WwA: Worldwide Attitudes [ WwA Home | Feedback | Search | International Survey Center Home] Job Complexity:
Australia 1989-1995 and Hungary 1992M.D.R. Evans and Jonathan Kelley
Worldwide Attitudes Volume1996-06-24:1-8 ISSN 1323-9589
© Copyright M.D.R. Evans and Jonathan Kelley 1996. All rights reserved.
How skilled are the jobs that people work at today? Champions of labour market reform in the 1980s and 1990s argued that excess government regulation and petrified work practices were strangling economic growth, in part by making inefficient use of the workforce’s skills. They proposed that de-regulating the labour market would make fuller use of workers’ capabilities to undertake complex and demanding work thereby increasing productivity and fuelling economic growth. Similarly, champions of corporate restructuring argued that many firms could increase productivity by slimming down middle management -- reducing or even eliminating whole layers of supervision and control -- and leaving many of their former functions to be performed by the workers themselves. Both these reforms assume that workers are able to undertake more complex and demanding tasks than they have done previously.By contrast, proponents of the "deskilling" thesis have argued that labour market de-regulation and technological change would combine to polarize the workforce into a highly skilled elite, a tiny wasp-waist sized middle class, and a de-skilled proletariat at the bottom. They see this as a cause of the increasing income inequality in deregulated economies like the USA and the emerging market economies of Eastern Europe (Danziger and Gottschalk 1994).
Data and Measurement
To assess these issues, we collected data in the International Survey of Economic Attitudes in three Australian surveys in 1989-90, 1993, and 1995 and a 1992 Hungarian survey (Kelley, Evans and Bean 1993; Kelley, Bean, Evans and Zagorski 1994, 1996; Robert et al 1993).
The analysis is restricted to full-time workers, working 35 hours or more. With this restriction, there are 4,523 respondents (1,995, 1033 and 957 in the three Australian surveys respectively , and 583 in Hungary).
Earlier research exploring job complexity used observers’ ratings (Kohn et al. 1983). From that we developed a set of four questions for respondents to complete themselves on the assumption that workers themselves are good informants about the work they do. These items are highly correlated among themselves and stand out as a separate concept in a factor analysis of a variety of job conditions and characteristics (Evans 1995).
Job Complexity in Australia, 1995
We introduced the questions on complexity with the phrase "Is this true of your job, or not ..." and then asked what their tasks were like:
"Is your job complex and difficult?" Yes, definitely (100) 19 Yes (75) 48 Neutral, yes & no (50) 14 No (25) 17 No, definitely not(0) 2 ---- Total 100% Mean 67 Cases 915Thus, a substantial majority -- 67% -- "agree" or "Yes, definitely" that their jobs are complex. This suggests that many modern employers trust their employees with complex tasks rather than routinizing their work, as the deskilling hypothesis suggests. Scoring these on a conventional points out of 100 basis -- 0 for No, definitely not, 100 for Yes, definitely, and other answers at equal intervals in between -- gives the average is 67, about half way between "Neutral, yes & no " and "agree".
From a slightly different angle, we also asked about the skills their job requires:
"Does your job require special talents and abilities that most people do not have?" Yes, definitely (100) 21 Yes (75) 51 Neutral, yes & no (50) 12 No (25) 14 No, definitely not 2 ---- Total 100% Mean 69 Cases 915
Overall, 72% of people feel that their jobs require special talents and abilities, so that in this sense at least, the market seems to be matching up people and jobs well. It is also noteworthy that this utilization of talent and skill is so widespread: contrary to the de-skilling thesis, engagement of talent and ability is not restricted to an elite -- almost three quarters of the population "agree" or "Yes, definitely" that their job requires such talents. The mean is 69 points out of 100.
Probing especially into the kind of cognitive skills that will be needed if Australia really is to become the "clever country", we also asked:
"Does it require a lot of thinking?"
Yes, definitely (100) 33
Yes (75) 55
Neutral, yes & no (50) 6
No (25) 6
No, definitely not 0
----
Total 100%
Mean 78
Cases 917
Most workers think their job requires a lot of thinking, with a mean of 78 points out of 100.
And respondents also report that wisdom is equally important, with a mean of 84 points out of 100:
"Does it require good sense and sound judgment?"
Yes, definitely (100) 42
Yes (75) 53
Neutral, yes & no (50) 4
No (25) 1
No, definitely not 0
----
Total 100%
Mean 84
Cases 917
In sum, by 1995 most full-time Australian workers reported felt that their jobs involved substantial skill and complexity, requiring special talents and a lot of thinking, good sense, and sound judgment.
