--------------------------------------------------------------------- ISSP INEQUALITY-1999 MODULE: OVERVIEW --------------------------------------------------------------------- Australia (convenor) J. Kelley M.D.R. Evans Czech Republic P. Mateju Norway K. Knudsen K. Skjak Poland B. Cichomski P. Morawski Slovenia B. Malnar (c) Portions copyright 1984-1997 by J. Kelley, M.D.R. Evans, P. Mateju, K. Knudsen, K. Skjak, B. Cichomski, P. Morawski, B. Malnar, and the International Social Survey Programme. All rights reserved. (NOTE: THIS ONE FILE COMBINES ALL 8 FILES THE DRAFTING GROUP SENT SEPARATELY ON 17 NOVEMBER. OTHERWISE IT IS UNCHANGED – INCLUDING REFERENCES IN TEXT TO ‘SEPARATE FILES’ WHICH NOW JUST REFER TO DIFFERENT SECTIONS OF THIS ONE FILE) ------------------- INTRODUCTION This is our draft proposal for the ISSP Inequality-1999 module. It is based on the discussions at the last ISSP meeting (May, the Netherlands), on comments and suggestions by many groups in the subsequent months, and on our analysis of the Developmental Pretest. Please review it and get comments back to us as soon as possible. We will then revise this proposal in light of the comments and get a further draft to you just after Christmas, ready for the Manila meeting in January. -------------------- FILES YOU SHOULD HAVE RECEIVED FROM US The proposal is long and so we have sent it in several separate e-mail messages. These files (version 3) REPLACE ALL PREVIOUS FILES we have sent to you. They all have explicit names (in the “subject” line of the e- mail) so you can (hopefully!) kep track of them all. Please let us know if you do not receive all of these in the next day or two. (e- mail to the Drafting Group convenor: Jonathan.Kelley@anu.edu.au) To be sure everyone has everything, we are RE-SENDING ALL FILES you should have. Some are new but others are only slightly modified versions of files we sent you over the last few weeks – you do not have to read them twice! But it seemed safer to send all again, at one go, with consistent names and this file as guide to what you should have: YOU SHOULD HAVE: "Begin_here" – This file, an overview & table of contents (new) "Core-3: Part 1 -- core questions, part 1 of 3" (re-sent) "Core-3: Part 2 -- core questions, part 2 of 3" (re-sent) "Core-3: Part 3 -- core questions, part 3 of 3" (re-sent) "Background-3" -- family background questions (re-sent) "Econ-3" -- economic benefit/harm questions (new) "Equal-3" –- norms about equality of opportunity (new) "Discrm-3" – discrimination and related questions (new) ------------------ SUMMARY OF RECOMMENDATIONS This is a summary of the recommended module, together with notes on where each topic is discussed. (ones marked with * are not yet documented in our discussion so far): 1. CORE QUESTIONS, WHICH WE DEFINITELY WANT TO INCLUDE IN THE MODULE: 47 core questions [“core-3, part 1; part 2; part 3”] 2. HIGH PRIORITY, ALTHOUGH NOT CORE: 2 family background Qs: fathers education, cultural capital [“background-3”] 4 economic system benefit/harm you & nation [“econ-3”] 3. MEDIUM PRIORITY: 7 QUESTIONS TO COME FROM AMONG THESE: 4 perceived discrimination (re women, minorities, poor)[in “discrm-3”] 2 Norms about equal opportunity/equal outcome [“equal-3”] 2 Perception: meritocracy in society [in “discrim-3”] 2 Perception: crime is required to get ahead [in ”discrim-3”] *1 (or 3) subjective social mobility Qs *3 tax questions 4. LOW PRIORITY: VALUABLE BUT WE DON’T REALLY HAVE SPACE IN THE MODULE *(several) functionalism & related *4 Qs on how much “shopkeepers” and “sales clerks” do earn & should earn *1 pyramid diagram Q: where ‘farm laborer’ fits in 5. TO DROP 13 Attribution Qs (but see perceived discrimination) (several) assorted single items (as decided in the May meeting) (all other items in the 1992 module which are not mentioned above) --- end of this message --- -------------------------------------------------------------------------------- -------------------------------------------------------------------------------- (this separates two files) -------------------------------------------------------------------------------- -------------------------------------------------------------------------------- (e-mail: split into core3-1; core3-2; core3-3) To: ISSP From: Inequality Drafting Group Subject: ISSP INEQUALITY MODULE: “Core-1, part 1 of 3”: Core questions. ------------------------------------------------------------------------- ISSP INEQUALITY-3 1999 DRAFTING GROUP: PRELIMINARY REPORT, VERSION 4, PART 1. (document ID: “Core-3” ) ------------------------------------------------------------------------- Australia (convenor) J. Kelley M.D.R. Evans Czech Republic P. Mateju Norway K. Knudsen K. Skjak Poland B. Cichomski P. Morawski Slovenia B. Malnar These are the “core” questions for the Inequality module. They were discussed at some length at the meeting last May in the Netherlands. There is unanimous agreement among the Drafting Group that these core questions should all be included in the questionnaire, a view that got general assent from the ISSP as a whole at the May meeting in the Netherlands. The more controversial issues about “candidate” items that are not in the core, but not on the list to be dropped, are reserved for a future document. -------------------------------------------------- CORE ITEMS DEFINITELY TO BE INCLUDED IN THE MODULE -------------------------------------------------- OVERVIEW #1: PERCEIVED EARNINGS OF OCCUPATIONS (what pay “is”) – 10 Qs #1: LEGITIMATE EARNINGS OF OCCUPATIONS (what pay “ought to be”) – 10 Qs #2: RELATIVE INEQUALITY (More or less inequality in society as whole?)- 2 Qs #3: LEGITIMATE BASES OF REWARD (what justifies higher pay?) - 5 Qs #4: LEGITIMATE RETURNS TO HUMAN CAPITAL (education justifies pay?)- 3 Qs #5: FAIRNESS OF YOUR EARNINGS (are you, yourself fairly paid?) - 2 Qs #6: IMAGES OF CLASS (“pyramid” diagrams; now & in past) - 8 Qs #7: SUBJECTIVE SOCIAL CLASS (now & in past)- 4 Qs #8: CLASS CONFLICT - 3 Qs ---------------------------------------------------------------------------- ---------------------------------------------------------------------------- SECTION #1: PERCEIVED EARNINGS & LEGITIMATE EARNINGS OF OCCUPATIONS - CORE QUESTIONS FOR THE 1999 INEQUALITY MODULE ###1.1 QUESTIONS ON PERCEIVED EARNINGS (WHAT "IS") 10 Qs: ---Concept: Perceived earnings ("is") of manual occupations V26 Actually earn: skilled factory worker V32 Actually earn: a farm worker V35 Actually earn: unskilled factory worker ---Concept: Perceived earnings ("is") of professionals V27 Actually earn: doctor in general practice V30 Actually earn: a solicitor ---Concept: Perceived earnings ("is") of government elite V34 Actually earn: an appeal court judge V36 Actually earn: cabinet minister-nat government ---Concept: Perceived earnings ("is") of business elite V29 Actually earn: chair -large nation-wide company V33 Actually earn: owner-manager of a factory --- Concept: Perceived earnings of your occupation (V301-New item) "And someone who works in YOUR usual occupation -- about how much do they actually earn? ###1.2 QUESTIONS ON LEGITIMATE/IDEAL EARNINGS (WHAT "OUGHT TO BE") 10 Qs ---Concept: Legitimate earnings ("ought") of manual occupations V41 Should earn: skilled factory worker V47 Should earn: a farm worker V50 Should earn: unskilled factory worker ----Concept: Legitimate earnings ("ought") of professionals V42 Should earn: doctor in general practice V45 Should earn: a solicitor ----Concept: Legitimate earnings ("ought") of government elite V49 Should earn: an appeal court judge V51 Should earn: cabinet minister-nat govern ---Concept: Legitimate earnings ("ought") of business elite V44 Should earn: chair -large nation-wide company V48 Should earn: owner-manager large factory --- Concept: Legitimate earnings of YOUR occupation V302 (New item) "And someone who works in YOUR usual occupation -- how much ought they earn?" KNOWN USAGE -- Perceived income ("is"): Known uses: B&K91, B&K94A & B; B&U90; EK&K92, F&H&J..96, G96, H91, K96, K&Z97, T92b; Sv93a; Sv93b; Sv93d; S96; S97. KNOWN USAGE -- legitimate income ("ought") questions: B&K91, B&K94A & B; B&U90; F&H&J..96; G96, H90, H91, K94, K96, K&E90, K&E93 K&Z97, T92a, T92b; Sv93a; Sv93b, Sv96; S96; S97 ### --- (Wording issue: In some countries there is a big difference in earnings of a “doctor” in private vs government employment. We may want to specify which. There are some Polish data with both, which should provide guidance, but they are not yet analyzed.) ###1.4 OVERVIEW These questions on ideals and perceptions of occupational earnings form the core of the module. They have been widely used and form the basis of numerous publications, some in very visible places bringing the ISSP to the attention of a wide audience. Their origin lies in a search for genuinely comparable measurement (inspired in part by a dissatisfaction with the "more" and "less" inequality formulations used in much prior research on the grounds that these "relative" measures usually did not assess how much more earnings inequality there should be, and that they made comparisons over time and between countries treacherous), and in the worry that from the relative inequality items we did not know whether proponents of inequality were more concerned with raising the earnings of ordinary folk OR with reducing earnings at the top. The module retains a "relative inequality" scale to allow comparisons with other surveys, but focuses on the magnitudes of ideal earnings -- a flexible formulation that allows researchers to define inequality in a wide variety of ways suitable to their purposes. This "magnitude" approach also has the important feature that it allows us to assess differences in the sources of legitimacy/ illegitimacy of high earnings for different groups at the top. Indeed, an important evolution of the module from round-1 to round-2 was building on indications from round- 1 that assessments of high status occupations were more differentiated than originally expected, with some marked differences in the determinants of ideals concerning the ideal earnings of professional, business and government elites. The