Chapter 2 Using Computers To Monitor Office Work

Chapter 2

Using Computers ToMonitor Office Work

CONTENTS

PageIntroduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27What Kind of Work Gets Monitored? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28How Widespread Is Electronic Work Monitoring?. . . . . . . . . . . . . . . . . . . . . 30

Work Monitoring in the Private Sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30Work Monitoring in the Federal Government . . . . . . . . . . . . . . . . . . . . . . . 33

How Technology Assists Selected Business Applications . . . . . . . . . . . . . . . 34Workplace Issues Related to Work Monitoring . . . . . . . . . . . . . . . . . . . . . . . 36

What Is Fair . . . . . . . . . . . . . . .Standards of Performance . . . .Quality of Work. . . . . . . . . . . . .Job Design and Work Process .Supervision and Evaluation . . .Personal Computer MonitoringMonitoring and Stress . . . . . . .Computer Pacing and Stress . .Feedback and Motivation . . . . .

The Future of Work Monitoring .

Figure No.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

FiguresPage 3

2.

3.

4.

5.6.

7.

8.

- Example of a System Status Report for an Automatic Call DistributionSystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35Example of a Supervisor’s Display in an Automatic Call DistributionSystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36Example of Agent Statistics From an Automatic Call DistributionSystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36Two Models for Electronic Work Monitoring . . . . . . . . . . . . . . . . . . . . . . . 37Example of Individual Work Monitoring Report {Performancesummary). . . . . . . . . . . . . . . . . . . . ● . . . . . . . . . . . . . . . . . . . 41Example of a Work Monitoring Report for a Group (Volume of WorkAccomplished by Group in One Week). ... , . . . . . . . . . . . . . . . . . . . . . . . . 42Example of Individual Work Monitoring Report (Hours Summary) . . . . 43

TablesTable No. Page1. Some Office Jobs Currently Subject to Electronic Work Monitoring . . . 292.3.4.5.6.

8.

Conditions Most Favorable to Electronic Work Measurement . . . . . . . . . 29Workplace Issues Related to Work Moontering. . . . . . . . . . . . . . . . . . . . . 38Key Issues and Problem Areas in Monitoring Worker Performance . . . . 39Comparison of Monitored and Nonmonitored Workers . . . . . . . . . . . . . . . 49Rates of Frequent Health Problems Related to ComputerizedMonitoring of Work Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52Rates of Frequent Health Problems Related to Production Quotas orProductivity Standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52Review of Oregon Worker Compensation Claims Involving MentalStress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ........4 53

Chapter 2

Using Computers To Monitor Office Work—

INTRODUCTION

Electronic work monitoring is the computer-ized collection, storage, analysis, and report-ing of information about employees’ produc-tive activities. Within this broad definition, theprimary focus of this report will be on obtain-ing data about employees directly throughtheir use of computer and telecommunicationequipment. 1 This type of information gather-ing has been called ‘‘electronic monitoring,‘‘electronic work measurement, ” or “telephonemonitoring. “2 This chapter will also include adiscussion of “service observation, ” the prac-tice of listening in on conversations betweenemployees and customers to evaluate the em-ployees’ courtesy or competence. Service ob-servation is often used in conjunction withelectronic work measurement for telephoneoperators and customer service workers.

Some people have warned that use of elec-tronic monitoring leads to creation of electronicsweatshops. ‘‘Electronic sweatshop’ ‘–the termconjures up images that combine the worst fea-tures of both the factory and the office: bor-ing, repetitive, fast-paced work that requiresconstant alertness and attention to detail, alldone under the pressure of constant supervi-sion and demands for faster work. Worst ofall, the supervisor isn’t even human. Employ-ees must labor at top speed under the view ofunwinking computer taskmasters that recordevery item of work completed, along with everymistake, rest break, and deviation from stand-ard practice. A person’s job depends on thecomputer’s comparison of performance to thestandard. Interaction with fellow workers is

‘This definition is very similar to one found in: Andrew Clem-ent, “Electronic Management: The New Technology of WorkPlace Surveillance,” Canadian Information Processing SocietJ’Session W Proceedings, Calgary, Alberta, May 9-11, 1984, pp.2.59-266.

‘In this chapter, ‘‘work measurement” is used to include thewhole process of developing procedures and standards for jot)performance, collecting data on actual performance and com-paring actual performance to standards. “Work monitoring”refers only to collection of information about actual performance

impossible because of the pace of work; jobsatisfaction is low, and stress and stress-relatedhealth problems are the inevitable result ofhaving to work under such conditions.

Electronic work monitoring is already a dailyreality for millions of U.S. workers. They seemto be having varied experiences with it. Someview electronic monitoring as a useful tool thathelps them get better control of their work,ensures that their supervisors give objectiveevaluations, and helps their company be moreproductive. Others believe they indeed do workin electronic sweatshops as described above,and that monitoring is an unfair surveillanceused to control them. Still others have mixedfeelings: they may not mind monitoring perse, but they feel it isn’t being used in the fairestor most effective way.

Electronic monitoring is usually used in con-duction with a work measurement system. Workmeasurement systems usually do four things.First, they set standards for the time it shouldtake to produce certain units of work. Second,they monitor the actual time it takes to pro-duce each unit of work. Third, they analyze thevariance of the actual time from the standard.And finally, they provide data for use in plan-ning, cost estimates, and productivity improve-ment.3 As more employees use computers intheir jobs, computer software is increasinglyused to monitor actual performance, compareperformance to standards, and provide plan-ning data.

As more capabilities of the computer andtelephone are being explored for office work,it is probably natural that some of the samecapabilities are found useful in supervisingwork as well. Both work and supervision arebecoming automated. Work done- on a com-

SEW U.S. Department of Defense, Office of The 1 n.specterGeneral, W-ork Measurement b:ystems and li’ngineer~d Labm-.Standards (Lf’ashington, IX: October 1986).

27

28 ● The Electronic Supervisor: New Technology, New Tensions— —

puter is very abstract. Many information work-ers no longer handle concrete items like docu-ments, file folders, paper clips, orders, invoices,rubber stamps, or checks. Those physical ob-jects have disappeared into the computer andhave become abstract analogs of their formerselves. The various stages of work that trans-form raw materials into final products takeplace inside the computer, too. A supervisorwho could only observe the physical activityof people at their computer terminals, with-out knowing what was going on inside the com-puter, would know little about the work beingdone.

The tool that gives a picture of what is hap-pening inside the computer is the computer it-self. Computers’ capacities for recording andstoring information make it possible to keepdetailed records about all aspects of the pro-duction process. And their ability to sort thatinformation in different ways means that theinformation can be put to many uses.

Computer records can give a picture of thetotal performance of a work group, or a depart-ment. Statistics on historical patterns of pro-

duction can be used to estimate future work-loads, to plan for new personnel, or to justifynew equipment. Performance statistics mayalso be compared with budgeted costs to de-termine the cost-effectiveness of an operation.Such complete and up-to-date information isnecessary in a cybernetic model of manage-ment that requires immediate feedback tomanagers about current activities to be usedas the basis of future decisions.

Computer work monitoring can also give in-formation on individual performance. The fo-cus of this chapter is computer monitoring ofindividual performance. Computer-generatedstatistics can be used as a tool to increase ormaintain levels of employee performance. Theymay be used in individual personnel decisions—pay, promotion, retraining, and discharge.They can be a feedback tool to help employeesgain more control of their own work; conversely,they can be used to limit employee decisionsabout the work process. Like most technologi-cal tools, work monitoring per se is neither badnor good. Its effects depend on how it is used.

WHAT KIND OF WORK GETS MONITORED?

Most of the electronic monitoring found byOTA and other researchers affects office work-ers with short-cycle “production” jobs, thatis, jobs where a limited number of standard-ized tasks are performed repeatedly to producesome information-based end-product. Mostsuch jobs are considered clerical, for exampledata entry or insurance claims processing.However, monitoring can also be applied toprofessional jobs with a quantifiable output,for example computer programmers or insur-ance underwriters.

Data-entry jobs are perhaps the epitome ofroutine, standardized information-processingwork. The operator reads information from apaper source and enters it on a keyboard orkeypad to be recorded on computer tape ordisk. Most key-to-disk and key-to-tape data-entry systems are equipped to count operator

keystrokes, and in these high-production jobs,counting keystrokes is an obvious way tomeasure performance.

However, key entry is not the only job whereproduction can be monitored electronicallyusing computer counts. Table 1 lists some ofthe office jobs that are often subject to workmeasurement from production data gatheredthrough electronic monitoring. The list is byno means exhaustive. The table summarizesa few of the aspects of work that can be elec-tronically monitored for each job.

What do all the jobs in table 1 have in com-mon? Why is it that they lend themselves tocomputer monitoring? Jobs that are subjectto electronic monitoring are generally thosethat are subject to work measurement tech-niques. In work measurement systems, man-

Ch. 2—Using Computers To Monitor Office Work • 29—

Table 1 .—Some Office Jobs Currently Subject to Electronic Work Monitoring

Job

W o r d p r o c e s s o r s . .D a t a - e n t r y c l e r k s . . . . ,Telephone operators . . . . .Customer service workers .

Telemarketing/other sales

Insurance c la ims c lerksMail clerks . . .Bank proof clerks . . . .

What is measuredspeed, errors, time workingspeed, errors, time workingaverage time per calltime per customernumber and type of transactiontime per customer; sales volume

number of cases per unit timeletters or packages per unit timechecks processed per unit of

timeSOURCE Off Ice of Technology Assessment ;967

agement sets standards of production and thenmakes records of actual performance in orderto compare them to the standard. For manykinds of work, manual recordkeeping systemsor physical counts of completed work are usedto gather performance information, but thegrowing availability of computer-based sys-tems in offices has led many employers to col-lect this information electronically.

Table 2 identifies some of the work condi-tions most favorable to the application of workmeasurement. The first set of conditions is thatthe work be routine and require the repetitiveperformance of a small range of tasks. Whenthis is the case it is possible to time those tasksand establish a standard amount of time inwhich a competent worker can be expected to

Table 2.—Conditions Most Favorable to ElectronicWork Measurement

Routinized work:● smalI number of tasks performed by each employee● large volume of work. relatively continuous supply of work

Interchangeable workers:● relatively low training requirementŽ relatively smalI difference in the productivity of experi-

enced and inexperienced workers, or short time neededto bring inexperienced workers to full capability

Ž tolerance for turnover● ample labor supply

Simple data collection:● employees use information technology as part of their work. information about transactions is already being CoIIected

for other purposes● data collection is transparent to the users, and making use

of it is simple for supervisorsSOURCES Adapted from Robert E Nolan, “W;rk Measurement, ’:ln Robert N

Leherer, White Co//ar Producflwty (New York, NY McGraw Hill, 1983),also personal communication from James Rule, August 1986

How obtained

keystrokes counted by computerkeystrokes counted by computereach call timed by call distribution systemeach calI timed by call distribution system,

transactions counted by computereach call or transaction timed. sales tabulated by

computertime spent on each form tabulated by computercollected by letter of package sorting machinescollected by proof machine

perform them (more about work measurementstandard setting below). Measurement is alsoeasier and more meaningful when there is alarge volume of work from which to draw, ora relatively continuous flow of work.

Routine tasks can be performed by inter-changeable workers. Individual employeesmay come and go (turnover) but the work stillgets done. These jobs do not require a workerto have rare personal qualities, extensiveprofessional training, or highly specializedskills. The training required for most routinejobs is minimal, and the amount of time neededfor a newly trained worker to reach full com-petence is usually short, Training for sometypes of data-entry jobs, or for such jobs asproof machine operators in banks, can be ac-complished in as little as a few days. Other rou-tine jobs, however, require more skills andlonger training. For example, training for tele-phone customer service representatives atfirms interviewed by OTA ranged from a fewdays to 6 weeks, depending on the firm andthe complexity of the services offered.

Although work measurement is most easilyapplied to less skilled jobs, it is increasinglybeing directed to higher level, more skilled tech-nical, professional, and managerial positions.Even the most complex work has its routineelements, and given sufficient analysis, thoseelements can be identified, grouped together,and counted. The jobs of commodities broker,computer programmer, and bank loan officer,for example, could lend themselves to moni-toring. They all have a high proportion of rou-tine elements. But these jobs also require

30 ● The Electronic Supervisor: New Technology, New Tensions

higher levels of training and experience. Workmeasurement and electronic monitoring canbe used in jobs like these, but if workers re-sent them, the costs of resistance might be un-acceptably high for the employer. A firm’s per-ception of the interchangeability of certaintypes of workers, the ampleness of replacementlabor, or their own tolerance for turnover areall relative. They can change over time withvariations in corporate goals, job markets,managers’ personalities, or internal corporatepolitics.