Australia: Trends Since 1989-90
It short the evidence suggests that complex jobs requiring talent, thought, and sound judgment are more the rule than the exception in the Australian workforce. But, for all that, information on the present cannot tell us whether such jobs are getting to be more or less common over time, and the crux of the controversy over deregulation and deskilling is whether such jobs are multiplying or dwindling. To answer this, we made a complexity scale combining the four items additively. This increases reliability of measurement and abstracts away from idiosyncratic features of any one item. All the items are highly correlated in both Australia and Hungary and show strong loadings in a factor analysis (table 5). In fact, the evidence suggests that de-regulation has probably fostered such jobs. The average complexity score rose from 69 points in 1989/90, to 72 points in 1993 and 74 points in 1995. This is a rate of change of about 1 percent per year (0.84 with a standard error of 0.11, by OLS). These differences are statistically significant (t=7.6, p<.001), and suggest quite rapid change in a short period of time. The same pattern holds for the items individually (see the regression analyses in tables 1 to 4).
International Comparison: Hungary
Another striking comparison to a highly regulated labour market is with Hungary in 1992, immediately at the end of the Communist era. Hungarians worked in less complex and difficult jobs, 55 points compared to around 65 for Australians; in jobs that required many fewer special talents and abilities (42 points, compared to 64 for Australians; in jobs that required much less thinking (50 compared to 78); and in jobs that required a little less good sense and sound judgment (72 compared to 82; see tables 1 to 4). All these differences are highly significant statistically. About a third of these differences reflect the lower level of economic development in Hungary compared to Australia. Hungarians have less education than Australians, work in lower status jobs, and are less likely to be in supervisory positions. Regression standardization (using the methods of Kelley and Evans 1993) shows that if Hungarians had the same education, occupational status, supervisory responsibilities, age, labor force participation rates, and were equally urban, equally often self-employed, equally often employed outside the government sector, and had the same levels of labor force experience, their average level of job complexity would rise from 55 points to 61, still well below Australia’s 72 (table 6).Thus it seems that the communist command economy kept levels of job complexity well below those of Australia’s market economy, even adjusting for differences in the two nation’s level of economic development. We would expect that as the Hungarian economy becomes more market oriented in coming years, the complexity of workers’ jobs will rise.
Does Complexity Pay?
That employers value complexity is shown by the fact that Australian workers in highly complex jobs earn, on average, about 31% more than workers in very simple jobs (table 7). Hungarian employers seem to reward complex jobs at least as highly. This result is net of many other characteristics of these workers -- their education, their labour force experience, their supervisory responsibilities, their location in private or government sector, whether they own their business or are employees. Importantly, it is also net of their job titles, suggesting the complexity is associated with greater productivity even within particular occupations.
REFERENCES
Bean, Clive S. 1991. "Comparison of National Social Science Survey Data with the 1986 Census." National Social Science Survey Report 2(6):12-19. [ISSN 1031-4067]Danziger, Sheldon and Peter Gottschalk (editors). 1994. Uneven Tides. New York: Russell Sage Foundation.
Evans, M.D.R. 1995. "Job Characteristics" Worldwide Attitudes 19950605: 1-8.
Gregory, Robert. 1995. "The Macroeconomy and the Growth of Ghettoes and Urban Poverty in Australia." Address to the National Press Club, available from the Economics Program, Research School of Social Sciences, Australian National University.
Kelley, Jonathan and M.D.R. Evans. 1994. Australia, 1993: International Survey of Economic Attitudes, Round 1. Codebook and Machine Readable Data File. Canberra: International Survey Center, Institute of Advanced Studies, Australian National University.
Kelley, Jonathan, Clive S. Bean, M.D.R. Evans and Krzysztof Zagorski. 1996. Australia, 1995: International Social Science Survey. Codebook and Machine Readable Data File (Preliminary). Canberra: International Survey Center, Institute of Advanced Studies, Australian National University.
Kelley, Jonathan, M.D.R. EVans and Clive S. Bean. 1993. Australia, 1989 & 1990: International Social Science Survey -- Lifestyles and Family Values. Codebook and Machine Readable Data File. Canberra: International Survey Center, Institute of Advanced Studies, Australian National University.
Kohn, Melvin L. and Carmi Schooler, with Joanne Miller, Karen A. Miller, Carrie Schoenbach, and Ronald Schoenberg. 1983. Work and Personality. Norwood, NJ: Ablex.
Mortimer, Jeylan T., Jon Lorence, and Donald S. Kumka. 1986. Work, Family, and Personality. Norwood, NJ: Ablex.
Murphy, Kevin M. and Finis Welch. 1994. "Industrial Change and the Rising Importance of Skill." Pp 101-132 in Uneven Tides. Edited by Sheldon Danziger and Peter Gottschalk. New York: Russell Sage Foundation.
Robert, Peter, Tamas Kolosi, M.D.R. Evans, and Jonathan Kelley. 1993. Hungary, 1992: International Survey of Economic Attitudes, Round 1. Codebook and Machine Readable Data File. Budapest: TARKI.
