‘SECRETARIAL JOB ENRICHMENT IN A SCANLON PLAN ORGANIZATIONALCLEMATE Thesis for the Degree'of M. A. MICHIGAN STATE UNIVERSITY RODNEY L. LOWMAN 1-9 7.5 "‘ “WW\\|\\u\|\1\\g\1\1}u\u§m\ my I f 93 rn n l 'v:. ”:7; .- Lu ,. ' Q " 3"“an y- " '~*Vé—Jen u 12%.":isors +‘*~~~ h-» 3'0’ . ”n .. :DQJEQH 130, 'P‘ 1 I. n‘ n \ C‘Ii Pa 3 n - b ' L:viq\,.c. In "v" . * ‘- ...C. Cage } I , F _ ——- — — —" —— ABSTRACT SECRETARIAL JOB ENRICHMENT IN A SCANLON PLAN ORGANIZATIONAL CLIMATE By Rodney L. Lowman A field study was conducted to test the effects of a job enrichment training session held for secretaries and their , supervisors from western Michigan Scanlon Plan organizations. The Solomon Four Group design was used for both supervisory and secretarial groups to assess the effects of training. As hypothesized, the job enrichment sessions, which attempted to inculcate higher level job responsibilities for the secretaries, resulted in experimental groups' achieving a higher enriched job duty composite, but, contrary to expec— tations, in no statistically significant differences between trained and untrained secretarial groups in terms of the number of enriched job duties assumed. The hypotheses were also supported that the training would produce no significant differences for non-enriched job duties, or for the relatively stable organizational dependent variables of job satisfaction, motivation, perceived work effectiveness, and perceived leadership of the supervisor. Implications of the results for job enrichment and u ’49-. '0 w‘“ (‘1 ‘ '5 9.597 4 ‘ ' "eye" 5"] 9 1 ' c ‘V‘n V 1‘.“ «Lu ‘ . q I! .d“:';fie 11;; 1.. gusty-5 \A‘ :‘rezeric . Rodney L. Lowman for training are discussed, as are the limitations of the study. Approved by Thesis Committee: Dr. Frank L. Schmidt, Chairman ‘3 Dr. Eugene Jacobson Dr. Frederic R. Wickert RTHV‘Y" ' D' 1 N" dVLI-U‘o SECRETARIAL JOB ENRICHMENT IN'A SCANLON PLAN ORGANIZATIONAL CLIMATE by Rodney L. Lowman A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psychology ’1 975 o's this -vv-‘v-fi ... of the < " .fl |"-~ V‘l . U.’ L'JEE-ce ACKNOWLEDGMENTS My sincere thanks to Dr. Frank Schmidt for his ex- tensive help with the numerous statistical and design prdblems this research raised, and for his exemplary mod— eling of the characteristics of an effective researcher; to Dr. Eugene Jacobson, for his contributions to the experi— mental design and literature review, and whose many discussions with me regarding this research proved excep- tionally valuable; and to Dr. Fred Wickert for his helpful assistance with reviewing the training literature, with the paper's style, and for his enthusiastic encouragement throughout the research. Special thanks go to Dr. Carl Frost for facilitating access to Scanlon Plan organizations and for his indefati- gable efforts in my behalf throughout and beyond this study; to Anne Mbntgomery and Kathye Kubica for their cooperation with the research, and to Sue weesner for her unfailing typing and clerical support. ii _,. Vol .... {a II- ...do U. Q.— Aqvvq ‘-fi" .‘;'d {J And-U1“ O LWL‘JE Aug n—g- Aapmv"fi . 7 \I‘ ‘\ ‘1, ".1. “‘5‘. U ~r~nfiqfl .- _‘ -e-LJU 0.00.. 437 ‘3 “‘H‘ "*:u‘ ‘N = N :e‘.‘ ‘ (- say“. r. .. n is “19:3“ v'“ H \u v v- . . M*Pe:19: \J \:?A‘~ ‘“~¢~n n'..‘ “~44- 1' I . , ‘ ~“d‘L D L ‘. ‘h W4, I: Q 3 i"-.\~ \\|5V~ Q TABLE OF CONTENTS Page LIST OF TABLES . ........................ '. .............. iv LIST OF FIGURES .. ..................................... vi INTRODUCTION ... ....................................... 1 REVIEW OF THE LITERATURE . ............................. 8 STUDY SETTING AND HYPOTHESES .......................... 4} METHODS ............................................... SO The Experimental Design ... .................... .. SO Subjects .. ...................................... 51 The Instrument . ................................. 55 Data Analysis ... .............. . ........... . ..... 72 RESULTS ....... ...... . .......... . ............... . ..... . 75 Pretest Analyses .... .......................... .. 75 Reliability of the Measures ......... ............ 80 Relations Among the Dependent Variables ......... 90 Management Support Group Hypotheses .. .......... . 97 Supervisor Hypotheses .... ...................... . 104 DISCUSSION ................ ..... . ..... ... .............. 111 APPENDIX A, Specimen Pretest, Secretaries ...... ....... 118 .APPENDIX B, Specimen Pretest, Supervisors . ........... . 126 APPENDIX C, Specimen Posttest, Supervisors .. .......... 154 AITENDIX’D, Specimen Posttest, Secretaries ...... ..... . 144 APPENDIX E, Cover Letters and Followup Letter ........ 154 REFERENCES ...... ..... ... ....... ........... ........... . 165 iii — I AI 4" r“ .LS-Uo-.l‘ In ? ‘ nt-a v 36343 - Wr‘flr; 'G‘cho» O f .- "A ..C 2 '1 -Vy-rk. - a . I'~rv‘ ! 1' I (T) I) ( 9 (n F. ‘1“ ‘4‘ 7 . V‘N ‘, ‘“~Aef_ 'U'““h U.&H - - , ‘ LBS"; C 2 (4 fl ' . 3 (D ’ J , (D $ )1 3 J £A, CD t4 (D l v I I . .“‘ a“ on V ‘- I4-” 5.! (A L :‘v I“ vfifihc‘ v“ V‘ CA UH“? vv‘ Table 10 11 12 13 14 15 LIST OF TABLES Page Summary of Returned Questionnaires . ............ 54—55 .3 Tests for Pretest Differences on All Variables, .Matched Samples, Secretaries ......... ........ 76 .E Tests for Pretest Differences on All Variables, Matched Samples, Supervisors ... ............ .. 77 t_Tests for Pretest Differences on All Variables, Expanded Samples, Secretaries ................ 78 §_Tests for Pretest Differences on All Variables, Expanded Samples, Supervisors ................ 79 Correlations and Reliabilities for the Secretarial Job Duty Scales .. ................ 81 Correlations and Reliabilities for the Supervisors' Job Duty Scales .. .............. . 82 Reliability Estimates for the Job Motivation Ind-ex O......0...0.00.0.......OOOOOOOOOOOOOOO. 86 Correlations and Reliability Estimates for the Leadership Dimensions, Secretaries Rating SuperViSOrs ......OOOOOOOOOOOOOOOOOOOOOOOOO... 87 Correlations and Reliability Estimates for the Leadership Dimensions, Supervisors Rating selves .....OOOOOOOOOOO ..... .....OOOOOOIOOOOOO 88 Inter—item Correlations with Coefficient Alphas, Hypothesized Job Efficiency Cluster .......... 91 Correlations Among All Dependent Variables, secretaries coo... ..... ooooooooooooooooooooooo 92-95 Correlations Among All Dependent Variables, superViSOI‘S oooooooooooooooooooo000.00.000.00. 94-95 Univariate Analysis of Variance, Enriched Job Duties Composite, Secretaries ................ 98 Univariate Analysis of Variance, Number of Enriched Job Duties, Secretaries .. ......... .. 101 iv var“ , fix .- l1*_: DA \ ‘0‘.- ~ war J4... MA; . J Table 16 17 18 19 Page Multivariate Analysis of Variance, secretaries ....O............OOOOOOOOOOOOOO... 105 Univariate Analysis of Variance, Enriched Job Duties Composite, Supervisors ............ 105 Univariate Analysis of Variance, Number of Enriched Job Duties Performed by Secretaries as Reported by Supervisors ....... .......... .. 107 Multivariate Analysis of Variance, Supervisors.. 110 '— LIST OF FIGURES Figure Page 1 Graphical Representation of the Solomon Four-Group DBSign o. ooooooooooo coo oooooo 0.00000 50 vi , Tze a- w!" v: "' :L‘rll .. '13}: a" y‘.a0 1 "no.- +3? - A; \ Hévuc '- .,; , :4. 0f LEE: | '0; n 1 mmsmall , I “- F55“, -"-'""A: C ' h m “ «NV 0‘ ~ ‘- . In L975} 1“,. “I“ -Q ~;- I u" . ‘i R . ‘uha‘ . .Mv.~: Jone “ e _‘ ‘F | I‘: In: 1 ‘ N .L‘*e of ; W s . u 'L .‘a rh‘ ‘ H ~ INTRODUCTION The allegedly doleful plight of the worker has long been a pOpular theme among the literati, from Coleridge ("Work without hOpe draws nectar in a sieve . . .") to Camus ("Without work all life goes rotten. But when work is soulless, life stifles and dies.") It creates both good copy and lively discussions to lament the horrific state of drudgery in which workers are said to toil. The alienation of the contemporary worker is hardly a new or original concept, however. The dissatisfaction and ennui of the workers, assigned as they supposedly are to intrinsically worthless taSks, has been a common theme per— vading the writings of many social critics and philosophers for some time. The "blue collar blues" and "white collar woes" said to characterize the modern American worker, are not fundamentally different, in concept, from the anomie of Durkheim or the alienation of Marx and Fromm. Indeed, it was Karl Marx's early (1840's) and persistent writings on the alienation of the industrial worker that have been the cornerstone of much of the recent ideas on alienation (Blauner, 1964), although the concept of alienation itself, outside of its original pathological meaning, was introduced by the philOSOpher Hegel, not by Marx (Branden, 1971). It is a small step, conceptually, from Marx's contention that a worker cannot achieve self—fulfillment without control ':'- “'5 QI‘OC‘JC ,‘u- ' o l . :;:::e of ti _wu 9A .- '13133) 30 3 77.16 NONI; I y A. ‘ ”fit "" 0 0| 0 Vahh’d- ‘- 4. v 2.3:..31y in: .v 0 More re .: 4+ :? .. ...e wor? .n'n‘. :5 ~24 “4‘ ,. 1 savt‘e war WT}. . ‘48 m ‘~“h '\ rs ad WC) “. ~ 6, .‘E‘: over the product he is producing, to Fromm's decrying the "anonymity of the social forces . . . inherent in the structure of the capitalistic mode of production," (Fromm, 1955, 158) to Blauner's assertion (1964) that freedom is at its lowest in the assembly line industries of the 20th century, to the recent HEW Secretary's Task Force's statement that " . . . employment in meaningless work is creating an in— creasingly intolerable situation" (Work In America, 1975, 186). More recently, the bandwagon headed for improving the lot of the working person has been joined by psychologists, sanetimes taking a more empirical approach. Workers, said the human relations advocates of the 1950's, are happiest—- and most productive--when they are both treated as human beings and assigned to jobs that are meaningful (Ash, 1975). Heererg and his followers, in the 1960's, contended (and do contend) that a two-factor phenomenon is Operative in the employees' work: once an individual's basic "hygiene" needs are met (salary, working conditions, etc.) motivators (self-esteem, achievement, etc.) become important (Heererg, 1966). ‘Worker dissatisfaction is both understandable and unavoidable When.higher level "motivators" are deficient. The image that emerges from much of the contemporary musings on the subject of worker dissatisfaction suggests widespread worker discontent. The Work In America study, for example, states that " . . . significant numbers of American workers are dissatisfied with the quality of their :nxtzvity C' -q-vnfi “(‘+ C ".53- LJ- . ..fiuflos, and I ..- .t“ u i». Self .n '«s § 4‘ *\ eyCI :‘s‘ ‘H D. IN". th Q'"\ '3i:¢. ...‘ed Q“ Nu: .m "firth working lives. Dull, repetitive, seemingly meaningless tasks, offering little challenge or autonomy, are causing discontent among workers at all occupational levels . . . the productivity of the worker is low--as measured by absenteeism, turnover rates, wildcat strikes, sabotage, poor quality products, and a reluctance by workers to commit themselves to their work tasks," (Work In America, 1975, xi-xvi). Similarly, Herrick (1972) reporting on a 1500 subject study commissioned by the United States Department of Labor Employment Standards Administration to the Survey Research Center of the University of Michigan, reports "The overall survey shows that the chance to do meaningful work and to achieve and grow on the job is of great importance to the average American worker--perhaps even overshadowing financial considerations. It also appears that this chance is sadly lacking in the average job" (Herrick, 1972, 7). But is the American worker all that dissatisfied? Perhaps the most comprehensive look at job satisfaction is provided by the recent Department of Labor monograph (19b Satisfaction: Is There a Trend?, 1974). This study demon- strates that, viewing the overall research picture, the 1958-1973 time period saw very little change in the percentage of selfereported "satisfied" workers. And, considering the fact that the percentage of workers who consider themselves satisfied stays relatively constant at about the 90th percentile, the phenomenon of widespread worker dissatis— faction is probably more illusionary than real. 4 Whether real or illusionary, the problem of worker dissatisfaction has not been without its suggested panaceas. One of the most frequently touted of the current nostrums is "job enrichment," an intuitively appealing concept that is anything but pleasant to define. As Judson Gooding remarks, "Job enrichment is a diffuse, open ended kind of concept. It is more an attitude or a strategy than it is a definable entity. In fact there is no one term for it that is accepted by all the experts" (Gooding, 1972a, 24). What, then, is job enrichment? Essentially, it is an attempt to bring responsibility down to the lowest level possible, to allow individual workers a great deal of initia- tive and responsibility in their jobs, to group tasks into meaningful units, in short, to make work meaningful. It is to be distinguished from "job enlargement" (though the two terms are increasingly used interchangeably) which refers to early attempts to "improve" the boredom of the worker's job by adding a greater number of the same taSks to his job, rather than adding higher level responsibilities. Again, from Gooding: "Central [to job enrichment] is the basic idea of giving the worker more of a say about what he or she is doing, including more responsibility for establishing procedures, more responsibility for setting goals, and more responsibility for the excellence of the completed product" (Gooding, 1972a, 24). The intuitive appeal of job enrichment in America is obvious. In a culture which finds it desirable to espouse A85 A ::::ies and aafechniqxe tintinty e I 1 ‘ H W Hf“? ‘ wk -.. 131‘"? Sex ' '0: ":19; 955‘?“ I" .-"V uv%-€ ‘ I H. ‘9 he.“ a] "- ’7‘ ?.r\.- . ..L: “41+; 8 5:, ‘1 .22. , . L93." Wer \EZ‘ of CH v.’.' . ~ ‘ ":‘Z a .. the success motif, the inherent pleasures of hard work, and the right of all citizens to self-actualization, job enrich— ment strikes a harmonious chord. More recently, European countries and England have begun to turn to job enrichment as a technique to combat seemingly endemic problems of low productivity and worker angst. How widespread is the use of job enrichment in industry? Several surveys have been reported to date in the literature. Reif and Schoderbek (1966) and Schoderbek (1968) report that 80.5 percent of a 210 company sample indicated that they were not using job enlargement. Of the 41 com- panies that responded positively, the three main reasons for use were: reduction of costs, "enriching" the worker's job, and decreasing job specialization. However, when queried as to the number of job enlargement projects in— stalled during the five years preceding the study, the number of projects reported is quite low (only 14 firms even indicated a figure). The respondents were also asked to list advantages and disadvantages of job enlargement, though unfortunately the authors did not use open-ended response format, thus potentially biasing the responses. The three most frequently mentioned advantages were increased job satisfaction, cost reduction, and increased work quality, whereas the three major disadvantages were overcoming resis- tance to change, the fact that some workers were not capable of growing with the job, and increased training time. .. . ‘ «nu. .‘ '5 ‘2 _7‘i —¢—V¢U.',“ -‘ V ‘ ’-v-.“, ‘ 1" .a.‘-’:, L--e a I. :79 T‘D‘W ‘ “rl .- --y ‘Vl.~“l“ 'l ‘ 1 .:"‘"n J. =-~-5 «16 t ...! _ _ . A ~. ...-g: 156k: 1 v‘l ..pa t; “- ---. 448 C ‘2‘: "fl“*-.' . Aerv.‘ L. ‘- l “: - uvsi‘e ‘ V‘. , ‘4 ..‘u ‘ II ‘ ".‘Q‘Y‘I‘ .- . \ .’~_ _ I‘:‘~R‘.j “‘6? ~ V — .\. ‘b SCI Ev '-C-h -‘\\ ‘ U” S ”.1 -V. U 4 .‘.“‘ xv," ' 'L‘ ‘ ~‘~::n£ \Vufi ‘Q Although Reif and Schoderbek conclude, in their original report of the study (1966), that " . . . the advantages far outweigh the disadvantages" (25), this conclusion is not clearly supported by the authors' data, particularly in view of the small number of respondents even using job enrichment. In a similar, more recent study, Reif, Ferrazzi, and Evans (1974) report questionnaire results from a 125 company sample. Forty—eight percent of this sample reported that they did not use job enrichment, and of the 25 percent indicating that they planned to use job enrichment in the future, the authors report that, based on the comments written on the returned forms, these companies seemed cautious re- garding the practice. Only 29 percent of the rather small sample used in the study reported employing job enrichment at the time of the study, with only 4 percent of the respon- dents reporting a formal job enlargement program. The enthusiastic, over-generalizing tone of the previous surveys is lacking in the Reif, Ferrazzi, and Evans study, and they conclude, in part, that probably the majority of firms practicing job enrichment have a rather limited understanding of the concept. A great deal has been written about job enrichment, especially in the last decade, most of it, regrettably, of little scientific value. Typically published are glowing success stories (for example, Rush, 1971, and Butteriss, 1971) with little information provided that would enable a dispassionate reader to draw his own conclusions. Instead, :.. . I I ..I {CHEFS "”5"" c o r. ..-.v‘ :5: Leyc :"‘lfi.r"p,r+- —-V-—wh‘.', (’11 o g . . ... ”.Q‘yv‘.” r“ “'--L“¢6 SUV; 121‘s of mg; one is typically offered proselytizing rodomontade from the already converted, odes of praise, case histories, or "how to" articles, usually with the implicit assumption that job enrichment is the cure for all industry's ills, especially for workers' boredom and alienation. More recently, however, serious researchers have turned their attention to job enrichment, and the results of their more controlled and less idolizing studies have somewhat dampened the pious pronounce- ments of praise of the earlier job enrichment advocates. (‘3 .3 (‘J .9 | J 0.. .. 3‘ d ' ;L35 ;26 E. --'-- Mxm +r~ Q ..1. are.“ U» -o4 g 7‘. ,H,‘ “r. N 1“»...- .....u. u- .-., Apr: , A N“ 7: grecisely '.'II." ‘ i #44» L35!“ .N— rrL ._‘~“ s “Eek-4‘10 . Q 0" A ‘ ‘1 ‘T‘I S ere P'x'é' .' ‘ 'A 3-3 :dvg‘-bav‘c vic~ . ~54 ”PF/K“ . 7‘ vadue‘ ‘b x. h L \'Ovt “‘ a 33"» I Q 21’1“ \ N ~ “C :7)” m“ Ifi‘. ‘. ‘ ‘H\ ‘4 ”h we :.::VQ’N ‘ s“‘ u . IA ‘\:I‘S' " .b .“ ...,“ “‘2“ 5‘3 F V. ,— '§ -. 4" eke REVIEW OF THE LITERATURE Concern with job design has been pursued at least since the delineation of "job specialization" (i.e., breaking jobs down to their simplest components) by Adam Smith in 1776 (Hulin and Blood, 1968). Later, Lillian Gilbreth, Frederick Taylor, and others, promulgated the same fundamental concept of specialization into "scientific management," an attempt to precisely specify the Optimal combination of job com- ponent tasks and methods (Reif and Schoderbek, 1969). Men, being fundamentally interchangeable units, especially in the immigrant-dominated, low intellectual level America of Taylor's era, were of less importance than the "scientific" Specification of job performance techniques. Jobs, not people, were the variables of concern (Gilmer, 1971; Reif and Schoderbek, 1969). Experimentation with job enlargement, the opposite of job specialization, is reported as early as 1944 at the Endicott plant of IBM (walker, 1950). Walker reports that the IBM experiment, which consisted primarily of adding skills and responsibilities to the single Operation worker, including the inspection of the final product, resulted in decreased costs, improved quality, less idle time, "enriched" workers' jobs, and enhanced social relations between the workers and the foremen. Interestingly, despite hosannas for the results, Walker suggests a limited application of 8 “1“" he: 5" _,vrbV U ‘ r:?