decision to ask fewer items about low status occupations and more questions about high status occupations was based partly on the research findings that there was substantial consensus about the ideal pay of low status occupations. In analysis of the 9 nations in the 1987 inequality module Kelley and Evans concluded (American Journal of Sociology 1993: 75-125): "Comprehensive data on public beliefs about the legitimacy of income inequality from large, representative national samples in nine nations, conducted by the International Social Survey Programme, show broad agreement on the legitimate pay of low status, ordinary jobs; agreement that high status, elite occupations should be paid more than the minimum; but disagreement over how much more they should get. This disagreement is linked to politics and social structure, with older, high SES, politically conservative respondents preferring markedly higher pay for elite occupations, but usually not lower pay for ordinary jobs. Differences among nations in the legitimate earnings hierarchy are small but differences about the amount of inequality thought right are substantial, even after adjusting for differences in social structure. Most strikingly, communist countries are much more equalitarian than capitalist ones. Among capitalist nations, Austria is the most inequalitarian; the U.S., Britain, Germany, the Netherlands, and Switzerland are similar; and Australia stands out as a de Tocquevillian equalitarian New World society, preferring only three-quarters the inequality of Western Europe." Table 1.4A. Legitimate earnings, 1987 (ratio to actual income of unskilled workers) --------------------------------------------------------------------------- USA Britain Australia Hungary Austria Germany Poland --------------------------------------------------------------------------- Unskilled worker 0.89 1.22 0.98 1.28 1.00 1.13 1.12 Farm worker 0.99* 1.29* 1.18* 1.38* 1.13* 1.62* 1.30* Secretary 1.04* 1.45* 1.09* 1.51* 1.17* 1.02 0.96 Bank clerk 1.04* 1.80* 1.19* 1.83* 1.24* 1.21* 1.10 Bus driver 1.09* 1.67* 1.18* 1.61* 1.16* 1.65* 1.74* Brick layer 1.45* 1.62* 1.35* 1.67* 1.43* 1.83* 1.91* Skilled worker 1.42* 1.78* 1.47* 1.78* 1.35* 1.63* 1.51* Shop owner 1.39* 2.20* 1.50* 2.40* 1.64* 2.22* 1.78* Doctor 3.97* 4.44* 2.80* 4.49* 3.08* 1.95* 2.05* Cabinet Minister 3.34* 6.04* 3.35* 5.06* 2.84* 3.16* 3.28* Chairman 5.22* 5.93* 5.14* 5.74* 3.72* 2.69* 2.70* --------------------------------------------------------------------------- For details see attached file 5Pay_Ine.doc. The improved range of questions in the 1992 inequality module provides a much firmer basis for measuring the earnings of elite occupations of several types (professionals, government elite, business elite -- the distinction among these, although not always large, is clear in LISREL analyses for several nations). The fall of communism and emergence of a market economy in the nations of Eastern Europe has had a profound impact on norms about pay (Kelley & Zagorski, paper presented to the World Congress of Central and East European Studies Association, Warsaw, 1995). In an analysis of Poland, Hungary and Australia we found that normative support for inequality increases dramatically: In communist times the Polish and Hungarian publics favored much less inequality than citizens of Western capitalist nations thought right; within two or three years after the fall of communism they favored almost as much inequality as Westerners; and within a further few years favored as much or more inequality than Westerners think right. These normative changes did not arise from socioeconomic or demographic change. Rather they arose in large part from changes in perceptions of actual income inequality. Table 1.4B. How much should high status occupations earn compared to low? (Column A; adjusted for differences in social structure and perceived pay, column B). How has this changed over time (annual percent change, column C). Poland, 1987-1994. ------------------------------------------------- Chairman of a large national corporation A: Ratio B: Adjusted C: % change 1987 1.82 1.82 -- 1991 2.97* 2.66* 10% 1994 3.74* 2.95* 4% A cabinet minister A: Ratio B: Adjusted C: % change 1987 2.14 2.14 -- 1991 2.69* 2.25* 1% 1994 3.56* 2.86* 8% The owner/manager of a large factory A: Ratio B: Adjusted C: % change 1991 3.10 2.51 -- 1994 5.70* 3.74* 14% A judge in the High Court A: Ratio B: Adjusted C: % change 1991 2.66 2.25 -- 1994 3.86* 2.92* 9% A doctor in general practice A: Ratio B: Adjusted C: % change 1987 1.32 1.32 -- 1991 1.77* 1.67* 6% 1994 1.86* 1.79* 2% A lawyer A: Ratio B: Adjusted C: % change 1991 1.97 1.68 -- 1994 2.69* 2.23* 10% ------------------------------------------------- For details see the attached file 5EEurop.Doc. ###1.5 EXTENSION: JUST EARNING FOR YOUR OWN OCCUPATION Two new related items are included to tie the overall issue of earnings inequality directly to respondent's personal situation: "And someone who works in YOUR usual occupation -- about how much do they actually earn?” V301 and "And someone who works in YOUR usual occupation -- how much ought they earn?"(V302). These are to be asked with other questions in the earnings perceptions and ideal sections respectively. These items permit analysis of your occupation compared to other occupations. This: (1) links the general ideals about earnings inequality (as measured by the other items in the battery) to respondents’ personal situation; (2) allows an analysis of discrimination – for example an analysis of how much women in particular occupations think they should earn, compared to men, and compared to what they actually earn. Similar analyses are possible for old versus young, urban versus rural residents, those born into low status families versus those born in high status families, and the like. --------------------------------------------------------------------------- --------------------------------------------------------------------------- SECTION #2: RELATIVE INEQUALITY - CORE QUESTIONS FOR THE 1999 INEQUALITY MODULE ###2.1 ------------------------------- RELATIVE INEQUALITY SCALE: V56, V57 V56 Differences in income in country too large V57 Government: responsibility to reduce differences high low income KNOWN USES: [B&K94A & B; B&U90; Sv93a, Sv93b; Sv93d; Sv96] SUMMARY: These items make a good scale empirically: They are very highly correlated, have similar correlations with criterion variables and strong links to political views. For ISSP purposes they also have the advantage of links across surveys (they appear in the Role of Government module, too). Their meaning is less clear than for the "legitimate incomes" questions, but they form an interesting complement to the core questions. In a way, the "fuzziness" may be an advantage because it approximates the actual rhetoric of political dispute over these issues. ###2.2 -------------------- Results: Descriptive The general pattern is that the public thinks income inequality is too great and the government should reduce it. Differences between nations are substantial, with post- Communist Eastern Europeans much more likely to think there is too much inequality. For example: Table 2.2A. Government: responsibility to reduce differences high low income. Three nations, circa 1992. ----------------------------------------------- Poland Hungary Australia ----------------------------------------------- strongly agree 30 33 10 agree 43 42 33 neither 7 14 20 disagree 15 9 29 strongly disagre 5 3 8 ------------------------ 100% 100% 100% Cases 1334 1233 2091 ----------------------------------------------- --------------------------------------------------------------------------- --------------------------------------------------------------------------- SECTION #3: LEGITIMATE BASES OF REWARD: - CORE QUESTIONS FOR THE 1999 INEQUALITY MODULE ###3.1 -------------------- responsibility and skill (one scale) QUESTIONS: V77 Important for pay: responsibility V78 Important for pay: years of education V79 Important for pay: supervise others KNOWN USES: B&K94A & B; K96; R94; SV93B These questions tap which aspects of jobs and persons the public sees as ideally playing a role in the allocation of pay. This three-item scale focuses on responsibility and skill (which go together well empirically, although conceptually one might prefer a separate scale for each concept). The items are well correlated among themselves, and have similar correlations with criterion variables, so them perform well in factor analysis, too. The basic hypothesis here is that people who think that responsibility and skill are legitimate bases of reward will endorse higher levels of inequality than will their peers who hold that responsibility and skill should not be rewarded. Preliminary evidence from regression analyses supports this hypothesis. ###3.2 -------------------------------------------welfare-needs scale QUESTIONS: V80 Important for pay: support family V81 Important for pay: support child KNOWN USES: [B&K94A & B; K96; R94; SV93B] Assessing the legitimacy of redistribution through the pay system is the goal of these items. They are intended to tap the legitimacy of non-market claims in the allocation of pay. The original idea was that people who endorsed the legitimacy of these welfare-oriented claims would want to pay workers in low status jobs more (one bonus of the magnitude-approach to measuring ideal inequality is the serendipitous discovery that proponents of the welfare-approach to pay particularly want to pay less to occupations at the top!) . They perform very well empirically. In general the ISSP prefers 3-item scales but this one works so well in such a variety of places that we don’t really feel another item is needed. ###3.3 DESCRIPTIVE RESULTS Here are some typical results, these from West Germany: Table 3.3A. Frequencies, West Germany 1992. -------------------------------------------------------- REWRESP V77 IMP DECID PAY: RESPONSIBILITY Value Label Value Frequency Percent Essential 1 616 27 Very important 2 1319 57 Fairly important 3 270 12 Not very important 4 13 1 Not important at all 5 4 0 . 