Some Efiects of the Changing Work Environment in the Office
Floyd C. Mann and Lawrence K. Williams'
Electronic data processing equipment is rapidly altering the en- vironment in which thousands of white collar employees earn their daily living. A series of scientific breakthroughs in the late 1940's greatly modified the conceptualization, quantification, and processing of information. This revolution in the handling of information initiated a number of trends which are culminating today in major technologi- cal changes in the office. In our industrial mental health program we are interested in the impact of these changes on both the organization and its personnel. In a society as dynamic as the one in which we now live, a program such as ours must be concerned with change in the structure of tasks and the work environment and the effects that these changes have upon the willing or unwilling participants.
This paper will present some of the findings from a study of a change-over to electronic data processing (EDP) in the accounting functions of a large electric power company. The study began long before our program attempted to develop a systematic and theoretical approach to the effects of the industrial environment on mental health (French, Kahn, Mann, & Wolfe, 1959). As one of the first studies in this new area of office automation, the research has had to be con- cerned with both chronicling qualitatively what changes occurred (Mann & Williams, 1959; Mann & Williams, 1960) and attempting to measure quantitatively, where possible, the changes and their effects. The aims of this specific paper are ( a ) to describe the changes in the objective job environment in an office, and ( b ) to indicate some of the effects of these changes on employee perceptions, attitudes, and adjustments to the situation.
A Brief Account of the Change-over The installation of EDP equipment affected the work of two
major divisions in the light and power company in which this research
1 Now with the New York State School of Industrial and Labor Relations, Cornell University.
WORK ENVIRONMENTS 91
was done. These two divisions, handling the accounts of more than a million customers, were Accounting-the division responsible for all customer billing, bookkeeping, and records, and Sales-the division responsible for all direct contacts with customers relating to service and the payment of their bills. The primary impact of this change- over occurred in the Accounting Division where there were about 700 employees and supervisory personnel. Our study therefore concen- trated on the change as experienced in this division.
The change-over, from beginning to end, spread out over a period of five years from 1954 to 1959. Technically, it involved going from equipment in the IBM 400 series to two small scale computers (IBM 650s) and ultimately to the installation and operation of an IBM 705 complex. The consolidation of record keeping and processing was ac- companied by a major reorganization of functions within and between the two divisions. In terms of pace, the whole process of change ap- peared to start off slowly, gradually accelerate to a period of high ac- tivity, involvement, and pressure, and then decelerate as the organiza- tion arrived at a new state of equilibrium and system maintenance. This was one of the first major changes to EDP in the industry, and one of the most ambitious in terms of the degree to which the compa- nys accounting and customer contact activities were rethought and redesigned. Consequently, there were a number of transitional prob- lems. These included the problems typical of any major change- extensive overtime in the evenings and week ends, training and re- training, errors and repeated errors because of unfamiliarity with the new system, and mounting pressures to meet deadlines that had been established when there was little understanding of the magnitude of the envisaged change. There were also the tensions and concerns that accompany a top management decision to invite a consulting firm to evaluate the effects of the new system on customer relations and rela- tions among units within the company. These problems of transition, along with those of final assignment of personnel and full conversion to the new equipment, were met and handled successfully, and the organization settled again into the familiar patterns of a steady-state equilibrium.2
Since we as researchers had even less information about what to expect than the managers who were pioneering this change, we used a combination of methods to try to capture the story of this change-over
2 This paper focuses on the effects of the introduction of EDP in this situa- tion. An account of the problems faced in managing such a change-along with some of the post facto insights about what should be kept in mind during such a period-is given in Managing Major Change in Organizations (F. C. Mann and F. W. Neff), Ann Arbor, Michigan: Foundation for Research on Human Be- havior, 1960. See especially Case IV: Completing and Stabilizing Changes and A New Role: The Change Catalyst.
92 FLOYD C. MANN AND LAUrRENCE K. WILLIAMS
and the meaning it had for those who were directly or indirectly af- fected by this extended period of disruption and transition. Just be- fore the change got under way, we asked the nonsupervisory em- ployees, supervisors, and department heads to fill out paper-and-pencil questionnaires about their work situation. By informal interviews and discussions with people at all levels, we then followed the course of the change through the conversion. After the new system was in place and the new equilibrium was emerging, we again asked all levels to fill out questionnaire forms. These after measures contained questions which we had asked in the before period as well as those which we had not had the foresight to ask. To test some hypotheses we there- fore have both before and after measures, for others we have only after measures or after measures in which we asked the respondent to try and recall how he felt at three different periods-before, during, and now (after) the change. These attitudinal data were supple- mented with ratings from managerial and supervisory personnel con- cerning the characteristics of the most predominant jobs which existed before and after the change. Since some of the departments within each division were not affected by the change-over, we have used these as a type of control group. A full account of the change-over and the research methods employed will be available in a subsequent monograph on this study.