As was pointed out in OTA’s earlier reporton office automation, the change in work proc-ess that takes place when certain types ofprofessional or technical workers start mak-ing use of computers, sometimes leads togreater standardization or routinization oftheir work.’ Some researchers hold that theincreased use of computers to assist profes-sional and managerial work will lead inevita-bly to the de-skilling of mental work and thecreation of “intellectual assembly lines.“5

Many employers are greatly concerned withgetting higher performance from highly paidprofessional and managerial workers—the“last great frontier” of productivity improve-ment. Computer monitoring can offer a wayto make them more accountable and to meas-ure their performance against performancegoals. There will be further discussion on elec-tronic monitoring of professional, managerial,and technical workers later.

In table 2 the final group of characteristicstypical of monitored jobs is “simple data col-lection. ” As noted below, performance statis-

‘U.S. Congress, Office of Technology Assessment, ALztorna-tion of America Offices, OTA-CIT-287 ( Washington, DC: U.S.Government Printing Office, December 1985), p. 105.

‘)Judith A. Perrolle, “Intellectual Assembly Lines: The Ra-tionalization of Managerial, Professional, and Technical Work, ’paper presented to the American Sociological Association, Wash-ington, DC, August 1985.

tics can be collected about many routine of-fice jobs that are not computerized, usually byhaving the employee or supervisor keep paperrecords. Care must be taken to design a meansof data collection which does not unduly bur-den workers or supervisors–otherwise themeasurement system may decrease produc-tivity because it takes too much time or re-duces morale. Further, a work-measurementsystem that requires people to take an extrastep to keep performance records maybe sub-ject to error or fraud. With electronic moni-toring, manual recordkeeping can often be re-duced or eliminated, even while much moredetailed measurements are being made.

Jobs involving telephone contact with thepublic are often subject to “service observa-tion, ” that is, having a supervisor or qualitycontrol specialist listen in on employee tele-phone calls to evaluate courtesy, accuracy, orcompliance with company guidelines. Serviceobservation is a common practice in telemar-keting firms, direct sales outlets, market re-search firms, companies with large customerservice departments, and of course telephonecompanies. Service observation is not new; tele-phone companies have been practicing it forover 80 years, as have many other firms. Serv-ice observation is also not automatic. It re-quires a human listener to make judgmentson the content of a call.

However, new information technology hasbeen transforming service observation by de-veloping systems that integrate service obser-vation with other, more automatic, monitor-ing techniques and also by improving thequality of new equipment. In older telephonesystems, for example, a drop in volume or aclick might be heard when an observer cameon the line. Most modern service-observationequipment is perfectly silent and does not in-terfere with the operation of lines.

HOW WIDESPREAD IS ELECTRONIC WORK MONITORING?Work Monitoring in the Private Sector puter use as part of computer security programs,

audit trails, or cost allocation programs. Thus,It should be noted that many computers re- in nearly every organization that has a main-

cord information about individual workers’ com- frame, minicomputer, or integrated word proc-

Ch. 2—Using Computers To Monitor Office Work . 31

essing system, computer-generated records aremade when a terminal logs on or off; when adatabase is accessed; when a file is created, re-vised, or deleted; or when a remote computeris accessed through telecommunication lines.Some privacy implications of computer secu-rity systems will be addressed separately ina companion report. G However, for the pur-poses of the present report, such records arenot considered to be work monitoring.

There are no reliable figures on how exten-sively employers are applying computer-basedsoftware to monitor individual employee per-formance, or to what extent they are using itto make judgments on individual pay, promo-tion, or discipline. No one has such figures, andno available basis for calculating them is athand.

Before trying to estimate how many officeworkers are subject to work monitoring bycomputer, it is important to clearly define workmonitoring. In this report, work monitoringwill refer to the computerized collection oftransactions on performance statistics used inindividual work evaluations. Based on this def-inition it appears, as will be discussed in thissection, that only a minority of office workersare monitored.

According to one work-measurement expertin the health insurance industry, some sort ofstatistics are collected about the computertransactions of nearly everyone who uses acomputer—about 80 percent of the people inthe industry. However, he estimated that onlyfor about 20 percent of the people were thesestatistics actually used to measure individualperformance.7 This 20 percent representedprimarily the low-skill end of the clerical workforce.

Some information on monitoring was col-lected in the 1984 National Survey on Womenand Stress, conducted by 9 to 5 National Asso-ciation of Working Women. This survey re-

‘See Office of Technology Assessment, “Federal Policy onFllectronic Information Security: Emerging Issues and Tech-nological Trends, ” forthcoming, 1987.

‘Fred Friedman, Director of Operations Strategy, Blue CrossAssociation, interview, hlay 23, 1986.

ceived responses from 40,000 readers of fourmajor women’s magazines— Working Woman,Ms., Glamour, and Essence. Of the 4,500 re-sponses randomly selected for analysis, 43 per-cent reported that they used visual-display ter-minals (VDTs), cathode-ray tubes (CRTs), orpersonal computers (PCs). Of these users, 25percent were in managerial jobs, 30 percentwere professional and technical workers, and44 percent were in clerical jobs. The clericalcategory combined secretarial, customer serv-ice, data entry, and similar job titles.

One question on the survey asked:

Is your work measured, monitored, “con-stantly watched” or controlled by machine orcomputer system.

Seventeen percent of all office automationusers answered “yes” to this question. Whenbroken down by occupation, 20 percent of cler-icals answered “yes,” compared with 15 per-cent of managers and professionals and and13 percent of technical workers.

Some critics have noted that the sample inthis survey is self-selected, and that the resultsmay not be representative of all women in theU.S. work force. On the other hand, the posi-tive replies to the question on monitoring maybe lower than the actual incidence of electronicmonitoring in the United States. In the courseof doing this study, OTA staff and contrac-tors often found it difficult to ask about workmonitoring with a simple yes or no question.Some people simply did not understand thequestion without further explanation. Work-ers in some locations did not know that theywere being monitored. In other firms, computeruse information was collected but not used forbut individual evaluation.

One small survey of office automation useat 45 large New York firms, conducted forOTA, suggests that electronic monitoring isstill not widespread at those firms. Only eightfirms (18 percent) reported using informationcollected through electronic monitoring as abasis for individual performance evaluations,and six firms (13 percent) used it for team orwork group appraisal. Fourteen firms (31 per-


cent) used automatically collected data forplanning work force requirements.8

On the other hand, another survey of 110organizations in 1982-84 found that the greatmajority of firms (80 to 90 percent) collectedindividual performance statistics for at leastsome of their workers.9 Most of the jobs af-fected were the clerical jobs, but some wereprofessional or technical. About one-quarterof the firms collecting performance statisticssaid they did so only for assessing group per-formance, to plan for peaks and valleys of workdemand, and to cost-justify their use of the of-fice systems. The remaining three-quarterswere using individual operator statistics tomake some sort of individual evaluation—whether for base pay, incentive pay, promo-tion, or training-for some of their work force.In some cases, machine statistics were “almostthe entire basis” for such judgements, and inothers it was ‘‘one factor in five or six factorsused to evaluate performance. ”

A survey of the same 110 organizations in1985-86 revealed no increase in the percent-age of employers using computer measurementfor personal evaluation. However, a majorityof the firms now reported that they had cre-ated a‘ ‘more formal system’ for setting workstandards and letting employees see the resultsof monitoring if they wished.l0

Those organizations not collecting statisticsat all were usually either:

1. organizations using older word processorsor microcomputers that did not have soft-ware for measurement;

2. organizations with new applications inearly implementation;

3. non-profit organizations or universitieswho “just don’t do that”; or

4. State and local government agencies which“saw no need to compile those records. ”

“Hay Group, Inc., “Analyses of Customized Items for the Of-fice of Technology Assessment, ” 1986 Hay Office Systems Sur-vey, September 1986.

‘Alan Westin: “Privacy and Quality of Work Life Issues inEmployee Motoring,” contractor report prepared for OTA,1986.

‘“ Ibid.

Service observation, which is usually donefor jobs where employees have a great deal oftelephone contact with customers, is oftencombined with electronic measurements ofproductivity as well. As mentioned above, serv-ice observation is a standard practice in thetelephone industry, and most of the Nation’s226,000 operators and service representativesare evaluated in this way. In addition, the greatgrowth of telemarketing and telephone cus-tomer service in the past decade means thatan increasing number of employees are affectedby service observation. A few firms inter-viewed by OTA reported that they had service-observing capabilities in their telephone sys-tems but did not use them.

Westin estimates that the great majority ofclerical employees working on computer ter-minals-in the 65 to 80 percent range-are notcurrently being monitored by computer andevaluated for pay, promotion, or discipline onthat basis.11 Further, most professional, tech-nical, and managerial workers—95 percent ormore—are not currently evaluated based oncomputer statistics. However, if 20 to 35 per-cent of clerical employees are being monitored,this means that 4 to 6 million employees arebeing evaluated in this manner. The additionof professional, technical, and managerialworkers could add another million or two tothe total, and this number could grow stead-ily larger over the next 5 to 10 years. To thistotal also should be added retail sales workersand grocery clerks, whose speed and sales vol-ume are sometimes monitored via electroniccash registers.

The clerical work force is predominantly fe-male, and the low-skill end of the clerical workforce has a disproportionate number of minor-ity women. l2 Similarly, women are morelikely to be employed in the lower levels ofprofessional work, such as routine computerprogramming or routine insurance underwrit-ing, rather than in higher levels of those profes-sions. Because monitoring is most likely to be

‘‘ Ibid.IW. S. Congress, Office of Technology Assessment, Automat-

ion of America Offices, OTA-C IT-287 (Washington, DC: U.S.Government Printing Office, December 1985), pp. 300-304.


applied to precisely these lower level jobs, workmonitoring is a topic that especially affectswomen and minorities.

Work Monitoring in the FederalGovernment 13

The work-monitoring practices in the Fed-eral Government tend to follow some of thesame patterns as the private sector. There areno reliable statistics on how many Federal em-ployees have their work monitored by computer.

In general, Westin’s survey, combined withinterviews by OTA, found the same sort of dis-tribution in the use of work monitoring in thegovernment as in the private sector. Someagencies, or specific work groups within agen-cies, used performance statistics only at theaggregate level and only for planning or bud-geting purposes. Some used monitoring sta-tistics as part of individual evaluation of someworkers, usually clerical workers. Some, espe-cially small agencies and those with olderequipment, did no monitoring at all.

The following are a few specific examples ofapplications of work monitoring in the FederalGovernment. For example, performance sta-tistics are automatically collected for data tran-scribers at the Department of Agriculture’sNational Finance Center. Statistics include atotal time on machine, keystrokes per hour,and errors. Supervisors get daily, weekly, andmonthly reports; operators get feedbackmonthly and some also maintain manualrecords of their own performance. 14 Similarperformance criteria are used for data tran-scribers at the Bureau of the Census, but be-cause of the design of the computer system,keyers must record the time manually. Whena new system is installed in 1988, all informa-tion will be collected automatically. ”

I :lThi~ section is based on a survey performed by Westin ‘n

1984 of 44 Federal agencies. See Alan Westin, “Privacy andQuality of Work Life Issues in Employee Monitoring, ” contrac-tor report prepared for OTA, 1986, pp. 39-63. In addition, OTAstaff interviewed a number of Federal agency managers andunion officials.

14 Based on OTA interviews.‘sBased on OTA interviews.

The Department of Labor’s Office of Work-ers Compensation monitors the work of claimsexaminers and bill examiners. For claims ex-aminers the statistics include time elapsedfrom case creation to case adjudication, num-ber of adjudications, number of wage lossclaims processed, and elapsed time from receiptof claim to decision. Bill examiners statisticsinclude number of bills paid per day, timeelapsed from receipt of bill to payment. Claimsexaminers are in the GS 5- I 1 range; bill ex-aminers are typically GS 4-6. 16

The Internal Revenue Service, which has re-cently implemented its Automated CollectionService, employs about 2,300 contact repre-sentatives who speak with delinquent tax-payers by telephone, negotiate payment sched-ules, and update taxpayer files. Performancedata is collected by computer (time per trans-action, time logged on and available for work);in addition supervisors are required to listenin on calls to monitor for courtesy and correct-ness of information. Employees know thatservice observation is performed, but do notknow specifically when they are being listenedto. According to IRS sources, service obser-vation is fairly infrequent and used primarilyfor training of new employees.

At the Social Security Administration’s Tele-service Centers, service observation capabil-ity of the new telephone system has becomea matter of dispute between SSA and the unionrepresenting 1,500 teleservice workers (Amer-ican Federation of Government Employees).The union is attempting to negotiate specifictime periods during which service observationwill be used. At present it can be applied atanytime and employees do not know whetherthey are being monitored.

Given similar levels of computerization inthe Federal Government and the private sec-tor,17 it seems likely that the number of mon-itored office workers in the Federal work forceis similar to the private sector—20 to 35percent.

1 6 B a s e d on OTA hIkrVkWS.

‘iSee OTA, Automation of America Offices, op. cit.