APPENDIX: TABLESTable 1. "Is your job complex and difficult?"[1] Percentages read down. Australia, 1989-90, 1993 and 1995 and Hungary 1992.-------------------------------------------------- Oz Oz Oz Hungary 1989-90 1993 1995 1992 ----------------------------- No, definitely not(0) 2 2 2 3 No (25) 23 17 17 25 Mixed, neutral (50) 15 16 14 26 Yes (75) 45 48 48 38 Definitely yes (100) 15 17 19 8 ----------------------------- Total 100% 100% 100% 100% Mean[2] 62 65 67 55 Change per year b= 0.8 per year, t=4.9, p<.001 Cases 1919 882 915 574 -------------------------------------------------- Notes [1] Kelley, Bean, Evans and Zagorski 1995: page 27 question 3f. [2] Mean points out of 100 (scoring shown in parentheses). Change in mean over time for Australia is estimated by OLS. Hungary is significantly lower than any Australian year. Table 2. "Does it [your job] require special talents and abilities that most people do not have?"[1] Percentages read down. Australia, 1989-90, 1993 and 1995 and Hungary 1992. -------------------------------------------------- Oz Oz Oz Hungary 1989-90 1993 1995 1992 ----------------------------- No, definitely not(0) 2 2 2 8 No (25) 26 18 14 45 Mixed, neutral (50) 17 19 12 22 Yes (75) 41 43 51 22 Definitely yes (100) 13 18 21 4 ----------------------------- Total 100% 100% 100% 100% Mean[2] 59 64 69 42 Change per year b= 1.5 per year, t=9.1, p<.001 Cases 1920 885 915 572 -------------------------------------------------- Notes [1] Kelley, Bean, Evans and Zagorski 1995: page 27 question 3e. [2] Mean points out of 100 (scoring shown in parentheses). Change in mean over time for Australia is estimated by OLS. Hungary is significantly lower than any Australian year. Table 3. "Does it [your job] require a lot of thinking?"[1] Percentages read down. Australia, 1989-90, 1993 and 1995 and Hungary 1992. -------------------------------------------------- Oz Oz Oz Hungary 1989-90 1993 1995 1992 ----------------------------- No, definitely not(0) 1 1 0 6 No (25) 9 4 6 30 Mixed, neutral (50) 7 9 6 32 Yes (75) 57 56 55 26 Definitely yes (100) 26 31 33 7 ----------------------------- Total 100% 100% 100% 100% Mean[2] 75 78 78 50 Change per year b= 0.7 per year, t=5.1, p<.001 Cases 1922 886 917 573 -------------------------------------------------- Notes [1] Kelley, Bean, Evans and Zagorski 1995: page 27 question 3h. [2] Mean points out of 100 (scoring shown in parentheses). Change in mean over time for Australia is estimated by OLS. Hungary is significantly lower than any Australian year. Table 4. "Does it [your job] require good sense and sound judgment?"[1] Percentages read down. Australia, 1989-90, 1993 and 1995 and Hungary 1992. -------------------------------------------------- Oz Oz Oz Hungary 1989-90 1993 1995 1992 ----------------------------- No, definitely not(0) 0 2 0 2 No (25) 2 1 1 7 Mixed, neutral (50) 4 5 4 10 Yes (75) 60 58 53 62 Definitely yes (100) 34 36 42 19 ----------------------------- Total 100% 100% 100% 100% Mean[2] 81 82 84 72 Change per year b= 0.4 per year, t=3.8, p<.001 Cases 1927 887 916 570 -------------------------------------------------- Notes [1] Kelley, Bean, Evans and Zagorski 1995: page 27 question 3i. [2] Mean points out of 100 (scoring shown in parentheses). Change in mean over time for Australia is estimated by OLS. Hungary is significantly lower than any Australian year. Table 5. Inter-item correlations among items[1] measuring job complexity and principal axis factor analysis.[2] Australia, 1989-90, 1993 and 1995 surveys pooled and Hungary 1992 separately. ----------------------------------------------------------- Factor JobTALENT JobTHINK JobCOMPLX JobGDSENS loading ----------------------------------------------------------- AUSTRALIA: JobCOMPLX 1.00 .70 JobTALENT .52 1.00 .61 JobTHINK .56 .46 1.00 .79 JobGDSENS .40 .38 .56 1.00 .58 HUNGARY: JobCOMPLX 1.00 .68 JobTALENT .46 1.00 .63 JobTHINK .58 .56 1.00 .77 JobGDSENS .43 .44 .50 1.00 .57 ----------------------------------------------------------- Notes:[1] JobTALENT = job requires special talents and abilities; JobTHINK = job requires a lot of thinking; JobCOMPLX = job is complex and difficult; JobGDSENS = job requires good sense and sound judgment. [2] Loadings are from a larger analysis including a further half-dozen items measuring supervisory skills, extrinsic rewards, dirty work, and danger. Loadings on the first factor are shown. Varimax rotation. Table 6. Job complexity in Australia and Hungary compared: Regression standardization.