Changes in Job Content While there is not much question but that the introduction of a
computer system means a more rationalized organization and job structure (Hardin, 1960; Hoos, 196l), the approach we are using in our industrial mental health program of obtaining objective measures of the working environment independent of the individuals affected by that environment dictated that we try to determine quantitatively changes in the structure of jobs. To do this, we undertook a type of job analysis.
Our approach to this analysis of the task structure was different from the usual job analysis in that we gave considerable attention to the systemic nature of the work flows involved. We attempted to look at the social-psychological framework and environment of the job, as well as the specific task components of jobs. The dimensions we used were designed to get at the extent of the rationalization of the system, the risks involved in each job, the interdependence of the structure of activities, and such system related facts.
Department heads and supervisors were asked to rate those jobs with which they were most familiar from a roster of fifty jobs. These fifty jobs had been chosen from a population of 150 jobs. They were representative of jobs both integral to the new 705 system and those
WORK ENVIRONhlENTS 93
not directly integrated or linked with the new 705 system; and a few were selected to represent the jobs that existed before the new system was installed. The fifteen dimensions on which each job was rated on a five or six point scale were as follows:
Serious error-estimate of the minimum cost in dollars and cents to correct a serious error (alternatives from less than one dollar to more than one hundred dollars) Typical error-estimate of the maximum cost in dollars and cents to correct a typical error (alternatives from less than one dollar to more than one hundred dollars) Slow-up production-estimate of maximum cost in time that a typical error will slow up some part of the operation (alternatives from no time at all to more than one work day) Effect of others errors on job time (within group)-cost in time to make correction for a typical error of another member of own work group (alternatives from no time at all to more than one work day to make corrections) Effect of others errors on job time (outside the group)- cost in time to make correction for a typical error of other persons outside of own work group (alternatives from no time at all to more than one work day to make corrections) Absence effect-how much inconvenience or extra work a persons absence causes others (alternatives from no incon- venience at all to a very great amount of inconvenience if absent) Choice of means-extent person has acceptable alternative ways of doing job (alternatives from always has more than one acceptable way to rarely or never has more than one
Work checked by people-extent work of people in this job is checked by other people (alternatives from work always checked to very seldom checked by other people) Necessity for understanding others jobs-how necessary it is for person in this job to have general understanding of several other jobs (alternatives from not at all to highly necessary to have a general understanding) Work pace set by people-to what degree pace is set by fellow workers (alternatives from not determined at all to completely determined by fellow workers) Work pace set by machines-to what degree pace is set by machines (alternatives from not determined at all to com- pletely determined by machines) Contact with others (within the group)-frequency of com- munication with others in own work group required by job
FLOYD C . MANN AND LAWRENCE K. WILLIAMS
(alternatives from never to consistently required that a per- son communicate with others in group) Contact with others (outside of group)-frequency of com- munication with others outside of own work group required by job (alternatives from never to consistently required that a person communicate with others outside of own group) Chance of detection-proportion of time a typical error can be detected (alternatives from 20 percent or less of the time to almost all or all of the time the error could be detected) Chance of attribution-proportion of time a typical error can be attributed to a spec&c person (alternatives from 20 per- cent or less to almost all or all of the time an error could be attributed).
The order of the alternatives for each of the above dimensions shows our general hypothesis with regard to change in the job structure: the more rationalized the job or a group of jobs, the higher the antici- pated score on each dimension.
Using the above dimensions, profiles of different classes of jobs were computed and comparisons were made to test our basic as- sumptions concerning the effect of EDP on the work environments of white collar workers. Comparisons were made between categories of jobs which existed prior to the installation of the customer billing system and the current job groups, and between different classes of contemporary jobs as a function of their integration with the new customer billing complex. Our systemic approach to the job analysis had led us to feel that it would be useful to distinguish between groups of jobs that were highly interlocked or integrated with the new system and those that were only moderately or nonintegrated. Two individuals responsible for the major part of the change-over were asked to rate all of the jobs in terms of degree of integration. The reliability of the judgments of these two individuals at one time and over a nine month interval was very high (rank order correlations exceeding .95).
