Sized Insurance Company," Journal of Applied Psychology 43 (Dec. 1959), p.
Hoos, Automation in the Office, p. 64; William Faunce, "Social Stratification and
Attitudes Toward Change," Social Forces 39 (Dec. 1960), pp. 140-48; William Faunce,
Einar Hardin, and Eugene Jacobsen, "Automation and the Employee," Annals of the
American Academy for Political and Social Science 340 (1962), pp. 66-67; U.S.
Department of Labor, Adjustments to the Introduction of Office Automation, pp. 28-34;
"Effects of Mechanization and Automation in Offices: Part II," pp. 264-65; Dreyfack,
"How Seven Companies Prepared Their Employees for Automation," pp. 28-31, and
Virgil K. Rowland, "There'll Be Some Changes Made," in Men, Machines, and Methods
in the Modern Office, Management Report No. 6 (New York: American Management
Association, 1958), pp. 102-06. Ben Miller found that older employees, contrary to
popular belief, were superior in coping with change. See Miller, Gaining Acceptance
for Major Methods Changes, pp. 8-10.
Harry Braverman, Labor and Monopoly Capital: The Degradation of Work in the
Twentieth Century (New York: Monthly Review Press, 1974). See especially Chapters
9 and 20.
David F. Noble, Forces of Production: A Social History of Industrial Automation
(New York: Knopf, 1984). See also David Montgomery, Workers' Control in America
(New York: Cambridge University Press, 1979). Although Montgomery does not focus
exclusively on technological change, he considers scientific management and
accompanying changes in production technology as a fundamental management assault
on workers' control over production.
Peter B. Doeringer and Michael J. Piore, Internal Labor Markets and Manpower
Analysis (Lexington, Mass.: D. C. Heath, 1971). See especially Chapter Six.
For examples, see Montgomery, Workers' Control in America, and Case Studies on
the Labor Process, Andrew Zimbalist, ed. (New York: Monthly Review Press, 1979).
While the majority of the case studies in the Zimbalist anthology examine industrial or
craft work, several case studies of women's work are included. See Evelyn Nakano
Glenn and Roslyn Feldberg, "Proletarianization of Clerical Work," and Maarten deKadt,
"Insurance: A Clerical Work Factory." Margery Davies also argues that a major
technological change, the introduction of the typewriter, was important in the
feminization and proletarianization of clerical work. See Margery Davies, Woman's
Place is at the Typewriter (Philadelphia: Temple University Press, 1982).
U.S. Department of Labor, Adjustments to the Introduction of Office Automation,
Ibid., pp. 4 and 32.
U.S. Department of Labor, Bureau of Labor Statistics Technological Trends in
Major American Industries, Bulletin No. 1474 (Washington, D.C.: U.S. Government
Printing Office, 1966), p. 252.
U.S. Department of Labor, Impact of Office Automation in the Insurance Industry,
Robert L. Caleo, "What's Happening in the Office?" Administrative Management
25 (Jan. 1964), p. 26, and U.S. Department of Labor, Impact of Office Automation in the
Insurance Industry, p. 20.
Technical Committee on Mechanization of Check Handling, Automation of Bank
Operating Procedure, Bank Management Publication No. 142 (Jan. 9, 1958), reprinted in
U. S. Congress, Joint Economic Committee, New Views on Automation (Washington,
D.C.: U.S. Government Printing Office, 1960), pp. 347-478; Statements of A. R. Zipf and
Leonard P. Chamberlain in U.S. Congress, New Views on Automation, pp. 479-88, and
Statement of Everett J. Lindsey in U.S. Congress, Joint Economic Committee, Recent
Trends in Automation (Washington, D.C.: U.S. Government Printing Office, 1957), pp.
John H. Humphrey, "IDP in the Pre-Computer Age," in Men, Machines, and
Methods in the Modern Office, pp. 7-8.