[1] ------------------------------------------------------ Job complexity (scale) ------------------------------------------------------ 1. Hungary 1992, actual 54.8 2. Hungary, adjusted to Australian social structure 61.3 3. Australia 1993, actual 72.3 4. difference, (3)-(2) = 11.0 5. Alternative estimate (OLS on pooled sample, b coef for Hungary; compare to line 4) 12.8 ------------------------------------------------------ Notes: [1] Whole population standardization, following the methods of Kelley and Evans 1993. Includes adjustment for differencs in education, occupational status, labor force experience, sex, age,residence, self-employment, farm occupation, supervisory responsibilities, and government employment. Table 7. Job complexity and earnings: OLS regression predicting log income. Full-time workers. Australia 1993 (N=960) and Hungary 1992 (N=525). --------------------------------------------------------------- Australia: Variables, means and standard deviations --------------------------------------------------- Mean Std Dev Label JB#SKIL 72.259 17.552 Job complexity (scale: 0 to 100) MALE .736 .441 AGE 40.280 11.487 Age (years) URBAN .904 .295 EDUC2 11.548 3.163 Education (years) STAT2 55.640 27.112 Occupational status NFSELFEM .107 .310 Non-farm self employed FARM .038 .191 Occup: Farm SUPER .498 .500 Supervises others at work GOVT .386 .487 Govt employee vs all other LF0_4 .068 .252 Labor force exp: 0-4 years LF5_9 .086 .281 Labor force exp: 5-9 years LF10_19 .272 .445 Labor force exp: 10-19 years LF20PLUS .574 .495 Labor force exp: 20+ years LNEARNR .120 .477 Earnings (natural log) AUSTRALIA:Regression Variable B SE B Beta T Sig T JB#SKIL .003049 8.4185E-04 .112301 3.622 .0003 MALE .250230 .031278 .231672 8.000 .0000 URBAN .044296 .050383 .027451 .879 .3796 EDUC2 .022321 .005460 .148148 4.088 .0000 STAT2 .003948 6.7691E-04 .224637 5.833 .0000 NFSELFEM .116608 .046260 .075818 2.521 .0119 FARM .125442 .084640 .050198 1.482 .1387 SUPER .169996 .028930 .178452 5.876 .0000 GOVT -.026562 .030091 -.027152 -.883 .3776 LF10_19 .257725 .043832 .240787 5.880 .0000 LF20PLUS .343290 .041573 .356437 8.258 .0000 (Constant) -1.160069 .094318 -12.300 .0000 Adjusted R Square .32616 Standard Error .39118 Hungary: Variables, means and standard deviations --------------------------------------------------- Mean Std Dev Label JB#SKIL 54.766 19.320 Job complexity: scale (0 to 100) MALE .575 .495 AGE 38.978 10.843 Age (years) URBAN .395 .489 EDUC2 10.640 2.329 Education (years) STAT2 41.567 24.901 Occupational status NFSELFEM .070 .256 Non-farm self employed FARM .062 .241 Occup: Farm SUPER .208 .406 Supervises others at work GOVT .588 .493 Govt employee vs all other LF0_4 .063 .244 Labor force experience: 0-4 years LF5_9 .086 .280 Labor force experience: 5-9 years LF10_19 .262 .440 Labor force experience: 10-19 years LF20PLUS .588 .493 Labor force experience: 20+ years LNEARNR .142 .461 Earnings (natural log) (Table 7 continued). Job complexity and earnings: OLS regression predicting log income. Hungary: Regression ------------------- Variable B SE B Beta T Sig T JB#SKIL .004232 9.6079E-04 .177378 4.405 .0000 MALE .172868 .035116 .185572 4.923 .0000 URBAN .070187 .035308 .074483 1.988 .0474 EDUC2 .060718 .009775 .306757 6.212 .0000 STAT2 .001330 9.6034E-04 .071860 1.385 .1666 NFSELFEM .001874 .067732 .001041 .028 .9779 FARM -.073976 .074857 -.038662 -.988 .3235 SUPER .203949 .043848 .179843 4.651 .0000 GOVT -.033544 .035982 -.035844 -.932 .3516 LF10_19 .032706 .052797 .031243 .619 .5359 LF20PLUS .054978 .048198 .058748 1.141 .2545 (Constant) -.977594 .098834 -9.891 .0000 Adjusted R Square .35276 Standard Error .37084 ----------------------------------------------------------------
Number of
readers since 4 June 1999:
| WwA: Worldwide Attitudes | [ WwA Home | Feedback | Search | International Survey Center Home] |