Analyses of the profiles of different categories of jobs, with job grade controlled, produced the following kinds of findings:
1. Highly integrated, nonclerical jobs in the present system-coders, programmers, librarian, console operator, and accounting ma- chine operator-had a job profile indicating greater rationaliza- tion, risk, and interdependence than nonintegrated jobs in the present system-senior bookkeeper, accountant, multilith opera- tor, special clerks.
9. The composite profile for highly integrated clerical jobs-unit clerk and customer accounts and bookkeeper jobs-was also
WORK ENVIRONMENTS 95
more rationalized, more interdependent, and fraught with risk than the composite profile for the nonintegrated clerical jobs of the present system-accounting clerk, disbursements auditor, stenographer, etc. The profile for the new highly integrated clerical jobs when com- pared with the profile for a series of clerical jobs that were dis- placed by the change-over-essentially a before and after com- parison-indicates that there is again a significant difference across nearly all dimensions in that the new jobs appear closer to the more rationalized and more demanding ends of the con- tinua than do the jobs which were eliminated.
These findings, together with other specific analyses of particular pairs of jobs before and after the change-over, demonstrate that auto- mation in the office has affected the structure of tasks that workers are asked to perfom. In general, the jobs require a greater amount of risk-the typical and serious errors cost more in time and dollars. The jobs require a greater understanding of the total system. There is a greater degree of interaction among the members of any work group and a greater degree of dependence upon fellow workers. There also appears to be a greater opportunity to detect and to at- tribute specific errors that are made to the individuals who made them.
Changes in Employees Perceptions of Jobs To investigate how these changes in the structure of tasks were
seen by clerical employees, we used two populations of employees for whom we were able to match before and after questionnaires. One group of 88 were now working in the highly integrated jobs; another group of 80 were working at jobs quite independent of the 705 system or on nonintegrated jobs.
Significantly more of the highly integrated job occupants, in con- trast to the nonintegrated job occupants, reported greater satisfaction with the amount of their job responsibility, more variety and change in their jobs, and greater opportunities to develop and learn new things. A significantly higher proportion of those on the highly inte- grated jobs also saw their jobs as more important now than before, while there was no significant change in this respect for those who were on nonintegrated jobs. In spite of these changes in how the em- ployees working at the highly integrated jobs saw their work, there was no corresponding change in their over-all liking for their jobs. When they were asked to consider their jobs as a whole and indicate how well they liked their jobs, employees on highly integrated jobs were no more satisfied with their jobs after the change than they were with the jobs they worked at before the change.
96 FLOW C. MANX AND LAWRENCE K. WILLIAMS
Greater job responsibility, variety, and opportunities for gro\\rtli are characteristics associated with job enlargement, and job enlarge- ment is usually related to increased job interest and satisfaction (hlann & Hoffman, 1960). To understand this lack of increase in over-all liking for the job, it was necessary to turn to employees perceptions of how the installation of the high speed computer and the new billing system had changed the standards of performance. Significantly more of the employees working on the integrated jobs than those on the noninte- grated jobs stated that performance standards are higher now than they were before, that standards are more rigidly enforced than be- fore, that there are more deadlines and that these deadlines are more important than before, and finally that the company expects more of their employees now. These findings came from questions which were asked on only the after questionnaire, but several different types of analyses confirm this general tightening of standards of performance.
With regard to group functioning, other types of analyses of em- ployee perceptions confirmed that in those departments in which the greatest changes had occurred, work groups had to rely more now than before on other groups, and another workers absence caused employees to have to work harder.
In general, these data lead toward the conclusion that while employees in the new work environment had more interesting and challenging jobs than before, the greater exposure to risk and the tighter performance standards negated the attractive aspects of job enlargement. The new work structure appeared to strip away further the anonymity of the white collar workers performance and to place even higher demands on him regarding the quality and quantity of his work.
Changes in Employees Feelings To get at how the employees in the Accounting Division felt
after this extended period of change and how they felt when they found their work environment to be more demanding than what it had been earlier, three different types of questions were asked. One group dealt with how employees felt about top management, their future in the company, and the company and their job in general; another with how frequently they found themselves worrying about different aspects of their work life; a third touched on their psycho- logical and physical anxieties.
After the change was over, the percent of all nonsupervisoiy em- ployees saying they felt top management was less interested in the employees now than a few years ago was twice as large as it was be- fore the change-over started (36 as compared to 18 percent), More- over, 42 percent of the individuals in the change departments felt this
WORK ENVIRONMENTS 97
way, as compared to 29 percent in the departments where there had been little or no change. The number of employees who felt the company was more interested in cutting costs than in its people rose from two out of five before to almost three out of five after the change- over. Again, employees in the change departments felt that the cutting of costs was being stressed over the welfare of people more often than did those in the nonchange departments.