Drawing on Durkheim's theories of the relationship between alienation and the
division of labor, numerous sociologists argued that automation represented a new
stage in industrial development that would reverse the ever-increasing specialization of
labor. For examples see Robert Blauner, Alienation and Freedom (Chicago: University
of Chicago Press, 1964), Charles R. Walker, "Life in the Automatic Factory," Harvard
Business Review 36 (Jan./Feb. 1958), pp. 111-19, and William Faunce, "Automation and
the Division of Labor," Social Problems 13 (Fall 1965), pp. 149-60. For critiques of this
assumption see Gray, "Problems of Adjustment in the Automated Office, p. 47; "Effects
of Mechanization and Automation in Offices: Part II," pp. 266-67, and William Eastman,
"Effect of Automation on Personnel Administration - A Look at the Future," 1957
LOMA Conference Proceedings (New York: LOMA, 1957), pp. 168-72.
"Effects of Mechanization and Automation in Offices: Part II," pp. 352-53; Albert
Kushner and Dallas H. Dobelbower, "Punched Card System Control," Business
Automation (Apr. 1962), pp. 26-29; John E. Bond, "Time Money Measurement in a
Keypunch Section," Administrative Management 25 (Apr. 1964), pp. 44-45, and "Desk
Work Gets Faster," Business Week (Oct. 30, 1965), pp. 91-94.
Hoos, Automation in the Office, p. 16.
U.S. Department of Labor, Adjustments to the Introduction of Office Automation,
p. 5; "Effects of Mechanization, Part II," p. 266; Testimony of Walter Buckingham in U.S.
Congress, New Views on Automation, p. 6; Jack Steiber, "Automation and the White
Collar Worker," Personnel (1957), p. 14, and H.C. Lee, "Automation and Skill
Requirements," Personnel Administration (1966), p. 53. One notable exception to this
conclusion was the study of George E. Delahanty, "Office Automation and the
Occupational Structure," Industrial Management Review 7 (Spring 1966), pp. 99-109.
In his study of five insurance companies, Delahanty found strong evidence that
substantial upgrading resulted from the elimination of the most routine jobs and the
increases in the demand of higher level clerical workers.
For examples see Delehanty, "Office Automation and the Occupational Structure,"
pp. 99-109; U.S. Department of Labor, Adjustments to the Introduction of Office
Automation; and Lee, "Automation and Skill Requirements."
Bagby, "How to Win Active Support for an Electronics Program," p. 46.
Hoos, Automation in the Office, pp. 68-70, and Kushner, "People and Computers,"
Hoos, Automation in the Office, p. 69.
For an example, see Craig, Administering A Conversion to Electronic Accounting,
Kushner, "People and Computers," pp. 32-34.
Mann and Williams, "Observations of the Dynamics of Change to Electronic Data
Processing Equipment," pp. 249-50.
James R. Bright, "Does Automation Raise Skill Requirements?" Harvard Business
Review (July/Aug. 1958), pp. 85-98.
American Management Association, Establishing an IDP System, pp. 109-11.
Ibid., p. 113.
Clayton M. Nicholson and Richard Kelly, "Are Standard Clerical Skills Suited to
Today's Automated Office?" Administrative Management, (Nov. 1961), p. 12.
Ibid., p. 13.
Life Insurance Fact Book for 1963 (New York: Institute for Life Insurance, 1963), p.
99; U.S. Department of Labor, Impact of Office Automation in the Insurance Industry, p.
26, and U.S. Department of Labor, Adjustments to the Introduction of Office
Automation, p. 52.
Goldstine, The Computer from Pascal to von Neumann, and Philip Kraft, "The
Industrialization of Computer Programming," in Case Studies on the Labor Process, pp.
U.S. Department of Labor, Impact of Office Automation in the Insurance Industry,
p. 31, and U.S. Department of Labor, Adjustments to the Introduction of Office
Automation, p. 5.
George E. Wallace, "Operating a Data Automation System," Proceedings of the
1958 Annual Conference of the Life Office Management Association (New York:
LOMA, 1958), pp. 58-60.
Edward A. Robie, "Selection and Training of Systems Personnel, Programmers and
Operating Staff," 1957 LOMA Conference Proceedings (New York: LOMA, 1957), p. 153.