“‘?‘c“" ‘ " RAM...“ W - . n . n 1' Iowa“ "0"“ , ~ . 2 .. uckvv-h-ZAVL tnénrer .: .-, Ir . ‘Vvl .1 r u ....e \J. :1- - ...”;- A \ '~-:..U' J81 Wake ‘-.‘.Qfi-L-y "1 Of "1‘! ' ‘:, 3" n h. 9" ‘ ‘V‘ .3 2,. ‘ l -~"'\~..1 . U“ .4 \ 1' 'L .5. .‘§. I‘n “ . d‘“ . "'e '4 . ; . ‘\ ~_CP\ I ' “° w' “ _ .. ‘« s, N *3. ‘. u. 5 ": In“. ‘u'yt ‘Lfi “‘4 N . 7‘ ~‘ V. .‘mfi+ I job enlargement, perhaps to a half million.American workers' jobs, and concludes that job enlargement should be avoided unless a number of factors are favorable, including shOp practices and company policies. It was not until the 1950's and later, however, that job designers and management personnel seriously began to experiment with alternatives to specialization in any kind of systematic manner. Early 1950's pOpular magazines reported the then revolutionary ideas aimed at decreasing the monotony of the (typically) blue collar jobs: job enlargement and job rotation (Davis, 1957). Walker and Guest, (1952a) cite one of the first attempts to deal systematically with the problems associated ‘with job specialization. They suggest that mass production technology has developed these characteristics: (a) stan— dardization; (b) interchangeability of parts; (0) orderly progression of the product through the plant; (d) mechanical delivery of parts to work stations at the right time and mechanical removal of the assembled product or Subproduct; (e) minute subdivision of the product; (f) severe limitations on social interaction; and (g) the requirement of only surface mental attention being needed for the worker to accomplish his job. While these characteristics have produced high production levels and low unit costs, the results have also been expensive in terms of enormous social costs, including the fact that, according to Walker and Guest, the average production worker is dissatisfied with his job. 28 2’. inepmln‘ . ‘V‘JV .ule“.“; '1 :.o:JJ-$LC b ‘In' :.;‘a a +h ‘ .1 ‘ -‘Vu n 1 ' :23". v ‘ ~ ANi-U-J .. .. :"‘72‘1CGC ”0.-....- VHO .. 2&3 notes, :gr-zvement 1 . . . inc; 12 which eq Thu)... q “‘13, DU: I 2,;n “32:: A? I ...“ U; I " ‘- .. fl“ 10 One of the first empirical research attempts aimed at determining the effects of job enlargement in a rigorous, scientific manner is reported by Marks (1954). His study deals with the introduction of job expansion in a formerly assembly line production operation that manufactured hospital appliances. In the redesigned job, the workers in the experimental groups controlled the sequence of production steps and provided their own quality control. The results, Marks notes, were quality improvement and employee attitude improvement, and, according to one summary of the research, " . . . increases in productivity above the Group Job Design [in which employees performed the assembly line production tasks, but rotated among the various tasks)" (Davis and Canter, 1956, 279; Davis, 1957). The data provided, however, indicate that the "enriched" job design resulted in an average of 2.5 percent lower output than the control group's production. In addition, the period of the experimental treatment (16 and 27 days in the two experimental groups) was so short as to constitute, at best, a Short term test of the changes attributed to the new production methods. In another study, Rice (1955) suggests that job design on the premise of placing all inter-dependent workers in the same geographical area in the plant accounted for a 15 percent improvement in worker efficiency in a 60 day trial period. This suggests that functionally arranged jobs (i.e., placing all of those workers with the same job duties in the same place) are perhaps less efficient than unit-of-production 11 arranged jobs, which later became a fundamental tenet of job enrichment principles (Ford, 1975). However, the short term nature of Rice's study, and the comparatively low level of improvement, suggests that the results are only suggestive and must be interpreted cautiously. Further empirical research supporting the job enrich— ment/enlargement principles is provided by Davis and Werling (1960). In this study the authors report the quantitative changes in production associated with a company that had enlarged employees' jobs in three departments (maintenance, distribution, and operating), all of which had been enlarged from 2 1/2 to 5 years prior to the study. The researchers report performance improvement in terms of increased pro- duction volume and decreased costs. They do not, however, make a convincing case that the improved performance is attributable to the job changes themselves and not to other, extraneous, sources. This is a particularly relevant criticism, since this was not a controlled study, but rather one that relied almost exclusively on post hoc analysis, without benefit of apprOpriate control groups. They do attempt, however, on the basis of questionnaire data, to determine the specific job factors associated with each of four criterion variables: mean quantity of output, improve- ment of quality, reduction in operating costs, and mean quality of output. On the basis of correlational data, they identify a number of job factors associated with each of these criteria, as summarized below: i .2 an F oi 0" Me 2» a. 3: w l 1 .. . \ ,— 2. .e 12 Criterion Associated Job Factor(s) Mean Quantity Restricted, closely specified job of Output Improvement in Fully Specified work assignment Quality and work rate Reduction in Has or perceives as having a full operating costs work assignment Mean Quality of Perceives job as being important Output Identifies high quality needs; worker control over quality; relates success to high perfor- mance Worker control of work organization, high evaluation of fellow workers Peer communication Improvement in Full work assignment and some quantity of out— worker control over activities put preparatory to work Relates job success to management fairness; meets specified minimal standards of performance In spite of the above attempt at specification of specific job factors associated with specific changes, the criteria were found to be highly intercorrelated (range of correlations: .777 - .964) so the value of delineating separate factors seems of little importance. Warren (1958) asserts that the two criteria relevant to evaluating job enlargement are productivity and cost changes. He further argues for a long-range cost evaluation since, over the long run, the costs savings of job special- ization may be usurped by boredom and alienation. Finally, he decries the lack of empirical studies done to that time regarding job enlargement, and suggest that generalization regarding its efficacy was premature. “:2“: -H .-.-v :1". {7 T3 |.. #8.. in 8213 H 1 -. '--Av-.r b” I_‘- ' J.‘ ~~.A-I- AU "1“"? TV” ...;‘L'J L233“ q ';"1’ Av”. -..-v-._ n. ' :.:":+ .MU .4) F a h: 3“ - .5 | '. ~§us ‘v .‘ V .‘- I ll) Ir! ‘ -,—‘ +- 6.: L 15 Few such empirical studies were promptly forthcoming, however. Kennedy and O'Neill (1958) provided a look at the effects of job enlargement on job satisfaction. They found no significant differences in the job attitudes of the workers in repetitive and in varied jobs. Assembly Operators, who were assigned a specific task to be performed at an assembly line station, were compared in job attitudes with utility men, workers who relieved the assembly Operators and performed training duties as well. The researchers observe that "The biggest difference in the assembly Operator's job and the utility man's job was that the former performed a single, routine and repetitive task while the latter performed a wide number of the same routine tasks. . . ." (Kennedy and O'Neill, 1958, 575). This is hardly very conclusive evidence against job enrichment, however, since one may argue that the difference between the assembly and the utility positions was slight, though the study does perhaps, argue against job enlargement (the horizontal addition of a larger number of similar job duties). Another study found job satisfaction less related to the specific work than to geographical location. Katzell, Barrett, and Parker (1961), using correlational and centroid factor analyses, found that job satisfaction was associated with a "small town culture" more than with an "urban culture." In addition, they found, in their samples, that job satis- faction was not significantly associated either with quality or production or with turnover. Rather, employee satisfaction n l 'qh‘ .‘Ar- PDT ‘I\.‘. , ‘ ~é-u. in a 4 ‘ :_ 1"!“ :zlzaulo ." _ \- :.T'wv‘ghfl 3".V‘A- ‘ \~:‘g:4_;h~ ...yS _“‘K .‘ ‘ 7 \ "‘6 e“ l . “vn\- . L ‘ ~ ~ '1‘” ‘ N. ‘v\“:\e 2 "’ a. '..%A‘; . ‘ ‘.,~ h“ v‘ ' ,U ..v ..I s‘ - I ,“ ' .. ‘~ 1 :r" ‘i‘ ‘Vv‘i In '\ \\ .Qhfi‘. V\ \I 14 was found to be a function of size of work force, wage rate, unionization, and percentage of work force that is male; these variables were roughly characterized as expressing an urbanization dimension. Another empirical study, essentially a case study, is reported by Conant and Kilbridge (1965). The criteria they apply to studying an enlarged job are cost analysis and social interaction consequences. They investigate, ex post facto, the general hypothesis that bench work (enlarged job) is superior to traditional work. The study is set in a home laundry equipment manufacturing concern, located in a small, rural location in the Midwest. The authors present data which Show that bench work (essentially, assembling the bulk of the machine by the indi- vidual worker, rather than working on a small part of it) apparently resulted in cost savings over line work, in fewer rejects, and slightly higher efficiency levels, but also in increased production time. Unfortunately, no statistical Significance levels for the group differences are provided. Interestingly, also, social interaction was markedly reduced in the enlarged jobs, although this appears to be primarily a function of the physical arrangement of the new work sta— tions, and not to the work changes themselves. Conant and Kilbridge also report that the workers responded with im— proved job attitudes toward the enlarged jobs. Finally, a comparison via correlational analysis of preferences for the Old versus the new (enlarged) job on various personal ":9 wiih C 3. “’5'.ng CO: Unfo: 3,..- “NWDT‘fl‘ .7... “win-- V 2- . ‘ ’5 r .D-Avflo . , :.:va 04- '..".'L 9 - 3....-uy OJ. 1 F - --~'-u v.0 - . . JPPLCCE" 7‘; "'"“‘-U h‘ .. Q- ‘ .-...1 a . -..“V:\U.‘b .- ... 4 Davi- I "r . - , ch '3‘“? w- : Ufi~* . .... . I ‘An‘ ‘uh— Nwel 1 52";61 Wi‘ H", :“fi4. dB‘V‘. \ ._~ “ - .-.-V “Dag ‘ ' “My [ | I“? - ‘ ‘\,“ ‘05 ““NOW . \A‘ . 4‘Q‘ ”ha, rip wkv‘ .. ":\‘\ \.\‘1 ‘ ‘1 2V- ., ‘4‘.“ .Qh:\“‘ \‘u f 15 characteristics (e.g., age, years of education, length of time with company, number of children, etc.) found no sig- nificant correlations on any of the variables. Unfortunately, the Conant and Kilbridge study suffers from numerous flaws. The use of after—the-fact data and the absence of a control group seriously limits the generaliz- ability of the data. The questionnaire used by the authors to measure worker attitudes is of uncertain validity, nor are any data supporting the reliability of the instrument provided. Finally, the fact that over half of the workers expressed neutral or favorable attitudes toward the old (line) work suggests that job enrichment is not a universally applicable panacea. Davis and Valfer (1968) provide evidence supporting job enrichment principles. They reject the still dominant managerial belief that the supervisor's job Should'be designed with a primary emphasis on production requirements, without much regard for the wider view of the final output (i.e., production testing quality control output.) They test the general hypothesis that production and attitudes will improve as greater responsibility for the final output is assigned to the supervisors. Specifically, they hypothe- size that higher economic productivity (i.e., lower total costs) and greater needs satisfaction for group members and supervisors will result from supervisory job designs in the direction of increasing authority and responsibility by including direct control over all Operation and inspection Joint the c: g‘vL 'U k 2“" '11“: t3 5. ...-9.:- . '- AR.- 1 4., limit...» IV 0 . ‘ “:.: 'Epf- w "chow-lynx} .... ‘i‘j’ n 4 Dun-L, ‘.“ C} 'iE’.‘W"':‘ ~' '1 ""1'4‘45‘ a... ...; a. 0‘. “ ‘dblhgfi :.:-H." 1' Prod 9’ CU "‘f_m 53m (LI—H (1'! rn 1.1.”! 16 functions required in the work unit's production. Two experimental treatments were employed in the Davis and Valfer study: one set of work groups and comparison groups were given a "product responsibility" treatment in which the supervisor's job design was changed to provide authority and responsibility for overall functions required to complete the products produced in the shop. The second treatment method was called "quality responsibility" treat- ment, and consisted of the addition of quality control responsibility added to the product responsibility treatment. Objective dependent variables measured were produc- tivity, direct production costs (labor + materials), product quality, personnel costs (absenteeism, lateness, grievances, transfers, injuries) and time distribution of supervisor's activities. Subjective dependent variables were "changes in attitudes" and "changes in perceptions of supervisors and workers" which were obtained by "questionnaire, interviews, and ratings," by instruments and/or protocols of an un- Specified nature. The results of the study are summarized below: 1. Production costs/Productivity A. Product Responsibility groups:» No statistically Significant changes in pre- or post-periods for either experimental or comparison groups. B. Quality Responsibility groups: One quarter of the experimental groups showed a statistically sig- nificant productivity improvement. Costs showed a significant decrease in one-half of the ex— perimental groups and a non-significant decrease in the other half; controls showed a cost increase, significant in only one shop. “I 4.‘ 'V A A. v... P . —~/ S. Ct Eric 8.1.1 E n S ...l Uh .C n6. V f.\ e. F..\ a o. ‘l C.“ \v _A .“‘ u "u .91“ § s 17 2. Personnel Costs A. Product Responsibility groups: No change. B. Quality Responsibility groups: No change. 5. Time Distribution of Supervisor's Activities: A. Product Responsibility groups; B. Quality Responsibility groups: For both groups, supervisors were reported to be more concerned with the technical aspects of the job, to have more autonomy, and to have less free time available. 4. Attitude Changes: A. Supervisors (both treatments): mild to vigorous support for changes (interview data), with negative evaluation of only the decreased time available in the new system for personnel management. B. Workers: Responses were directly related to the degree to which reSponSibility and authority were granted to them. That is, Product Responsibility workers showed neutral attitudes toward the changes, whereas Quality Responsibility groups showed favorable attitudes to the changes. 5. Perception Changes: The major variable here was leadership: Product Responsibility Groups viewed their supervisors as decreasing in initiating structure and participation; Quality Responsibility groups viewed their super— visors as increasing in both dimensions. (For both groups, initiating structure and participation were "positively correlated" at an unspecified level.) The authors conclude on the basis of their studies that supervisors' jobs should be designed to increase re— sponsibility and authority for all the functions required to complete the product, including quality acceptance, and that responsibility should be delegated to the lowest organizational level performing the work. Such generalization from this study seems unwarranted. The lack of specification of the post-matching group comparability leaves serious doubts about the actual similarity of the groups, and hence questions .I "- 'z'hfiz-‘r ' n '1' * w J“ a I‘ . ‘- ...- O ‘l 131‘. JJ-f‘w‘ Afr :p-.L.\JM, V :......'1.1 . .::.vn:1U*- ...-... .5 c .,;.\.’u \J .Awnzfi u" . I L -~::-|—‘va UL .SH22figr- . . -.--.‘VOAV .. .- 1 . t F: “g ‘n ~ '-v .3-..” '. . J A - lA . .3 ,:“"M .3 , Q . U. :Q“"“ V. J'L-¢t \ . \.:‘a. \‘l‘ ‘h -‘\,, II .“ sf‘ 1‘ 1‘. I Vs ‘ 5 TN “A ‘2‘ ‘ I ‘ ~_“ ‘- .‘ S “H e ;“ . .p . t‘ . :‘tCQY‘s \' N 2‘; V. ‘Qj . a, v\v .‘ .\ . wipes-1; h“ 18 as to whether group differences are attributable to between group variance rather than to the experimental treatment. In addition, the absence of control groups for the Product Responsibility groups is no minor flaw. Also, the specifi— cation of changes in attitudes and perceptions as being measured by "interviews and questionnaires," with no further amplification, leaves extremely serious doubts regarding the reliability and validity of the instruments employed. Lastly, one cannot but be impressed by the promptness of the authors to attribute non-significant findings, or findings against predictions, to hypothesized causes, that may or may not be valid. One of the key advocates of job enrichment, Robert Ford, reports a series of predominantly successful imple- mentations of job enrichment (Ford, 1969, 1975). He details 19 experiments at A.T.& T., many of which were reasonably well controlled. The studies were conducted in a variety of settings within the company: Treasury, Commercial, Traffic, Plant, Comptroller, Engineering, and Traffic depart- ments. A total of ten of A.T.& T.'s companies and nine different jobs were involved in the changes. Perhaps the most scientifically acceptable of the Ford studies is the initial one, involving customer service representatives responsible for answering customers' letters and telephone inquiries (Ford, 1969; Janson, 1971). The experimental group changes included the assignment of greater control over the letters written, less pressure for increased fiction: 3f " su‘: ....