75 3 ------- ------- Total 2297 100 - - - - - - - - - - - - - - - - - - - - - - - - - - - REWEDUC V78 IMP DECID PAY: YEARS OF EDUCATION Value Label Value Frequency Percent Essential 1 272 12 Very important 2 1043 45 Fairly important 3 737 32 Not very important 4 164 7 Not important at all 5 21 1 . 60 3 ------- ------- Total 2297 100 - - - - - - - - - - - - - - - - - - - - - - - - - - - REWSUPR V79 IMP DECID PAY: SUPERVISE OTHERS Value Label Value Frequency Percent Essential 1 131 6 Very important 2 640 28 Fairly important 3 944 41 Not very important 4 399 17 Not important at all 5 56 2 . 127 6 ------- ------- Total 2297 100 - - - - - - - - - - - - - - - - - - - - - - - - - - - REWFAM V80 IMP DECID PAY: SUPPORT FAMILY Value Label Value Frequency Percent Essential 1 335 15 Very important 2 1099 48 Fairly important 3 583 25 Not very important 4 140 6 Not important at all 5 43 2 . 97 4 ------- ------- Total 2297 100 - - - - - - - - - - - - - - - - - - - - - - - - - - - REWKIDS V81 IMP DECID PAY: SUPPORT CHILD Value Label Value Frequency Percent Essential 1 351 15 Very important 2 1207 53 Fairly important 3 500 22 Not very important 4 114 5 Not important at all 5 36 2 . 89 4 ------- ------- Total 2297 100 - - - - - - - - - - - - - - - - - - - - - - - - - - - ###3.4 MEASUREMENT PROPERTIES These items show good measurement properties in ISSP 1992. For example, here are the Australian inter-item correlations (N=2,090) for the 3 “responsibility” measures (responsibility, education, supervision); the 2 welfare items (family, children); and the 2 item effort scale (do job well; hard work – the items in the “effort” scale are not part “core” questions in the module): Table 3.4A. Inter-item correlations & factor analysis. Australia 1992 (N=2,090). ---------------------------------------------------------------------------- CORRELATIONS RewRESP RewEDUC RewSUPR RewFAM RewKIDS RewWELL RewHARD RewRESP 1.00 RewEDUC .46 1.00 RewSUPR .51 .50 1.00 RewFAM .09 .20 .22 1.00 RewKIDS .05 .18 .17 .82 1.00 RewWELL .38 .24 .29 .11 .11 1.00 RewHARD .33 .27 .29 .17 .18 .71 1.00 --------------------------------------------------------------------------- The factor analysis is also clear for Australia, and also for the USA and Germany. Table 3.4B. Factor analysis: Australia, USA, and Germany. (Factor loadings, varimax rotation) ------------------------------------------------- Factor 1 Factor 2 Factor 3 Australia 1992 (N= 2090). ------------------------ RewSUPR .81 .13 .13 RewEDUC .80 .14 .08 RewRESP .77 -.06 .28 RewKIDS .06 .95 .08 RewFAM .12 .94 .06 RewWELL .19 .03 .91 RewHARD .18 .12 .89 USA 1992 (N= 1250) ---------------- RewRESP .69 -.12 .27 RewEDUC .76 .18 -.01 RewSUPR .78 .07 .08 RewFAM .06 .92 .06 RewKIDS .06 .92 .07 RewWELL .15 -.00 .89 RewHARD .09 .14 .88 WEST GERMANY 1992 (N=2,200) --------------------------- RewRESP .63 -.10 .27 RewEDUC .73 .19 .03 RewSUPR .75 .05 .06 RewFAM .08 .94 .08 RewKIDS .05 .93 .13 RewWELL .13 .06 .85 RewHARD .12 .15 .81 ------------------------------------------------- [--------- ^Message “Core: Part-1” ENDS HERE ---------- -------------------------------------------------------------------------------- -------------------------------------------------------------------------------- (this separates two files) -------------------------------------------------------------------------------- -------------------------------------------------------------------------------- To: ISSP From: Inequality Drafting Group Subject: ISSP INEQUALITY MODULE: “Core-1, part 2 of 3”: Core questions. [--------- ^Message “Core: Part-2” BEGINS HERE ----------] --------------------------------------------------------------------------- --------------------------------------------------------------------------- SECTION #4: LEGITIMATE RETURNS TO HUMAN CAPITAL - CORE QUESTIONS FOR THE 1999 INEQUALITY MODULE QUESTIONS: "Nowadays in {COUNTRY} workers with a primary school education earn about {$22,000} a year by the time they are about 30 or 35 years old. Now please think about those who spend longer in school... V308. How much do you think someone about 30 years old who has finished secondary school should earn?. . . . . . . $ ___________ per year V309. How about a typical 30 year old {university graduate / who completed a BA degree} ? . . . . . . . . . . . . . . $ ___________ per year V310. How about someone who has completed a [doctoral/PhD] degree?. . . . . . . . . . . . . . $ ___________ per year Notes on question wording:(1) If people in your country are not used to viewing their income in before-tax terms, after-tax is OK (as for ISSP income questions generally). (2) If people in your country are more used to the idea of earnings per month than the idea of earnings per year, that is OK too; simply change the question to "per month". (3) Questions should not be sex specific. ###4.1 INTRODUCTION: Around the world and throughout history, leaders of government and industry, owners of large firms, and skilled professionals earn more than ordinary workers. Discovering the objective causes of this earnings inequality has long been a central focus of research in social stratification and labor economics (e.g. Becker 1964; Blau and Duncan 1967); governmental policies to redistribute income have been a central locus of political conflict in many nations, contested and re-contested generation after generation (e.g. Franklin, Mackie, and Valen, 1992); and the moral basis of this inequality has attracted philosophical analysis and prescriptive argument (e.g. Aristotle; Rawls 1971; Sen 1973). More recently, a flourishing tradition of empirical research on the origins and development of people's norms about the distribution of income has developed under labels such as "social justice", "equity", or the "Legitimation of inequality", especially in sociology (e.g. Alwin 1987; Jasso 1980; Kelley and Evans 1993) and social psychology (e.g. Berger, Zeldich, Anderson and Cohen 1972; Moore 1992). A central finding is that in all countries studied so far -- rich as well as poor, socialist as well as capitalist -- there is near consensus among the general public about how much ordinary workers should earn, consensus that elite occupations should be paid more than ordinary workers, but widespread disagreement about how much more (Headey 1991; Kelley and Evans 1993; Svallfors 1993; Zagorski 1994). A small part of this dissensus reflects self-interest, politics and social structure, but most remains unexplained. We suggest extending the empirical social justice-equity-Legitimation line of research by investigating the reasoning underlying these normative judgments: on what basis do people evaluate earnings differences and how widely shared are these principles? These are key questions in fields ranging from equity theory to political sociology, as well as for more pragmatic matters of job evaluation, pay administration, and public policy (Dornstein 1991: Ch 1; Milkovich and Newman 1987: Ch 9). Indeed non- market approaches to pay, for example "comparable worth", are built on the assumption that people agree on how to distribute pay fairly. ALTERNATIVE THEORIES Aristotelian Hypothesis: People will think that pay should be proportional to contributions. So people will believe that workers with education ought to be paid in proportion to the greater contributions they provide; Hence occupations that require higher education should be proportionately better paid. Human Capital Hypothesis, with the aid of Homans: People will think that pay ought to be proportional to investments in training; Hence occupations requiring higher education should be paid in proportion to the investment required to acquire the training. In a competitive economy, returns to investment will reflect economic contributions, leading to the same prediction as the Aristotelian Hypothesis. In Eastern Europe's command economy, however, returns were much lower, leading to a prediction of lower returns to education than in free market economies Equalitarianism: People will think that everyone should get the same pay. ###4.3 MEASUREMENT PROPERTIES: The items have good measurement properties in all three countries, suggesting the existence of coherent attitudes in the general public. Responses to the three questions are highly correlated with each other -- between .5 and .8, well above the usual range for attitude measures; they show a similar pattern of correlations with other variables; and they have satisfactory loadings in a LISREL confirmatory factor analysis. Attitude Stability. The questions are also stable over time. The correlation between what people said was reasonable pay for a secondary school graduate at one time and what the same people said three or four years later is .27. Views about the proper pay of university graduates are even more stable, r=.41, as are those about PhDs, r=.44. This is clear evidence that well formed, stable attitudes do exist in this area. ###4.4 DESCRIPTIVE RESULTS: These results are based on three large, representative national samples. Almost everyone thinks that education should receive some reward, but there are a variety of views about how much. For example, in Australia, on average people suggested that about: $31,000 is right for year 12 graduates; $44,000 is proper for university graduates (42% more than year 12 graduates); and $52,000 is right for PhDs (68% more than year 12 graduates). Interestingly, there are few differences between Australia (capitalist) and Poland & Hungary (formerly communist), despite the very different ideology prevailing in communist times in Eastern Europe, and the very low actual returns to education there: Table 4.4A. How much pay does education warrant? Means (geometric) ------------------------------------------------------------- 1. Legitimate return for a BA: Australia = 42% -- i.e. BA should earn 42% more than 2ndary school graduate Hungary = 41% Poland = 39% 2. Legitimate return for a PhD: Australia = 68% Hungary = 89% Poland = 81% (Number of cases: Poland 1448, Hungary 1162, Australia 1835) -------------------------------------------------------------- At least in Australia, bonus people think proper for university