HOW TECHNOLOGY ASSISTS SELECTEDBUSINESS APPLICATIONS

The technology used in computerized workmeasurement is not especially esoteric. Mostpeople whose work is electronically measuredare working on mainframe or minicomputersor are using telephone systems controlled byminicomputers. Stand-alone personal com-puters generally do not have the power to main-tain sophisticated work-measurement softwarewhile also carrying out the desired application.[See later section on “Personal Computer Mon-itoring. Office systems that interlink a num-ber of computers may sometimes have onecomputer dedicated to monitoring and meas-uring workflow.

Some work-monitoring software systems areavailable commercially, but in many cases,firms with large data processing departmentsdevelop their own work-monitoring softwarein-house, or with the help of work-measurementconsulting firms. Commercially available sys-tems in common use include those associatedwith key-to-disk data-entry systems, auto-matic call distributors, and “back office com-puters” for travel agents, which will be dis-cussed later.

Individual work-measurement statistics canbe developed using information that is alreadybeing collected for some other purpose. Devel-oping work-measurement statistics from thisinformation is simply a matter of being ableto retrieve some or all of this information, storeit in a separate file, and perform statisticalanalyses.

Because it is based on analysis of readilyavailable information, work monitoring caneven begin “accidentally,” without a specificmanagement plan to introduce it. For exam-ple, in one Midwestern bank, the data proc-essing department installed a “black box” torecord the time, type, duration, originating ter-minal, and user for every inquiry or transactionin on-line databases. The purpose was to mon-itor the speed of computer response time andto make sure other departments were gettingthe level of service promised by the data proc-

essing department. Itemized reports were sentto managers, who found that these reportswere useful for other purposes, such as plan-ning personnel schedules, justifying requestsfor new staff or equipment, and measuring thetime it took individual employees to completetransactions. 18 They are now being used aswork-monitoring tools in at least some of theparticipating departments.

In most locations, work-measurement toolswere developed explicitly for the purpose, butthe idea is still the same. They collect and re-analyze information that is already beingrecorded about computer utilization or busi-ness transactions. For example, travel agentsat some large agencies work on terminals con-nected to a network that includes a nearbyminicomputer and a mainframe computer atsome central location. The applications soft-ware in the mainframe allows them to checkschedules and make or cancel reservations. Theminicomputer, called in some organizations the“back office computer, ” records details aboutcomputer utilization. It notes who logged onto which terminal at which time, and it alsomakes records of the time, type, and amountof each transaction that the agents performon the mainframe. Thus the back office com-puter provides a local audit trail and sales rec-ord for the whole office. This information canbe used in a number of ways, for example, itallows the local printing of tickets and itiner-aries for customers. It also can be used to de-velop individual performance histories, sinceit has a complete record of all the computeractivities of each agent.19

In other firms interviewed by OTA, workmonitoring software was an integral part ofthe application software. That is, the samecomputer software package that helps an in-surance claims examiner to key in client infor-mation and calculate the amount of a payment,

180TA interview with Senior Vice President for Operations,a rnidwestern bank, May 1986.

190TA interview with three travel agency managers, June1986.

also automatically tallies the number of claimsof each type that each examiner completes ina day. In several cases these tallies were thentransferred to another computer program, per-haps in a personal computer, which does sta-tistical analyses, compares performance toestablished standards, and prints reports forsupervisors.

Collection of data about employee perform-ance can in many cases be made transparentto the user, that is, information can be collectedwithout interfering with the work that is be-ing done. From the viewpoint of the user, anautomatic call distributor (ACD) simply routesincoming calls to individual telephones. In ac-tual fact, however, the distributor is also auto-matically recording the type of call (inside oroutside line), the time the call arrived, the iden-tity of the employee to whom it was routed,the number of seconds before the employeepicked up, the time the call started, the timethe call ended, the number of times the callerwas put on hold and for how long, the exten-sion to which the call is transferred, the num-ber of seconds before that person picked up,and so on. In addition, it can show the super-visor at any moment which operators are busy,which are waiting for work, which are on break.At the end of the day it can provide summariesof individual and group activities.

Properly organized, this can be very usefulmanagement information. For example, anACD can report the number of seconds cus-tomers were “delayed” before someone wasavailable to help them or the number of cus-tomers that “abandoned” calls-hung up with-out speaking to anyone. If these figures gettoo high it may indicate the need for more tele-phone lines and more customer agents. Anal-

Ch. 2—Using Computers To Monitor Office Work ● 3 5— — . —

ysis of daily or monthly work volumes can helpmanagers better understand cycles in theirbusiness so they can predict busy periods whenthey must hire temporary workers or offerovertime.

The example in figure 2 shows such a statusreport. Service level (“serv level’ is the per-centage of incoming calls that were answeredwithin the specified time (typically 20 seconds).Calls offered is the total number of calls, in-cluding those that were lost, delayed, ordiverted as shown in the following columns.Positions manned (“pos manned”) means thenumber of agents jacked in and ready to work;the following columns indicate whether theagents are on incoming or outgoing calls. Aver-age delay is the average time in seconds fora call to be processed, whether answered, de-layed, or lost. The next columns show the num-ber of calls waiting (CW), the maximum num-ber of calls that were waiting at any time inthis period, and the number of seconds thatthe current longest call waiting has waited.

Here again, reports about individual per-formance are fairly simple to develop based oninformation that must be collected in any case.In order to route calls, the computer control-ling an automatic call distributor must keeptrack of which telephones are busy, which areavailable, and which are unattended at anygiven time. However, this information mustbe sorted and averaged in order to be of use.A supervisor could make no sense of all thedetailed information that a computer collectsabout each call. The work-monitoring softwaresorts, totals, averages, and summarizes the in-formation so that a supervisor can see activi-ties of the entire work group, or totals and aver-ages of each individual’s activities for a given

Figure 2.— Example of a System Status Report for an AutomaticCall Distribution System

Serv Cal ls Cal ls Cal ls Cal ls Pos. CalIs Calls Avg. Max OdlyGate level offrd. dlyd. Lost div. manned in out delay CW CW time

1 47 “ 15 8 7- 1 6 0 2 1 0 1 0 -

2 43 7 4 4 0 3 1 2 2 3 3 914 67 6 2 1 1 6 0 0 7 0 1 0

52 28- 14 12 2 15 1 4 ‘3 3SOURCE Ada-pled from Sold State Syslems I nc ‘Tie Smart Tele~hone System ACD Supervisor User s Gultie Mar;etla

Genrqla 1982


timed period, or exception reports, that showparticular calls or particular employees thatare far out of the average range.

Figure 3 shows the type of information typi-cally available on the VDT screen of the super-visor of a group of telephone customer serviceworkers using an automatic call distributionsystem. Status options here include vacant(“vent”) or on break, talking (“talk”) or avail-able for calls (“avlb”).” Work” means the agentis doing other work related to a previous call,perhaps updating the database or preparinga letter to the customer, and is not acceptingcalls. The agent named Joe is talking and hassignaled for ‘help, ” requesting the supervisor

to come on the line. The time column showshow long (in minutes) the individual has beenin the current status. The last column showsthat three agents have calls waiting for themon other lines.

The example in figure 4 shows the sort ofinformation that might be included in produc-tivity reports summarizing periods of a fewhours, or a day, or a week. This example reportsthe number of calls, the total handle time(“hndl”) in minutes (the sum of “talk time”and “work time’ ‘), and the total time availableand waiting for calls, in minutes. In additionthe report shows the average handle time, talktime, and work time per call in seconds.

Figure 3.— Example of a Supervisor’s Display in an Automatic Call Distribution System

AGENT AGENT STATUS TIM E 10:32 5/10/87SUPERVISOR GATE NO NAME SUPVS 1

1 11112223

513512514560570510513512

SAMJANEJOEBILL

FREDSUESAMJAN E

VCNTTALKHELP

WORKAVLBTALKVCNTTALK

6.3 CW 40.31,24,76.31,26.30.3 CW 1

SOURCE Ad t~~ t W ‘ I I IrII S I I j SI ate S} St F rTIs I r+ The Smart Tel! phor]e Svsterr ACD Supervisor Uwr s Guide Mar!ettaGt( r [] I d 1 ’102

Figure 4. —Example of Agent Statistics From an Automatic Call Distribution System

14:32 10/14/82 AGENT STATISTICS

AGENT NO. HNDL TALK WORK AVLB AVG AVGGROUP 4 GATE CALLS TIME TIME TIME TIME HNDL TALK WORK

GEORGE 454 1 0 0,0 0.0 0.0 0.0HARRY 455 1 0 0.0 0.0 0.0 0.0JERRY 582 1 6 8.4 8.4 0.0 6.9 84 84 0SUPV1 580 1 5 3.5 3.5 0.0 5.7 53 53 0S U S A N N A H 5 0 1 1 5 10.0 10,0 0.0 5.3 120 120 0BILL 503 1 4 3.6 3.6 0,0 3.2 54 54 0

SOURCE Adapted from Solld State Systems Inc ‘The Smart Telephone System ACD Supervisor User’s Guide, ’ MariettaGeorgia 1982

WORKPLACE ISSUES RELATED TO WORK MONITORING

People in some organizations perceive work ment at all levels to put out a good product.measurement and service observation to be At other organizations, use of the same tech-useful tools that help employees and manage- nologies is resented and feared as “Big Brother

UNIONMODEL

TAYLOR

PRODUCT IONMODEL

Indlvldualperformance

sampling

Constantmachine

montoring

Figure 5.– Two Models for Electronic Work Monitoring

All dataavailable toemployee

Data notreadiIy

available

PRIVACY-RELATED ISSUESintrusiveness ” and Subject Access’”

Known pro-cedure forChalIenglng

record

Employee canpace work

Machine paceswork

Group produc-tion quotas

Individualquotas

No procedurefor challenge

— —I

i

— —

Quality/problemfactors

recognized

Quantity/speedstandard

I Standard pay

Piecework payor bonus

Individual shortfall leads to

training or groupI discussionI

Individual shortfall leads todiscipline

—

SOU R( c A I ah F West I n Prlvdc y and Qual!ty of Life ISSUPS n Employee Mnn Itor!ng contractor report for OTA 1986

LABOR RELATION/EMPLOYEE RELATIONS ISSUES‘Fair Work, “Fair Pay. “Fair Performance Evaluation” I

I—.

I

38 . The Electronic Supervisor: New Technology, New Tensions

Table 3.—Workplace Issues Related to Work Monitoring—

Privacy and Labor relations Health/qualityaccess related or “fairness” of life

Is monitoring constant or intermittent? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x xCan employees see their own records? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xCan the employee challenge, explain, or correct records?. . . . . . . . . . . . . . . x xDoes the employee or the machine pace the work? . . . . . . . . . . . . . . . . . . . . x xDo employees understand performance criteria and use of information? . . x xAre quotas set on an individual or group basis? . . . . . . . . . . . . . . . . . . . . . xAre quotas fair, allowing work at a reasonable pace? . . . . . . . . . . . . . . . . . . . x xIS pay standard or based on performance? . . . . . . . . . . . . . . . . . . . . . . . . . . . . xWhat happens to employees falling short of quota? . . . . . . . . . . . . . . . . . . . . x x—SOURCE Off Ice of Technology Assessment, 1987

surveillance. ” The difference seems not to beso much in the specific measurement technol-ogy, but in the politics of how it is introduced,how it is used, and what is done with the in-formation collected.

Although many workers’ complaints aboutmonitoring focus on its intrusiveness, a closerlook shows that privacy is only one of a com-plex of issues raised by electronic work moni-toring. Table 3 outlines some possible charac-teristics of a work-monitoring program andindicates the kinds of issues that are raisedby them. Privacy and access issues cover suchquestions as whether employees know they arebeing monitored and whether the employeeshave access to records about their own per-formance. The second set of issues relates tothe perceived fairness of the monitoring sys-tem and the way the employer uses it in evalu-ating and rewarding employees; these are ques-tions of employee relations. The final set ofissues, overlapping the other two, relate tostress, health, and the quality of working life.

Whether the effect of monitoring is perceivedas intrusive, unfair, dehumanizing, or un-healthy often depends on how managementstructures the work-monitoring program, whatit does with the data it collects, and how thoseactions are perceived by employees.

What Is Fair

Westin used some of the same elements dis-cussed in table 3 to construct two models ofwork monitoring shown in figure 5.20 Westin

‘“This section draws heavily on Alan Westin, “Privacy andQuality of Work Life Issues in Employee Monitoring, contrac-tor report prepared for OTA, pp. 103-112 (draft).

chose to call the first model the “union’ model,since it represents a blend of features includedin model contract language suggested by someU.S. and international unions. The second hecalled the “Taylor production” model; it isbased on an extreme form of an industrial engi-neering approach to work measurement, onewhich places virtually all information andpower in the hands of management. In apply-ing these models to the organizations inter-viewed for OTA, no pure examples of the “unionmodel” were observed. The “Taylor produc-tion model” in its pure form was observed inaction in a few government and private sectororganizations. Most organizations used meth-ods representing a blend of the features of thetwo models, with about two-thirds of the orga-nizations interviewed by Westin tending towarda modified version of the “Taylor” model.