- .1 $- " 37:81-3- U .d " £32135 were 1:";1‘. WC): pun-Iv -7 “WA aheq,‘ . - p: -— u—Vh‘ O u Date (I: "2 49 production, full accountability for output, and the appoint— ment of "subject matter experts" in each unit for employees to consult prior to consulting the supervisors. Control groups were treated in the usual (i.e., highly supervised) manner. working conditions (wages, policies, hours, training) were maintained for all groups at the pre-experimental levels. To control for the Hawthorne effect, neither first line supervisors or employees were told that an experiment was in progress. Data regarding the results of this study are presented primarily in chart and qualitative form; there is inadequate quantitative data provided to determine the significance of differences between groups. Specific results included: a. Customer Service Index (a measure of quality) increased for the experimental groups; b. Turnover was "significantly reduced" for the experimental groups; c. Production levels were increased for the experimental groups; d. More promotions were made for girls in the experimental groups; e. Verification costs were reduced from 100 percent to 10 percent for the experimental groups; training costs were reduced for the experimental groups; f. Job satisfaction scores increased for experi- mental groups more than for control groups. The remaining 18 experiments generally reported similar results, with the following typical gains noted: productivity improvement,decreased employee grievances, improved employee (flib satisfaction, improved customer satisfaction, and de- creased turnover. Overall, the experiments are rated by Ford * f" x I 1 a n 'th. lie the at abl e 5 W1 or? to S are V plic a; ..by For rib .L J «fills 110 W Several' ‘ dies at ” ..vv-‘vl-B ~ run U .Nugr H“, "I 99& J ea? 1 . «Ian 3 ”moiestlfi’ u ‘- ,Av-m‘ '15.. iv“. - on: 4 I I - ”to. . . .I' on E e a o C a H «11 A~ v q y La 1 - § mm ..u 20 and the applicable supervisors as 11 "quite successful," six "modestly successful," and one "not successful." Several problems, of varying severity, must temper the reader's interpretation of Ford's findings. Although the studies are reasonably well controlled experiments, at least as far as field work is concerned, there are several problems with the experimental design and with the data reported. Ideally, workers should have been assigned at random to the control and experimental groups; as it is, taking intact groups for experiments assumes (if changes are to be attributed to the experimental treatment) that the groups are virtually identical. However, no data are pro- vided by Ford as to the mean group characteristics. There is thus no way of determining if the results attained are attributable to the experimental treatment or to some other characteristic of the groups. The unusually high education of the groups in the Treasury Department (over 70 percent college educated) severely limits any generalization of the results. While more representative working groups are pro- vided in the subsequent 18 studies, Ford refuses to report (and apparently to employ) any statistical tests of signi— ficance regarding group differences. He provides instead raw data or his own interpretations of the results. While he points out that the changes appeared real enough to management and to the experimenters, this is no substitute for employment of statistical tests, particularly those tests resilient against assumption violations. Thus, at best, 1"" 5“ “LVN- "" a! v“ u v... . :“2"an\ . 21 Ford's studies provide data pointing toward a general trend supporting his job enrichment efforts, but the enthusiasm with which Ford in this and subsequent (1975) writings proffers job enrichment as a "proven" technique is unjustified. A series of similar studies was conducted in England ‘by Imperial Chemical Industries (Paul and Robertson, 1970; Paul, Robertson, and Herzberg, 1969; Cotgrove, Dunham, and Vamplew, 1971). Although the studies involved a variety of widely divergent jobs (sales representatives, design engineers, experimental officers, draftsmen, production and engineering foremen) there are central themes to both the changes and the results. Essentially, the changes were consistent with job enrichment tenets: more responsibility and less super- vision was given to the incumbents. For example, the sales representatives no longer had to write up reports on every customer call; they decided for themselves the frequency of calling on clients; in the event of customer complaints, the salesman had the authority to make immediate settlement for up to 100 pounds; authority was given the sales representative to buy back unwanted stock; and a discretion range of 10 per- cent of product prices was given the salesmen. Results were, for the most part, in the direction predicted by the job enrichment enthusiasts. Production generally increased, as did job satisfaction and quality of work, with no reduction in profit margins. Continuing the sales representative example, Paul and Robertson report that sales for the experimental group increased by 18.6 percent, -= frecjaentlv 7.3.125 t-lat dc L. .AAbdl&C t EEC :.:"; respect t ..T 312:..802 01a: -_.':’lst at .85 ~-:‘Ctrg -‘Vvv 22 while decreasing 5 percent for the control group. Similarly, job satisfaction increased more for the experimental group than for the control group. Generally these studies at ICI suffer from very small n's, frequently questionable control groups, at times changes that do not really seem to be "job enrichment," and an endemic tendency to over—generalize and too enthusiasti— cally tout the results as being universally applicable. ‘With respect to the last criticism, for example, Paul and Rebertson claim that the cultural differences between the United States and Great Britain do not render the results inapplicable to both countries, and suggest that virtually all jobs are amenable to job enrichment efforts. Similarly, Paul, Robertson, and Herzberg (1969) claim that the scope of job enrichment's applications is enormous, and that they have not encountered any situations in which the passing down of increased responsibility would be inadvisable. An additional study performed at ICI is reported.by Cotgrove, Dunham, and Vamplew (1971) in which the jobs of nylon spinners were enlarged. In this study, however, the job changes were primarily horizontal additions of similar level tasks. Generally, manning savings were realized, and workers reported much less boredom in their new jobs. In addition, changes of various dependent variables are not clearly interpretable since other changes that are not really job enrichment were also introduced, e.g., work teams. Several older studies may be viewed in a job enrichment 32:5 05 refer vav‘ 11" ' ‘: #4 . ' ”113185- . 9,...“ :.:.ity. Fin 22:32.1 voic efore it "9‘4... ....‘IS‘IS (19?0 w: ¢lter9 +7) 71 “but; 25 frame of reference. Katz and Kahn (1966) point to Trist's classic coal mining study as being evidence for job satis- faction's increasing with an increase in the meaningful cycle of activities. Another well—known study, Morse and Reimer (1956) demonstrates the consequences of changing female insurance clerks' jobs in the direction of increased respon- sibility. Finally, a persistent, but essentially non— empirical voice advocating the importance of meaningful work long before it became a cause célébre is that of Whiting Williams (1920, 1921, 1967). There are numerous other case histories reported in the literature, for example, Taylor (1972), Butteriss (1971), Myers (1970) and Maher (1971). However, a detailed review of each of these many cases is both needlessly tedious and unnecessary. Virtually all of these case histories are strongly favorable toward job enrichment in their results, and also, for the most part, of very little use since they have not employed scientifically acceptable methodologies. More recently, there has been a flurry of anti-job enrichment studies in the literature, and of studies tempering the erstwhile paens of praise of job enrichment's virtues. Hulin and Blood (1968) after an extensive review of the jdb enrichment literature, conclude that the job enrich- ment studies are, for the most part, poorly done. "The studies . . . appear to be of two types. Those which have 'used acceptable methodology, control groups, appropriate analyses, and multivariate designs have generally not 24 yielded much evidence which could be considered as supporting the job enlargement thesis. Those studies which do appear to support such a thesis frequently contain a number of deviations from normally accepted research practice." (Hulin and Blood, 1968, 50). The researchers also conclude that job enlargement cannot be held to positively correlate with job satisfaction. Rather, there are moderating influences that make the effect variable. The job level of the indi— vidual (White collar, supervisory, and non-alienated blue collar personnel are amenable to job enlargement efforts), and the alienation of the worker from middle class work ethic values (roughly measured by a rural versus urban distinction). Another study, Alderfer (1969), demonstrates that job enrichment efforts may have negative as well as positive results. Although workers whose jobs were enlarged in this study reported a higher satisfaction with pay and the ability the job offered them to utilize their skills and abilities, they also reported a decreased satisfaction with reSpect received from their superiors after the changes. Although there were some design problems in.Alderfer's study (e.g., nonerandom assignment of subjects to control and experimental groups and the choosing of the "best possible" workers for the experimental group), their main finding that job enrich— ment can be a mixed blessing is of note. Lawler (1969) attempts to tie job enrichment into an expectancy theory framework. He argues that the motivation .. 3151‘me effe 552:": will re» states, three :;er order n :5: be percei :55 that he 1:- Lfi. the worker 2::1‘31 over s t: ac‘rieving t Lexi-3r then go is effective , aiiiition of mc 51539-71) and VE ties into t} 2:: horizon? 252mg thEy : 25 to perform effectively is determined by the likelihood that effort will result in expected reward. There are, Lawler states, three characteristics necessary for jobs to arouse higher order needs: (a) meaningful feedback, (b) the job must be perceived as requiring the individual to use abili- ties that he possesses which he regards as important, and (c) the worker must feel that he has a high degree of self control over setting his own goals and defining the paths to achieving these goals. With this framework in hand, Lawler then goes on to suggest that job design changes, to be effective, must be enlarged both horizontally (the addition of more tasks similar to the ones presently as- signed) and vertically (the integration of higher level duties into the job). He suggests that jobs enriched on both horizontal and vertical dimensions will be motivating because they provide the three characteristics necessary for effective motivation. Equally important, Lawler (1969) discusses the im- portance of considering individual differences in job design. Unless, he argues, one is enriching jobs of individuals who have higher order needs, one is wasting time in job enrich— ment. And further, in line with expectancy theory, jobs that are enriched must be fit to individuals whose valued abilities are challenged by the job. Lawler goes on to address the issue of quantity and quality of production as dependent variables, concluding that, for two reasons, job enlargement is more likely to lead to 26 quality rather than quantity increases. This is because, (a) in Lawler's scheme, performing well is a sine qua non of experiencing feelings of accomplishment; performing well, in turn, rests on turning out a high quality product, but not necessarily in large numbers; (b) Lawler states that because many job enrichment schemes involve decreased re- liance on machinery, individual workers may be working harder to produce less; hence, the tangibly measured job change improvements will be reflected as improved quality. It is impossible here to discuss the many exigencies of expectancy theory. Lawler's work rests heavily on the theory, but regardless of its validity, the criticisms that he raises in this study that challenge job enrichment are of note, particularly the point that job enrichment does not appear to be universally applicable to all workers. In ad— dition, Lawler provides an at least feasible explanation of the typical finding in the job enrichment studies that quantity does not usually improve as the result of job redesign. Approaching job enlargement from a similar expectancy theory framework, Hackman (1969) classifies the performance process into a conceptual model, suggesting four types of impact by which the task itself may influence the performance process: (a) Influence through "hypothesis control" (i.e., what people think they ought to do); (b) motive arousal, as determined by the task; (0) task impact on cognitive and physiological activation level; and (d) "process outcome . I f; :5 \i-e-s real‘. from a .4... f 1:85. 2..-: ' ‘- 91‘ 4» 2. (he EALeCL .,.' '- +- ' $-33 at 15 :s‘. be cons: :ayhave. Tr n k.- d 3.71) consc :rier need 5 ;;'om~' .‘ ...-5‘, 7‘ 1?“; \J : ~04 C x "H‘ a"; ‘V‘, pa -.‘C"'\ _ ‘v... C .~ ‘3‘ \ 1 ‘Y 4.“ i ‘0 ‘ ‘ .‘ —A :~ \V‘ “Qa‘ - V ‘. - I‘d“: ‘V 5‘ 3 -, +~ ‘V . st}; ‘\ \‘ ~ '\‘.a ‘7‘ *\.Q V 27 links" (i.e., the specification of exactly what outcome will result from an action). Hackman maintains that these four kinds of task impact on worker behavior enable the prediction of the effects tasks will have on performance. Thus, job enrichment is not a simple cause—and-effect affair, but must be considered in light of the differential impact tasks may have. In a subsequent, empirical, study, Hackman and Lawler (1971) consolidate their previous theories to posit higher order need satisfaction as, essentially, a moderator variable determining the effects of job enrichment efforts. They test, in this important study, the specific hypothesis that if employees are desirous of higher order need satisfaction there should be a positive relation (correlation) between the four task dimensions of variety, autonomy, task identity, and feedback, and the dependent variables of motivation, satisfaction, performance, and attendance. This is to say, jobs high on the four dimensions (e.g., prOperly "enriched" jobs) will result in employees who are desirous of higher order need satisfactions tending to be highly motivated, satisfied, and rated by supervisors, and to have a low absenteeism rate. In general, Hackman and Lawler's hypotheses were supported by their study, although many of their correla- tions, although statistically significant, were of such low magnitude as to be of questionable "practical significance." In addition, the mean higher order need strength of their ::"195 was so safe restrlc musive 1n ~bfiiu . r *‘r mently . 1:512:15 are :.: +1 Ratner "iiéd to t}: E? “4:0“? 11’1ch V - a ... I‘\. \‘~andl \i‘ 4! “wise: to d“. .‘H 28 samples was so high (6.01 of 7.0 maximum) as to suggest a severe restriction of range problem in the study. What is impressive in this research, however, is its soundness of design and its connection to an integrated theoretical base, so frequently lacking in the typical job enrichment study. In addition, the implications of this research for job design are important: to the extent that Hackman and Lawler's con- clusions are valid, there is no one best way of designing a job. Rather, the psychological demands of jobs must be matched to the personal needs of workers for Optimal results. The same contingency approach to job enrichment is detailed by Monczka and Reif (1975) and by Morse (1973). The former study attempts to provide a conceptual model of job design, identifying those factors most important to job en- richment's success. The authors identify ten key job characteristics manageable by job designers: variety, autonomy, interaction, knowledge and skills levels, responsi- bility, task identity, feedback, pay, working conditions, and cycle time. More importantly, they point out three fundamental questions, frequently ignored in piece-meal approaches to job re-design: (a) where are the workers positioned on Maslow's hierarchy of needs scale; (b) are workers capable of handling the increased requirements of enriched jobs; and (c) do the workers themselves want job enrichment. This last question is of particular relevance, since it is so frequently ignored, at least in print, by job enrichers. Finally, Monczka and Reif point to two other ,jge'sions in, sate of tech zazlcinery are gilcsophy a: tzsfile manag Morse :zzsiieratio: 3925331 Motor n" ' .n I :ndmdue ‘1 ("“0 n ‘”$')1 (53 :ensions re "‘57 "'3‘; and (e ., 4} Q ' *3. WE vile I V. "4.:5 " Critic-i d a“ 29 dimensions influencing the success of job enrichment: the state of technology (jobs dominated by expensive fixed asset machinery are not easily amenable to change); and management's philosophy and style (job enrichment efforts may fail in a hostile managerial climate). Morse (1975) again calls for individual differences considerations in job re-design. He points, for example, to General Motors' celebrated Lordstown, Ohio, plant, where union officials wanted eliminated some of the jobs manage- ment had consolidated, in part, to alleviate boredom. "Humanizing work," he states, "is work that is motivating to the individual and is suited to his behavioral preferences" (Morse, 1975, 74). Morse provides a conceptualization of dimensions relevant to job enrichment: (a) task and technical variables; (b) individual personality variables; (c) organiza- tion and job design attributes; (d) organization effectiveness level; and (e) level of individual motivation. He suggests that successful job enlargement rests on both fitting the predispositions of workers to the jobs being re-designed and fitting the re-designed jobs to the applicable technology. Merse criticizes the job enrichment studies to date for failing to provide a Specification of either the worker's personality attributes or the specific, concrete, job attributes being changed- There are, he holds, four person- ality tendencies of relevance to job designers: (a) attitude toward authority; (b) attitude toward being and working alone or in.highly coordinated groups; (c) tolerance for ambiguity, .L TTEEDLS, are , .‘ " ! gamers , 1 Cf which w ,7..3 POE-1118 191.50 0 taijl, raised (i. e. , iiL’im the P ..ZI‘EZEI‘S reSpo 3523:; Workers I‘Eiggnd to El“ : :Cb aCCOQF ‘ :5“ :v .' VIV'LYV‘e . Ely V -~ “.1 N':‘fl h “‘ QSN ‘ kern 4-K. a ‘ . {‘lfl‘, - *Nk‘Lt‘. ‘ “LN,- "Eels...“ s“- \ ‘ K v” tr}? ‘ ‘1 9h -- "'s N ‘ ac‘iw “V ¥ ‘ ..‘6: fl 50 and (d) cognitive complexity. Relevant task attributes, he suggests, are: (a) clarity of information about the task; (b) the ability of the task to be programmed; and (c) the time span before performance feedback is available. Susman (1975), challenges Hulin and Blood's (1969), and others', contention that a rural/urban dichotomy is a valid way of depicting amenability to job enrichment efforts. In an empirical study dealing with 26 manufacturing plants, 11 of which were in rural locations (defined as less than 50,000 population) and 15 of which were in urban locations, he also obtained the community in which the §s had been raised (i.e., rural v. urban). Susman's results did not confirm the previously suggested hypothesis that rural workers respond more favorably to job enrichment than do urban workers. Rather, Susman concludes, rural employees respond to greater discretion in jobs with increased pride in job accomplishment and lower instrumental work orientation. Urban workers, and transitional workers, however, respond to increased discretion with increased general job interest. Further, he maintains that current residence is a more effective guide to this differential effect than the child- hood residence of the worker. Additional empirical evidence showing the mixed blessings that job enrichment can bring is presented by Lawler, Hackman, and Kaufman (1975). In a study set in a telephone company, the authors examine the effects of the enrichment of the directory assistance operator's job. s ' M“, ‘Iora""' . . . V‘h ‘ .....