Greenbaum, In the Name of Efficiency, pp. 85-96; Philip Kraft, Programmers and
Managers: The Routinization of Computer Programming in the United States (New
York: Springer Verlag, 1977), and Robie, "Selection and Training of Systems Personnel,"
Hoos, Automation in the Office, pp. 41-42; Robie, "Selection and Training of
Systems Personnel," 153-55, and Lt. Col. C. R. Gregg, "Personnel Requirements of
Government Agencies in Machine Computation," in Proceedings of the First Conference
on Training Personnel for the Computing Machine Field, Arvid Jacobsen, ed. (Detroit:
Wayne University Press, 1955), pp. 13-14.
U.S. Department of Labor, Adjustments to the Introduction of Office Automation,
"A Profile of the Programmer," Communications of the ACM 6 (Oct. 1963), p. 592.
Greenbaum, In the Name of Efficiency, pp. 86.
"A Profile of the Programmer," p. 592.
U.S. Department of Labor, Bureau of Labor Statistics, Automation and
Employment Opportunities for Officeworkers, Bulletin No. 1241 (Washington, D.C.:
U.S. Government Printing Office, 1958), p. 11, and Beth Parkhurst and Joan L. Richards,
"History of Women in Computer Programming," unpublished paper (December 1985),
Jay Mettler, "Profile: Computer Personnel Characteristics - 1964, Part 2," Business
Automation (April 1964), pp. 30-33, and "A Profile of the Programmer," p. 592-94.
Diebold, "False Starts in Office Automation," p. 84.
Valerie Rockmael, "The Woman Programmer: A Subjective Appraisal," Datamation
(Jan. 1963), p. 41, and Edwin C. Wallace, Jr., "The Selection and Training of Men and
Women Programmers in a Bank," Computers and Automation (Apr. 1965), pp. 24-25.
"Automation Session: Discussion," 1957 LOMA Conference Proceedings, pp.
Esther R. Becker and Eugene F. Murphey, The Office in Transition: Meeting the
Problems of Automation (New York: Harper & Row, 1957), p. 109.
Gerald H. F. Gardner, "The Status of Women in the Field of Computing,"
Computers and Automation (Jan. 1970), pp. 57-58; U.S. Department of Labor,
Automation and Employment Opportunities for Officeworkers, p. 11, and Rockmael,
"The Woman Programmer," p. 41.
Wallace, "The Selection and Training of Men and Women Programmers in a Bank,"
"The Bright Young Men of Information," Dun's Review and Modern Industry 82
(Sept. 1963), pp. 95-96, and 136-46; James P. Moore, "One Man's Opinion: Management
Viewpoints on Men and Machines," The Automatic Office (August 1958), pp. 3-4;
Hattery, "Management Impact of Electronic Systems," in Electronics and Management,
pp. 10-11; Robert M. Gordon, "Management's Failure to Use Information Technology,"
in American Management Association, Computer-Based Management for Information
and Control, Management Bulletin No. 30 (New York: AMA, 1963), pp. 3-4, and Robie,
"Selection and Training of Systems Personnel, Programmers and Operating Staff," p.
Samuel Cohn, The Process of Occupational Sex-Typing: The Feminization of
Clerical Labor in Great Britain (Philadelphia: Temple University Press, 1985), pp.
Parkhurst and Richards, "History of Women in Computer Programming."
U.S. Department of Labor, Impact of Automation on the Insurance Industry, pp.
35-40; U.S. Department of Labor, Technological Trends in Major American Industries,
pp. 249-54; Leonard Rico, "The Staffing Process and the Computer," Management of
Personnel Quarterly (Autumn/Winter 1962), p. 33; Edith Conner, "End of Personnel
Shortage Sighted Even if Economy Enters Another Boom," Office Management,
Yearbook Issue (1958), pp. 18-19ff; Wisconsin State Employment Service, Automation
Manpower Services Program, A Large Life Insurance Company Automates
(Washington, D.C.: U.S. Department of Labor, Bureau of Employment Security, April,
1964), pp. 1-3, and Wisconsin State Employment Service, Automation Manpower
Services Program, Changing Workforce Characteristics of an Automated Insurance
Company (Washington, D.C.: U.S. Department of Labor, Bureau of Employment
Security, June 1964), pp. 6-8.