In most cases, employers introduce electronicmonitoring unilaterally, only informing em-ployees of the change after all decisions havebeen made. Often, too, monitoring is only oneof a number of changes in work process or jobdesign that take place when new office equip-ment is purchased. As was discussed in detailin OTA’s report, Automation of America Of-fices, employee participation in design and im-plementation is often a key to successful im-plementation of new office systems.

Ensuring employee participation can requireeffort on the part of managers, as few U.S.workplaces have mechanisms for employee in-put in areas of technological change or evalu-ation procedures. Nevertheless, Westin foundthat the difference between employees protest-ing over “Big Brother surveillance” and em-ployees perceiving work measurement as rea-

Ch. 2—Using Computers To Monitor Office Work ● 3 9-— —

sonable, often depends on whether they agreeon: 1) the fairness of the standards set; 2) thefairness of the monitoring process employed;and 3) the fairness of the way measurementsare used in employee evaluation. This agree-ment was usually made through worker-man-agement discussions before monitoring wasimplemented. Such agreements are possiblewhere there is genuine involvement of employ-ees—either through joint labor-managementcommittees in unionized organizations, orthrough employee involvement techniques innonunion settings. Where management startedwith the trusting assumption that almost allemployees were ready to put in a fair day’swork for a fair day’s pay, and where topics suchas work standards, work measurement, andproductivity recognition were matters for opendiscussion, introduction of monitoring wasusually relatively painless.

One impression that emerged from OTA’sinterviews is that the way managers and em-ployees deal with monitoring often closelyparallels the way they deal with other work-place issues. Firms whose “corporate culture”tends toward authoritarianism tend to usemonitoring in an authoritarian way. In orga-nizations where relations between employeesand managers are antagonistic, the monitor-ing system is a source of antagonism, but onlyone of many. In organizations where coopera-tion is the norm, people worked together to develop a fair system.

Recognizing that employee involvement inthe design, testing, implementation, and con-tinuing adjustment of work monitoring is cru-cial to a successful process, it is also necessaryto deal with the substantive issues, to be con-sidered in designing such a program. Table 4shows some of the issues to be considered. Themain categories, and the specific questions inthis chart represent recurring themes in a num-ber of interviews with monitored workers andtheir managers.

Westin’s sample found that only about one-third of the firms in his sample using electronicwork monitoring for individual evaluationswere following what he called ‘fair work evalu-

Table 4.—Key Issues and Problem Areas inMonitoring Worker Performance

Key issues/problem aspects

Fairness of work standards:● Do standards fairly reflect the average capacities of the

particular work force?. WiII they create unhealthy stress for many employees?● Do they take into account recurring system difficulties and

other workplace problems?. Do they include quality as well as quantity goals?● Do they represent a “fair day’s pay’” for a‘ fair day’s work?’. Do employees share in any productivity gains achieved

through introduction of new technology?

Fairness of the measurement process:●

●

●

●

●

●

●

Do employees know and understand how the measure-ments are being done?Can the measurement system be defeated easily, therebyimpairing the morale of those willing to “follow the ru!es?”Do employees receive the statistics on their performancedirectly, and in time to help them manage their work rate?Is the relation between quality, service measures, and workquantity communicated by supervisors when they discussproblems of performance levels with employees?Do supervisors communicate clearly that they are takingsystem and workplace problems into account?Are group rather than individual rates used when particu-lar tasks make such an approach more equitable?Is there a formal complaint process by which an operatorcan contest the way work data has been used by the su-pervisor?

Fairness in applying measurements to employee evaluation:● Are there meaningful recognition programs for these em-

ployees?. Is work quantity only one of a well-rounded and objective

set of performance criteria used for employee appraisals?Ž Does the employee get to see and participate i n the per-

formance appraisal?. Is there an appeal process from the supervisor’s perfor-

mance appraisal?● Is there a performance-planning system that identifies em-

ployee weaknesses in performance and identifies ways toremedy such problems?

SOURCE Alan R West In, “ P r i v a c y a n d Quallty of L14e issues I n E mploym

Monitoring, ” contractor reporl for OTA, 1986

ation policies’ along the lines of a positive an-swer to most of the questions in table 4.

Of the 34 case examples submitted to OTAby unions, 28 dealt with electronic monitor-ing of office workers (unionized and nonunion-ized) like the ones studied by Westin. In nearlyall the case examples, employees had little in-put concerning the monitoring system, and inonly a few cases was it clear that they had ac-cess to information about their own perform-ance or the ability to contest wrong informa-tion. In nearly all these cases the workers were


described as considering the monitoring sys-tem unfair.21

Standards of Performance

Work measurement systems are usually ap-plied to jobs for which standard end products,or surrogates for end products, can be clearlyidentified. That end product might be “cus-tomers served, ” “claims paid, ” “programswritten, “ “interviews completed. ” Generallyspeaking, electronic work monitoring primar-ily measures the quantity of work performed.Other methods, discussed in the next section,measure quality.

An important element in measurement ofalmost all kinds of work is time. In almostevery case in table 2, the purpose of measure-ment is to measure the time it takes the em-ployee to do something, and then to comparethe result to a standard. Robert Nolan, an ex-pert in developing work-measurement systems,defines a work-measurement system in thisway:

In the most simple terms, it is a means ofestablishing what a fair day’s work should be.It has two main components, a measure of thevolume of work, and a measure of the employeetime used up. These two factors can be ex-pressed in their only common denominator:the time required to produce one unit of work,or what we call a standard.22

Thus, measurement alone is often of little useas a management tool, unless its purpose isto compare the individual or group perform-ance to a standard.

Standards may be established in a numberof ways. Many are arrived at rather informallyor arbitrarily, perhaps based on supervisors’or managers’ estimates of how long it oughtto take to complete certain tasks. In some casesstandards are set based on historical perform-ance levels; managers may take an average ofsome past period, and expect that it be main-tained as an average in the future.

‘*’’AFL-CIO Case Examples, ” November 1987.‘*Robert E. Nolan, “Work Measurement, ” in Robert N. Le-

herer, White Coflarl%ductivit.y (New York, NY: McGraw Hifl,1983), p. 111.

Sometimes “standards” are really goals orideals. In one firm interviewed by OTA thestandard of 50 completed transactions peragent per week had recently been establishedby the national office. The standard was de-veloped by dividing the average revenue pertransaction into the total revenue the firmhoped to generate at each branch office.23 Thenew standard was not related to past perform-ance levels or analysis of the best way to dothe job, but rather to the amount that mustbe sold in order to meet revenue projections.In this case, 50 transactions per week was farabove past performance; office managershoped that introducing incentive programswould inspire agents to achieve the new goals.

A more formal method of standard settingis the “engineered standard. ” The Methods-Time Measurement (MTM) system or the Ad-vanced Office Controls (AOC) system havebeen used in many office settings. In thesemethods, a trained analyst, usually an indus-trial engineer, observes a work task, selectsthe most efficient method of performing thetask, and then will time the actions of averagepeople performing the task under averageworking conditions. General MTM and AOCstandards have been developed for nearly everyimaginable motion in an office workplace. Forexample, the MTM standard for fasteningsheets of paper with a table model stapler is41 time measurement units (TMU), or about2.9 seconds. Opening an envelope and remov-ing the contents takes 198 TMU or 14.2 sec-onds. A trained analyst can combine a num-ber of these general standards, develop newones, and adapt them to the special circum-stances of spacial arrangement, work process,or equipment use in a given office.24

A well-designed standard, according to ex-perts, is not one that makes people work asfast as possible, but one that encourages goodaverage work. It should include time for per-sonal breaks and allow for personal variabil-

230TA i~~rviews, ~SiSt~t manager, Washington office! ‘a-tional travel firm, June 1986.

24 ExmpleS from Robert E. Nolan, “Work Measurement,in Robert N. Leherer, White Collar Productivity (New York,NY: McGraw Hill, 1983), pp. 142-146.

Ch. 2—Using Computers To Monitor Office Work . 41— —

ity — working a little faster at some times andslower at others. Fair standards must be real-istic, taking into account system downtime,slow response time, varying levels of complex-ity of different tasks, and so on. When stand-ards are not realistic, or when they are not per-ceived as fair by employees and managers, theycan easily lead to declines in morale, increasedturnover rate, and ultimately a decrease in pro-ductivity.

Since AOC, MTM, and other “predeterminedtime’ systems provide a standard time for thecompletion of each task, an employee’s actualperformance can be compared to that stand-ard. If the standard time for examining a cer-tain type of insurance claim is 10 minutes, thenan employee who completes 6 of them will havedone 60 standard minutes or 1 standard hour’sworth of work. An employee who completed48 such cases in an 8-hour day would be saidto be working at 100 percent of the standard,that is, his or her paid hours would exactlyequal standard hours. Faster employees mightwork at 110 or 120 percent of standard, whileslower ones work at 80 or 90 percent of stand-

ard. The determination of an “acceptable’ pacedepends on the firm, but a well-designed stand-ard is one where most trained, experienced em-ployees will work in the range of 85 to 100 per-cent of standard most of the time. 25

Figure 6 shows part of a weekly work moni-toring report for an insurance employee work-ing under a‘ ‘predetermined time’ work-meas-urement system. The report lists the types oftasks done, the standard time to do the taskonce, the number of times the employee actu-ally completed the tasks, and a calculation ofthe “earned’ hours. Figure 7 summarizes workof a group of insurance employees. Figure 8integrates the work monitoring system witha time and attendance report. It shows thenumber of hours each employee was availablefor work (’‘avail work”). The time available formeasured work (“avail meas’ ‘—employees mayhave other duties that are not captured by thesystem) and the number of earned hours (’‘earnhrs”) worth of work completed. Note that there

~~see Robert E. Nolan, “Work Measurement, ’ in Robert N.Leherer, White Collar Productivit.v (New York, NY: Mc(~rawHill, 1983), p. 121.

Figure 6. —Example of Individual Work Monitoring Report (Performance Summary)

RUN NUMBER ARM1B140 WORK MANAGEMENT PROGRAMDATE OF RUN - 10/09/84 INDIVIDUAL PERFORMANCE REPORTTIME OF RUN -18:00

WEEK ENDING 10/6/84

EMPLOYEE EMPLOYEE DDIVISION : G CLAIMS AND SERVICEDEPARTMENT: GB CLMS PROC

---------------------------------------------------------------------------CALCULATION OF EARNED HOURS -------------------------------------------------------------------------- -

KVI DESCRIPTION UKVI AVERAGE TIME ITEMS EARN EDDV DP CODE NAME FLAG TO DO ONE ITEM x COMPLETED – HOURS

G GBG GBG GBG GBG GBG GBG GBG GBG GBG GBG GB

00400050007000800140015001750176019002100185

EOB STUFFSUBSCRIBER PREPSUBSCRIBER CODINGPROVIDER CODINGMEP ADJUDICATIONVISION ADJUDICATIONPRIMARY GSCR ROSTERGSCR CHK-ADJUDICATORCOB PROCESSINGCOB RETURN LETTERSPULLING CLAIMS

00 MIN 16 SEC01 MIN 48 SEC03 MIN 54 SEC01 MIN 03 SEC02 MIN 47 SEC02 MIN 33 SEC00 MIN 10 SEC00 MIN 08 SEC04 MIN 57 SEC00 MIN 49 SEC03 MIN 00 SEC

148228

48268

621454

2050

8

0 60,00.10.1

22.42.91.81,01,60.60.4

TOTAL EARNED 32.0

SOURCE Adapted from James S Hogg Manager Professional Producttwty Blue Cross & Blue Shield of Maryland

63 -982 0 - 87 - QL : 3 ‘ - 2


Figure 7. —Example of a Work Monitoring Report for a Group‘(Volume of Work Accomplished by Group in One Week) “

RUN NUMBER - ARM1B130 WORK MANAGEMENT PROGRAMDATE OF RUN 10/09/84 VOLUME SUMMARY FOR WEEK ENDING 10/06/84TIME OF RUN - 18:13

DIVISION: G CLAIMS AND SERVICEDEPARTMENT GB CLMS PROCUNIT x x xSECTION YYY

KVICODE

001000200040005000510060007000800090010001100120013001400150016001750176018001810182018301850190019502000210

KVIDESCRIPTION

MAIL SORTPROCESS IN COMPLETESEOB STUFFSUBSCRIBER PREPPROVIDER PREPMEDICARE CLAIMS PROCSUBSCRIBER CODINGPROVIDER CODINGSUBSCBR DATA ENTRYPROVIDER DATA ENTRYVISION PREP & CODEVISION DATA ENTRYPAF REPORT UPDATEMEP ADJUDICATIONVISION ADJUDICATIONPHONE INQUIRYPRIMARY GSCR ROSTERGSCR CHK-ADJUDICATORADJUSTMENT CODINGPROC. CODED ADJUSTMTSHND. PROC ADJ/NO FILEART COMP-HANDLG REFDPULLING CLAIMSCOB PROCESSINGWRITTEN CORRESPONDENCEPROC OCC DELETESCOB RETURN LETTERS

VOLUME

1938219

2105216425391457

1496293

1072164168232846

70382

1740926

15141415

836214355

PERFORMANCEREFERENCE

0,00680.0085000460028000045003850,0657001790,02240.01850024600236001000.04660,042601015000300,00230.09270045400102012840.0496008230.38010105400135

EARN EDHOURS

132199.76 019

151300268

6.61984 04 02 3

3943 0

3885 221140 601190 43 08.04 50 7

is a wide variation in effectiveness (“eff %"),which calculates the earned hours as a percent-age of measured hours. This particular firmdoes not use an incentive system so all employ-ees are paid for their regular hours despitethese differences. The report also shows totalsfor the whole group.