education has declined in recent years. (1) In 1989-90, these same respondents believed that the bonus paid to university graduates should be 44%, fractionally higher than what they thought proper three or four years later. Although this difference is small, only a 0.5% drop per year, it is statistically significant (t=2.59, p<.01). And if it persists for a longer period, it would produce a noticeable erosion in the legitimate pay of university graduates. Over a generation (30 years), the drop would be about 14%, reducing the legitimate bonus for university graduates from 42% to 28% . (2) The decline in the bonus thought proper for a PhD has been even clearer: about 1% a year (t=4.54, p<.001). Over a generation, it suggests a drop in the bonus thought legitimate for a PhD of 35%, from 68% to 33%. That would leave people with a PhD entitled to little more than someone with a BA. ###4.5 ANALYSIS: CAUSES We estimate the structural equation and measurement models simultaneously in a single maximum likelihood solution, using multiple-group LISREL models. There is little class conflict over the proper returns to education, instead class differences are mostly small. Occupational status, self-employment, supervision, and subjective class are irrelevant in all three countries. Government employment, being a farmer, and family income have at most small effects, varying unsystematically among countries. The exception is education. In formerly communist countries, the well educated think education warrants substantially higher rewards. Self- interest may or may not play a role in this: all aspects of market reform -- including big increases in actual rewards to education and business assets -- are much more strongly supported by the intelligentsia. ###4.6 ANALYSIS: CONSEQUENCES Views about the legitimate returns to education have only a small impact on how much people believe blue collar and routine white collar workers should earn, with standardized effects between zero and .12. These modest effects presumably reflect the modest contribution of skill, education and training to successful performance in these jobs. Views on legitimate returns to education have a larger impact on what people think appropriate for owners of small businesses or farms. Views about the legitimate returns to education strongly affect the pay people think legitimate for professional occupations, elite government officials, and the economic elite. Standardized effects are around .30 in formerly communist countries and .20 in Australia, net of a wide range of class and demographic factors. The dissensus about elite, together with clear class and demographic differences and substantial percent of variance explained, are standard findings (Kelley and Evans 1993: 97-110). Despite 40 years of communist Equalitarianism, views about the legitimate size of returns to education have a strong impact on the pay Poles and Hungarians think legitimate for elite occupations. Indeed, the effect is about twice as strong in Hungary and Poland as in capitalist Australia. The finding that ideal returns to education influence legitimate pay for elite occupations in all three societies is consistent with the Aristotelian theory's prediction that education's evident productivity should legitimate rewards in all societies, communist as well as capitalist. But it is inconsistent with the other theories: inconsistent with radical Equalitarianism; inconsistent with the credentialist thesis that equalitarian elites would not reward education; inconsistent with existential arguments that communism's relatively equalitarian distribution of pay would in time come to be normatively accepted; inconsistent with hegemonic power of communist elites to impose their equalitarian ideology on their society; and inconsistent with human capitalist arguments that education must be rewarded only because it is an alternative to investments in physical capital. The results from these three diverse nations strongly suggest that views about the legitimate of educational attainment would turn out to be important in other ISSP nations. Attachment: "Human Capital and the Legitimation of Inequality: Public Opinion in Three Nations" (M. Evans and J. Kelley) Presented to the 13th World Congress of Sociology (Bielefeld). File: 5Educ$.doc ------------------------------------------------------------------------ ------------------------------------------------------------------------ SECTION #5: FAIRNESS OF YOUR EARNINGS. 2 Qs - CORE QUESTIONS FOR THE 1999 INEQUALITY MODULE V311 Are you paid fairly? We are asking about what is morally right – not about what you do earn, nor what you would like to earn – but what you feel you deserve, what you ought in fairness earn given your skills and the like. ... All in all, how fair is your pay? >> If not working now, please tell about your last job. Very low Low Fair High Very high V312 [[in the questionnaire, put this question far away from V311]] Would you say that you earn... >> If you are not working now, please tell about your last job. Much less than I deserve Less than I deserve What I deserve More than I deserve Much more than I deserve INTRODUCTION: These questions on fairness of respondent’s own earnings will allow us to find out what kind of people think their earnings are fair and what sorts think their earnings are too low. It speaks directly to issues of discrimination: for example, do women think their earnings are fair? How about migrants? Youth? Country dwellers rather than city dwellers? People born into working class families? The questions also speak to a very old issue: Is there a "just price" or a "fair wage" for different kinds of work? Or should all workers get the same pay? Or are labor markets intrinsically fair, with just rewards determined by the operation of supply and demand for different types of labor? These issues have long concerned social philosophers (e.g. Rawls 1971) and sociologists (e.g. Alwin 1987; Homans, 1974; Haller, Mach, and Zwicky 1991; Jasso 1980). Prior research has shown that people in the developed world reject equal pay (Jasso and Rossi 1977; Kelley and Evans 1993). They feel that, in general, pay should differ on several key dimensions including complexity (or level of skill required), workplace authority, and working conditions. How do workplaces measure up to those standards? Do workers actually feel fairly paid? To find out, we developed and pre-tested a series of questions on the topic of fairness of pay and fielded them in Australia, Hungary, and Poland in the early 1990s to representative, nationwide samples. This report is based on the subsamples of respondents who were currently employed and who answered the questions in the fairness module -- 1136 Australians, 595 Hungarians, and 834 Poles. The situation of the formerly socialist countries is of particular interest because one of socialism's clarion calls was its aspiration to greater fairness or "social justice" (Kolosi 1989; Hout and Wright 1992). ###5.2 RESULTS: DESCRIPTIVE 1. Measurement properties: Almost no-one thinks their income is too high. Most Australians think theirs is fair, or if not fair then “too low” (but not “far too low”). But in Eastern Europe, dissatisfaction with pay is far greater: only small minorities think their incomes are “fair”, the rest thinking their incomes are “too low” or “far too low”. The results: Table 5.2A. “All in all, is your income fair?” Three nations, circa 1992. ----------------------------------------- Poland Hungary Australia ----------------------------------------- far too low 35 27 6 too low 50 52 32 fair 15 19 58 too high 0 2 4 far too high 0 0 1 -------------------------- 100% 100% 100% Cases 1084 595 1250 ----------------------------------------- The two items have satisfactory measurement properties in the Developmental pretest. They are highly correlated and have very similar correlations with other relevant variables: Table 5.2B. How fair are your earnings? Inter-item correlations (Australian Developmental Pretest 1997; N=249). ---------------------------------------------------------- V311 V312 V312 .51 1.00 MALE .24 .20 AGE -.00 .01 Education .24 .15 Family Income .28 .22 Occupational status .20 .28 Subjective class (pyramid, v94) .36 .31 Class conflict -.18 -.21 Legitimate earnings: High status occupations .02 .09 Low status occupations -.16 -.05 ---------------------------------------------------------- Men, the well educated, those with higher incomes, those with higher status occupations, and especially those who feel themselves subjectively of high status all fell their pay is relatively fair. Those who feel their earnings are fair see somewhat less class conflict. But they do not have distinctive views about the legitimate earnings of either high or low status occupations. ###5.3 RESULTS: ANALYSIS Multiple regression analysis reveals some systematic differences (see the table below). Table 5.3A. Is your pay fair? Regression analysis for Australia, Hungary, and Poland. ----------------------------------- Metric Regression predicting overall fairness of pay ----------------------------------- Country: Australia (baseline) 0 Hungary -24 Poland -33 Age (years) ns Gender: Male -3 Owner(non-farm ns Education (years) ns Farmer -6 Supervises ns Earnings 5 Government employee -3 Subjective class: Working (baseline) 0 Middle 13 Residence: Rural ns Occupational status ns (Constant) 72 (R Square) 26% Std Error 32 ----------------------------------- Class position matters little. Aside from other influences, people working in high status occupation feel no more or less fairly paid, nor do the well educated differ from the poorly educated. Owners do not differ from employees, and supervisors do not differ from those they supervise. However those who identify with the middle class feel substantially more fairly paid than the working class. Farmers feel less fairly paid. Interestingly, government workers feel slightly, but significantly less fairly paid than do private sector workers. The old feel no different than do the young. People living in urban places feel no different from those living