Toward the end of the questionnaire, the employees were asked whether they could say their future in the company looked better or worse than a few years ago. The proportion who felt their future looked either much worse or somewhat worse increased significantly from before to after the change-over period. After the change was completed, three out of ten felt their future looked worse, a third about the same, and another third somewhat or much better. To test whether these feelings meant that employees would be leaving the company, a question was asked that would determine the employees propensity to leave the organization. Significantly more accounting employees after the change than before (but only one in eight) said they would leave the company for any other j o b or for another job with the same pay and other benefits. Eighty-four percent stated that they would leave the company only if a particularly good op- portunity came along. This suggests that while these employees were less satisfied than earlier with their company and their jobs they were not so dissatisfied that they were about to translate their dis- satisfaction into the behavioral act of resigning to take another job.
Perhaps the best description of what has happened to the over-all level of satisfaction of the nonsupervisory employees is the trend of answers they have given to a summary question about their job and the company during a period of eleven years which included the change-over to EDP. In the three biennial surveys taken in 1948, 1950, and 1952 (before there had been any consideration of the pos- sibility of a change to a new accounting system), the percent of em- ployees who said they were not satisfied in general with the company and their jobs dropped from nine to seven to five percent. In 1954- just after the announcement that a feasibility study had been made and the process of changing over would soon start-the percent not satisfied went back to nine percent. But by 1959-after the chnnge- over had been completed-seventeen percent of the almost 600 non- supervisory employees in the Accounting Division were not satisfied and felt there were a great many things that could be changed or were not satisfied but felt they could see no way things could be changed.
To understand more about other factors that might be con- tributing to this change in the general level of satisfaction, w7e asked
98 FLOYD C. M A N N AND LAVVRENCE K. WILLLAMS
employees to indicate how often they were ever bothered or worried about certain personal and job problems. These ranged from how good a job they were doing through layoffs to money and health problems. Analyses indicate that significantly higher proportions of employees in the change departments than employees in the depart- ments where there was little change were worrying about the pos- sibility of temporary layoffs or the possibility of losing my job. With respect to other worries, there was little difference between these two groups. Before and after analyses for matched samples also confirm this finding. More employees working on highly integrated jobs were worrying more after the change than before about tempo- rary and permanent layoffs.
While it is probably not possible to determine with any precision whether the extent and depth of these worries and dissatisfactions were dysfunctional or functional for most of the pcople experiencing these changes (Tyhurst, 1958), there is little evidence from the after measurements about how the psychological or psychosomatic symptom patterns of some of our groups compare with roughly similar occupa- tional groups from a nationwide study. Viewing the responses of a national sample of men and women to a symptom checklist as dif- ferent ways of expressing emotional stress and disturbance, Gurin, Veroff, and Feld (1960) used a factor analysis to identify four basic symptom patterns. Two of these served to represent psychological expressions of dysfunction, where symptoms are experienced as a psychological incapacity of which the respondent has some degree of awareness; the other two seemed more related to a somatization of stress. Psychological Anxiety was a kind of free-floating anxiety; Im- mobilization a general psychological inertia. Physical Anxiety and Physi- cal Health appeared to involve more of the extreme physical symptoms that usually accompany intense anxiety. Using several cri- teria, Gurin, et al., found that for each factor two symptom items emerged which would adequately describe each dimension. In the after questionnaire of our EDP study, we incorporated the three pairs of questions for measuring psychological anxiety, physical anxiety, and immobilization. It is these measures which we draw upon to learn more about what may have been some of the effects of the changeover on oiir white collar population.
Since the findings presented in Americans Viczu Their Mental Health demonstrate that such social groupings as sex, age, and educa- tion are related to the differential prevalence of symptom factors, we attempted to increase the appropriateness of our comparisons by con- trolling for both sex and occupation. This was a costly refinement in that we found that the number of men in the nationwide study work- ing at clerical jobs was quite small ( N = 3 7 ) , and the niimber of gain- fully employed clerical women was too small to provide any norms at
WORK ENVIRONMENTS 99
all. Our analyses of these health relevant variables thus had to be re- stricted to men.
Given these restrictions, the following findings appear to be re- lated to the stresses of the change-over and to the new job structure.