The work process-the set of procedures thatgovern what tasks are done and how tasksinterrelate—also has a major impact on manypredetermined-time work-measurement sys-tems. Because the “standard” for each taskor set of tasks depends on a close study of thework performed, any changes in the work re-quire a change in the standard, if the stand-

ard is to be fair. Changes in the work mightarise when a new product is introduced (say,in an insurance company, a new kind of pol-icy), or when regulations change (e.g., requir-ing a change in the kind of information banksmust supply to customers), or when the tech-nology changes. Work-management specialistsat firms using predetermined-time systemsnote that “maintenance” is a major need ifwork measurement is to be applied conscien-tiously. Work must be periodically re-analyzedand standards must be adjusted.

Standard setting is often combined with jobdesign, work simplification, or proceduralchanges, because it is difficult to establish a

Ch 2—Using Computers To Monitor Office Work ● 4 3.

o

1- 03w

u

wn


set time for performing a task if everyone isfree to do it in a different way. Some ways arebetter than others, whether faster, less fatig-uing, or more reliable. Experts in productivityand work measurement usually counsel thatemployees should be included in the processof changing procedures and establishing stand-ards. They argue that employee involvementnot only short-circuits resentment to stand-ards that are imposed from outside, but usu-ally leads to the creation of better proceduresand fairer standards, since employees under-stand the work that needs to be done betterthan anyone, including their managers.

An example of employee involvement instandards development is the case of legal caseanalysts at an insurance firm.26 These highlyskilled workers handle correspondence and trackthe progress of legal cases on a computer-basedlegal diary system (LDS). Although the LDSsoftware created some internal statistics onthe transactions done, they had not been pre-viously used for individual evaluation becauseno standards had been developed for the caseanalyst’s work. The work-measurement spe-cialist assigned to develop standards found thework very complex and also saw that the le-gal diary software was very flexible, allowinganalysts to use several different procedures forcertain tasks. Instead of trying to prescribeprocedures and standards, the work-measure-ment specialist held a series of 24 two-hourseminars in which the analysts talked abouttheir work. They discussed different tasks,compared their approaches, and decided amongthemselves the simplest and most effectiveprocedures for each task. They also helped toset the standard times for the tasks. Interest-ingly, the productivity of this department, interms of dollars recovered through legal ac-tions, began to increase before the final workmeasurement program was in place, presum-ably because the case analysts voluntarily be-gan using the improved procedures as soon asthey were developed in the seminars.

‘GJarnes Hogg, “The Results of Technical and ProfessionalMeasurement in Insurance, ” in Proceedings, 1986 AOC UsersConference, May 15-18, 1986, Robert E. Nolan Co.; also, per-sonal communication, May 29, 1986.

Within many organizations, introduction ofwork measurement and the process of settingstandards can become a hotly contested labor-management issue and a major source of em-ployee discontent. Where employees are notinvolved in standard setting, they may viewa new standard as an unfair “speed up, ” anattempt by management to make them workharder for the same pay. Similarly, work sim-plification or procedural changes that are im-posed from outside can be viewed as remov-ing variety and autonomy from the job, andmaking it less interesting and more me-chanical.

For example, the changes in work standards,evaluation, and pay that accompanied workmonitoring for claims examiners prompted aunionization drive at Equitable Life Assur-ance. With the introduction of the measure-ment system, pay was changed from a straightsalary to an incentive program that was basedon performance. Examiners complained thatthey had to work much faster in order to makethe equivalent of their old salaries. A few ac-cepted transfers to lower paying jobs becausethey could not keep up the pace. The contractbetween Equitable and District 925 of TheService Employees International Union(SEIU), addresses some of the issues discussedin the section on “What Is Fair. ” Under thecontract, employees now have access to theirown performance records and a procedure forchallenging records. Evaluations are based 80percent on computer-based statistics and 20percent on supervisors’ judgments. In addi-tion, the contract changed several other work-ing conditions, such as leave policy, that hadbeen a subject of dispute.

When electronic monitoring allows a com-plete record of each worker’s performance, itbecomes easier to pay workers based on theiroutput. Some call this a revival of the “piece-rate’ system and decry it as a form of workerexploitation. Often, however, performance doesnot translate into pay directly on a per-piecebasis. For example, “incentive” plans pay abase rate for acceptable performance and bo-nuses for higher levels of performance.

Ch. 2—Using Computers To Monitor Office Work ● 45

Incentive programs appear to be fairly com-mon in data entry, where there is a wide rangeof performance.27 Operators who are very fastcan increase their income by 50 percent or moreabove the base rate, depending on how the in-centive plan is structured. In some firms wherestandards are based on a predetermined-timemethod (i.e., so many keystrokes equal onestandard hour of work), slower keyers can makebonuses if they are willing to work for a longertime. This also raises fairness questions andworker protection questions, for example, shouldemployees feel pressure to skip lunch or breaksin order to improve their performance?

Incentive programs have also been used byemployers to increase the performance of agroup of employees. In one bank, for example,industrial engineers studied the work of check-proof reading operators and found that theengineered standard was far above the currentlevel of achievement of the department. Ratherthan insist that operators begin to work to thenew standard, management began to pay regu-lar wages to those who met the old averageand bonuses to all whose work approached thenew standard. As departmental proficiency in-creases, the management expects to raise thestandard and adjust the bonus structure to en-courage even faster performance. 28

The practice of “rate busting” or increasingperformance standards over time is the basisfor many objections to monitoring. Ever-increas-ing standards do not have to be related to in-centive pay. Standards can rise due to newtechnology, revised productivity goals, or forother reasons that lead management to expectbetter performance from employees. In thewell-publicized case of one data-entry centeroperated by the Internal Revenue Service, em-ployees and their union were complaining aboutthe stress resulting from the increased paceof work. In this instance, workers were sea-sonal, and were invited back to work again each

“Data Entry Management Association, “Sixth AnnualMember Statistical-Compensation Survey, ” DE~A ~ews~et-ter, April 1985. The average rate for U.S. operators is 1 I ,400keystrokes per hour, but the fastest operators can do around25,000.

1kInterview with W’ork Nleasurement Manager of a southernbank.

year based on their previous year’s perform-ance. Since the number of available jobs haddeclined, only above-average keyers were in-vited back. However, performance standardswere also raised yearly, presumably to dealwith the workload. Thus, each keyer was re-quired to make an increasingly greater effortto remain “above average. ”29 An annual in-crease in standards has also been a cause ofcomplaint among key entry operators in DadeCounty, Florida.30

Complaints about job stress in the U.S.Postal Service, which received a great deal ofpublicity in 1984 and 1985, were directed pri-marily at fast pace and high work standardsrather than at automated equipment or thepresence of monitoring per se.31 Industrialengineers have noted since the beginning ofthe century that there is a limit to how mucha pace can be increased, even if incentives areoffered. Beyond a certain point the employees,either individually or as a group, will not per-form any faster on a regular basis, no matterwhat the inducement. Tolerance for perceivedunfair standards may depend on many factors,including the availability of other jobs. Onecase example noted that at three Internal Rev-enue Service Centers (where standards havebeen increased over the years), the turnoverrates for key entry operators are very high;presumably workers left due to the heavyworkload and fast pace required. However, theWilkes Barre Service Center, located in an areaof high unemployment, has a low turnover rate,but a high incidence of absenteeism.

Quality of Work

One problem with computerized work mon-itoring is that it focuses mainly on quantityor speed of work. Although a well-designedwork standard should allow workers time todo a good job, some standards require such

29Alan Westin, “Privacy and Quality of Y$rork Life 1 ssues inEmployee Monitoring, “ contractor report prepared for OTA,1986; also John Harris, American Federation of (governmentEmplo},ees, personal communication, February 1986.

j[’’’AFCIOCIO Case Examples, ” November 1987.]’For example, Peter Perl, “Monitoring by Computer Sparks

Emplo~ee Concerns, ” M’ashington Post, Sept. 2, 1984.‘J4’AFI.-C10 Case Examples, ” November 1987.


a fast pace that workers feel quality must besacrificed, or that the pressure to maintainboth speed and quality leads to excessivestress. In a number of cases, for example, tele-phone operators have objected that the pres-sure to complete calls within the standard timeprevents them from giving courteous, high-quality service. Some customers agree. On theother hand, because most operators are alsosubject to service observation (i.e., a supervi-sor sometimes listens in on calls to check foradherence to company procedures), they some-times feel stressed because of the conflict be-tween quantity and quality imperatives.33

In one mail order firm employee moraledropped and turnover rose to 80 percent aftermonitoring was introduced for VDT operators.“Everything was numbers, ” one executiverecalled, with ‘‘no attention to the downtimeand slow-response-time problems of the newsystem, or the changes in volumes operatorsfaced during peak periods, or of the differentlength and complexity of customer orders. ”In addition, the pressure operators felt to speedup their work led to mistakes and improperlyfilled orders. This productivity system wasscrapped after several years of operation, andreplaced with anew approach that still collectsindividual operator statistics, but has stand-ards geared to actual system operations andload cycles. In addition, as part of an overall“Quality First” campaign in this firm, the newperformance standards stress “order quality”over “sheer numbers. Several dozen long-termemployees interviewed for OTA said that thefirst productivity system was a “very badtime” at the company, but that the new ap-proach is “fair to both company and employ-ees, ” and they have no trouble in meeting boththe quantity and quality standards.34

Quality evaluation often requires inspectionby a human supervisor, but even here computertechnology can be of assistance. Some officesystems allow the supervisor to view on hisor her screen whatever transactions are tak-

‘3~’’AFL-CI0 Case Examples, ” November 1987.‘JAlan Westin, “Privacy and Quality of Work Life Issues in

Employee Monitoring, ” contractor report prepared for OTA,1986, p. 72.

ing place on an employee’s screen. Thus su-pervisors can view transactions as they aretaking place to check them for correctness.Computerized letter-sorting equipment usedby the U.S. Postal Service has similar capa-bility, so supervisors can periodically checkeach worker to be sure he or she is keying inproper zip codes.

For telephone service workers, qualitychecks are made by supervisors who listen inon calls to check that employees are courte-ous, are using proper procedures, and are giv-ing correct information. Correct informationis of interest to many firms whose representa-tives deal with the public, because employersmay be held liable for information their em-ployees give out over the telephone or for ac-tions taken as a result of telephone conversa-tions. In some cities “911“ emergency calls orutility company “trouble” calls are recordedso that there will be a record of time, address,or other information for possible future use.

In some organizations employees know whentheir supervisor is listening, either becausethere is a drop in volume or because a beep toneis heard. In other cases, the monitoring equip-ment is completely silent. One organizationwas so concerned that employees not knowwhen the supervisor was listening that super-visors were required to wear their headsets allday so that employees would not be able toguess whether they were listening or attend-ing to other duties.

In some firms quality assurance is consid-ered such an important function that a sepa-rate department handles it. At American Ex-press, for example, customer service supervisorslisten in on calls on an regular basis and ratethe quality based on established criteria. Inaddition, a separate quality assurance workerlistens in on calls of employees in any unit. Re-sults are always discussed with the employeewithin a short time after the call.

Some workers object to service observationprecisely because it is not necessarily objec-tive. Some firms, in fact, do not have firmlyestablished criteria for how often to listen orhow to rate quality. At one government agency


there were even stories of service observationbeing used punitively –i.e., a supervisor lis-tened to certain workers almost constantly, inorder to accumulate enough mistakes to dis-cipline them. This would clearly seem to be anabuse of service observation.

Service observation also invokes feelings ofinvasion of privacy, even though the conver-sation involved is not really a private one. Oneoperator interviewed for OTA said, “Whenthey are listening to me, I’m very upset be-cause you can’t stop it. ”35 The privacy aspectapplies more clearly to the customer’s side ofthe conversation. Some people may object tothird parties overhearing their conversations.Two States, West Virginia and California, at-tempted to legislate restrictions on service ob-servation. The West Virginia law required abeep tone when the supervisor is on the lineas well as a published notice to customers thatcalls might be observed. This law was passedin 1983 but repealed in 1986. A similar law waspassed in the California legislature but was ve-toed by the Governor.