in rural places. Surprisingly, men feel slightly less fairly paid than do women. Finally, even aside from all the differences in class position and other factors captured by the model, substantial national differences remain with Australians coming closest to feeling fairly paid, Hungarians feeling much less fairly paid and Poles feeling even less fairly paid. Whether this difference widens or narrows in the post communist period is an important issue we can address in future surveys. [--------- ^Message “Core: Part-2” ENDS HERE ----------] -------------------------------------------------------------------------------- -------------------------------------------------------------------------------- (this separates two files) -------------------------------------------------------------------------------- -------------------------------------------------------------------------------- To: ISSP From: Inequality Drafting Group Subject: ISSP INEQUALITY MODULE: “Core-1, part 3 of 3”: Core questions. [--------- ^Message “Core: Part-3” BEGINS HERE ----------] --------------------------------------------------------------------------- --------------------------------------------------------------------------- SECTION #6: IMAGES OF CLASS - CORE QUESTIONS FOR THE 1999 INEQUALITY MODULE ###6.1 TEXT OF QUESTIONS: The questions all refer to a set of diagrams depicting different types of society: "These five diagrams show different types of society. Please read the descriptions and look at the diagrams and decide which you think best describes Australia today ... [[The ASCII version below is a rough sketch -- all the diagrams were equal area in the original, and of somewhat different proportions than these.]] Type A. A Small elite at the top, very few people in the middle and the great mass of people at the bottom. XXX X X X X X XXXXXXXXXXXXXXXXXXXX Type B: A society like a pyramid with a small elite at the top, more people in the middle, and most at the bottom X XXX XXXXX XXXXXXX XXXXXXXXX XXXXXXXXXXX XXXXXXXXXXXXX Type C: A pyramid except that just a few people are at the very bottom X XXX XXXXX XXXXXXX XXXXXXXXX XXXXXXXXXXX XXXXXXX Type D: A society with most people in the middle X XXX XXXXXXX XXXXXXXXXXX XXXXXXX XXX X Type E: Many people near the top, and only a few near the bottom" XXXXXXX XXXXXXXXXXX XXXXXXXXX XXXXXXX XXXXX XXX X ----- Concept: Perception of society – how equal/unequal now & in past V90. First, what type of society is [ Australia ] today -- which diagram comes closest? Type A . . . . . . . . . . . . . . . . . . 1 Type B . . . . . . . . . . . . . . . . . . 2 Type C . . . . . . . . . . . . . . . . . . 3 Type D . . . . . . . . . . . . . . . . . . 4 Type E . . . . . . . . . . . . . . . . . . 5 (Can't choose) . . . . . . . . . . . . . . - V91. And what do you think [ Australia ] was like 10 or 12 years ago, in the 1980s -- just your best guess? Type A . . . . . . . . . . . . . . . . . . 1 Type B . . . . . . . . . . . . . . . . . . 2 etc... ----- Concept: Normative preference for society: how equal/unequal V93. What do you think [ Australia ] ought to be like -- which would you prefer? Type A . . . . . . . . . . . . . . . . . . 1 Type B . . . . . . . . . . . . . . . . . . 2 etc... Now we would like to know where you and other people fit in. Please look at the diagram you think best describes [ Australia ] today, the one you chose first... ---- Concept: Subjective position in society, now & in past V94. Where would you say you and your family actually are -- at the top, somewhere in the middle, or toward the bottom? 1 Top . . . . . . . . . . . . . . . . . . . 1 2 . . . . . . . . . . . . . . . . . . . . . 2 3 . . . . . . . . . . . . . . . . . . . . . 3 4 . . . . . . . . . . . . . . . . . . . . . 4 5 . . . . . . . . . . . . . . . . . . . . . 5 6 . . . . . . . . . . . . . . . . . . . . . 6 7 Bottom. . . . . . . . . . . . . . . . . . 7 (Can't choose). . . . . . . . . . . . . . - V307. Where would you say you and your family were 10 or 12 years ago? 1 Top . . . . . . . . . . . . . . . . . . . 1 etc... 7 Bottom. . . . . . . . . . . . . . . . . . 7 ----- Concept: Class position of various occupations [These questions, together with the perceived occupational earnings questions (v26, v27, and v29) show how much people think each class ACTUALLY earns. Together with the questions on legitimate earnings (v41, v42, and v51) they show how much people think each class OUGHT to earn.] V95. Where would you say an unskilled worker in a factory fits in? 1 Top . . . . . . . . . . . . . . . . . . . 1 etc... 7 Bottom. . . . . . . . . . . . . . . . . . 7 V96. The chairman of a large national corporation? 1 Top . . . . . . . . . . . . . . . . . . . 1 etc... 7 Bottom. . . . . . . . . . . . . . . . . . 7 V98. Where would a doctor in general practice fit in? 1 Top . . . . . . . . . . . . . . . . . . . 1 etc... 7 Bottom. . . . . . . . . . . . . . . . . . 7 (NOTE: These were optional questions in 1992. They attracted substantial support in discussions leading up to 1992 and were actually done in 8 or 9 nations in 1992, as well as in Hungary and Australia in 1987. So we already have a considerable over-time perspective.) ###6.3 INTRODUCTION: These questions provide a more visual approach to the inequality issue, allowing respondents to readily report on perceived changes over times, and ideals for a just society. All in a framework that relies on exactly comparable stimuli across nations (not just "more" or "less"). The placement of occupations within societies is also an important clue to class politics (who is "us" and who "them"). Changes in these perceptions have been rapid in Eastern Europe since the fall of communism -- for example, in communist times only 8% of Hungarians thought their society was of the very inequalitarian Type A but by 1992 no less than 50% thought it was that inequalitarian. Other post- communist Eastern European nations are also seen as very inequalitarian, very different from Western nations: Table: ------------------------------------------------- (Inequalitarian) Equalitarian) Type Type Type Type Type A B C D E (Cases) ------------------------------------------------- Eastern Europe: Hungary 1987 8 27 25 33 7 100% 2400 1992 50 33 7 7 3 100% 1155 Bulgaria 59 24 7 9 1 100% 995 Czech Repub. 28 39 12 19 3 100% 1066 Poland 1992 59 22 7 9 4 100% 1403 1994 53 28 9 8 3 100% 1821 Western industrial: Australia 1995 10 28 30 32 1 100% 2217 1994 12 34 27 26 1 100% 1378 1993 15 32 27 25 1 100% 2092 1989 14 33 31 21 1 100% 4158 Canada 23 33 21 20 3 100% 949 Italy 14 34 21 23 9 100% 996 Finland 13 24 24 38 1 100% 1658 New Zealand 24 39 23 13 1 100% 1189 Norway 8 16 25 49 3 100% 1397 Developing: Philippines 44 27 10 12 8 100% 1179 ------------------------------------------------- Indications from the existing data are that these "image" questions capture important aspects of the inequality issue. For example preliminary analysis indicates that respondents who see their societies as inequalitarian place themselves (and the various rated occupations) lower in the class structure than do people who perceive more "room at the top". They complement the perceived and ideal occupational earnings questions (which focus on the magnitude of differences between occupations) by assessing how much "social space" there is at the top. The questions on where respondent places "unskilled worker", "chairman", and “doctor” in the hierarchy are important in linking these questions with the questions on the actual and legitimate earnings of different occupations that are the core of the module: (1) They tell us where in the society as a whole respondents place the occupations and (2) (conversely) how much respondents think each segment of society does earn, and ought to earn. Also (3) there have been changes in where respondents place occupations since the fall of communism in Eastern Europe. However, we have tentatively dropped the question on where “farm laborer” fits in the hierarchy, in order to reduce the number of questions. See "Images of Class: A Comparative Analysis of Public Perceptions in Hungary and Australia". M.D.R. Evans, Jonathan Kelley, and Tamas Kolosi. American Sociological Review, 57 (1992): 461- 482. (file Pyr1.doc, a Microsoft Word7 file). NEW QUESTION: The addition "you and your family 10 or 12 years ago" is important in assessing the effects of economic change -- an especially important topic in Eastern Europe, but also important in the West. Attachments: 5Pyr1 -- Images of class ASR article 5Pyr2 -- Images of class, many nations, ISSP 1992 WWA article. --------------------------------------------------------------------------- --------------------------------------------------------------------------- ###7.5 SUBJECTIVE SOCIAL CLASS – NOW & 10 YEARS AGO - CORE QUESTIONS FOR THE 1999 INEQUALITY MODULE ### -------------- Concept #1: Your subjective class position NOW (V73 & V305) ### -------------- Concept #2: Your subjective class position 10 OR 12 YEARS AGO (ie prior to the fall of communism & introduction of market reforms (V304 & V306) QUESTIONS: V73. In our society there are some social groups which are higher and some which are lower. Where do you think you are on this scale? High . . . . . . . 1 2 3 4 5 6 7 8 9 Low . . . . . . . 10 (Can't choose) 99 V304. And where were you 10 or 12 years ago? High . . . . . . . 1 2 etc. Low . . . . . . . 10 Subjective social class is of obvious and long-standing importance in sociology. SEE, for example: ATTACHMENT: file "5Class.doc", text of Kelley and Evans, "Class and Class- conflict…", American Sociological Review, 1995: 157-178.) The original question (v73) was originally suggested by Tom Smith to give a second subjective class measure, to add to the traditional "working class/ middle class" question that is a standard part of the ISSP demographic battery. But the traditional question does not work reliably in many nations -- it uses highly politicized words and concepts in many European and formerly Communist nations, and so is not suitable for cross-national comparisons (see the discussion in Evans, Kelley and Kolosi, "Images of Class…", American Sociological Review, 1992: 461-482). We therefore add a second question in a format similar to V73 to provide a second item measuring this important concept (in the questionnaire, this will be placed far away from the first question): ###7.6 NEW ‘STANDARD OF LIVING’ QUESTIONS: V305. Thinking about your income and standard of living, are you and your family near the top of society, near the bottom, or somewhere in between... Top . . . . . . . 