1. The proportion of nonsupervisory men operating the EDP sys- tem who had high scores on each symptom factor was signifi- cantly greater than the proportion of clerical men in the national sample who had high scores on psychological anxiety, immobili- zation, and physical anxiety. (Higher scores were characteristic of higher anxiety. )
2. The proportion of first-line men supervisors in the accounting de- partments who had high scores on each symptom factor was also significantly greater than the proportion of managers in the national sample who had high scores on these three measures. Particularly noteworthy in view of the findings in the preceding paper by Kasl and French on status was our finding that more than two-thirds of the accounting supervisors scored high on the psychological anxiety scale, while only a fourth of the managers in the national sample had equally high scores. Supervisors who were evaluated by their superiors as more effective during the change-over process than their supervisory peers reported a lower incidence on a combined symptom score than those who were evaluated as less effective by superiors.
3. The proportion of reserve employees-a group who after the change were unsuited for the jobs that remained-who had high scores on each symptom factor was also significantly greater than the proportion of clerical males in th,e national sample who had high scores on psychological anxiety, immobilization, and physi- cal anxiety. This group of individuals displaced by the change was comprised of an approximately equal number of men and women and all of them would be classified as older workers. Being put on the reserve list did not mean they were being dropped from the company rolls as employees, but rather that there was no job for them to which they could be assigned which was consistent with their former pay grade.
Summary It is our expectation that the Institutes Intercenter Research Pro-
gram on Mental Health in Industry will frequently be concerned with problems of change such as described in this paper. As members of a scientific society of which one of the principal features is an ac- celerating rate of change, we have a unique environment in which to investigate the impact of technological change on the organization of work and the individuals living in this relatively unstable social order.
100 FLOYD C. MANN AND LAWRENCE K. WILLIAMS
This study of some of the effects of the changing work environ- ment in the office has provided us with an opportunity to begin the development of measures for quantitatively demonstrating ( a ) changes in the objective work environment, and ( b ) the relation of these changes to employee perceptions, attitudes, and anxieties.
The findings of this specific study indicate that the introduction of electronic data processing equipment means a general tightening of the task structure of the office. Jobs require a greater amount of risk, a greater understanding of the total system, and a greater degree of interdependence. While employees working on jobs that are a part of this more highly integrated system of tasks are more satisfied with the amount of their job responsibility, the variety and change in their jobs, and the opportunity they have to develop and learn new things, they do not express increased satisfaction with their jobs as a whole. They are keenly aware that performance standards are now tighter than before, that these standards are more rigidly enforced, that there are more deadlines and these are more important than before.
These office workers felt that top management not only expected more of them, but was less interested in them as individual employees than before the change-over to more integrated, electronically con- trolled, accounting procedures. Significantly higher proportions of these white collar workers believed their future looked somewhat worse, and they were worried about the possibilities of temporaray layoffs or losing their jobs. The individuals-nonsupervisory and supervisory men-who had been at the very vortex of the extended and intense process of change-over to the new accounting systems showed evidence of more psychological and physical anxiety than for similar white collar groups in the nation.
While automation in the office has meant further rationalization of the work environment and the clerical worker has become less satisfied in general with his job and company, this dissatisfaction has not been so intense as to cause these employees to be actively search- ing for other jobs.
REFERENCES FRENCH, J. R. P., JR., KAHN, R. L., MA", F. C., & WOLFE, D. The Effects of the
Industrial Enoironwient on Mental Health: Working paper I , 1959. GURIN, G., VEROFF, J., & FELD, SHEILA. Americans View their Mental Health: A
Nationwide lnteruiew Suroey. New York: Basic Books, 1960. HARDIN, E. The reactions of employees to office automation. In Impact of Auto-
mation. Bureau of Labor Statistics Bulletin No. 1287, 1960. Hoos, IDA R. Automation in the Ofice. Washington, D. C . : Public Affairs Press,
1961. MA", F. C., & HOFFMAN, L. R. Automation and the Worker: A Study of Social
Change in Power Plants. New York: Henry Holt, 1960.
WORK ENVIRONMENTS 101
MA, F. C., I% WILLIAMS, L. K. Organizational impact of white collar automa- tion. Proceedings of the eleuenth annual meeting of the Industrial Rekz- tiom Research Associotwn. Chicago, December 1958. Publication No. 22,
. Observations on the dynamics of a change to electronic data processing equipment. Admin. Sci. Quart., 1960, 5, 2, 217-258.
TYHURST, J. S. The role of transition states-including disasters-in mental ill- ness. In Symposium on Preventiue and Social Psychiatry. Washington, D. C . : US. Government Printing Office, 1958.