In the case of West Virginia there is evidencethat both operator productivity and customersatisfaction remained high during the periodwhen “secret service observation was not per-mitted. 36 However, several employers, par-ticularly AT&T objected to the legislation.AT&T’s threat to build its new credit manage-ment center in another State was instrumen-tal in the repeal of the West Virginia monitor-ing law.37

Job Design and Work Process

As discussed in greater detail in OTA’s re-port Automation of America Offices, new in-formation technologies sometimes offer firmsmore flexibility in the way office work is de-

i“Michael J. Smith, Pascale Carayon, and Kathleen Miegio,“Motivational, Behavioral, and Psychological Implications ofElectronic Monitoring of Worker Performance, ” contract re-port prepared for OTA, July 1986.

‘h’’ Results Summary, Key Service Indicators, ” C&P Tele-phone Co. of West Virginia as of Sept. 11, 1985, supplied toOTA by Communications Workers’ of America.

‘Testimony of John D. Landers, AT&T, before the Judici-ary Committee, West Virginia House of Representatives, Feb.12, 1986.

signed. While many firms use computers tocontinue or intensify the assembly line work-ing conditions of the industrial style of workorganization, some others have experimentedwith new forms of organization that reinvestthe jobs of individual workers or teams withmore variety and responsibility.38

Some of these experimental organizationsmake use of what is called a semi-autonomouswork group-a team of workers who are re-sponsible for not only doing the work, but man-aging some aspects of their own work as well.In these cases, work monitoring data may stillbe collected, but is used by the work group asa tool for assessing its own progress.

One well-publicized example is the HOBIS(Hotel Billing Information System) office atTempe, Arizona, in an experiment worked outjointly by AT&T and the CommunicationsWorkers of America (CWA) in 1982.39 Thisoffice of 100 operators was reorganized accord-ing to the autonomous work group principle.It had no first-line supervisors and only onesecond-line supervisor in the role of advisor.Operators assumed the responsibility of super-visors, rotating through administrative duties.

The employees changed the traditional workmonitoring practices. They eliminated individ-ual measurement and remote secret serviceobservation. Average work time (AWT) wasmeasured only for the whole group. Service ob-servation was performed by small groups ofpeers by the old-fashioned “jack-in” method,where the observer sits beside the person be-ing monitored, listens to a few calls and thendiscusses the results with the employee.

It was generally agreed by CWA and AT&Tthat the Tempe experiment was a success: to-tal office AWT was equal to or better than thatof traditionally supervised HOBIS offices;

MU s Con=ess, Office of Technology Assessment, A uLoma-. ,tion of America Offi”ces, OTA-CIT-287 ( Washington, DC: U .S.Government Printing Office, December 1985), ch. 4.

]9Thi~ description is based On Ronnie Straw and GregoryNichlas, “Office Automation and Autonomy: .4 Comparison ofChoices, ” Communications Workers of America, 1985; AlanWestin, “privacy and Quality of Work Life Issues in EmployeeMonitoring, ” contractor report prepared for OTA, 1986; ThomasTaylor, Mountain Bell Telephone, telephone interview, Apr. 16,1986.


there were fewer customer complaints; the em-ployee grievance rate was lower and absentee-ism was lower. In addition, there were consid-erable savings in management salaries, andsome of the money was spent on training foremployees. The Tempe office was closedshortly after the AT&T divestiture for reasonscompletely unrelated to the experiment. Otherjoint labor-management experiments in alter-native methods of work organization are be-ing sponsored by CWA, AT&T, and some ofthe local telephone companies. For example,further experiments with semi-autonomouswork groups are being carried out amongAT&T operator groups in Columbus, Ohio, andsouth-central Florida.40

In a financial services firm interviewed byOTA, autonomous work groups were also con-sidered successful. Workers were taken out of“functional” areas and organized into teamsservicing the needs of certain large clients orclient groups. Employees “cross-trained’ oneanother in different jobs so that each could doa variety of work and understand the wholeprocess. The group met together to establishproductivity goals. Although this firm main-tained a more traditional management struc-ture within each group and still applied indi-vidual measurement to some jobs, officials andemployees believed that the reorganization,teamwork, and greater diversity of work greatlyimproved both productivity and quality ofworking life.41

Supervision and Evaluation

A few researchers have attempted to com-pare how perceptions of closeness of supervi-sion, emphasis on performance measures, andjob satisfaction differ in monitored and non-monitored workers. This is a difficult task be-cause so many other cultural, job design, andenvironmental factors can overshadow the ef-fect of electronic monitoring.

40Communications Workers of America and AT&T Commu-nications, The Emergence of Second Generation Qud”ty of WorkLife Models in AT&T Communications: A Pilot Study, Febru-~y 1986.

“OTA interviews, November 1985.

One researcher who studied data-entry oper-ators, claims processors, and data collectors(telephone interviewers and collection agents)and their supervisors, found no significant pat-tern of differences between the monitored andnonmonitored sites.42 The differences she didnotice were between unionized and non-union-ized locations. The workers in unionized loca-tions were better informed about VDT healthissues and more willing to ask questions andstate opinions during informal workshops heldafter their survey forms were complete. Shefound, however, their concerns encompasseda variety of VDT health issues, including vi-sion problems, workstation design, and repro-ductive hazards; monitoring did not emergeas the major focus of concern.

Another study found that in both monitoredand nonmonitored sites, roughly half the work-ers (47.8 and 46.3 percent respectively) ex-pressed satisfaction with the evaluation proc-ess.43 Among the monitored workers, 17.3percent were not satisfied and 34.7 percentwere neutral. Among the nonmonitored work-ers, 28 percent were not satisfied and 25 per-cent were neutral (see table 5). In reviewingthe supplementary comments made by inter-view subjects, the authors found clear differ-ences in the causes of dissatisfaction. At themonitored sites, nearly all dissatisfaction wasdirected at the electronic monitoring system;at the nonmonitored sites it was directed ata number of causes, including supervisors, lackof standards, unfair evaluations, and the like.

This study also found that workers at moni-tored sites tended to believe that their evalua-tions overemphasized quantity and under-emphasized quality, tended to see their re-wards as closely tied to their evaluations, andthought that level of supervision was too close.The majority of workers in both groups feltthey had little participation in workplace de-

42Elaine J. Eisenman, “Employee Perceptions and Supervi-sory Behaviors in Clerical VDT Work Performed on SystemsThat Allow Electronic Monitoring, prepared for EducationalFund For Individual Rights and submitted as a contract re-port to OTA, 1986.

43R. H. Irving, C.A. Higgins, and F.R. Safayeri, “Computer-ized Performance Monitoring Systems: Use and Abuse, Com-munications of the ACM, August 1986.

Ch. 2—Using Computers To Monitor Office Work ● 4 9

Table 5.— Comparison of Monitored and Nonmonitored Workers

Dimension Monitored

Satisfaction:Not satisfied ., 8Neutral ... 16S a t i s f i e d . , . . . 22

T o t a l 46

Emphasis on performance measures:Quantity

U n d e r e m p h a s i z e d 0A p p r o p r i a t e 19O v e r e m p h a s i z e d 31

Total . . . 50

Quality:Underemphasized . . . . . . 22A p p r o p r i a t e 19Overemphasized . 5

Total 50

Characteristics of feedback:Amount of feedback:

Not enough . 21E n o u g h . . 22T o o m u c h . , 6

Total . 49

Usefulness of feedbackN o t u s e f u l 8Adequate 13Useful . . 26

Total ., 47

Importance of evaluation for rewards:Not Important . . ., 2M a r g i n a l i m p o r t a n c e 3I m p o r t a n t 41

T o t a l 46

Closeness of supervision:Not close ., . ., . . . 16Acceptable . 10T o o C l o s e 23

Total ., 49Participation in evaluation process:

Low participation ... . . . . . . 30Average ... ... . . . . . . . . . . 8High participation ., . . . . 8

Total . . . . . . . . . . . . . . . . 46

Percent Not monitored Percent

173347478

38062.0

44.03808 3

428448124

17027655,3

4.36 5

891

326204469

65,2173173

23213882

16451980

1545

174

4141

991

132058

91

188

55

81

383617

91

501119

80

28025.646.3

20.256.223,7

20.256.2

1,4

45,045.0

9.8

14221.9637

2229.8

67.9

41739.518.6

62.513,7237

Total

313760

128

166450

130

3764

6

124

626315

140

213384

138

201196

127

544640

140

801927

126

Chi-Sq

2.25

2379

15,3

0.20

0.93

8.08

1314

0.84

(df) (p)

2

2

2

2

2

2

2

2

p O 3247

p 0001

p 00005

p O 9048

p O 6281

p --00176

p 00014

p 0.6570

cisions. This study found little evidence thatworkers opposed computer monitoring in prin-ciple; their chief problems were not with thetechnology itself but with the way it was usedby management.

Personal Computer Monitoring

Most electronically monitored work is per-formed on workstations attached to main-

frames or minicomputers. Yet personal com-puter (PC) use is growing rapidly, especiallyamong professional and managerial workers.OTA did not find examples of production mon-itoring of workers using PCs, but there wasconsiderable interest and controversy overprivacy of an employee’s PC files and the rightof employers to inspect them.

Three primary areas of employer interest inPC monitoring have already surfaced:


1. abuse of PCs (using company resourcesfor personal purposes),

2. confidentiality (security breaches), and3. violation of legal/regulatory duties in use

of client or employee data.

Of the 10 firms interviewed for OTA regard-ing PC monitoring, none were doing any in-spections or searches of PC-user disks or files.All of them, however, said they felt they hadthe legal right to do so, and would not hesi-tate to do so if a specific rule violation or com-promise were suspected. A typical commentby one information system director was:

We have issued a policy guide for privacyand security compliance in PC use, and havestated that the company reserves the right toinspect all PC files and materials bought bythe company and used here for our businesspurposes. But we haven’t felt it wise or neces-sary to swoop down on people and demand tosee what they have on their disks.

On the other hand, representatives of all 10firms said their organizations audited trans-actions done by PCs interacting with data-bases on other computers. These were part ofthe regular, user-password-based security pro-cedures of mainframe/database management.The monitoring consisted of: 1) following upon any unusual use patterns indicated in regu-lar audit-trail records; or 2) ad hoc inspectionsof audit records to identify use levels and pat-terns. End users are informed (in all the orga-nizations) that such auditing is conducted.

As for assuring end-user compliance with le-gal and regulatory rules governing an organi-zation’s handling of client or employee personaldata, 4 of the 10 organizations reported theyhad issued written policies to PC end-usersrestating such requirements (e.g., Fair CreditReporting Act; State employee access to per-sonnel records laws; confidentiality of medi-cal information laws; etc.). However, none ofthe 10 firms reported having done surprise orannounced inspections of disks or other desk-held file materials. Interviews with officials ofthe Inspector General’s office of General Serv-ices Administration (GSA) indicate that inspec-

tors from GSA and other agencies’ InspectorGenerals have inspected PC disks of govern-ment employees. These audits have been doneboth to determine that computers are beingused for official purposes and to ensure thatconfidential information is being properly usedand properly protected.44

Monitoring and Stress45

One area in which electronic monitoring mayhave far-reaching implications is in the areaof health effects. A number of authors havenoted the likelihood of a link between electronicmonitoring and physical and psychologicalstress. Many of the published stories of oppres-sive, heavily monitored workplaces cite theoverwhelming fear, anxiety, hatred, and lossof self-image that workers suffer. Manyauthors have stated that there must be a linkbetween monitoring, stress and health prob-lems, absenteeism, and lowered produc-tivity.46

Stress is now recognized as a major occupa-tional health problem. Stress-related symp-toms have been estimated to cost U.S. indus-try $50 to $75 billion per year in absenteeism,company medical expenses, and lost time. 47

Statistics indicate that claims for worker com-pensation, based on disability due to gradualaccumulation of stress, have been growing rap-idly during the 1980s.48 For workers under.~r~ew with Don Sheridm, Office of Inspector General

General Services Administration, Dec. 16, 1986.‘sThis section draws heavily from Michael J. Smith, Pascale

Carayon, and Kathleen Miezio, “Motivational, Behavioral, andPsychological Implications of Electronic Monitoring of WorkerPerformance, ” contractor report prepared for OTA, 1986.

“See, for example, Tim Healy and Peter Marshall, “BigBusiness is Watching You, ” In These Times, Feb. 26-Mar. 11,1986; Arlene Hershman, “Corporate Big Brother is WatchingYou, ” Dun Business Month, January 1984; Robert Howard,Brave iVew Workplace (New York, NY: Elizabeth Sifton Books-Viking, 1985); Peter Perl, “Monitoring by Computers SparksEmployee Concerns, ” The Washington Post, Sept. 2, 1984; PeterPerl, “Watching the Workplace: High Tech Methods BoostProductivity, But at a Cost, ” The Washington Post, Sept. 3,1984.