10 9 8 7 6 5 4 3 2 Bottom . . . . . . . . 1 (Can't choose) 99 V306. And where was your income and standard of living 10 or 12 years ago? Top . . . . . . . 10 9 etc Bottom . . . . . . . . 1 To evaluate subjective changes in the past 10 or 12 years -- a period that covers the fall of communism in Eastern Europe -- we propose to ask both the class questions retrospectively. The retrospective question would go just after the corresponding "now" question, as indicated above. These questions would allow us to evaluate perceived changes due to economic changes in the West as well as the even larger changes in Eastern Europe. ###7.7 MEASUREMENT PROPERTIES People's subjective class identification is typically with the middle classes or just below, with very few people identifying with the top or the bottom classes. Residents of prosperous Western nations generally identify with somewhat higher classes than do Eastern European residents. Interestingly, changes in Eastern Europe since in the fall of communism seem to have shifted ordinary people’s class identification toward the lower classes, while changes in the West have had little effect. For example: Table 7.7A. Class identification, 1987 and 1992 --------------------------------------------------- Hungary Hungary Australia Australia 1987 1992 1987 1993 --------------------------------------------------- High: 1 0 0 1 1 2 0 0 1 1 3 2 1 8 8 4 6 4 19 21 5 30 9 43 40 6 22 23 14 15 7 16 24 8 9 8 12 20 4 4 9 5 11 1 1 Low: 10 6 8 1 1 ------------------------------------ 100% 100% 100% 100% Mean 6.3 7.1 5.1 5.1 Cases 2542 1239 1529 1870 --------------------------------------------------- The original subjective class measure (V73) worked well (see, for example, Kelley and Evans, "Class and Class-conflict…", American Sociological Review, 1995: 157- 178.) In the Australian Developmental Pretest, the new item “income and standard of living” item (v305) also performed well: it was highly correlated with the existing item and had very similar correlations with relevant other variables. The new measures of class 10 years ago also seem to have worked well (V304 & V306) in the in the Developmental pretest: Table 7.7B. Subjective class now & 10 years ago. (Australian Developmental Pretest, 1997; N=249) --------------------------------------------------------------- Now: 10 years ago: ------------- ----------------- SUBCLS INCCLS SUB10AGO INC10AGO V73 V305 V304 V306 V73: SUBCLS 1.00 V305: INCCLS .78 1.00 V304: SUB10AGO .58 .41 1.00 V306: INC10AGO .43 .50 .62 1.00 MALE .09 .07 .11 .08 AGE .01 -.08 .08 .14 Income .43 .57 .12 .11 Education .44 .39 .22 .21 Occupational status .46 .37 .26 .17 Legitimate pay: Low status occupations -.24 -.18 -.16 -.15 High status occupations .22 .22 .09 .13 --------------------------------------------------------------- ###7.8 POSSIBLE ANALYSES WITH THESE SUBJECTIVE CLASS MEASURES 1. What shapes subjective class position in different nations? These items, in conjunction with existing ISSP demographic and background questions, allow many valuable analyses. The influence of objective class on subjective class in different nations is an important question – is education a more important determinant of class in richer countries or in poorer, or in Asian nations rather than in the west, or among Protestants rather than Catholics? Is money a more important determinant of class in Western capitalist nations than it is in the formerly communist Eastern Europe? Such questions can be estimated from: (Eq. 7.1A) Subjective class identification now = R’s education; Income; R’s occupation; supervisor; self-employed; government employee; size of workplace; union member; Father’s occupation; father’s education; urban-rural; age; sex; religion; 2. Discrimination & disadvantage in subjective class Discrimination against women, minorities, those born into lower status families, the young, rural residents, and the like can also be with these variables. (1) For example, if a society devalues women and their work, or values urban life- styles over rural, this will show up as a negative coefficient for “sex” or “rural” in equation 7.1. (2) An even more powerful analysis is possible, catering for the possibility that (for example) education and occupational status pay off less for women, or minorities, or those born into low status families, or the young. This is done by estimating equations predicting subjective class separately for two groups: GROUP A (eg men, majority ethnicity, old, high status family background): (Eq. 7.2A) Subjective class identification now = R’s education; Income; R’s occupation; supervisor; self-employed; government employee; size of workplace; union member; GROUP B (eg women, minority, young, low status family background): (Eq. 7.2B) Subjective class identification now = R’s education; Income; R’s occupation; supervisor; self-employed; government employee; size of workplace; union member; 3. Changes over time The effects of economic change over the last decade – notably, but not exclusively the fall of communism – can be estimated using the information on subjective social class in the past (V304 & V306). A first issue is how much things have changed in different nations, estimated either as a simple difference (subjective class now - subjective class 10 years ago) or from a regression equation: (Eq. 7.3) Subjective class identification now = SUBJECTIVE CLASS IDENTIFICATION 10 YEARS AGO; R’s education; Income; R’s occupation; supervisor; self-employed; government employee; size of workplace; union member; Father’s occupation; father’s education; urban-rural; age; sex; religion; 4. Who gains & who loses from economic change? There are strong reasons to think that the market oriented economic change that has taken place in many nations over the last decade, and especially in Eastern Europe, has benefited some more than others – eg men more than women; the young more than the old; the better educated more than the poorly educated; urban more than rural residents. This too can be measured systematically with these data: GROUP A (eg men, majority ethnicity, old, high status family background): (Eq. 7.4A) Subjective class identification now = SUBJECTIVE CLASS IDENTIFICATION 10 YEARS AGO; R’s education; Income; R’s occupation; supervisor; self-employed; government employee; size of workplace; union member; GROUP B (eg women, minority, young, low status family background): (Eq. 7.4B) Subjective class identification now = SUBJECTIVE CLASS IDENTIFICATION 10 YEARS AGO; R’s education; Income; R’s occupation; supervisor; self-employed; government employee; size of workplace; union member; --------------------------------------------------------------------------- --------------------------------------------------------------------------- SECTION #8: CLASS CONFLICT - CORE QUESTIONS FOR THE 1999 INEQUALITY MODULE ###8.1 -------------------- Class conflict scale: v67, v68, v303 QUESTIONS: In all countries there are differences or even conflicts between different social groups. In your opinion, in [country] how much conflict is there between... V67 Poor people and rich people? 1 Very strong conflicts 2 Strong conflicts 3 Not very strong conflicts 4 There are no conflicts - (Can't choose)" V68 The working class and the middle class? V303 (New item): People at the top of society and people at the bottom? KNOWN USAGE: B&K94A & B; B&U90, E&K95b, K96, K&E95 INTRODUCTION: Marx thought that perceptions of class conflict would be a strong force impelling people towards left-wing politics. Empirical assessment of the class conflict scale in ISSP Inequality-1987 strongly bear this out: people who see lots of conflict do embrace left-wing politics. However,the level of conflict ordinary people see is not high, especially in Central European nations: Table 8.1A. Percentage Distribution of Perceived Consensus or Conflict between Classes: Six Western Democracies, 1987-1988. ISSP 1987. ------------------------------------------------------------------- Not Very Very Strong Strong Strong No Conflict Conflict Conflict Conflict Cases ------------------------------------------------------------------- All nations Working class vs middle 20 63 15 2 100% 6,962 Poor vs rich 8 46 35 11 100% 6,963 (Scale) 12 55 27 6 100% 6,801 Anglo-Celtic nations (Scale) United States 6 52 34 9 101% 1,171 Great Britain 8 52 30 10 100% 1,125 Australia 10 56 28 6 100% 1,580 Central European (Scale) Switzerland 14 61 22 3 100% 836 Austria 16 57 22 4 99% 857 Germany 19 54 22 5 100% 1,232 ------------------------------------------------------------------- We strongly advocate keeping these items as we think that they may be among the "items to watch" during the marketization reforms. In particular, work that has been done on these items so far finds only a moderate connection between perceptions of conflict and actual location in the class structure, a result that may reflect the distorting perceptual effects of homogeneous reference groups. But that "marketization" may increase people’s perceptions of conflict remains a live issue (and one that animates some political elites’ fears of impending social disorder). Because the concept is important, we add a new item to strengthen the scale. The empirical properties of the existing two-item scale are good, but three-item scales are much to be preferred when possible as greatly reducing measurement error, so we recommend a third item for the scale (V303). In the Australian Developmental Pretest, the new item performed well: it was highly correlated with the existing items and had very similar correlations with relevant other variables: Table 8.1B. Class conflict: Correlations (Australian Developmental pretest, 