- . Maia-3:1 ...U- ""3"."2 7117 u a. .d—'. I - . \ ":"£“0"‘Q/' I ..vouvdob . vwfl set “V ""~V-O‘A‘ ._‘ GE‘- ' A " MVMH. UL. "F‘. H - . . . “‘3' £1 - ‘. ’r. r‘ “ 1 ‘Ir .L .Le 81“. '9; . ~ sfvp‘. ‘g‘r‘. - ,_ ‘ u‘" IVA- :7». I “I 3‘s“; +0 V‘ng Ur .-:. ‘9' . "f‘ ‘Ve W ‘ ‘ v‘gect 7'? so. V 3 ‘2‘:., x“ n K. ;’\."‘. = '- \‘V-\QV\‘+ v4“. I I "QC-*1- ." .‘. .‘t V IA" A F “Vi-r: ‘ ‘ l'\ s \‘7 h ‘. .A - Vvhn 'L '~ :.“3‘. ‘ -\ A r v [a (f) 51 Although improvements were noted in the amount and variety of decision making allowed the operators, there was also a significant and negative effect on interpersonal relation- ships, with older employees reporting less post—change satisfaction with their interpersonal relationships and the Supervisory Assistants (the intermediate supervisors of the operators) reporting less job security and less inter- personal satisfaction. But, while the point that job enrichment of one job may mean.job disenchantment of another is not insignificant, the emphasis placed by Lawler et al. on this finding seems too great. Job enrichment proponents in fact, have fre- quently cited decreases in the number of jobs needed for the same work to be done as one of job enrichment's accomplish- ments. In fact, another interpretation of the findings is that the supervisors were made redundant by the job changes, and that this resulted in a costs savings. The authors' statement that " . . . some [operators] even suggested that the office might function more effectively if the job of the [intermediate supervisor] were eliminated," (61) implies, incorrectly, that the number and type of jobs are to be taken as a given, clearly an approach antithetical to job enrichment. Another job enrichment critic, Mitchell Fein, offers littlelnnxe for job expansion. In somewhat virulent writings that border at times on diatribe, Fein (1975, 1974) excoriates job enrichment as an unworkable, ill-conceived technique that citings to (1'- 57.~i'.es to SUE . I V “w ~-o:> “’ ;:;..; ‘ L..— - U :rltioizes 2339:; with ass Ivy-V‘ u‘. -4a—E¢ , kle (:: :-:..t ,job enr. .1... with co 7‘. ;. -a_ 'T‘N \"C Q‘Fr \u~;k; I ‘_"r ._\‘:I:?" (41’NI ‘1 J ‘13‘ .W‘ fr‘n _ .. JOb i ~:,"‘ "‘E: (34 :L ("n s; Q .~ “‘5. ..x 1C h ' «\ ‘-a . a" ‘ v 3 x'; a .V v y.‘ ‘ \J‘ ‘2 ‘1 . '«.C~h y 52 is of little value, use, or importance. Unfortunately, Fein indulges in a great deal of non-scientific, non—empirical criticisms that offer little scientific evidence to support his views. He is particularly fond, in his published writings to date, of quoting single individuals or case studies to support his points, while simultaneously casti- gating job enrichment studies for using identical techniques. He criticizes numerous job enrichment "success stories," often with assumption-ridden, contrived criticisms. For example, he claims the often cited General Foods Topeka plant job enrichment was a "controlled experiment in a small plant with conditions set up to achieve desired results . . . The plant and its operations are not typical of those in industry today . . . what makes this plant so unique is not only the management style but the workers themselves who were handpicked" (Fein, 1974, 72). Or again, Fein throws out Proctor and Gamble's job enrichment efforts because "[Proctor and Gamble] is an unusual company with a history of concern for its employees that is matched by few other firms in the country" (1974, 75). Apparently, to Fein, the only suitable test for job enrichment would be with a hostile, antagonistic management. Fein also spends a great deal of effort discarding "job enrichment" efforts that do not meet some unspecified definition of the term. He says, for example, that Polaroid Corporation's experiments involved only job rotation, and are therefore not job enrichment, and that the famous A.T.& T. 2 AC efirlChfll ,\ ..P .U Exit elf ,. ”1 ,..-1 1? e13 no V 4, ..-.AJ‘ ,. f ‘ eeif eln ': S, even 59 sierson, 19 ‘ 1 5.8.... :‘fi‘fl‘ 1 ...lv all] ”I n, a ,- u b 4. 1 - ‘ -- uEC' olo, \ 15 «19 this 1" costs C.“ a. \A. he cl .11: ’ . v \ _ \ 55 job enrichment studies (Ford, 1969, 1975) merely re-designed jobs "Which had been ineffectively set up in the first place" (Fein, 1974, 74). But even if Fein's anti-job enrichment tone and relatively non-empirical, frequently ad hominem, comments must be approached cautiously, one should not discard his views simply for their apparent hostility. In fact, several of the points he raises are relevant job enrichment criti- cisms, even if only of heuristic interest. Specifically, Fein notes the previously cited limits to job enrichment imposed by technology (also cited by Anderson, 1970), claiming, with some overstatement, that job structure in the United States today is "dictated largely by the technology employed in the production process" (1974, 75). While this is true to some extent in manufacturing, one cannot ignore the fact that the American economy is today more service oriented than manufacturing oriented. He notes additional problems as well. Job enrichment can result in higher costs rather than lower, there are relatively few jobs (he claims) with higher skill requirements, group norms may work against job enrichment efforts, and.many workers express fear that they will be penalized if they improve their productivity. Claiming to present "a more balanced approach," Fein suggests that "there are no data which show that re— structing and enriching jobs will raise productivity" (1974, 80). He suggests an alternative, more correctly, the status rings outsid tie lack of it were true, :11: see their 13: very well which they pr; . .,, “1, ' u 54 quo: workers self—select themselves into higher level jobs; those dissatisfied with boring jobs get out of them. While Fein admits that some workers do want larger jobs, he main— tains that "Most workers want more freedom to act on personal things outside of their work place" (1974, 86). Ignoring the lack of any supporting data for this statement, even if it were true, another interpretation is possible. Workers who see their jobs as hopeless, but inescapable, dead ends, may very well turn their energies to outside activities, which they perceive as the only avenue available for creative work. It is impossible here to detail any more thoroughly the many deficiencies in Fein's reasoning and writings. He remains less a formidable critic of job enrichment than a purveyor of questions job enrichment theorists must legiti— mately address, but on the basis of hard core research findings, not on unsupported generalizations. Levitan and Johnston (1975) offer a bit more balanced, though also essentially noneempirical, criticism of job enrichment. They argue that dull, tedious jobs still exist because there is still an economic demand for them. They claim that the job enrichment literature is filled mostly with success stories since the failures are, understandably, not publicized. In addition, the production benefits fre— quently attributed to job enrichment efforts are typically not controlled against gains from alternative sources. They also point out that attitude changes attributed to job ehargelent hall that te lie-sip, and factlring d1 investment 1 :m'nl “‘4. m cu Bull 3 1t .. T116 11". . b~rt2~rCe " C h: . ('31:;er , 55 enlargement cannot be guaranteed to be non-transitory. They hold that technology is the most important factor in job design, and cite the statistic that the average plant manu- facturing durable goods averages over $25,000 fixed capital investment for each worker. As for those who suggest a social efficiency model approach to job enrichment (i.e., make jobs more interesting regardless of whether production gains result), Johnson and Levitan state: "Improved social efficiency cannot proceed along Opposite paths to industrial efficiency, but must parallel it. Without the tremendous affluence generated in large part by efficient mass produc— tion, there would be no alternative life styles or occu— pations for workers to envy, and no time to invest in the education which has contributed to some workers' dissatis- faction with their jobs" (Levitan and Johnston, 1975, 59). Reif and Tinnell (1975) address the frequently ignored problem of which jobs are most suitable to job en- richment efforts, acknowledging at the outset that all are not. They prOpose an eighteen.element scheme by which to rate the suitability of jobs to enlargement efforts. The elements they prOpose fall into four categories: (a) the job itself; (b) technology; (0) the workers; and (d) manage— :ment. The specific elements suggested are: (a) job: im— portance of quality, flexibility, extent of coordination required, specialization benefits, conversion and training costs, wage payment plan, and role of job satisfaction in affecting production rates; (b) technology: role of fixed l'cestment ' tibility of satisfactic man local net of max of manager to tolerat and observ the diagnc schemata : Which one of Parana: 56 investment in equipment and technology; (c) workers: suscep- tibility of workers to change, job security, present job satisfaction levels, skill levels, education levels, rural vs. urban location, and unionization; (d) management: commit- ment of managers to job enrichment, experience and training of managers in job enrichment, and willingness of management to tolerate a time lag between job enrichment implementation and observable results. Reif and Tinnell do not suggest that the diagnostic instrument they provide based on the above schemata is ideal for all situations, especially in cases in which one factor (e.g., technology or union opposition) is of paramount importance. What is of particular value in this presentation is the gestalt view it provides to would-be job enrichers. A non—empirical job enrichment broadside is provided by Schappe (1974) in an article auspiciously entitled: "Twenty Two Arguments Against Job Enrichment." Many of Schappe's criticisms have already been discussed above. IEssentially, he argues, and generally without supporting -—Knows job well . On your feet __Bad ::}rustrating __Intelligent Simple ::Leaves.me on my own :Endless Around when needed _Gives sense of :Lazy accomplishment 64 Validity of the Job Satisfaction Measures Extensive evidence for discriminant and convergent validity for the JDI is presented by Smith, Kendall, and Hulin (1969). In addition, the JDI has been shown to correlate significantly with a number Of other variables (e.g., age, absences, performance ratings, etc.), as well as, in at least one study, showing a "substantial relation- ship" with turnover over a 12 month period. .Motivation Patchen's Job Motivation Index was used to assess changes in amount of energy respondents devoted to job tasks. This is a simple, four question, Likert response format instrument consisting of the following items: (Robinson, Athanasiou, and Head, 1969) On most days on your job, how Often does time seem to drag for you? 1) About half the day or more 2) About one-third of the day 5) About one—quarter of the day 4) About one-eighth of the day 5) Time never seems to drag Some people are completely involved in their job--they are absorbed in it night and day. For other peOple, their job is simply one Of several interests. How ginvolved do you feel in your job? (1) Very little involved; my other interests are more absorbing gag Slightly involved 5 Moderately involved; my job and my other interests are equally absorbing to me 24; Strongly involved 5 Very strongly involved; my work is the most absorbing interest in my life How often isn't rea E?) Once 1) About Would ym same as < organize ratings Of I PEIatively Eeasured by In addition Promotion the laLtter groups (R0) 65 How often do you do some extra work for your job which isn't really required of you? 5) Almost every day 4) Several times a week 5) About once a week 2 Once every few weeks 1 About once a month or less Would you say you work harder, less hard, or about the same as other peOple doing your type of work at your organization? 5) Much harder than most others 4) A little harder than most others 5) About the same as most others 2) A little less hard than most others 1) Much less hard than most others Validity of the Job Motivation Index Moderate to slight correlations between supervisors ratings of "concern for doing a good job" are reported; relatively strong correlations between.motivation (as measured by this scale) and absence rates are also indicated. In addition, the scale correlates with mixed results for production volume, but fairly strongly with job satisfaction, the latter correlation being substantiated over 90 work groups (RObinson, Athanasiou, and Head, 1969). Leadership Dimensions A wealth of attention has been directed through the years to the definition and measurement of leadership dimen- sions. The present research adOpted the results of a recent study (Johnson, 1975) Which determined, using multiple dimen— sional scaling analysis, a third component to be added to the usually delineated factors of initiating structure and consideration york, however students, us: source of da' addition to afield test of Johnson's questions w five select ”a‘inost non Teponse 81 order from by 5 to 1), Wows (t: 66 consideration: participative decision making. Johnson's work, however, involved laboratory studies with college students, using hypothetical leaders, at best a limited source of data for field leadership analysis. Thus, in addition to its other purposes, the present study enabled a field test (to the experimenter's knowledge, the first) of Johnson's three dimensional leadership scales. The questions were presented with Likert response format, with five selection alternatives, ranging from "greatly" (1) to "almost none" (5). To ameliorate potential response bias, response alternatives were presented in alternate scaling order from one question to the next (i.e., 1 to 5, followed by 5 to 1). The question stems for each dimension are as follows (the secretarial question stems are presented; bosses were asked to rate themselves): Consideration: How friendly and easily approached is your boss? How much appreciation does your boss express when you do a good job? When you first began working with your present boss, how much did he (she) facilitate adjustments to your new work setting? How much at ease do you feel in talking with your boss? How much does your boss look out for your personal welfare? Participation: How much does your boss allow you to modify the pro— cedures required in your job? How much does your boss have you share in decision making? To What extent does your boss assign you a task, then let you handle it? To What extent does your boss allow you influence equal to his (her) own in decisions Which affect you? How much does your boss let you work the way you think‘best? Initiatl How muc of yo How muc EcW‘muc place How mu: perfc Bow mus , mate Validity of The behavior it Stodgill, ‘ initial 33] considerat to '15 iten graduate 5 items werq was Obtain 67 Initiatinngtructure: How much does your boss let you know what is expected Hogfngg?does your boss schedule the work you have to do? How much does your boss make sure you understand his place in the work group? HOw much does your boss maintain definite standards of performance for you? How much does your boss see to it that you have the material you need to work with? Eglidity of the Leadership Instrument The items listed above were generated from leader behavior items on standard leadership questionnaires (e.g., Stodgill, 1963; Fleishman, Harris and Burtt, 1955). An initial sample of 56 items (13 for participation, 12 for consideration, and 11 for initiating structure) was reduced to 15 items (5 for each dimension) by industrial psychology graduate student judges using a Q sort technique. The final items were those on which at least .80 interjudge agreement was obtained. Because of the method of generating the questions and the absence of field testing of the instrument, no strong claim can be made for the construct validity of the instru- ment. lAlthough extensive validation studies have been conducted with.most of the questions, it was in their original form, as a separate leadership instrument that these questions were employed, and hence no mention is made here of the results of such studies. The only "hard core" validity evidence yet available for these questions as a separate instrument, and it is admittedly pretty soft "hard core," is a study done with college etude four depender activation, t 19273). In tt variables wei however, aga: of initiatin the dimensio satisfactior the supervi: other studi more faithf a defect i1: View that J restrictiOJ W The an 9110mm fiOI‘k, in ‘ produCed 1 amts to SSCI-etari such meas inflated! Evolve " 'r . \‘E‘O Pt 1 C an. 68 college students dealing with the instrument in relation to four dependent variables: satisfaction with the supervisor, motivation, task competence, and interpersonal style (Johnson, 1973). In the case of each dimension, the four dependent variables were found to positively correlate. This was, however, against the experimenter's prediction in the case of initiating structure, for which the prediction was that the dimension would negatively correlate with subordinate's satisfaction with the supervisor and the taSk competence of the supervisor. Although there is conflicting evidence from other studies, this finding against the prediction probably more faithfully reflects a deficiency in hypothesizing than a defect in the construct validity of the instrument. (The view that American workers automatically want freedom from restrictions and not to be told what to do dies slowly.) §2p_3erformance Effectiveness Measures The measurement of production in white collar jobs is an enormously complicated undertaking. Unlike production work, in which one has a tangible product whose amounts produced can be assessed relatively easily, secretarial work admits to no ready measure. In certain very repetitive type secretarial clerk jobs one might conceivably deal with some such measure as number of letters typed or phone calls answered, but any except the most basic secretarial positions involve a variety of tasks beyond that. This is especially important in a study such as the present one, in which to ite "‘2 U) 69 results of a job enrichment scheme are being studied. To the extent the training is successful, secretaries exposed to it should assume a number of additional duties, indeed, should.be on the premise of accepting as many of the boss's job duties as possible. Thus, no "hard" measures of pro- duction effectiveness are possible. A subjective, self-perceived measure of job production was therefore adOpted. The best measure encountered for this purpose was a set of two questions from Mott (1972) intended to assess the quantity and quality of work produced. These items are: Thinking now of the various services produced by you and your boss, how much are you producing? How good would you say is the quality of the services produced by you and your boss as a team? As