‘TRobert Arndt and Larry Chapman, “Potential Office Haz-zards and Control, ” September 1984, report prepared for OTAproject on Preventing Illness and Injury in the Workplace, p. 30.

~~National Council on Compensation Insurance, “EmotionalStress in the Workplace–New Legal Rights in the Eighties, ”1985.

. . .

age 40, claims related to stress exceeded claimsrelated to other occupational disease in 1985.49

These claims are from workers of all kinds, in-cluding managers and supervisors, who are lesslikely than other workers to file claims forphysical injuries. To the extent that electronicmonitoring is associated with stress, then itmust be viewed as contributing to an impor-tant health hazard.

According to the view most frequently citedin the literature, the presence of stress can beinferred in an individual from a very general-ized physiological response pattern. 50 Symp-toms include increases in adrenaline secretion,the dumping of sugar into the bloodstream,and other related physiological processes.These symptoms can be provoked by a vari-ety of environmental agents and situations,such as drugs, fear, and job ambiguity. Whilethere is nothing wrong with physiologicalarousal per se, it can, if chronic, produce seri-ous degenerative effects due to wear and tearon the body. Thus, stress provides a basis forthe development of various illnesses called“diseases of adaptation, ” since they are nota direct function of the agent or situation thatelicited the response pattern, but a conse-quence of the body’s adaptive reaction.

A 1982 journal article suggests that job fac-tors can create stress and lead to chronic dis-orders. 51 The author states that individualsmay perceive the demands imposed by theenvironment as either stressful or not stress-ful, depending on factors such as prior experi-ence, current emotional status, health status,and genetically predisposing features. If de-mands are perceived to be stressful, then acutebiological and emotional responses occur,which, if they continue to occur with some con-sistency over a prolonged period of time, caneventually lead to disease. Various interven-ing factors, which determine the potential fordisease to develop, include individual coping

“’Ibid.“’llans Selye, The Stress of I.ife (New York, NY: hlcGraw-

Ilill, 1956).11,. I,e\i, “Methodology Considerations in Psychoendocrine

Research, Acta ,Ifedica Scandinatria, 1982, 191, Supplement528, pp. 28-!54,


style, genetic predisposition to disease, andemotional support from others.

Although there has been some research onthe health effects of office automation, therehas been little research attempting to draw adirect link between electronic monitoring andstress. There are theoretical grounds forpostulating a link between monitoring andstress, and the few studies that have been donesuggest that monitoring may be stressful. Un-fortunately, none of these studies have success-fully separated the effects of computer-basedmonitoring from the combined effects of otherstressors.

Although there is no clear scientific valida-tion of a link between electronic monitoringand stress, several surveys have found higherincidence of stress among people in monitoredjobs. One survey that attempted to lookdirectly at stress and health outcomes of workmonitoring was the 1984 National Survey onWomen and Stress, conducted by the 9 to 5National Association of Working Women. Asnoted above, this survey includes one questiondirectly related to monitoring: “Is your workmeasured, monitored, ‘constantly watched’ or‘controlled’ by machine or computer system?’When the health problems experienced bywomen who answered ‘‘yes’ to this questionare compared to those of all respondents, asshown in table 6, they show a consistentlyhigher experience of stress-related illnesses.Respondents whose work was subject to com-puterized monitoring were also more likely torate their jobs as “very stressful. Forty ninepercent of them rated their jobs as very stress-ful, compared to 33 percent of all respondents.Seventy-four percent of the monitored work-ers reported strain, stress, or pressure ‘‘oftenor always” in the previous month, comparedwith an overall rate of 63.5 percent for all re-spondents. 52

A related question in the 9 to 5 survey askedabout production quotas. Almost half (47.4 per-cent) of the women working under production

“:9 to 5 National Association of W’orking Women, ‘‘The 9 to5 National Sur\ey on W’omen and Stress—Office Automation:Addendum, ” 1984, pp. 4-5.

52 ● T h e Electronic Supervisor: New Technology, New Tensions

I1-

‘ o3

Cn—

(n E1

.—

1

.

standards reported that their work was meas-ured or monitored by a computer. For theothers, work is presumably counted by super-visors or by the workers themselves. Womenworking under production standards weremore likely to rate their jobs as very stressful(48.1 percent if always under standards, 41 per-cent if often, 29.2 percent if never). Their ex-perience of stress-related illness, as shown intable 7, was higher than the experience of allrespondent s.53

These results are consistent with other re-search suggesting that monitoring inducespressure to perform. Some managers may feelthat this is a desirable effect, since it implieshigh production. But occupational stress re-search indicates that excessive work pressureis not conducive to good long-term perform-ance and brings about adverse health conse-quences.” In fact, there are a range of stress-ful working conditions that may be related toelectronic monitoring of employee perform-ance. These include heavy workload, especiallyof repetitive or machine-paced tasks; routinizedwork activities; lack of control over timing,speed, and variety of tasks; and social isola-tion, including lack of peer social support, re-duced supervisory support, and fear of job loss.

There are many potential causes of stressin the workplace, and it is not clear from workercompensation claims that work monitoring isa dominant one. However, a review of mentalstress worker compensation claims from the

531bid.“C.L. Cooper ~d J. Marshall, “Occupation~ Sources of

Stress: A Review of the Literature Relating to Coronary HeartDisease and Mental 111 Health, ” Journal of Occupational Psy-chology 49, 1976, pp. 11-28.


State of Oregon shows that a little under one-fifth of the total claims were made by peoplein occupations where monitoring is common.Worker compensation records do not releasethe detailed cause of injury or the detailed jobdescription of the claimant, so it is impossibleto determine if electronic monitoring was ac-tually a factor. Of the 542 cases listed, about102 (18.8 percent) were in occupations whereelectronic monitoring is fairly common. Theseoccupations include clerks (of various kinds),insurance adjustors, bank tellers, telephoneoperators, dispatchers, and retail sales work-ers. The rate of acceptance and denial of claimsis shown in table 8. The acceptance rate forpotentially monitored office occupations wasroughly the same as for all jobs, 34.2 and 35.2percent respectively.

Other studies have found a high incidenceof stress-related illness among workers mostlikely to experience electronic monitoring, eventhough monitoring itself was not examined asa variable. For example, a study by the Na-tional Institute of Occupational Safety andHealth found that secretaries had the second-highest incidence of stress-related illnessamong 22,000 workers. The Framingham heartstudy, released in 1985, found that women cler-ical workers develop coronary heart disease atnearly twice the rate of other women work-ers.” Researchers have commented that thestress-provoking factors in these jobs are rapidwork pacing, including machine pacing, monot-onous or repetitive work, and lack of discre-tionary control.

55Working Women Education Fund, 4’Health Hazards for Of-fice Workers, ” April 1981.

Table 8.—Review of Oregon Worker Compensation Claims Involving Mental StressJanuary 1985 Through September 1986

Number of PercentOccupations claims Accepted accepted Denied

All occupations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 542 191 35.2 351

Possibly monitored occupations:Office a (percent of total 13.4) . . . . . . . . . . . 73 25 34.2 48Retail sales (percent of total 5.3) . . . . . . . . . . 29 11 37.9 18

%lccupations, In order of decreasing frequency, are clerk (39), Insurance adjustor (10), dispatcher (9), admlnlstrat!ve support(5), computer operator (4), data entry (2), bank teller (2), telephone operator (2)

SOURCE Oregon Worker’s Compensation Department, Research and Statistics Sect Ion, “Accepted and Dented Clalms Involving Mental Stress, Oregon, 1/85.9/86 ‘


Time pressure, such as having to meet dead-lines, is another significant factor in stress.Studies have shown increases in stress levelas difficult deadlines draw near.56 The experi-ence of deadline pressure on a constant basis,as might be the case in a fast-paced monitoredjob, may be more damaging than deadline pres-sure experienced on an occasional basis.

Two organizational factors have been shownto be of special significance for increased jobstress and decreased worker health. These are:1) job involvement or participation; and 2) or-ganizational support, as reflected by supervi-sory style, support from managers, and chancesfor career development. Lack of participationin work activities has been demonstrated toresult in increases in negative psychologicalmoods.57 In terms of organizational support,it has been shown that close supervision anda supervisory style characterized by constantnegative performance feedback are related tohigh levels of stress and poorer worker health.58

The implication of these findings is that ex-cessive, impersonal electronic monitoring ofemployee performance that produces close su-pervision and constant negative performancefeedback could promote worker stress.

It has also been demonstrated that workers’feelings of lack of involvement are related tostress and that prolonged stress can be relatedto health complaints.59 Electronic monitoring

56M. Friedman, R.H. Rosenman, and V. Carroll, “Changes inthe Serum Cholesterol and Blood Clotting Time in Men Sub-jected to Cyclic Variation of Occupational Stress, ” Circulation,1958, pp. 852-861.

“B. Margolis, W.M. Kroes, and R. Quinn, “Job Stress: AnUnlisted Occupational Hazard, Journal of Occupational Medi-cine 16, 1974, pp. 654-661. R.D. Caplan, S. Cobb, J.R. P. French,R.V. Harrison, and S.R. Pinneau, Job Demands and WorkerHealth (Washington, DC: U.S. Government Printing Office,1975). M.J. Smith, B.G. Cohen, and L.W. Stammerjohn, “AnInvestigation of Health Complaints and Job Stress in VideoDisplay Operations, ” Human Factors 23, 1981, pp. 387-400.

58R.D. Caplan, S. Cobb, J.R.P. French, R.V. Harrison, andS.R. Pinneau, Job Demands and Worker Health (Washington,DC: U.S. Government Printing Office, 1975). M.J. Smith, B.G.Cohen, and L.W. Stammerjohn, “An Investigation of HealthComplaints and Job Stress in Video Display Operations, ” Hu-man Factors 23, 1981, pp. 387-400.

‘gWorld Health Organization, Psychosocial Factors andHealth: Monitoring the Psychosocial Work Environment andWorkers’ Health (Geneva: 1984). J. Rutenfranz, W. Colquhoun,P. Knauth, and J. Ghata, “Biomedial and Psychosocial Aspectsof Shiftwork, Scandinavian Journal of Work Environment

has a propensity for reducing worker feelingsof job involvement and may in this way in-crease worker distress. The chances to partici-pate and be involved in the job process maybe diminished in work systems that are drivenby employee performance monitoring.

Reduced coworker support can also contrib-ute to stress. Monitored workers in severalstudies, and those interviewed by OTA statedthat, due to their production standards andthe electronic monitoring system, they had noopportunity to interact with coworkers.60

One study of work monitoring in the tele-communication industry suggests that the pos-sible connection between monitoring and job-related stress is through the changed struc-ture of the work. In this study, no direct cor-relation was found between electronic moni-toring and stress-related illness. However, acorrelation was found between monitoring andlow job control which has been found, in otherstudies, to be associated with stress-relatedillness. The conclusion reached by the research-ers is that when jobs are redesigned to facili-tate computerized monitoring of work perform-ance, they are also reshaped in ways thatincrease the degree to which managementdirects both the pace and the method of work.

and Health 3, 1977, pp. 165-182. R.A. Karasek, Jr., “Job Deci-sion Latidute, Job Design, and Coronary Heart Disease, inG. Salvendy and M.J. Smith (eds.), Machine Pacing and Oc-cupational Stress (London: Taylor & Francis, 1981), pp. 45-56.R.D. Caplan, S. Cobb, J.R.P. French: R.V. Harrison, and S.R,Pinneau, Job Demands and Worker Health (Washington, DC:U.S. Government Printing Office, 1975). M.J. Smith, B.G. Co-hen, and L.W. Stammerjohn, ‘‘An Investigation of Health Com-plaints and Job Stress in Video Display Operations, ” HumanFactors 23, 1981, pp. 387-400. B. Garden, “Technology Aliena-tion and Mental Health, ” Acta Sociolo@”ca 19, 1976, pp. 83-94.B. Margolis, W.M. Kroes, and R. Quinn, “Job Stress: An Un-listed Occupational Hazard, ” op. cit. S.G. Haynes and M. Fein-leib, “Women, Work and Coronary Heart Disease: ProspectiveFindings From the Framingham Heart Study, American Jour-nal of Public Health 70, 1980, pp. 133-141. M.J. Colligan, J.J.Smith, and J.J. Hurrell, “Occupational Incidents Rates of Men-tal Health Disorders, ” Journal of Human Stress 3, 1977, pp.34-39.

bOFor example see R.H. Irving, C.A. Higgins, and F.R.Safayeri, “Computerized Performance Monitoring Systems: Useand Abuse, ” Communications of the ACM, August 1986. In-terviews in Michael J. Smith, Pascale Carayon, and KathleenMiegio, “Motivational, Behavioral and Psychological Implica-tions of Electronic Monitoring of Worker Performance,, ” con-tract report prepared for OTA, July 1986.