1997; N=249) ---------------------------------------------------- CPOOR CTOP CCLASS v67 v303 v68 CPOOR 1.00 CTOP .71 1.00 CCLASS .59 .50 1.00 MALE -.07 -.12 -.10 AGE -.02 -.06 -.11 Income -.08 -.13 -.12 Education -.08 -.10 .06 Occupational status -.07 -.10 -.03 Subjective class: Now -.16 -.16 -.12 Subj class: 10 years ago -.11 -.14 -.07 Legitimate earnings: Low status occupations .09 .05 .08 High status occupations -.19 -.20 -.18 Political party (right) -.22 -.16 -.14 ---------------------------------------------------- ------------- end of document --------------- [--------- ^Message “Core: Part-3” ENDS HERE ----------] -------------------------------------------------------------------------------- -------------------------------------------------------------------------------- (this separates two files) -------------------------------------------------------------------------------- -------------------------------------------------------------------------------- To: ISSP From: Inequality Drafting Group Subject: ISSP DISCRIMINATION MODULE: “discrm-3” perceived discrimination ------------------------------------------------------------------------ POSSIBLE NEW QUESTIONS FOR THE ISSP INEQUALITY MODULE: DISCRIMINATION, ATTRIBUTION OF SUCCESS -- POOR, WOMEN, MIGRANTS ------------------------------------------------------------------------ Australia (convenor) J. Kelley M.D.R. Evans Czech Republic P. Mateju Norway K. Knudsen K. Skjak Poland B. Cichomski P. Morawski Slovenia B. Malnar This is suggestion for new questions for the inequality module (to add to the ‘core’ questions that have already been circulated). This discussion presumes that the 2 family background questions which are discussed elsewhere (see file “background-2”) are in the final questionnaire – if not, they would have to be added here to allow the analyses discussed below. ###-------------------------- QUESTIONS ANALYZED Discrimination on the basis of family background: FAPROF 64 246 p47q8b father professional --> better ch RICHFAM 67 243 p47q8a rich parents --> better chance FASECUR 61 245 p47q8c father secure --> better chance FAMBACK 47 243 p47q8d achieve in life depends family ba DPOORALL 64 243 p45q1a Discrimination: peole from poor f DPORJOB 46 243 p46q3b Poor kids discrim - jobs DPORED 55 244 p46q3a Poor kids discrim - education DPORPAY 39 245 p46q3c Poor kids discrim - pay Discrimination against Women: DWOMALL 59 241 p45q1c2 Discrimination: women DWOMJOB 48 246 p46q4b Women discrim - jobs DWOMPAY 54 246 p46q4c Women discrim - pay D2WOMJOB 61 244 p47q7a Women: job opportunities D2WOMPAY 65 243 p47q7b Women: income & wages, compared w DWOMED 31 246 p46q4a Women discrim - education Discrimination on the basis of ethnicity, (i) in the labor market: DMIGALL 63 244 p45q1b2 Discrimination: migrants DMIGEMPL 49 243 p46q2b Migrant Discrim - steady work? DMIGJOB 51 242 p46q2c Migrant Discrim - good jobs? DMIGPAY 45 243 p46q2d Migrant Discrim - pay DMIGED 38 245 p46q2a Migrant Discrim - education? BOSSPRJ 21 231 p47q6 Migrants: employers discriminate? (ii) by government DMIGWELF 31 244 p46q2e Migrant Discrim - welfare DMIGHLTH 31 244 p46q2f Migrant Discrim - health care DMIGLAW 36 242 p46q2h Migrant Discrim - courts DMIGHOUS 36 242 p46q2g Migrant Discrim - housing DMIGTALK 33 243 p46q2i Migrant Discrim - freedom of expression Perceived meritocracy -- rewards reflect inputs: ISSKILL 62 246 p48q2d In Australia: rewarded for intelligence ISEFFORT 60 245 p48q2b In Australia: rewarded for effort ISNEED 58 245 p48q2c In Australia: people get what they need ISEQOPP 58 242 p48q2a In Australia: equal opportunities Perception: Poor can (or can not) help themselves: CANRAISE 62 245 p49q3b can raise standard living by work IMPRVLOT 55 242 p49q3d Poor people improve if they tried DECIDHOW 64 245 p49q3c have freedom in deciding how to l NOTTRY 39 244 p49q3a poor because of the lack of effort NOTRAISE 48 242 p49q3e Many poor people can do little to Perception: People need to do criminal things to get ahead ORGCRMBZ 25 244 p49q5f Bz success: link to organized crime CROOKED 47 245 p49q5d1 Bz success: crooked dealing CORRUPT 27 246 p49q5e Bz success: be corrupt MAFDEAL 29 244 p49q5d2 Bz success: made deals with Mafia ORGCRMEN 30 244 p49q5g Bz success: Relying on organized DISHONST 30 245 p49q5b Bz success: have to be dishonest CONSCIEN 43 246 p49q5c Bz success: & conscience dont mix TYCONADV 58 243 p49q5a Bz success: taking unfair advantage GAORGCRM 14 244 p49q4f Bz success: with organized crime? GACHEAT 20 245 p49q4e Bz success: cheat and steal? GATAKEAD 42 245 p49q4d Bz success: take advantage of others Attribution: Meritocracy ASMART 74.07 246 p48q1e v8 Get ahead: Natural ability? ASTRIVE 81.96 245 p48q1f v9 Get ahead: Hard work AAMBIT 81.54 247 p48q1d v7 Get ahead: Ambition Demographic & background variables; other scales MALE 249 p58q1 What is your sex? AGE 48 246 p58q2 What is your age in years? EDUC 14 236 p57q2b Education OCCSTAT 61 235 Occupational status: current or previous LNFAMINC 4 229 Family income (log) SUB#NOW 56 242 Subj class & $ (0-100): Now D#YOU 50 241 Pyramid: Your class, now & future EYOU_NOW 46 242 p55q6b YOU: economic sys NOW: benefits? EALL_NOW 44 245 p4q8b AUSTRALIA: Econ system NOW benefic PYRNOW 44 240 Pyr: Society NOW (equal high) D#LOW 27 241 Pyramid: Class of low occups (1-7) D#HIGH 76 242 Pyramid: Class of high occups (1-7) CLSCNF# 45 241 Perceived class conflict O$YOU 11 223 p8q9a ought $: you yourself earn FAIR#YOU 67 245 Your earnings fair & deserved O#LOW 10 232 Legitimate earnings: Ordinary workers O#HIGH 12 231 Legitimate earnings: Elite occups (bz, govt, prof) PYRIDEAL 73 241 Pyr: Ideal Society (equal high) DV_PREF 48 237 p53q4c Economic development: you prefer? TRIGHT 38 248 For right wing party (thermometer) B#AUTHR 74 249 Legit basis pay: Authority, decisions B#FAM 65 249 Legit basis pay: Kids, wife ###---------------- OVERVIEW: RECOMMENDED QUESTIONS (MODERATE PRIORITY) ###-----Concept: Perceived discrimination against minorities ###-----Concept: Perceived discrimination against women ###-----Concept: Perceived discrimination against those from poor families -- ALL THREE RECOMMENDED FOR INCLUSION IN THE MODULE WITH MODERATE PRIORITY In getting the things one needs to lead a good life in Australia, do these groups receive equal treatment with others, or do they suffer discrimination and unfair treatment... V321 (dMigAll): a. Migrants -- compared to others are they treated Much worse Worse Equal Better Much better V322 (dWomAll): b. Are women treated Much worse Worse Equal Better Much better V323(dPorEd): c. How about children of poor families, do they suffer discrimination in access to education? Yes, definitely Yes, probably Might or might not No, probably not No, definitely not V324 (dPorJob): d. Do people who grew up in poor families suffer discrimination in getting good jobs? Yes, definitely Yes, probably Might or might not No, probably not No, definitely not ###-----Concept: Perception—country is a meritocracy (ie skill, effort, & intelligence are rewarded): -- RECOMMENDED FOR INCLUSION IN THE MODULE WITH MODERATE PRIORITY V325(IsEffort): In Australia people get rewarded for their effort Strongly agree Agree Neither agree nor disagree Disagree Strongly disagree v326(IsSkill): In Australia people get rewarded for their intelligence and skills Strongly agree Agree Neither agree nor disagree Disagree Strongly disagree ###--- Concept: Perception -- Crime, dishonesty needed to get ahead. -- RECOMMENDED FOR INCLUSION IN THE MODULE WITH MODERATE PRIORITY v328 (CROOKED p49q5d1) Many business tycoons got there by crooked dealing Strongly agree Agree Neither agree nor disagree Disagree Strongly disagree v327 (ORGCRMBZ p49q5f) f. Behind the scenes in nearly every successful business, there’s a link to organized crime. Strongly agree Agree Neither agree nor disagree Disagree Strongly disagree ###---------------------------------------------------------------------- ###---------------------------------------------------------------------- ##---------------- ACTUAL DISCRIMINATION We can assess the actual extent of discrimination against those born into poor families, discrimination against women, and (in many countries) discrimination against minorities by using existing ISSP background and demographic data (including with the 2 new background variables proposed previously in file “background- 2”). The basis idea is to estimate a suitable status attainment model twice: once for the ‘under- privileged” group (eg women, migrants, those born into poor families) and once for the ‘privileged’ group. ###--- What A Model Of This Sort Can Do: This type of analysis addresses issues of discrimination/equal opportunity – issues raised by many ISSP members at the May meeting in the Netherlands. For example, using existing ISSP demographics in connection with this model we can address discrimination against women in education, occupation, and income. Similarly, we can address discrimination against religious and ethnic minorities. We can also address regional (sub-national) differences, urban- rural differences, and the like. This model also allows estimates of changes over time in the intensity of discrimination against different groups. For example, the links between family background and education for older cohorts show what happened many years ago, while the corresponding links for younger people show what is happening in more recent years. This allows us to assess the equity impact of changes educational policy (eg more market oriented university sectors in recent years in many nations). Similarly, these data would allow us to assess sex differences in education in different historical periods. Equally, these data would allow us to assess changes over time in the differences between migrants and the native born in access to good jobs and incomes. ###--- Specification of the Model: All these issues can be addressed with a version of a basic status attainment model estimated separately for men & women, majority & minority, old & young, etc: (one could also conduct a parallel analysis using Goldthorpe’s class typology, or Wright’s, or add other background variables