might be expected, these two items correlate fairly highly with each other. This is understandable, since it is unlikely that quality and quantity of work pro- duced would be perceived by respondents as two separate, unrelated dimensions. However, this is not an altogether unmixed blessing, since, if the correlation holds up in the present study, it permits a rough assessment of reliability of the measures, an assessment impossible with a single item. Validity of the Job Effectiveness Items Other than face validity, what evidence is there for content validity of these items? Mott (1972) reports the use of the questions (plus a good many more) in field studies fl, 70 with NASA, the State Department, an office of the HEW department, numerous hospitals in Michigan, and a state mental institution in Pennsylvania. These two items were found to correlate strongly with such measures as overall effectiveness, adaptability, and flexibility. In addition, the author validated his instrument against tOp management's rating of organizational effectiveness of the sub-units measured. Mott concludes that: ". . . the effectiveness index is a valid and inexpensive measure except when re- sponses reflect outmoded standards. Such situations can usually be revealed by comparing internal assessments with those of t0p management and other outside rankers and by examining disagreements through followup interviews," (Mott, 1972, 199). The Overall Instrument The final version of the pretest is presented as Appendix A and Appendix B for secretaries and bosses, respectively. The posttest for'bosses is presented in Appendix C and that for secretaries in.Appendix D. The pre- test was administered approximately one week prior to the training; the posttest was sent out about one month after the session. Specimen cover letters for the various question- naires, as well as the followup letter for recalcitrants, are presented as Appendix E. It will be noted that there are four job duties present on the supervisor posttest and tWelve on the secretary 71 posttest that are not present on the pretest. These represent job duties that were specifically covered in the training session that were not known at the time the pretest was com- posed. The Solomon Four Group design looks only at post- test results; because only a small number of items were added to the posttest that were not on the pretest, it is argued that this addition constitutes no threat to obscuring a pretesting effect. A stronger demonstration of this con- tention would be fulfilled by running analyses with and with— out the added items and noting any differences in the pre- test effect. However, both because of time and financial constraints, and because of the extremely small pretest effect, accounting for 0 percent of the variance (See Re- sults section), such an analysis was not run. It will also be noted that a series of six questions are asked regarding biographical data on the subjects. These represent hypothesized dimensions on which it is desirable that experimental and control groups be similar, since, especially in the case of secretaries, differences on these variables could represent different susceptibility to job enrichment efforts. Finally, it should be observed.that there are a num- ber of questions on each instrument that are not utilized in the present study (e.g., JDI supervision scores for bosses). These items represent responses collected in some cases for purposes irrelevant to the present study, and, in other cases, for an intention that was impossible to achieve. With reslDE cipated ti variable C bosses'and aptness f o the nature teams did : to make th Data Analz: Rel: determined correlatio: the diagone The training on assessed by “tivariat ilalySiS of :EOtheseS 2 72 With respect to the latter category, it was initially anti— cipated that it would be possible to look at dependent variable changes as a function of the discrepancy between bosses'and.secretaries' perceptions of the secretary's aptness for assuming new job duties. However, because of the nature of the returned questionnaires, boss-secretary teams did not complete the questionnaires in adequate numbers to make this type of analysis possible. Data.Analysis Reliabilities of the scales used in this study are determined by computing coefficient alphas by means of a correlational program permitting use of communalities in the diagonals. The determination of the effect, if any, of the training on the supervisory and secretarial groups is assessed by means of analysis of variance. Because of the multivariate nature of the dependent variables, multivariate analysis of variance (MANOVA) is essential for testing hypotheses 2 through 6 and 8 through 12 (Hummel and Sligo, 1971), the variables oh which no change due to the training session is predicted. Only if there is a significant over— all F ratio for the effects of the entire set of variables (which would be against the hypothesized direction of the results) is one justified in looking at the univariate F ratios. Therefore, to demonstrate evidence supporting the previously mentioned hypotheses, it is necessary to demon— strate that the overall F's are not significant. To test these sets 0- orogram was - On til" the secretar: and not assu not appropri iependent v: sized to ch all other V Schmidt , 1c 75 these sets of hypotheses, Jeremy D. Finn's Mhltivariance program was employed (Scheifley and Schmidt, 1973). On the single variable for which change is predicted (enriched job duties), a univariate analysis of variance was performed for both the number of job duties assumed by the secretaries and for an average of the duties assumed and not assumed for the enriched job duty cluster. It is not apprOpriate, despite the multivariate nature of the dependent variables, to include a single variable hypothe- sized to change in a multivariate analysis consisting of all other variables hypothesized not to change (Hunter and Schmidt, 1974). A note is in order regarding scoring of the job duty portion of the instrument. .A simple dichotomy was used to score the job duties for the secretarial samples (1 Do the task; 2 Do not do the task). Then the various scores were summated and divided by the number of duties in the category. Thus, the closer the cluster average was to 1.00, the more cluster duties the secretary reported accomplishing. The supervisors' scales were, through an unfortunate scoring error, scored in the Opposite direction. A "1" was assigned to job duties the supervisor indicated he did him- self or were not done, and a "2" to job duties that the supervisor reported were done by his secretary (or by both he and his secretary). Thus, the closer the cluster average was to 2.00, the more cluster duties, on average, the 74 secretary was reported by her boss to be accomplishing. As noted above, the present study consists of 2x2 analyses of variance with four groups: a matched and an expanded secretarial group, and a matched and an expanded supervisory group. Only those variables demonstrating no significant differences on the pretest are included in the posttest analyses. RESULTS Pretest Analyses Pretest comparability was tested for the matched and expanded groups by the appropriate t_test. Results for the matched samples are presented in Tables 2 and 3, and for the expanded samples in Tables 4 and 5. For both secretaries and bosses, no pretest group differences were noted in either matched sample, including the finding of no significant differences on the biographical variables. However, .05 alpha level significant differences did emerge for the participation leadership dimension for bosses in the expanded sample. This is not a particularly alarming finding, however, since in a group of 14 variables such as this five percent of the variables, or .7 or a variable, could be expected to be significantly different by chance. On the other hand, the expanded secretarial sample showed significant differences on three dependent variables: motivation, quantity of work produced, and par- ticipation of bosses. In addition, three of the biographical variables showed significant differences: age, time on job, and time with present boss. Dependent variables with significant differences on the pretests are omitted from posttest MANOVA analysis. In the case of the expanded secretarial sample, the relatively 75 76 TABLE 2. t Tests for Pretest Differences on.All Variables Matched Samples, Secretaries. Control Experimental 1 Variable Group Mean Group Mean t Dependent Variables: nF16 Routine Job Duties Compositea 1.522 1.417 .833 Enriched Job Duties Composite2 1.639 1.558 .743 Social/Personal Job2 Duties Composite 1.775 1.625 .898 Motivation 14.750 16.875 —1.747 Quantity of Work Produced5 2.166 1.833 1.000 Quality of Work Produced3 2.250 2.000 .608 Job Description Index-- WOrk 40.250 41.625 -.577 Job Description Index—- Supervision 45.500 44.500 .225 Consideration of Boss 11.625 10.125 .760 Participation of Boss 13.875 10.875 1.612 Initiating Structure of Boss 12.375 11.500 .622 Biographical Variables: Age4 5 5.125 3.875 -1.112 Time on Job 5 2.875 3.750 -1.830 Time with Boss6 2.571 3.571 -1.383 Marital Status 1.500 1.130 2.17 No. of Children 1.250 1.000 .386 1 2 §.test for matched samples, df='7 1 = Done by Secretary; 2 =Not Done by Secretary 51=Excellent; 2=Good; 3= Fair; 4= Not too good; 5= Poor “Under 2o=1; 2o-29= 2; 50—59: 5; 40-49: 4; 50-59: 5; 60-65= 6; Over 65= 7 5Less than 6 months =1; 6 months to 1 year= 2; 1 to 2 years= 3; Over 2 years=4 6Married= 1; Other: 2 '77 TABLE 3. t Tests for Pretest Differences on All Variables Matched Sampl es , Supervisors Control Experimental 1 Variable Group Mean Group Mean t Dependent Variables: n=14 Routine Job Duties Composite 1.300 1.223 1.000 Enriched Jo Duties Composite 1.425 1.344 .743 Social/Personal Job Duties Composite2 1.257 1.150 1.389 Motivation 16.714 18.571 -1.596 Quantit 3 2.000 1.715 .680 Quality 3.142 2.000 1.804 Job Description Index-- Work 45.428 43.143 .697 Consideration 12.285 12.570 -.257 Initiating Structure 11.142 11.856 -.442 Participation 12.285 13.999 -1.333 Biographical Variables: Age4 5 4.000 4.285 —.547 Time on Job 5 4.000 3.572 1.002 Time with Secrgtary 2.714 3.428 -1.110 Marital Status - 1.000 1.000 0.000 No. of Children 3.000 2.572 .891 1: test for matched samples, df= 6 2 1:. Done by Boss or Not Done; 2= Done by Secretary or By Both Boss and Secretary 31=Excellent; 2= Good; 3= Fair; 4= Not too good; 5= Poor “Under 2o= 1; 20—29= 2; 50—59= 5; 40—49= 4; 50—59= 5; 60—65= 6; Over 65=‘7 5Less than 6 months= 1; 6 months to 1 year==2; 1 to 2 years= 3; Over 2 years=4 6Married = 1; Other: 2 78 TABLE 4. t Tests for Pretest Differences on All Variables Expanded Samples, Secretaries Control Experimental 1 Variable Group Mean Group Mean t Dependent Variables: n=29 Routine Job Duties Composite2 1.437 1.409 .337 Enriched Job Duties Composite2 1.591 1.552 .956 Social/Personal Job Duties Composite 1.742 1.622 1.016 Motivation 15.368 17.000 -1.886* Quantity of Work Produced} 2.000 1.555 2.060* Quality of Work Produced} 1.789 1.888 -.292 Job Description Index-— Work 39.842 42.777 -1.093 Job Description Index—- Supervision 46.210 45.222 .258 Consideration 11.421 9.666 1.279 Participation 13.421 10.666 2.295* Initiating Structure 12.526 10.888 .959 Biographical Variables: Age4 5 2.947 4.000 -1.857* Time on Job 5 2.789 3.777 -3.454** Time with Boss6 2.473 3.555 -3.501** Marital Status 1.210 1.440 -1.350 No. of Children 1.000 .888 .214 1t test for independent samples, df==27 *p <.O5 **p'<.01 2 1= Done by Secretary; 2: Not Done by Secretary 31= Excellent; 2==Good; 3==Fair; 4: Not too good; 5: Poor “Under 20=1; 20-29 =2; 30-39: 3; 40-49 =5; 50-59: 5; 60-65: 6; Over 65= 7 5Less than 6 months= 1; 6 months to 1 year= 2; 1 to 2 years= 3; Over 2 years=l+ 6Married==1; Other: 2 79 TABLE 5. t Tests for Pretest Differences on All Variables Expanded Sample, Supervisors Control Experimental 1 Variable Group Mean Group Mean t Dependent Variables: n=18 Routine Job Duties Compositez 1.275 1.222 .486 Enriched Job Duties Composite2 1.431 1.343 0698 Social/Personal Job Duties Composite2 1.285 1.150 1.550 Motivation 16.727 18.571 1.414 Quantity of Work Produced} 1.916 1.714 .588 Quality of Work Produced3 2.416 2.000 .804 Job Description Index-- Work 43.833 43.142 .197 Consideration 11.666 12.571 .632 Initiating Structure 11.083 11.857 .506 Participation 12.000 14.000 2.127* Biographical Variables: Age4 5 5.916 4.285 .680 Time on Job 5 3.666 3.571 .191 Time with Secretary 2.818 3.428 0.000 Marital Status6 1.000 1.000 0.000 No. of Children 2.916 2.571 .552 13 test for independent samples, df= 16 *p <.05 21==Done by Boss or Not Done; 2=.Done by Secretary or By Both Boss and Secretary 31=Excellent; 2= Good; 3 =Fair; 4= Not too good; 5= Poor 4U'nder 20:1; 20-29: 2; 30-59:}; LPG—49:4; 50-59: 5; 60__65= 6; Over 65 = 7 5Less than 6 months=1; 6 months to 1 year: 2; 1 to 2 years= 3; Over 2 years=4 6Married= 1; Other: 2 80 large number of significant difference variables argues for extremely cautious interpretation of the results for this sample, despite the fact that there is no particular pattern to the direction of the differences. Reliability of the Measures The Job Duties Instrument Correlational analysis with coefficient alpha in the diagonal was performed on the a priori constructed job duty scales using FACTRB (Hunter, 1974), a FORTRAN program that permits, in its Multiple Groups routine, user specification of subscale components. Resultant coefficient alphas, as well as the between scale correlations, are presented in Table 6 for the secretaries and Table 7 for the supervisors. Coefficient alpha is an internal consistency measure (range: 0.0 - 1.0;trounding error). It provides a measure of the reliability of a test or scale, and is especially useful when alternative tests for the same phenomenon are not available. For the secretarial sample, it will be observed that thecoefficient alphas were reasonably high, indicating adequate reliability levels for the purposes of this study. This is especially true for the enriched and routine job duty clusters. The general stability of the measures is fairly good from pre- to posttest, although it should be noted that the posttest correlations are based on larger samples than the pretest correlations. This accounts for the general 81 TABLE 6. Correlations and Reliabilities for the Secretarial Job Duty Scales1 Scale Routine Enriched Social-Personal PretestL Matched Sample n=16 Routine {-85) Enriched .54 (.82) Social/Personal .93 .75 (.47) Pretest, Expanded Sample n=2 Routine (.82) Enriched .58 (.79) Social/Personal .87 .9O (.38) Posttest, Matched Sample n=32 Routine (.87) Enriched .91 (.90) Social/Personal .88 .87 (.51) Posttesti Expanded Sample n;4 Routine (.89) Enriched .92 (~91) Social/Personal .83 .90 (.60) qcorrelations corrected for attenuation with coefficient alpha's in the diagonals. 82 TABLE 7. Correlations and Reliabilities for the Supervisors' Job Duty Scales Scale Routine Enriched SOCial—Personal Pretest, Matched Sample n=14 Routine (.81) Enriched 1.01 (.95) Social/Personal .45 .53 (.62) Pretest, Expanded Sample n;18 Routine (.76) Enriched 1.01 (.91) Social/Personal .58 .50 (.51) Posttest, Matched Sample n=28 Routine (~71) Enriched .92 (.88) Social/Personal .52 .80 (.57) PosttestigExpanded Sample n=32 Routine {-71) Enriched .93 (.86) Social/Personal .66 .81 (.66) qurrelations corrected for attenuation with coefficient alphas in the diagonals. 85 increase in the coefficient alphas from the pre— to the posttests since the large sample sizes provide increased variance. In both pre- and posttests, both the control and the experimental groups are combined to enable a determina- tion of the coefficient alpha for a reasonably large sample, though in all cases, the n's are still smaller than desirable. For the secretarial samples, the between-scale correlations are moderate in the pretest samples, but, in the case of the routine/enriched correlation, jump from .50's level correlations to low .90's level correlations. There is thus some indication that, especially in the posttests, the routine and enriched job duty clusters are tapping a similar factor. However, because of the hypotheses generated prior to the data analysis, and because of the relative in- stability of the routine/enriched correlations from pre- to posttests, the analysis of variance procedures will still be carried out as indicated in the Methods section. As for the supervisors, essentially the same situation exists. The coefficient alphas, with the exception of the social-personal duty cluster are reasonably strong. Again, however, there are quite high interscale correlations between the routine and the enriched clusters (with little instabil— ity'from.pre- to posttest). These strong interscale correla- tions indicate that a similar, if not identical, factor is 'being tapped. This may indicate that the respondents did not discriminate between "enriched" and "routine" duties, i.e., to the respondents the level of the job task was an irrelevant 84 issue to whether or not they did the task. If true, this suggests a possible defect in hypothesizing: the idea of change on one cluster of job duties and lack of change on another, highly correlated, cluster, presents a contradic- tion. It should be recalled, however, that these inter— scale correlations are only estimates and are based on the combination of the experimental and control groups to in- crease the size of the n's. The extent that change occurred in the experimental group due to training but not in the control group complicates the picture by obscuring the real interscale correlations. In any event, the original analysis of variance procedures will be employed for the supervisors for the same reasons discussed above for the secretarial samples. Job Satisfaction Index Reliabilities The split—half internal consistency coefficients for the JDI are reported to be greater than .80 for each scale, and there is evidence for the stability of the measures over time (Robinson, Athanasiou , and Head, 1969). Smith, Kendall, and Hulin (1969) also report numerous correlations above the .70 to .80 range between the JDI and.other measures of job satisfaction, indicating, they state, the lower bounds of the JDI's reliability. In this study, a conservative .80 estimate of the two scales' reliabilities will be employed. 