Ch. 2—Using Computers To Monitor Office Work ● 5 5—

This lack of personal control, in turn, placesworkers at significantly greater risk of illhealth.”

Computer Pacing and Stress

Machine pacing is different from computermonitoring. The work of a directory assistanceoperator offers an example of a “low controljob, ’ one that is both paced and monitored bycomputer. A computer-controlled distributionsystem passes a call to an operator. He or shegreets the customer, hears the name to belooked up, and keys it into a computer termi-nal. Once the proper information appears onthe screen, the operator presses a key to re-lease the call. A voice synthesizer actuallyreads the telephone number to the customer.Once the call is released, the distribution sys-tem presents another call to the operator.

The job is monitored, in that records are kepton the operator’s performance within eachcall-the time to respond to the call, locate theproper information, and release the call (sum-marized as average work time or AWT). Inaddition, the job is also machine paced in thatthe cycle time between calls is controlled bythe computer, not by the operator. Work pres-sure increases if that cycle time is very short.

It has been noted that new technology hasturned the job of directory assistance opera-tor into a literally thankless task. Not only isthe pace hectic, but because the operator re-leases the call before the customer receives theneeded information, the operator never hearscustomers say “thank you. ” Job design fac-tors, along with the fast pace, probably greatlycontribute to stress in this job.

Machine pacing has been implicated as a sig-nificant factor in ill-health among factory work-ers. Computers—which can operate at highspeeds on a continuous basis—have increasedthe pacing impact on office workers. Recentresearch suggests that pacing produced bycomputer-driven video display systems may

‘)’ te\en P. Vallas and Jt’illiam Jr. Calabro, “OccupationalConditions and F$rorker Health in the Communications Indus-try, ” New York Institute of Technology, Human Resources De-velopment Center, no date.

have an even greater stress effect than tradi-tional factory pacing.62

Feedback and Motivation’ ]

Perhaps the best use of information aboutan employee’s performance is to give it backto the employee. One advantage that electronicmeasurement can offer to workers is accurateand timely information about their own per-formance. Studies of feedback, whether relatedto simple sensory feedback or to higher levelsof feedback related to knowledge of results, allindicate that people want to know about theirperformance and will seek out such knowledgewhen it is absent.64

Immediate sensory feedback helps employ-ees to exert better control over skilled actionsand to correct errors.65 For example, the feelof the keyboard and the display of the charac-ters on the video screen help a data entry orword processing operator to know that datais being keyed properly. This type of feedbackis continuous throughout the task.

A higher level of feedback, knowledge of re-sults, occurs when a task is completed andevaluated against some external standard, andthe results are fed back to the employee; “Youhave produced 10 percent over the productiongoal today, “ or “Your output had 2 percenterrors. This kind of feedback provides direc-tion to the worker about future output.

Feedback about one’s own activities can be

‘2M.J. Smith, B.G. Cohen, and I..W. Stammerjohn, “An In-vestigation of Health Complaints and Job Stress in Video Dis-play Operations, ” Human Factors 23, 1981, pp. 387-400. A.Cakir, H. Reuter, L. Von Schmude, and A. Armbruster, inves-tigations of the Accommodations of Human Psychic and Ph~w-ica) Functions to Data Displa~’ Screens in the Workplace (I?er-lin: Institute fur Arbeitswissenscharft der TechnicianUniversitat Berlin, 1978).

‘ ‘This section draws heavily from Michael J. Smith, PascaleCarayon, and Kathleen Miegio, “Motivational, Behavioral andpsychological Implications of Electronic Monitoring of J4’orkerPe~formance,, ” contract report prepared for OTA, July 1986.

‘tS.J. Ashford and L.L. Cummings, “Feedback as an ]ndi-\’idual Resource: Personal Strategies of Creating Information,Organizational Behavior and Human Performance 32, ]983, pp.370-398.

““K.U. Smith and M.R. Smith, Cybernetics Principles ofLearning and Education Design (New York, NY: HoIt, Rine-hart & Winston, 1966).


a very powerful motivator and has been foundto have a strong influence on productivity, andin some cases on job satisfaction as well. Peo-ple like to do a good job, but without informa-tion they often don’t know whether they aredoing one or not.

Because computer technology is adept atgathering and correlating information, it canbe very useful in giving timely feedback toworkers in a useful form. For example, officesystems can be designed to give performanceinformation to workers as well as to supervi-sors. In some firms, for example, customerservice representatives can get private accessto their own recent performance by keying theproper code into their workstations.66 Any in-formation that is available to their supervisoris available to the individual workers, as wellas comparisons with the group average andthe standards.

A recent study reviewed 42 case historiesand found that without exception people per-formed better when they were given some ob-jective, quantitative feedback about their ownperformance or output. This study did not fo-cus on electronic work measurement, butrather on both manual and electronic methodsin a variety of work settings, including banks,payroll offices, reservation offices, manufac-turing facilities, and health care facilities. Theform in which feedback was given also variedfrom one setting to another. The three meth-ods most commonly used were private individ-ual feedback, public individual feedback, andpublic group feedback. In some cases, objec-tive feedback was combined with other inter-ventions such as praise, public recognition, oradditional training; but positive results werealso noted where feedback alone wasprovided. 67

Feedback may serve as both a motivator andan instructional device. When people receivewhat they perceive as objective feedback, theycan compare what they actually did to: 1) what

b61nterviews at American Express Southern Regional Oper-ations Center.

‘TRichard E. Kopelman, Mana~”ng Productivity in Organi-zations: A Practical, People-Oriented Perspective (New York,NY: McGraw Hill, 1986), pp. 163-187.

they thought they did and 2) what they areexpected to do. In some cases it may correctmisconceptions or inaccurate perceptions aboutwhat they are doing. In one example, airlinereservation clerks were provided with profilesof their verbal behaviors based on samplerecordings of their telephone conversationswith customers. One clerk commented on see-ing the feedback:

When asked previously whether I used thecustomer’s name I would have said-andbelieved–”Of course, we were trained to dothat. ” I was really surprised when I saw ob-jective evidence on how little I was actuallydoing it.68

As a result of the feedback, use of the custom-er’s name by the clerks rose by 87.5 percent,while the clerks’ interrupting of customers (ahabit the employer wished to discourage)nearly disappeared.

Feedback is an effective modifier of behaviorif it is seen as a valued commodity by the re-cipient and if it is timely. It takes on value tothe individual when it is effective (relevant, un-derstandable, accurate, useful) and when itcomes from a trusted or highly regardedsource. Although feedback need not be imme-diate or continuous, it should be given fre-quently. The longer the delay, the less effec-tive it is in affecting performance.69 Anumber of the workers interviewed for OTAexpressed a desire for more frequent feedbackabout their work. They also thought that feed-back information from the electronic monitor-ing system could be better designed to helpthem gain more control of their work.

If employees perceive that rewards and/orpunishments could ensue from an evaluationof their performance, they are especially inter-ested in feedback. They want to understandthe basis of rewards and punishments, andfeedback helps to resolve feelings of ambiguityor uncertainty.

sBIbid., p. 176, citing Stephen A. AIlen, “Aer Lingus-Irish(B)” case #9-477-640 (Bostmn, MA: Intercollegiate Case Clearing-house, 1976), p. 7.

69M.J. s~th, B,G. Cohen, and L.W. Stammerjohn, “An In-

vestigation of Health Complaints and Job Stress in Video Dis-play Operations, ” Efurnan Factors 23, 1981, pp. 387-400.


At the same time, feedback of performancecan create feelings of anxiety, frustration, andlowered self-esteem. Since feedback fulfills anerror-correction function as well as a perfor-mance-appraisal function, it can indicate to em-ployees that they are not doing their job aswell as they would like to, or as well as theemployer would like them to. This can createstress, even when it resolves the stress asso-ciated with uncertainty and ambiguity. Inshort, feedback is a two-edged sword insofaras stress is concerned.

Feedback is necessary to groups as well asindividuals, and computer monitoring systemscan also give workers immediate informationabout the work environment letting them knowhow their work group is doing right now, andhow they can best contribute.

To take telephone customer service again asan example, some offices have clearly visibledisplays on the wall that show the number ofincoming telephone calls waiting to be an-swered and the age, in seconds, of the oldestcall. Such displays could be used as weaponsof callous management to keep constant pres-sure on agents of understaffed offices. How-ever, in a properly staffed office, where peaksof incoming calls occur for a few minutes ata time, a few times a day, status displays be-come tools in the hands of the work group. Peo-

ple know how to pace their work. When thedisplay shows all zeros, agents feel freer to takea little extra time with a difficult caller, to catchup on paper work, or to take a break. Whenmany calls are backed up, they can make anextra effort to finish a call quickly, or perhapsto defer a break for a few minutes rather thanabandon their colleagues in a crunch.

Team spirit and friendly competition can bepowerful motivators, and both employers andemployees can benefit if they are not abused.However, workers can also perceive employers’use of feedback and social pressure to be un-fair and manipulative. A Pacific Western Air-lines (PWA) productivity campaign drew unionprotests when company posters urged reser-vation clerks to:

Compare yourself with your friends. Com-pare yourself with ones who aren’t yourfriends. Are you pulling your weight at the of-fice? When the monthly statistics are pub-lished, ensure you’re not dragging down yourteam and your office.

The union newsletter charged PWA with set-ting workers against each other, and called thecampaign a “new low in . . . degradation.’’70

70 Lawrence Archer, “I Saw What You Did and I Know W’hoYou Are, ” Canadian Business, November 19S5, p. H 1.

THE FUTURE OF WORK MONITORING

The OTA report on Automation of America'sOffices pointed out some trends in the growthof computer-based office automation equip-ment that have implications for the future ofwork monitoring.

One trend was the inevitable movementtoward direct machine-to-machine communi-cation. Increasingly, data will be captured inmachine-readable form at the point of origin,customers will enter their own data (as withautomatic teller machines) information will berecorded using optical scanning and voice rec-ognition, and different computer systems will

talk directly to each other, thus reducing theneed to keyboard data.

Another trend was the growth in the intro-duction and use of office automation equip-ment and its rapid adoption by all sectors ofthe economy. It is estimated that by the year1990 there will be one computer terminal forevery three workers in the United States; bythe year 2000, terminals may be as commonin offices as telephones.

Both of these trends suggest possible changesin the population of workers that will be af-


fected by computer-based work-monitoringtechnology. For example, the prime exampleof the monitored job today is that of the data-entry operator, but over the next 20 years thegrowth rate of data-entry workers is expectedto slow or perhaps decline. Those that remainwill probably still be monitored, but they willbe a smaller proportion of the office work force.

The other trend—towards wider use of com-puter-based office equipment–suggests thatmore jobs will be at least partly automated ordependent on the use of a computer. As a re-sult, more types of jobs will be possible candi-dates for electronic monitoring. Although thecharacteristics of monitored jobs listed at thebeginning of the chapter (repetitive tasks, highvolume of work, low training requirement, hightolerance for turnover, ample labor supply) de-scribe ideal conditions for monitoring, they arenot absolutes. It is already possible to applyelectronic monitoring to some highly skilledprofessional and management positions. Sometimes monitoring has not worked well in high-level positions. Employee resistance may havecaused management to back down on imple-mentation plans, or, as in the case of bank loanofficers interviewed by Westin, employees mayhave found ways to “game” the system byfeeding it false information.71 The costs andimportance of employee resistance can change

“Alan Westin, “Privacy and Quality of Life Issues, ” AlanWestin, “Privacy and Quality of Work Life Issues in EmployeeMonitoring, ” contractor report prepared for OTA, 1986.

over time, however. If at some future time man-agement determines that the benefits to begenerated from monitoring a particular job cat-egory will outweigh possible costs in higherturnover, monitoring systems are likely be in-troduced. And while professionals may be ableto defeat their current monitoring system, asystem that automatically collects correct in-formation could be designed if their employerever decides it is worth the cost.

The growing use of computer-based manage-ment information systems also means thatmore managers will be subject to closer moni-toring, simply because more of their day-to-day decisions will be revealed to superiorsthrough the computer system, rather thanwaiting for monthly or quarterly reports.

If there is a growth in computer monitoring,or a spread to other types of work, it does notnecessarily mean a devaluation of office work.Computer-based monitoring can offer advan-tages to employees, for example, improvedfeedback and better control of their own work.Professional and managerial workers may beable to use their bargaining power with em-ployers to participate in decisions about theredesign of their jobs or the implementationof work measurement and monitoring, as toassure fair use of monitoring. As with otherexamples of technology in the workplace, manynontechnological factors, including manage-ment and employee attitudes, corporate cul-ture and relative power relationships, will gov-ern how the technology is used.

Documents

Chapter 2 Using Computers To Monitor Office Work