Wisconsin State Employment Service, A Large Life Insurance Company
Automates, p. 7, and Wisconsin State Employment Service, Changing Workforce
Characteristics of an Automated Insurance Company, p. 6.
"Revolution in Office Work -- Meaning for Jobs, Business," U.S. News and World
Report (May 13, 1963), p. 88, and Rico, "The Staffing Process and the Computer," p. 33.
Wisconsin State Employment Service, Changing Workforce Characteristics of an
Automated Insurance Company, p. 7.
"Effects of Mechanization and Automation in Offices," pp. 266-68; Thomas R.
Brooks, "The Labor Front: Automation and White Collar Unionism," Dun's Review and
Modern Industry (Jan. 1966), p. 59, and Rico, "The Staffing Process and the Computer,"
L. C. Guest, "A Temperate View of Data Processing and Management Information
Systems," in American Management Association, Administrative Services Division,
Advances in EDP and Information Systems (New York: AMA, 1961), p. 8; H. Ladd
Plumley, "Puzzles in Progress," Proceedings of the 1964 Annual Conference of the Life
Office Management Association (New York: LOMA, 1964), p. 29; "Office Productivity: A
New Path to Profits," Dun's Review and Modern Industry 76 (September 1960), p. 72;
Lowell Hattery, "Management Impact of Electronic Systems," in Electronics and
Management (1956), p. 3.; Stewart Toy, "Bugs in Automation," Computers and
Automation 10 (May 1961), pp. 10-12, and "Effects of Business Automation in the '60's,"
Ricco Martino, "Creating a Total System for Management," in American
Management Association, Administrative Services Division, Case Studies in
Computer-Based Management (New York: AMA, 1963), p. 3, and John T. Garrity, "Top
Management and Computer Profits," Harvard Business Review 41 (July/Aug. 1963),
Howard S. Levin, Office Work and Automation (New York: John Wiley, 1956), pp.
4-9; Jay W. Forrester, "Computer Applications and Management Problems," in
American Management Association, General Management Series, No. 178 (1955), pp.
22-31; Rico, The Advance Against Paperwork, p. 19; Eugene Boulanger, "There's More
to Computers Than Data Processing," Proceedings of the 1960 Annual Conference of the
Life Office Management Association (New York: LOMA, 1960), pp. 71-78, and Herbert
M. Cherry, "What Should Systems Planning Mean Today?" 1964 LOMA Conference
Proceedings, pp. 194-95.
Perrin Stryker, What Management Doesn't Know Can't Hurt," Fortune 56 (Nov.
1957), p. 284.
Cherry, "What Should Systems Planning Mean Today?" pp. 194-95.
"Effects of Business Automation in the '60's," p. 33.
Robert W. Harvey, "Preparing for Electronic Operation," Proceedings of the 1958
Annual Conference of the Life Office Management Association (New York: LOMA,
1958), p. 47.
Ibid., p. 49.
Garrity, "Top Management and Computer Profits," pp. 6-12.
Diebold, "False Starts in Office Automation," pp. 82-88, and Rico, The Advance
Against Paperwork, pp. 105-110.
"Effects of Mechanization and Automation in Offices, Part II," pp. 258-60; Kraft,
"The Industrialization of Computer Programming," in Case Studies on the Labor
Process, pp. 8-12, and Greenbaum, In the Name of Efficiency, Chapter Four.
"The Bright Young Men of Information," p. 146.
"The Insurance Man's Best Friend," Business Week (November 7, 1964), pp. 100-04;
Robert E. Slater, "Slater Sees Rise in Computer Use in Underwriting," National
Underwriter 70 (Nov. 5, 1966), p. 4; "Underwriting Clerical Duties Now Delegated to
Computers,' National Underwriter 71 (April 21, 1967), p. 10, and Rico, The Advance
Against Paperwork, pp. 82-83.
Harold J. Leavitt and Thomas L. Whisler, "Management in the 1980s," 36
(Nov./Dec. 1958), p. 43. See also Schultz and Whisler, "Information Technology and
Management Organization," in Management Organization and the Computer, p. 24.