85 Reliability of Patchen's Job Motivation Index Robinson, Athanasiou, and Head (1969) report test- retest reliability of .80 for a subset of two of the Moti— vation Index questions with a sample of 49. To augment this limited reported reliability, coefficient alphas were com— puted for the four item test and are reported in Table 8. Although fairly strong and consistent alphas are indicated for the secretarial sub-sample, only the posttest alphas for the boss samples are reasonably high. This is probably accounted for by the small sample size of the pretest samples (n's of 14 and 18). In any case, the posttest alphas for the supervisors are judged adequate, as are all alpha values for the secretaries. Leadership Dimension Reliabilities As previously mentioned, this study appears to embody the first field test for the Johnson (1973) three dimensional leadership scales. Accordingly, there are no published reliabilities yet determined, at least to the experimenter's knowledge. Interscale correlation matrices for the three dimensions are presented for the secretaries in Table 9 and for the bosses in Table 10. Please recall that the bosses were asked to rate themselves on leadership dimensions (with respect to their secretaries), and so the results of such ratings may be expected to be much less accurate than the secretaries' appraisal of the same leadership characteristics. The secretaries' coefficient alphas are all reasonably high (.80's and .90's) except for the moderate pretest 86 TABLE 8. Reliability Estimates for the Job Motivation Index Sample .u Coefficient Alpha Secretaries, Pretest, Matched Sample 16 .77 Secretaries, Pretest, Expanded Sample 29 .70 Secretaries, Posttest, Matched Sample 32 .78 Secretaries, Posttest, Expanded Sample 49 .73 Bosses, Pretest, Matched Sample 14 .41 Bosses, Pretest, Expanded Sample 18 .24 Bosses, Posttest, Matched Sample 28 .61 Bosses, Posttest, Expanded Sample 37 .62 87 TABLE 9. Correlations and Reliability Estimates for the Leadership Dimensions, Secretaries Rating Supervisors Consider- Partici- Initiating Scale ation pation Structure Pretest, Matched Sample n=16 Consideration (.85) Participation .91 (.80) Init. Structure .74 .64 (.42) Pretest,_Expanded Sample n=2 Consideration (.90) Participation .69 (.80) Init. Structure .80 .62 (.53) Posttest, Matched Sample n=32 Consideration (.84) Participation .42 (.82) Init. Structure .48 .37 (.67) Posttesti_Expanded Sample n=4 Consideration (.80) Participation .40 (.81) Init. Structure .53 .53 (.62) 1Correlations corrected for attenuation with coefficient alphas in the diagonal. 88 TABLE 10. Correlations and Reliability Estimates for the Leadership Dimensions, Supervisors Rating Selves Consider— Partici- Initiating Scale ation pation Structure Pretesti Matched Sample n=14 Consideration (.60) Participation .23 (.75) Init. Structure .47 —.09 (.42) Pretest, Expanded Sample n=18 Consideration (.73) Participation .50 (.79) Init. Structure .72 .41 (.48) PosttestpiMatched Sample n:28 Consideration (.76) Participation .29 (.87) Init. Structure .87 .63 (.46) Posttest, Expanded Sample n: Consideration (.76) Participation .34 (.86) Init. Structure .83 .60 (.47) qurrelations corrected for attenuation with coefficient alphas in the diagonal. 89 initiating structure reliabilities (.42 and .53). The intercorrelations among the three dimensions do not show adequate stability from the pre- to the posttest. This may simply represent an artifact of dealing with small samples in the pretest. Looking at the posttests, however, (which are probably the most reliable of the groups) we see moderate correlations among the scales, suggesting that, at least for these samples, participation does appear to be a viable third dimension of leadership. The boss coefficient alphas are reasonably strong (.60's to .80's) except for the initiating structure scale, which is only moderate. The boss samples do show more stability on the measures from pre- to posttest than did the secretaries, but there still exist a few problems, especially the participation, initiating structure for the matched samples. Again, participation appears to be a promising third dimension of leadership, even when dealing with self-reported scores. Job Effectiveness Measure Reliabilities Because single item questions were used to assess quantity and quality of work produced, no measure of the reliability of these questions can be reported. Unfor- tunately, Mott (1972) does not present any sort of relia- bility data for his questions. The best estimate that can be made for the questions' reliability, and it is admittedly a far from satisfactory 9O suggestion, is to look at the two questions together as a "job effectiveness" cluster. While it is fully recognized that quantity and quality are not necessarily responded to by subjects in a similar manner, the moderate correlations both reported by Mott (1972) and found in the present study suggest that there might be some similar response to the two items. As simply suggestive data, the correlations for the various samples for the quantity and quality dependent variables, along with the coefficient alphas for the hypothe- sized job efficiency cluster are presented in Table 11. Relations Among the Dependent Variables Tables 12 and 13 present the overall correlation matrices for all four samples for all dependent variables on the posttest. Although several variables (viz., secre- tarial job duty clusters, leadership dimensions, and job efficiency items) were "negatively scored" in the data coding (i.e., the lower the score, the more enriched job duties assumed, the higher the leadership score, etc.), this correlation matrix has been presented as if all items were "positively" scored (i.e., higher score=more of the trait). Thus, each variable's correlation is interpreted in the intuitive direction: positive correlations indicate the variables vary in the same direction. It will also be noted that this matrix provides a summary of the coefficient alphas for the various scales in its diagonal. Correlations in the parentheses in the matrix have been corrected for attenuation. 91 TABLE 11. Inter-item Correlations with Coefficient Alphas, Hypothesized Job Efficiency Cluster1 Quantity/Quality Coefficient Sample Correlation ,m_ Alpha Pretest, Secretaries, Matched Sample .37 16 .54 Pretest, Secretaries, Expanded Sample .42 29 .60 Posttest, Secretaries, Matched Sample .55 32 .71 Posttest, Secretaries, Expanded Sample .50 49 .67 Pretest, Bosses, Matched Sample .57 14 .73 Pretest, Bosses, Expanded Sample .60 18 .75 Posttest, Bosses, Matched Sample .41 28 .58 Posttest, Bosses, Expanded Sample .35 37 .52 qcorrelations are corrected for attenuation. 92 .mHoqowwfio 03p QH ohm m.msmam pdoaoawwooo mhpflfiflnmflaoasd How popoonnoo soon o>ms momonpdomwm ma mHOflpmHoHHOOV Arne Acovmo Acevmv Aomvmr Acovrm Ammvmv Amrvme Amevmr Avsvom Ammvmm Acmvom soaoataoos .re Arnv Ammvmm Ammvnm Amovmm Armvom Aomvmm Ammvmm Amnvos AerOV Aervmo seasons .0: Arnv Anovos Ammvmm Aomvmm Ammvnr Amovto Ansvmm AvaOV Amovso suspense .m Anov Anmvnm Awsvom Armvnm mowers Astemo Amrvmr Ammvme .ooosom .pasH .m Ammv Amsvom Anevsr Asmvss Amnvrm Ansvos Ammvmm cosponaoansom .n Aswv Asmvmm Assvmm Acovms Amsvms Avovmm soapssooancoo .o more Aosvmm Ammvmm Awmvmm Amovsm .satsdmque .m Aowv Ammvnm Arovmm Aocvmm nsoanne .s Armv Asmvmm Amwvmm .nsnn m\m .m Aomv Armvvw .nqu .tsom .m Amwv .oson .snm .r mmuq oaoamm ponent: vr Or m m .5 w. .m .d. m N r oflnmfimw> vmofiamponoom .moHQwHHm> psoosomom Hag mnoa¢ mnOprHonnoo .mr mamas 93 .Nfinpma nOHpmHoHHoo mamamm pooqmmxm map Ga wooofiosfl ohm pmoponm osp so moouoHoMMHo unrefiwflqmfim on oozonm pomp moapwflhm> hanom Anne Acnvms Aoovms Anovms Amsvmm Aoovms Ammvom Awmvmm seasons .w Ammv Ammvnm Ammvom Amwvss Anuvns Ammem Ammvmm .posmpm .pasH .n Aomv Ammvmn Aemvrs Aosvmm Amsvwm Ammvos coacesoeansoo .6 Acre Aoovms Ammvvs Aosvsm Aosvms .sntsdmuHme .m Aomv Ammvnm Aomvms Armvmm xsoanns .4 Aoov Aomvno Ampere .nacq m\m .m Acme Ammvmm .aaem .psom .m Ammv .ospn .som .v mean Nodmawm doodomxm m n o m s m m c od8n3> omfidflpnOOII .Nr mag .mamoommflp can as who m.msmam pdoaoaemooo mapfiaflomflaonqo How popomssoo soon o>ms memosonHmm ma macapmHoHsOOV Arcs Amovmo Avwvns Asmsse-aomvoe Amevoo Aemvmm Asmicr.avasv Avamo coanetacoz .0v Ammv Arsvsm Ammvmv Amovmo Amsvmm Asovmo Asmvrm Ammvom Anmssmu moaacsa .m Ammv “sovmo Amovmo Ammvrm Asmvmm Anovso Aomvmm Ameomv- snaocosa .m mosv Amovos Anwvrm Aomvmm Ammamr-amoVeo Asmvrm .posscm .pasH .3 Anne Ammvsm Awmvmm Asmoov-astmr Ammvne soacaoaoaonem .o Aonv Arcaoo-amovmo Armvmm Aomvmm scanssoeansoo .m Aowv Arowvo.amovmo Asovmo asozuHme .4 Anmv Aomvnm Ammvmm .nspm mxm .m Ammv Ammvmn .oatm .nsom .m Arnv .nspa .ssm .r wmno camawm concur: Or a m a o m s m m c odtr8> r whomfi>momom .moapmflaw> psoosomom Had waoa4 mdoapmaonnoo .mv mqmaa 95 .xfinpma soapmaonnoo mamaom dooqmmxm ms» ca ooosaoofl ohm pmopoam esp so moodosowmao pddoagadmflm on oozosm pomp moapmflaw> madom Amov Amovmo A4mvo4 mo4vmm Amrvmo Aomvmv Aomimr.armvmv Armv4r soapstapoz .m Ammv Ammev Aoovmo A4mvmv Ammvmr Ammome.ammvme Aoovoo steamed .m Ammv Ammvnm Aerve Amovmo Acraoo-am4vom Antone- suspense .6 A648 Amwvom Aomvmr Anmwmr-aoovoo Aomvmv .oosscm .panH .6 Aonv AeoVro A4ovmo Aorvme Amrvmo soanosooanooo .m Aowv Anewmo-aoovoo Aoovmo nsozane .4 Aoov Avmvmm Aoovnm .osom m\m .m Aomv Ammvmn .ospm .esom .m Arne .oeom .ssm .4 emu ofiosem noeoodwm m m 6 o m 4 .m .m r odtfi8> oesoaosoouu.mr mamas It. .1 () 96 For both the supervisors and secretaries, the pattern of correlations that emerges is of interest. Most relation— ships are as expected (e.g., quantity correlates with quality, job satisfaction with motivation, etc.). Of par- ticular interest is the pattern of relationships for the job duty clusters. For both secretaries and their bosses, the reporting of more job duties being accomplished by the secretaries correlates positively with job satisfaction, perceived leadership of the boss, and motivation, although a few exceptions to this generalization are noted, espe— cially in the case of the supervisors. The so—called "social—personal" job duty cluster is particularly unpre- dictable in the supervisory samples. But despite the exceptions, the findings with respect to the job duty clusters lends some evidence to the validity of the instru- ment. It is also worth noting that the moderate correlations noted for the quantity and quality measures lend support to the use of the "job efficiency" cluster used to test for the reliability of the measures of quantity and quality. Lastly, it will be noted that the correlations among the various dependent variables make obligatory the employ- ment of multivariate analyses for testing for group differ— ences, with the exception of the enriched job duty cluster, as noted above. N1 97 Management Support Groungypotheses Eyppthesis 1: Ex erimental groups (exposed to training sessions will self-report accomplishing more duties in the enriched job duty cluster post training than will control groups. Results of a two—way, two-by-two analysis of variance for the enriched job duty cluster are reported in Table 14. A significant F ratio is reported for the expanded sample training effect (.028 significance level), and a borderline significant effect for the matched sample was obtained (.083 significance level). The differences were in the predicted direction. In addition, it is of note that there is neither an effect for pretesting nor an interaction effect. Tempering these results somewhat, however, are the Eta2 values, which indicate that only a relatively small portion of the variance (9.5 percent for the Matched Sample and 9.9 percent for the Expanded Sample) is explained by the training session effect. The pretest effect accounts for virtually no variance, and there is a similarly small inter- action effect accountability for variance. Unfortunately, though not atypically, most of the variance is attributable to within cell variation, i.e., individual differences. It is also noted that there is a tendency for the standard deviations of the pretested groups to be lower than the standard deviations of those groups not pretested. ,If significantly different, this would imply that the effect of pretesting was to make the groups more homogenous in their responses, an artifact attributable to the pretesting. 98 TABLE 14. Univariate Analysis of Variance, Enriched Job Duties Composite, Secretaries Uni- Signifi— 2 Factor df MS variate F canoe Eta Matched Sample Pretest 1 .0004 .008 .928 .000 Training Session 1 .148 3.229 .083 .095 Interaction 1 .115 2.408 .132 .074 Within Cells 28 .046 .830 Cell Mean Cell Std. Dev. Pretested Experimental 1.580 .179 No Pretest, Experimental 1.454 .234 Pretested, Control 1.597 .170 No Pretest, Control 1.709 .261 Expanded Sample Pretest 1 .023 .482 .491 .009 Training Session 1 .248 5.138 .028* .099 Interaction 1 .069 1.430 .239 .028 Within Cells 45 .048 .864 Cell Mean Cell Std. Dev. Pretested, Experimental 1.500 .217 No Pretest, Experimental 1.483 .233 Pretested, Control 1.570 .177 No Pretest, Control 1.712 .244 99 Thus, a test for homogeneity of variance was employed to see if these standard deviations were significantly different. Because the formula* for this 3 test requires the use of a correlation coefficient, the test cannot be computed for the expanded experimental groups, in which the asymmetries of the design make a correlation coefficient impossible to compute. The.t test for the matched samples was .271 for the experimental groups and 1.096 for the control groups, neither of which is statistically significant. The control groups for the expanded sample resulted in a-t of .860, which was also not statistically significant. Thus, the standard deviations are not significantly different from each other, although there is a trend in the direction of increased central tendency responses with pre- tested groups, suggesting a potential problem in any future uses of this instrument. It will be recalled that the job duties composite is an average based on summation of the 1's (duty reported as being done by the secretary) and 2's (duty reported as not being done by the secretary). To test more directly the above hypothesis, an analysis of variance was performed based on the number of duties the subjects reported accom- plishing: i.e., with no consideration of the duties not being done. The results of this ANOVA are reported in 100 Table 15. Here it will be seen that the F ratio for the training session was not significant for either the matched or the expanded sample, although in the case of the expanded sample, the F does border on significance, and in the pre— dicted direction. This suggests that while the job duty composite was significant, that this significance was in- fluenced, in part, by the job duties mpEDdone, since they entered into the computation of the average. This implies that the hypothesis is only partially supported, that there is a tendency to assume more job duties as the result of training, but that it is not a clearcut, definitive trend. As in the composite ANOVA, the standard deviations for the number of enriched duties analysis suggests that the effect of the pretest for the control groups was to decrease the standard deviation. This hypothesis was again tested by the apprOpriate t test for homogeneity of variance, which yielded a t of 1.1915, which was not significant at either the .01 or .05 probability level. Despite this lack of significance, there is a trend for the standard deviation to decrease with the pretested group vis-a-vis the unpre— tested one. This suggests that there was a tendency on the part of those pretested to report in the posttest more central tendency responses, perhaps indicating a frustration with the questionnaire and a tendenCy to answer in a uniform manner on the second taking of the instrument. Thus, the data for the first hypothesis, in summary, suggest that the training session in job enrichment was 101 TABLE 15. Univariate Analysis of Variance, Number of Enriched Job Duties, Secretaries Uni- Signifi— 2 Factor df MS variate F cance Eta ‘ Matched Sample Pretest 1 10.125 .1546 NS .005 Training Session 1 91.125 1.391 NS .044 Interaction 1 112.500 1.717 NS .055 Within Cells 28 65.509 .896 .Cell Mean Cell Std. Dev. Pretested, Experimental 17.125 7.32 Not pretested, Experi— 19.750 6.41 mental Pretested, Control 17.500 6.89 Not pretested, Control 10.950 12.63 Expanded Sample Pretest 1 150.19 2.069 NS .041 Training Session 1 249.91 3.443 NS .077 Interaction 1 29.84 .411 NS .008 Within Cells 45 72.58 .884 Cell Mean Cell Std. Dev. Pretested, Experimental 20.45 8.80 Not pretested, Experi— 19.55 7.33 mental Pretested, Control 18.56 7.18 Not pretested, Control 12.33 10.28 102 effective in implementing some change, but the trend is not an unambiguous one. Hypotheses 2-6: There will be no significant differ- ences in experimental and control groups after training on: 2. Routine job duty or social/personal job duty composites; 3. Self—assessed job performance measures (quantity and quality of production); 4. Secretaries' ratings of their bosses' leadership dimensions, viz., consideration, participation, and initiating structure; 5. Job satisfaction variables (satisfaction with work and with supervision); and 6. Motivation. Results of the multivariate analysis of variance (MANOVA) for the above dependent variables are reported in Table 16. As hypothesized, multivariate F's for both the matched and the expanded samples show no significant results for the training factor, the pretest factor, or for any interaction. Because of this lack of overall significance, it is not possible to look any further at the individual variables to determine whether or not any of them changed. What is implied by this finding is that the training session did not have any effect on the variables listed above, which includes primarily relatively stable organizational dimensions. Further interpretation of this finding, as well as its limi— tations, will be made in the Discussion section below. 103 TABLE 16. Multivariate Analysis of Variance, Secretaries F Ratio for Equality Significance Factor of Mean Vectors (p less than:) 1 Matched Sample n=32 Pretest .802 .630 Training .722 .695 Interaction .986 .487 Expanded Sample2 n=49 Pretest 1.126 .367 Training 1.574 .172 Interaction 1.29 .283 1Based on all dependent variables except the enriched job duty cluster. 2Only variables that showed no significant differences on the pretest are included in this analysis. 104 Supervisor Hypotheses Hypothesis 7: Ex erimental groups (exposed to training sessions will report that their secreta- ries accomplish more duties in the enriched job duty cluster than will control groups. Results of a two-way, two-by-two analysis of variance for the enriched job duty composite are reported in Table 17. It is noted that there is a significant F ratio for the training session effect (p<.024) for the matched sample and an F ratio of borderline significanct (p<.070) for the ex- panded sample. Both of the effects are in the predicted direction (recall that the lower the score on the super- visory job duty composites, the more duties the secretaries were reported to have assumed). As with the secretarial samples, the Eta2 values indicated the significance of the effect are less high than desirable. Thus, the training session accounts for only 14.5 percent of the matched sample variance and 9.1 percent of the expanded sample variance. Again, most of the variance is represented by individual differences (i.e., within cells variance). This suggests that, while there is clearly some effect attributable to the training, it is, practically, less of one than the statistical significance would imply. It should be noted, however, that the Eta2 values obtained are quite typical of field research studies. There is a tendency in the supervisory samples for the standard deviation to be higher for the pretested con- trol groups than for the unpretested control group. In the 105 TABLE 17. Univariate Analysis of Variance, Enriched Job Duties Composite, Supervisors Uni- Signifi- 2 Factor df MS variate F cance ' Eta Matched Sample Pretest 1 .036 1.06 .312 .027 Training Session 1 .195 5.78 .024* .145 Interaction 1 .008 .24 .629 .006 Within Cells 24 .046 .822 Cell Mean Cell Std. Dev. Pretested, Experimental Group 1.47 .190 No Pretest, Experimental Group 1.51 .191 Pretested, Control Group 1.27 .200 No Pretest, Control Group 1.38 .150 Expanded Sample Pretest 1 .0002 .0045 .950 .000 Training Session 1 .123 3.510 .070 .091 Interaction 1 .068 1.952 .173 .051 Within Cells 33 .035 .858 Cell Mean Cell Std. Dev. Pretested, Experimental Group 1.480 .156 No Pretest, Experimental Group 1.409 .220 Pretested, Control Group 1.272 .200 No Pretest, Control Group 1.378 .150 —_¥ *1) <- 05 106 case of the experimental groups, the deviations are ap— proximately equal for the matched sample, but lower in the pretested group for the expanded sample. The significance of the difference for the control groups' standard devi- ations were tested by the apprOpriate p_test, which yielded a value of .6976, which was not significant. (Note that the control samples were identical in the matched and the ex- panded samples; thus, only one §_test was performed.) Observe too that the standard deviations for the expanded experimental groups cannot be tested for their difference's significance because of the asymmetries of the expanded sample cell frequencies. An analysis was also made of the number of enriched job duties reported by the supervisors to be performed by the secretaries. The results of this ANOVA are presented in Table 18. Here we see that there is a significant training effect for both the matched sample and the expanded sample, accounting for 16.7 percent and 11.0 percent respec— tively of the total variance. Again, these Eta2 values are less than ideal, but still high enough to make the statis— tical significance finding of interest. The finding of a significant training effect for the number of duties assumed by the secretary (as reported by the supervisor) suggests an apparent contradiction in light of the non-significance of the training effect for the number of enriched duties reported by the secretaries. One eXplanation might be that bosses are less aware of their 107 TABLE 18. Univariate Analysis of Variance, Number of Enriched Job Duties Performed by Secretaries as Reported by Supervisors Uni- Signifi- 2 Factor df MS variate F cance Eta Matched Sample Pretest 1 57.14 1.125 NS .037 Training Session 1 264.14 5.201 * .167 Interaction 1 14.29 .281 NS .009 Within Groups 24 50.786 .784 Cell Mean Cell Std. Dev. Pretested, Experimental 17.71 7.52 Not pretested, Experimental 19.14 7.71 Pretested, Control 10.14 7.52 Not pretested, Control 14.43 5.53 Expanded Sample Pretest 1 1.706 .050 NS .001 Training Session 1 159.38 4.659 * .110 Interaction 1 130.774 3.823 NS .090 Within Groups 34 34.207 .799 Cell Mean Cell Std. Dev. Pretested, Experimental 18.36 6.23 Not pretested, Experimental 14.92 8.26 Pretested, Control 10.14 7.52 Not pretested, Control 14.43 5.53 *p<-05 108 secretaries' activities than the secretaries themselves, seemingly a reasonable prOposition. It should also be re- called, however, that the scoring procedure for the super- visors' questionnaire was such that a job duty reported as being done by both the secretary and her supervisor was scored as a duty done by the secretary, thus providing a lenient interpretation of the secretarial duties. Apparent differences between cell standard deviations are noted for the matched and the expanded controls and the expanded experimental groups. Although the expanded experi- mental groups cannot be tested because of the unequal n's, the t.test for the control groups had a value of .7328, which was not significant. (A p_test run on the matched experimental groups had a value of .0567, also not significant.) In short, there is some evidence that bosses exposed to training report their secretaries as performing more "enriched" duties than those not exposed to training. The composite enriched cluster, which includes the duties done by the supervisor himself or not done, in the average, also resulted in a significant training effect. The number of duties analysis supported this finding by showing a signi- ficant training effect for the number of duties assumed by the secretary. Eta2 values for this effect, especially in the number of duties analysis, indicate that the effect was Alreasonably strong in terms of amount of variance accounted fcug but far from ideal. 109 Hypotheses 8—12: There will be no significant differences between experimental and control groups after training on: 8. Routine job duty or social/personal job duty composites; 9. Self-assessed job performance measures (quantity and quality of production): 10. Supervisors' ratings of their own leadership ratings with respect to their secretaries on three dimensions: consideration, participation, and initiating structure; 11. Job satisfaction with work; and 12. Motivation. Results of the MANOVA's for the above dependent vari- ables are reported in Table 19. As hypothesized, multi- variate F's for both the matched and the expanded samples show no significant results for the training effect, the pretest effect, or the interaction effect. Again, it is not justified to look at the results of specific dependent variables in a multivariate analysis when the overall F is not significant. As with the secretarial samples, the results of this finding will be further dis- cussed in the Discussion section. 110 TABLE 19. Multivariate Analysis of Variance, Supervisors F Ratio for Equality Significance Factor of Mean Vectors (p less than:) Matched Sample’1 n=28 Pretest .149 .996 Training .579 .786 Pretest x Training 1.370 .254 Expanded sample2 n=52 Pretest .508 .848 Training .719 .685 Pretest x Training 1.069 .434 1Based on all dependent variables except enriched job duty cluster. 2C)Illy variables that showed no significant differences on tile pretest are included in this analysis. DISCUSSION In general, the hypotheses of this study were con- firmed. The enriched job duty composite showed statisti- cally significant differences between experimental and control groups for both the secretaries and their super— visors, and in the predicted direction. The number of enriched duties analysis, however, showed statistically significant results only for the supervisory samples, but not for the secretarial samples, although the results for all samples were in the hypothesized direction. This, over- all, provides general, but not unambiguous, support for the efficacy of the training on the enriched job duty variables. Such a finding must be tempered by the size of the effect, however, which was far less than Optimal, as the Eta2 analysis attests, although not particularly atypical for field research. On the other hand, the multivariate measures showed .no significant overall F, and hence no change on these dfiflpendent variables can be attributed to the enrichment trxaining, as predicted. There is, as noted in the Results ISecrtion, a possibly serious defect in hypothesizing regarding tile? enriched job duty variables in relation to the multi— van:‘Ziate analysis. Because it is inappropriate to include VaI‘iables hypothesized to change in a multivariate analysis ()1? (dependent variables hypothesized not to change, univariate 111 112 ANOVA's were performed on the enriched variables. However, the strong correlations between the enriched and the other job duty clusters suggest that the enriched composite is not a distinct entity. Hopefully, future uses of the job duty instrument will be performed with sample sizes large enough to permit blind cluster or factor analysis of the duties, resulting in statistically distinct job duty dimensions. The preliminary conclusion that emerges from the present study is that job enrichment training similar to the type studied here is best suited for specific job duty changes, and that few changes should be expected in terms of attitude changes, perception of bosses' leadership, or per- ceived job efficiency. Such a conclusion must be moderated, however, by other considerations. To begin with, it may be argued that the effects of training are long-term and thus were not captured by the Present study. Although the posttest was conducted one month after the training, it is possible that the training's re- c3C>Inxnendations had not been implemented by then, but rather required a more leisurely beginning. In some training Se ssions, for example, units are built into the training that may not be needed until some time after the session. Thus, conceivably, a questionnaire administered six months, or eVGIl a year, after the training session could result in more Cunetilflges being attributable to the training than the one-month- after posttest. The emergence of any such time—delayed changes is, however, by no means a certain, or even probable 115 phenomenon, but it is certainly an unknown in the present research. In a future study, the addition of a second, or even a third, posttest some time after the original should be able to capture the emergence of any training effects over a more extended time period, as well as the stability of any changes over time. Even if repeated questionnaire administrations revealed no attitude or job duty changes, training such as that studied here is not necessarily of no value. Not to be ignored is the possibility of a cumulative effect of training. While one individual training sessions, isolated in time, may have comparatively little, if any, direct results, it is certainly arguable that a series of training sessions may, collectively, cause considerable changes. The research in this study was not, of course, designed to measure such a phenomenon nor were the companies involved here conducting a long term training series for their secretaries. Indeed, such a longitudinal study would likely be quite difficult to imp]. ement in a field setting. However, management contem- Plating the implementation of training should not ignore the POSSibility that seemingly ineffective training may be more efficacious than first appearances might suggest, especially if it is part of a series of employee develOpment exercises. Finally, there are those who would argue that a train— ing; session is merely a means of demonstrating management's authentic concern for the welfare and development of its employees. Under this view, it is relatively unimportant 114 that a given session shows no change in organizational dependent variables; it is sufficient that employees view their management as interested enough in them to undertake the expense of training sessions. Such an approach may be good.management, but its unscientific nature makes it :relatively immune to serious testing. What about the job enrichment aspects of this study? 1k; job enrichment, as the better studies to date have shown, Luilikely to result in any significant changes in perceptions of one's job, boss, and work attitudes? This is a difficult queestion to address directly. The design of this study looked at 'two sets of results: specific job duty changes resulting frconljob enrichment training, and, the motivational, produc- ticxn.effectiveness, leadership, and attitudinal changes re- sulxting from the training. Although differences were obtained in 'the predicted direction for the enriched job duty cluster, thcnugh not in all cases of statistical significance (parti— culzarly in terms of the number of job duties assumed), no difiTerences between control and experimental groups were notemd.for the other variables. Because of the small magni— 'UM162 of’the size of the group differences on the enriched varfieables as reflected by the relatively small variance accc>11nted for by the training factor, it is at least arguable tha13 the seminars were not particularly effective in imple— menting job enrichment, and hence offer no real test of Whether or not job enrichment results in changes in organi- zational dependent variables. 115 Thus, conclusions from this study about the efficacy of job enrichment assume a causal model: training causes enrichment, and enrichment causes (more accurately, does not cause) changes in other variables. Since the implementation of enriched duties was limited in this study, conclusions regarding the effects of enrichment must themselves be of liJnited s00pe. However, to the extent that job enrichment teas induced, there appears to be little association between gkib enrichment and the dependent variables studied here. The question of the means of implementing job en— rixzhment is an important one. In one sense, the results in thxe predicted direction for the enriched job duty cluster is a rather strong indication that the training resulted in chsanges. Because of the variety of job duties included in thee enriched cluster, to get an overall significant F ratio sugggests that the experimental groups must really have been afjfected by the training. However, this conclusion must be cozrsidered in the light of the analyses of the number of job dutcies adopted, which is less heartening. 'What seems to emerge from the overall picture is that job enrichment can be iiJnculcated to some extent by a seminar approach, but that Other means of training should also be investigated that miEflflrt be more significant in terms of the size of the effect. It Eslnould be noted, too, that enrichment by seminar has aPplicability primarily to service jobs, Where there is very litVCCLe equipment and where there are relatively few employees inVTDZLved. A high level of fixed assets in a job, or an 116 interface between many different kinds of jobs, would seem to contraindicate a seminar approach to job enrichment. The flaws in this study must not be overlooked or ‘underestimated. Matching is inevitably a second—best substitute for randomization. Even though no pretest dif- .ferences emerged for the matched samples, the researcher is still never fully sure of the group comparability on non— deilineated variables, ones that may pose an undetected threat tc> an otherwise strong experimental design. Small n's are a]_so an unfortunate, if not atypical, defect of this study. Tklis is particularly relevant to the j0b duty measures, wtieie small sample size necessitated some rather crude teuzhniques to create clusters and estimate their internal lnsistency. The job efficiency measures are appropriate as ikanr as they go, but it would have been desirable to have some 'Wlsard" production measures. Admittedly businesspeople are venqy reluctant to permit access to such measures to researchers, anti indeed, few such measures exist for a secretarial position. NOIT, finally, should the limited generalizability suggested for? the samples be ignored. As noted above, this study desilgt with participants from Scanlon Plan companies, which Pofsssess considerable built—in rewards for improving produc— tiv‘ity and augmenting growth. If a similar seminar approach wexrse employed to enrich the jobs of secretaries in less co- OPEEIPative or participative climates, entirely different res‘Lllts might ensue. But for all its faults--and they are inevitably many-- 117 this study is still a step in the right direction. It is in accord with a long line of literature advocating tight experimental designs for training evaluations. It is in line with a shorter collection of job enrichment studies advocating tight experimental designs and multivariate data analysis. And, it is in accord with the more recent better— done job enrichment research in its findings, tempering somewhat the all-too-frequent overly zealous magnificence ascribed to job enrichment. APPENDICES APPENDIX A 126 05:00... wndpoeowu 0» £030..” 09.0903 .om memo swoon scam mphomoh ouficmmao .3 0090009 nudges no 003.8 3300mm 50930: .3 _. 0.3.3300 50:09 mo 00300002038.“ 3300an .3 pacific“ 80.2.80.” 90 0005 0030M “00.3.8 05 Mom 095909.26 090A P820 2.8 Amnasoa 020 906V 0wcouuwsoa no 093; campsite .3 Hlll IHH lllll vcosuhemoc mamanflonmmm on 0.4503200 0000:0950... H0585 355353 003MB 50950: .3 mouammsm ooflmmo H0090 .3 5.3000 moo.» Mom 0950000 90359 05.. goox 6:0 .mfiuoflna Ho 5.85594 .2330...ng swam 000, 09335 .9 mohafiosomxo pmuanosuoms: no momsmso 8.0% .8300 Mo 50pmanmondmm oxen .600 mtoaon 90359 303909 08305.. .3 common ooflmo cocoons banana: coo 33:38 .2 WI llllll IIIIII IIIIII soapouacomao 05¢ canvas hpfinonuso mo onsvosnvm finance 0 moanmaanmpmm .ofi IIIIII. IIIIII. msuom hummmooos 0H090Hm 0:0 momsogxo 000Ca059 chooom .m IIIIII IIIIII. mownonocapa may» 0H0m0Hm .m IIIIII .IIIII. wcoapo>n0009 coavopnommsmha 0x02 .5 IIIIII _IIIIII 0:0H90>H000H H090: 0x02 .m IIIIII IIIIII. monsoooonm one modoaaom dormaapovmo woo msaahuoo cu hhomwoooc ocsmhosos was 00soonommonnoo ouoauacH .m IIIIII IIIII: IIIIII mnozpo hp 00ammo op 009»«3990 shes ouosaonooo 0:0 .0H3oosom .vmoauom .3 .IIIIII IIIIII .IIII wvcoeooqoosso Ho hmoo compo 00:03 no msoauwarn oaapsm .Amoo poemHmwwao wsausaosaw madmapu0>00 ho mswpahz 0>dp00h0 .m IIIIII IIIIII 002000 0 Ho manonoa Mom Hodnopds Moreno .N nIIIII. msoaponomo pecan no soapoom weapoommaH .H Hmsammomm we moz watemmomm an em mzom ms mmmeHmz wm mama we hm mzom .coavmspam xno: :30 use» on m0aamm0 ea 00 anowoado 0pmanmonmm0 one x0020 omooam .moavaflapamsommou non mo pmaa wsaxoaaow 0:» :H lemon H 002220 Hammad 8650me WHHHIHI HI I l H! l1 uQ H QH0000 000>00 0:0 mphom0u #00009 0000000: 0uhH0:< 9000:: 0:000:000 000mmo 0H0Q0Hm hHH0HoH:0 0:0 000H00H0o 000000: 5:00:00 Ho 000mmo 00M000mm :H0u:a0z 0:0000000000 000:000n Ho 00000>0¢00 00>00 0.0>0»:00X0 Mom xuo: 0H0m0um 000000: x0» 0:00:0 0.0>0¢:00x0 mo :odp0u0monm :0 000000 000:0009 H0:00000 0.0>Hp:00x0 0H0:0m 0000000000 00: 0:0 000: 0:0 0>Ho>:0 00:9 0:00»0::m H00000 000:0000 0:0 :0H0 mamx 000%0H000 000mwo Mom 0:00»0::m H00000 000:0muo 0:0 :0Hm 0H0m 00>0¢00:000nm0u 20:00: .0909000> 0oahmo Mom 000000 H0HH000H000 500000: weuom hn000000: 0H0m0nm 0:0 000:0:X0 000:0009 0.0>Hv=00X0 0Hoo0m 000:000:090 :00» 00000:: 0:oap0>0000n :ofia0vnomm:0nv 0x0: 0:00»0>H000H H000: 0x0: 0:0030 00 0009 :00: A00000:000 H0990 0H0«p«:« :30 H00» 0:0000V 000: 0.0009 :MHm 0:000H0 00:00:000 :30 000:: 0:0900H 0» 0:0: :mHm :owp0enom00 000000000 0:0 0>0: so» :00: 0>0p0ap«:w :3o :0 0:0000H H030:< 0005 0:0 :0 m0c0o0 0:00:0050 0:000:00: “00:00:0920 mo H0000 :0 .HH0S w:ano:0 mo 5:00:00 th00 0 0x0: H00: 0:000000 0H0:0: 0H00s m:0soo:« anew 0:0 000m 00:00:0080 000000esa mo 0:000 Mom 0H0m H000000 0 :00p:00: 0:000:0090 0.0009 Mom 0:000 mo 5009000: 0:00:0000 moaam ooammo :00p:00: 0:0 0: 90m 0x009 no .mn0pvoa .00H00090 mo 00090::00 no 000000 0H0a0nm 0000 00000000 :00p:00s Ho\0:0 000: 00 0>00000X0 How 00H00000 00H0 no #00: 0:00009090 000mmo 000>h0mzm on H 9490 .302 mHmB 0Q .mm .mm .0m .Nm .0m tr:- owu II \u nd. .11 I..- 420 H.zoo H oom< QMHmHA HO? Hum: Ho no consume wanuouono aonu connonowv mannumaonn vow uxoe .mm anon a no hao>nnoommo onQa connoaom emu no» so: unnanunov on moon can: nHo>nouofi noaowmn no yum: .nm Boon unannoa on Honnmnma manommn :nouanaz .om vuvoou nowaoa on moonnannumnsm oaowuaouono “momma ounmuo Janna no» «monumwma ou mannomnom .mm «Hanuno no vmon on wood mmnu an monoov on roan manwao on Amvmbnusouxo cu moonnwwwa mo ounonaoo mo nuanmu wumaaunno .om oonmwo an oaonuuaom monomoxwmsom .mm an nwuwo no omoao.woo .nn opnum “.0n0 .monnuma wanvau>_aonm anouuo no oomuou uxoz .Nm .Ao.mmon napo no ononaawno :30 no>ov omnooooona can umnonaou ounmnanoumw n50 manhnnwo on knoomwuu: «coonoama can ouoooaommonnou wnonnnam .nm vooumd up Hans muonnonuao so» conuqanomon now muaoaunauoo nonuo nonuon. .om m~o>n0uow noazwmn no amusn mwn>mn “monumo wnu now mobnuuofino owcmn unonm van nonuaoa Mam nm>ov omoonlwaoa «o manna anmunnuz .m. mnmsuo an o>nuoomxo on vununafiom xnos unmanvnoou odd .waooozuv .nomovom .m¢ ououaaoaoooaw no 5900 oowno «moo; no moonuuaon unanon .Amom oonmnmoonu monooauanv mononunm>va mo monunns m>nnmono .5, uduaunooov ouonnnonmou on munnoaoaoo oonoamnonoa Hansonu “nouaanoVw munmmo unmannmz .o. monauaou monumo noono .m. conuuoo nook now uuaaouuu nomwfin man @003 can .aHHonhma no .onuno:o> .oaonunowavun awn» can unmannaH .wa on H nwnuuaononoon Susana .mnonnun> monmuo non monuav Huwnuuonoom anownom .nuon any nmnmno on mannmuoo no connmanoman some as anuuno .nmoa mom nuns nounun> owns MuHAnmooa don: hHHmcomnoo monnuav new no anemone .unonwun> oonuwo wcn>nmuon an .moanu monmmo unannnoa was a: now .ounmmo an muonuonsm mannomxomaom . .conuannnumnv unease use mvnouon wnnmouxoenu mavens .mHHuu oaocnonon wsnowuao was mafiaoucn vacuum .nunuoomm vaunnnn no .mnouuua .munoqon wanunvo now annannnmaonmun Hanna o>mm .mnmnuo up o>nu=omxo on wouuaanau xnoa onuanvnooo use .oaavosou .nmoavom . «as use an nuance unannonan unwnanwnn .ooaounoaam mo novno an .Hnua mafiaouan no knoaaan Madam a 03m: .uammno an anon aoa3 Aonauoawnm nonum mamnunan :30 noon wanvvmv mama u.mmon swam .oann mo nuwcma oanwaouounn: no menu mnonwmn> cog: xwmu hunnonnm ucown: nuoa axon mo moon vanaom .msnauun awn uaooan n.u>nuaooxo mo conumnmmonm on umnmm< .munuananooeu vnooun was mxmz .mwsnuoua now monaana anemone can mono: mxma .uaonuanmeo uowuuo omn>nomam .mnooa on: new moon osu 0>Ho>uw nuns mdonuonbu Hanson oundwwno can amen mama. sameness announce a.o>«u=ouxo saunas .hd .oH .MH .cH .ma .NH O. O MGMO PIN .nofi :30 noon QM! managed on we hnowouoo oumnnnonmno man sound madman .monnndnannsommon anomaau unuamwoaua mo unnH wnwaoudou can on N EHOh N men<220nnmmoo maomu Hmonmam H2m2muwn:ooxo unoooM .unoannummv onunnnonnno on unanmanaou mouuauuunua Hosanna “nomamnauo oonwmo anmusnoz .mxoon no .mnmunoa .uoaonnnm mo monnmsaam no ummwnv unmounn .nooaonmmcoo new ownnnooa now endows onuoonm .un nonmm a: sound was .nn m>nom n.0uo .onnnuma wanvnu> sonu anuuno no moumoo use: .uonanaau monwmo novno .umnan Invnomxo vounnocuauaa no mowaosu mo nomn>nooaw omn>vo was mnnonon newts; unconnom unhaws< .Hnua wanowuso maven: manna monsoonn unom new comm .conuoom n50» now nuasooom nowvnn on» now: use .mHHonnmn no .onono=o> .uaonnnunaoon owns use onmnnnnn .oaoanoaou onu no>o no mangoes wannouonv aonu nonuuuonv onnnouumnu use oxma .ucownunoommo ammonmsn no wonun>nnuu on>no n.u>nu:ooxo non xnoz onmnonm .nonuoanOMnn vonnflunuonn no amonanoou now oonaom noaono on mnoaamo nowon «manuananu can: unomnaoh nuance on sonnua Inowan o>nw madame asonmoaou wanaooun noonom .vovoua newcoa on muonuonnomnam osannuoomnc “ovum: oonmuo Janna now moanuowoa on cannomasm H.200 n 04A0 .502 00 H.209 nnn n flaws. .302 00 H.209 H..ZQHQ H Emu—”3 «302 mHIH. 8 .mucoauoeaonua no mood newno oases no oaonumaon unapaa .Amuo vonmnmmmao wenvsuoenv wanmnnno>vu moannnnns o>nnoono .vovoun on HHfi3_ouuanonnau non connwanouen non mundannunov nonno noonnou .Am.unoa no>o no onnnucwnu :30 no>0v nonavoo Iona was nonunaon vocmnanouuo use wnnhnnuo an hnmouooun avdunoawa van ouaovaomnonnoo unmfiuuaH OQHSU .302 mnmh 00 .mm .cm .Nm .Hm .om .mm .mm .mm .oN .HN .oN .mH 1465 .munuooou no unnoaon now Honnonua nunnwo .uonnuanouan nonconvon man obs: non non3 o>nnmnnnon :30 no unannoa noauqd .wuov swoon sonu unnonon menu can sundowno .oanm wannmnsu anonnnga noxuuu noon on obnuaouxo now umaonuno undo no ans: .maohoagao ouwmmo now muonnuuSH Hanson onnoawno was none anew .mnOunnn> oonwmo m>nouom .nmon now connuanomnn naucnnnun sun3 mHHuo moanonw no noun w unwnonn wanna use a no monnSnon oHHuu o>un on owaunn< .ooohoanam 30a «0 nonumnaonno can mananmnn onu an nmnmw4 .unonno>nomon Hone: axe: .uaoaoouan can .wofl3on>noucn .wannnsnoon noncomnom .uonnunoannn annu onwmonm .nono: wanna no.oaonnoanuuen Huno.sonm.ononnon omooaoo .aoon wdnunm3 an Hunnonua mnnvmon cannons: .ouunnmonoaodn some use» unnoauanoamo wanmamon now hunannwunoauon o>u£ was ..uno .vonnauun man» no :nwsoH .ncuananoaom mo onoansn on we ooou>vm an donnmanouan anuuno .uuaoananonao wnnxma uH .omoona mo mounnoum now hunannnmnoa umon Hanna o>u: can nonuounanae now Hunnonma mane .uaonuupnomon nonnuunoewconn one: .uowvan wannunono ounwuo onoauna haduwnnan was onoaaoamo .Huwnunua unwnunaanu van noanmnonv axes 1: .ono>nonnn nudnwon no ouonn omn>un mounuuo can now uo>nuoofino omaun unonm use Ausnaoa Kan no>0v uncanlwaoa mo manna anuunnm: .noununuwaoo and uwanuoua now ovauwo can no naummu moanuuaou amen nnonnuuwwnm endows use: .unonnnu no vuun Ou raw: 5050 «a flvwovv 0n Evan aflnuaa 0n mmvubwuauoxo OD DQn—Hflflwa HO UHfiUQOU NO DQHDNU UUQHQOHHU H.200 H 0ws moon naom can 00% snap nos» unseen ”#000 u.mnon naom an neon manm mum wanvnaocn .oonmmo n00» an monnw Ham now :00H0 unnm= w anmnanmz .mnmonuon no .nonnuanoman .mnuvcnaon no moon on monon naanno> swan nonumnv connnn3 0202 .Aznnmzz cans nusumn :.moson hnmz:v onognmaou :0 mama Hflam ha meow hwnunmva .nao neon mnouuou Hanna can no no: .oosovnoamonnou no monuoo anm :0 hano Amnmnunon ouaonomonv nonmunwononm no nmnmnn no wamnnnun nsm .onnnmawnm m.mmon nom Amuuonv nwson swan nonnmnv monaoo Hucnm damn .mnonnmn on momaoamon munmonEou 0H .uanu ao>nm u an unseen mu moan onos o>mn no» 0023 manna ovn>nwun0m .2000 0:0 Eonu hose mnm 00% mafia oxn unwananc 00 nmvno an .mwcmnno nosno can no .mocnsome canneou 00 Hmnnoums 0200 on mofinn onmnummw no namnvn>nvan unmnuomm now oucmnn< . .Hmnnonue.wonmn mo monnou Hanna ammuono on Aocnnoma wananu onumaouamv swanmmooonm 0n03: Aon Hannuuufi nflanam nov on: .uunoaon manaouoo on nouanaou on: .0500 mommon :cwnn: on neon» onsuuawnu nonnsn on: .onanmawnm one nouns Amnmnnnan ouaunononv naunnnan noon wanna uncanns unannon on 0200 o.umon swam .wannnu new wanna: now hundnanmnonuun munvnnonn .momhoanam unnemnwocmnm noxvsm Haonnoau oun>naaam .BGUu u no haopnnoomuo onoa nonuunaw :00 00% 30: onnanonov 0n moon snn3 unannonan nwdnwon nu noes .mocunnoaan ouanvoaaa mo mawnn now cane Hanooam a anmndnma “nonuaonno «.0009 now maonn m0 hunnonnm mansnonun n.200 n 0¢mw .302 09 H.- ZOO one n mmnm; .302 00 H.200 .ann H mmHz on H n _Good quality (3) :Fair quality (4)_ ' *Their quality is not too good (5) _Their quality is poor 2. Even though additional work might be involved for you, how capable do you feel you are of assuming additional job duties? (1) I feel I am. extremely capable of assuming additional job duties (2) ____Veryc capable (3) ____Moderately capable (4) _;__Not very capable ' .. (5) ___.Almost no capability of assuming additional job duties 2= 3. How often do you do some extra work for your job which isn't really required of you? (5) _Almost every day .(4)~ ”Several times a weekf (3) *About once a week- 1{2):. Once every few weeks. ' r (1): About once a month or less 4. If you were to assume additional job duties, especially some of the things now done by your boss, how much more interesting would this make your job? (5) Wouldn' t make it any more interesting at all (4) Not very much more interesting (3) Moderately more interesting (2) Quite a bit more interesting (1) A great deal more interesting 5. How much do you think your boss would like for you to assume additional job duties, especially some of the tasks he now does himself? (1)“ I think my boss would be extremely willing for me to *take on additional duties , (2).___ Very willing (3) ____Mbderately willing (4) ___ Not very willing (5) ___ Very unWilling 6. Thinking now of the various services produced by you and your boss, how much are you producing? (l) ___ Our production is very high (2) It is fairly high (3) It is neither high nor low“. (4) It is fairly low (5) It is very low 7. 10. 449 How much would you like to assume additional job duties, particularly some of the tasks now being performed by your boss? (5) _‘__I would ng£_like to assume any additional job duties. (4) I would moderately dislike assuming additional job duties (3) Neither like nor dislike (2) I would moderately like to_assume additional job duties (1) I would very much like to assume additional job duties . On most days on your job, how often does time seem to drag for you? (1) ___ About half the day or more (2) ____About one third of the day (3) ___ About one quarter of the day (4) ___ About one eighth of the day (5) ____Time never seems to drag In the following two lists, put a "Y" (for Yes) beside those items that describe your work situation, an ”'” (for No) beside those items that do not describe your work situation, or a “?” if you can't decide. (When you are finished, each item in the lists should have either a "Y," "1“," or a H?" by it)o_ ‘ _Y = Yes N a No ? 8 Can't decide Your Work . Supervision Facinating Asks my advice Routine Hard to please Leaves me on my own Around when needed ,.Lazy Inefficient :::: Satisfying :::: Never satisfied ____Boring ____Impolite ___~ Good “__ Praises good work ____Creative *__ Tactful ____ Respected ___ Influential ____Hot .___ Up-to-date ___ Pleasant ___ Doesn't supervise enough ____Usefu1 ____Quick-tempered ___ Tiresome ____Tells me where I stand ____ Healthful ____ Annoying ____Challenging ‘_‘_ Irritating ____On your feet ____ Stubborn ___ Frustrating ___ Doesn't listen ___ Simple __~_Knows job well __' Endless a___ Bad ____Gives sense of accomplishment ____Intelligent Would you say you work harder, less hard, or about the same as other people doing your type of work at your organization?- (5) Much harder than most others (4) A little harder than most others (3) About the same as most others (2) A little less hard than most others (1) Much less hard than most others 11. 12. 13. 14. 15. l6. 17. 44{) Some people are completely involved in their job-~they are absorbed in it night and day. For other peeple, their job is simply One of several interests. How involved do you feel in your job? (1) _Very little involved; my other interests are more absorbing (2) “Slightly involved (3) _Moderately involved; my job and my other interests are : equally absorbing to me (4) ~_______Strongly involved (5) ”Very strongly involved; my work is the most absorbing interest in my life How much do you look out for your secretary's personal welfare? ' (l) I very much look out for her personal welfare (2) ____Quite a bit (3) Moderately (4) : Very little (5).____A1most not at all . .-'/ , - . .- How much appreciation do you express when your secretary does a good job? (5) ___ Almost none (4) ___ Very little (3) ___ A moderate amount (2) ___ Quite a bit (1) ___ A great deal How friendly and easily approached are'you?‘ (1) _Extremely friendly and easily approached (2) “Very friendly and easily approached (3) Mbderately (4)‘___ Not very (5) “Not at all How much do you allow your secretary to modify the procedures required in her job? (5) Almost none (4) ____Very little (3) ____Moderately (2) ____Very much (l) __ Greatly How much do you maintain definite standards of performance for your secretary? (5) Very little (4) ___ A little (3) *A moderate amount (2) :Quite a bit (1) _Very much How'much do you let your secretary do her work the way she thinks best? (5) To almost no extent - (4) _Very little (3) _A moderate amount (2)“ _Very much (1) “Almost totally (OVER) 18. 19. 20. 21. 22. 23. 24. 154 When you first began working with your present boss, how much did he (she) facilitate adjustments to your new work setting? (1) ____To a very great extent (2) ___ Quite a bit (3) ___ A moderate amount (4) __ A little (5) ___.Very little How much does your boss make sure that you understand his part in the work group? (1) ____Very much (2) ___ Quite a bit (3) ___ Moderately (4) ____Not too much (5) Very little How much at ease do you feel in talking with your boss? (5) Almost none (4) : Very little (3) ___ A moderate amount (2) __ Quite a bit (1) ____A great deal How much does your boss let you know what is expected of you? (1) To a very great extent (2) : Very much (3) ___ Moderately (4) ____To a little extent (5) ___ To a very little extent .How much does your boss have you share in decision making? (1) ___ Greatly '(2) ___ Very much (3) ___ Moderately (4)‘___ Very little (5) ___ Almost none To what extent does your boss allow you influence equal to his (her) own in decisions which affect your job? (1) ____To a very great extent (2) ___ Very much (3) ___ Mbderately (4) ___ To a very small extent (5) ___ To almost no extent How much does your boss see to it that you have the material you need to work with? ?H'- (5) Very little (1) ____Very much (2) __ Quite a bit (3) ____Moderate1y (4) ___ Not too much ‘152 25. How much does your boss schedule the work you have to do? (5) ____A.great deal (4) __ Quite a bit (3) _;_ A moderate amount (2) ____Not too much (1) ___ Almost none 26. To what extent does your boss assign you a task, then let you handle it? (5) ____A1most none (4) ____Very little (3) ___ Moderately (2) ___ Very much (1)'___ Greatly You have now completed this questionnaire. Now just slip it in the stamped, addressed envelope provided and mail. We again thank you for your valuable assistance with this project. “LEI! 145 25. How much do you schedule the work your secretary has to do? (5).____A great deal (4) ___ Quite a bit (3) ____A moderate amount (2) ___ Not too much (1) ___ Almost none 26. To what extent do you assign your secretary a task, then let her handle it? (5) ___ Almost none (4) ___ Very little (3) ___ Moderately (2) ___ Very much (1) ___ Greatly 27. Lastly (whew!) a few brief questions regarding yourself: Sex: Male ___ Female "_ Age: Under 20 __ 20-29 __. 30—39 ___ 40-49 __. 50-S9___ 60-65 __ Over 65 __ Length of time on present job: Less than 6 months __ 6 months to 1 year __ l to 2 years.__ Over 2 years __ Marital status: Never married ~__ married __. Divorced, Separated __ Number of children: (Only those living with you) 0‘__ 1___ 2.___ 3.__ Over 3.__ 28. How long have you worked with your present secretary? (1) ___ Less than six months (2) ____Six months to one year (3) ____One year to two years (4).___ Over two years - ._ _‘ You have now completed this questionnaire. Now just slip it in the stamped, addressed enve10pe provided and mail. We again thank you for your valuable assistance with this project. -~-~_ A _ —_.' — —— —— ‘— —----.C v— — -_-- APPENDIX E 154 INDUSTRIAL/OHGANIZATIONAL PSYCHOLOGY MICHIGAN STATE UNIVERSITY East Lansing, Michigan b882h January 7, 197h As you are aware, the Scanlon Plan Associates is sponsoring a Management Staff Support Training Seminar on January 17, l97h, which you will be attending. The Scanlon Plan Associates needs to determine if this Seminar is effective enough to extend its benefits to other personnel in the member companies. The Board of Directors of the Associates has requested us to carry out this evaluation study for them. Therefore, we need.you as one of the Seminar participants to assist us in making this evaluation for the Associates and your company. we need certain information before the Seminar, and we are asking you to answer the questions in the enclosed form. Answering the questions should take no more than about a half an hour. When you have answered all the questions, please send the completed form to us at Michigan State University in the stamped, addressed envelop provided. For the purpose of this project, we need.your form before January 15, l97h, so we will appreciate your cooperation in filling out the form.and returning it to us. (we realize that you have given information earlier to Mrs. Ann Montgomery, who is leader of the Seminar. Our request is an additional need for the Associates' Study.) You do not need to sign the form. The report of this project will combine answers from all the other participants so that your anonymity is assured. Several weeks from this time we will be sending you a similar form.and ask you to kindly complete it. The second form should require no more than about a half an hour to complete. Even though completing the two forms is all you are requested to do for this project, your cooperation is very important in carrying out the evaluation of the seminar. 'When the data are analyzed and the project is completed, a report of the total findings will be made available to companies in the Scanlon Plan Associates. Every effort will be made to see that each individual who participated in the research study receives a copy. we at Michigan State University are pleased to assist the Scanlon Plan Associates in such an evaluation project. we will certainly appreciate your assistance in making it possible for us to make the report helpful in deciding on future training programs for your Management Staff Support Group. I/fi ' / l/ .1...‘ ("L (.9074)?qu ’/¢L’//\-_— ")‘Z’Cj / Rod Lowman, Research Associate . F. Frost, Professor Indgstrial/Organizational Psychology Industrial/Organizational Psychology enc - Sincerely,‘g ‘155 INDUSTRIAL/ORGANIZATIONAL PSYCHOLOGY MICHIGAN STATE UNIVERSITY East Lansing, Michigan h882h The Scanlon Plan Associates, of which your Company is a member, is sponsoring a Management Staff Support Training Seminar in January. by the Dartnell Institute. 'we are aware that you personally will not be attending this first session. However, the Scanlon Plan Associates want to determine if this Seminar is effective enough to extend its benefits to other personnel in the Scanlon Plan Associates member companies, and has requested us to carry out this research study for them. Therefore, we need you as one of the Management Staff Support to assist in making this evaluation. we need certain information before the Seminar, and we are asking you to answer the questions in the enclosed forms Answering the questions should take no more than about a half an hour. When you have answered all the questions, please send the completed form to us at Michigan State University in the stamped, addressed envelop provided. For the purpose of this project, we need your form before January 15, l97h, so we will appreciate your cooperation in filling out the form and returning it to us. Ybu do not need to sign the form. The report of this project will combine answers from all the other participants so that your anonymity is assured. Several weeks from this time we will be sending you a similar form and.ask you to kindly complete it. The second form should require no more than about a half an hour to complete. Even though completing the two forms is all you are requested to do for this project, your cooperation is very important in carrying out the evaluation of the Seminar. ‘When the data are analyzed and the project is completed, a report of the total findings will be made available to companies in the Scanlon Plan Associates. Every effort will be made to see that each individual who participated in the research study receives a copy. we at Michigan State University are pleased to assist the Scanlon Plan Associates in such an evaluation project. ‘we will certainly appreciate your assistance in making it possible for us to make the report helpful in deciding on future training programs for your Management Staff Support Group. angers). 7 -4/ g, (ml imwacbx /