from the ISSP). GROUP A (eg men, majority ethnicity, old): (Eq. 1.1A) R’s education = Father’s occupation, father’s education; Family cultural capital; Birth cohort; (Eq. 1.2A) R’s occupation = Father’s occupation, father’s education; Family cultural capital; Birth cohort;R’s education; (Eq. 1.3A) R’s earnings = Father’s occupation, father’s education, Family cultural capital; Birth cohort;R’s education, R’s occupation; supervisor; self- employed; government employee; size of workplace; union member; urban-rural; GROUP B (eg women, minority, young): (Eq. 1.1A) R’s education = Father’s occupation, father’s education; Family cultural capital; Birth cohort; (Eq. 1.2A) R’s occupation = Father’s occupation, father’s education; Family cultural capital; Birth cohort;R’s education; (Eq. 1.3A) R’s earnings = Father’s occupation, father’s education, Family cultural capital; Birth cohort;R’s education, R’s occupation; supervisor; self- employed; government employee; size of workplace; union member; urban-rural; To estimate changes over time in discrimination, one would begin by comparing the predicted values for, say, women and men at different time points. To assess the more complex possibilities of change over time, one would need to add appropriate interaction terms to these equations. For example, to assess whether high status fathers are less successful in advantaging their offspring over time, one would add to each equation a variable computed as: ... (Father’s occupation) times (Birth cohort); ... ###---------------------------------------------------------------------- ###---------------------------------------------------------------------- ###---------------- PERCEPTIONS OF DISCRIMINATION: POOR, WOMEN & MINORITIES ###--- Perceived Discrimination Against Migrants: In the Developmental Pretest, we took discrimination against migrants as the prototypical form of discrimination and asked about many aspects of it (see page 46 of the questionnaire). For discrimination against women or against those born to poor families, we asked only about a few key aspects. An important question, then, is whether all aspects of discrimination go together (which makes it easier to measure!) or whether there are many empirically separate aspects. Perceptions of discrimination against migrants divide into two separate aspects (see table 1): Perceived discrimination in the LABOR MARKET, presumably mainly by private rather than government bodies – discrimination in getting employment, getting good jobs, and in pay. Levels of discrimination are perceived as being relatively high. (Perceptions of discrimination in education straddle between this and the government factor.) Perceived discrimination by GOVERNMENT – in welfare, health care, housing, the courts, freedom to express political views. Levels of discrimination are seen as relatively low. (A technical complication is that factor analysis is sometimes sensitive to differences in the means, and so may overstate the split between these to aspects of discrimination.) It is important to note that all aspects of labor market discrimination are thought to go together: there are NOT separate views about unemployment; separate views about getting good rather than poor jobs; separate views about pay. Instead people believe that there is discrimination in all of these, or in none. This has the important practical point that one, or a very few, questions can reasonably be used to cover the whole concept – we do not need separate questions about each aspect. We construct separate scales to measure each aspect. The scales are simple averages of the component items. Each item is scored from zero (low= little discrimination seen) to a high of 100 (high= much discrimination seen). The scales: dMigGov# "Migrants: Government perceived to discriminate" Scale = MEAN ( DMIGWELF, DMIGHLTH, DMIGLAW, DMIGHOUS, DMIGTALK ). dMigPri# "Migrants: Private business perceived to discriminate" Scale = MEAN (DMIGALL, DMIGEMPL, DMIGJOB, DMIGPAY ). Perceived discrimination in the labor market is a little more highly correlated with things of interest in the inequality module than is perceived discrimination by the government, but the differences are not large (see table 2, columns 2 and 3). Table 1. Perceptions of discrimination on the basis of family background, ethnicity, and gender. Factor loadings, varimax rotation. Australia 1997; n=249. ---------------------------------------------------------------- Factor1 Factor 2 Factor 3 Factor 4 Factor 5 ---------------------------------------------------------------- (Migrants: Private business perceived to discriminate = dMigPri#) DMIGALL .06 .28 .73* .16 .04 DMIGEMPL .32 .13 .81* .10 .10 DMIGJOB .38 .14 .76* .13 .11 DMIGPAY .42 .19 .58* .09 .15 DMIGED .57* .10 .45 .08 .20 BOSSPRJ .26 .13 .43 .02 .33 (Migrants: Government perceived to discriminate = dMigGov#) DMIGWELF .87* .11 .12 .03 .07 DMIGHLTH .86* .14 .17 .06 .07 DMIGLAW .73* .18 .26 .10 .17 DMIGHOUS .71* .12 .36 -.03 .12 DMIGTALK .67* .10 .15 .00 .04 (Women: Perceived discrimination = dwom#) DWOMALL .10 .77* .33 .12 .09 DWOMJOB .21 .78* .29 .04 .14 DWOMPAY .25 .79* .12 .04 .19 D2WOMJOB .05 .80* .17 .07 .04 D2WOMPAY .10 .77* -.01 .02 .08 DWOMED .52* .48 -.02 .08 .18 (Poor: Perceived discrimination = dPoor#) DPOORALL -.09 .11 .32 .24 .67* DPORJOB .26 .13 .08 .24 .78* DPORED .10 .20 .16 .22 .76* DPORPAY .42 .10 -.06 .29 .66* (Poor: Worse chances for unspecified reasons = dPoor#2) FAPROF .01 .01 .09 .87* .16 RICHFAM .06 .01 .21 .83* .20 FASECUR .04 .13 .04 .80* .15 FAMBACK .09 .09 .06 .54* .38 ------------------------------------------------------------- ###--- Perceived Discrimination Against Women: There is a clear factor reflecting people’s views about labor market discrimination against women. It is, highly correlated with views about discrimination against migrants and the poor (table 2, column 5). We measured it with a simple additive scale: dwom# "Perceived discrimination against women" Scale = MEAN ( DWOMJOB, DWOMPAY, DWOMALL, D2WOMJOB, D2WOMPAY). ###--- Perceived Discrimination against people born to poor families Empirically, perceived discrimination against people born into poor families splits into two separate factors, for reasons that are not clear (see table 1 above): One factor (dPoor#) is explicitly discrimination – eg “Do people who grew up in poor families suffer discrimination in getting good jobs?” and directly parallels the discrimination questions on migrants and women. The scale: dPoor# "From poor families: Perceived discrim" Scale= MEAN(DPORJOB ,DPORED, DPORPAY, DPOORALL ). The other factor (dPoor#2) seems to be asking for mere correlation rather than deliberate discrimination (it also includes some standard ISSP items). Example: “A person whose father is a professional has a better chance of earning a lot of money than a person whose father is an unskilled laborer”. Respondents could agree to this if there was real discrimination (for example, along Bordeaux’s line of elite families using cultural markers of class to discriminate in favor of the offspring of their class). Many people would find disadvantage stemming from this sort of process improper. But equally respondents could agree for other, relatively harmless reasons – for example, that professional families cultivate a taste for education and cognitively complex work that leads their children get high education and good jobs, and hence only indirectly high pay. Many people would find disadvantage stemming from this sort of process unobjectionable. So in that, it is very different. The scale: dPoor#2 "From poor families: alt perceived discrim scale"; Scale = MEAN (FAPROF, RICHFAM, FASECUR ). These two scales have quite different correlations with other variables (table 2, columns 7 and 8; see especially lines 15, 16, 17, 19, 21, and 22). dPoor# looks more ideological, more linked to party politics, to perceptions of conflict, to inequality, and of course to perceived discrimination against women and minorities. These sharp differences between scales that were meant to be conceptually similar is worrying. They cannot both be correct. Given the conceptual and empirical similarities between dPoor# and the discrimination questions dealing with women and migrants, it would seem that dPoor#2 is probably flawed and should be dropped from the questionnaire. Table 2. Perception of discrimination: Correlations. Australia, 1977. N=249. ------------------------------------------------------------------------------- DMIGALL DMIGPRI# DMIGGOV# DWOMALL DWOM# DPOORALL DPOOR# dPOOR#2 (1) (2) (3) (4) (5) (6) (7) (8) ------------------------------------------------------------------------------- 1. DMIGALL 1.00 2. DMIGPRI# .73 1.00 3. DMIGGOV# .33 .59 1.00 4. DWOMALL .52 .48 .35 1.00 5. DWOM# .43 .49 .40 .83 1.00 6. DPOORALL .35 .33 .18 .31 .27 1.00 7. DPOOR# .28 .40 .39 .34 .39 .70 1.00 8. DPOOR#2 .26 .28 .12 .21 .19 .41 .50 1.00 9. MALE -.08 -.09 -.03 -.22 -.22 -.02 .00 .06 10. AGE -.06 -.06 -.01 -.02 -.02 .01 .01 -.01 11. EDUC .18 .23 .16 .15 .15 .02 .12 .18 12. OCCSTAT .14 .10 .03 .10 .06 -.04 -.04 .11 13. LNFAMINC .05 -.02 -.09 .00 -.05 -.11 -.13 .02 14. SUB#NOW .10 .05 .06 .07 .04 -.18 -.14 -.09 15. EYOU_NOW -.09 -.12 -.15 -.13 -.12 -.28 -.30 -.10 16. PYRNOW .04 .11 -.02 -.03 .03 -.12 -.15 -.05 17. CLSCNF# .10 .15 .09 .17 .16 .23 .28 .13 18. O$YOU .02 .04 .02 -.11 -.09 -.12 -.06 .07 19. O#LOW .07 .13 .10 .08 .06 .16 .16 .02 20. O#HIGH .10 .02 -.10 .01 -.07 -.07 -.10 .08 ------------------------------------------------------------------------- ###--- Links between Perceived Discrimination & other Variables Perceptions of discrimination are, on the whole, little correlated with other variables in the inequality module (table 2). Perceptions of discrimination against migrants have few correlations of interest (column 2 & 3). Nor do perceptions of discrimination against women (column 5). Only perceptions of discrimination against those born into poor families show significant links, and that only with one version of the concept (column 7, but not column 8). Regression analysis, which addresses these issues more systematically, shows that perceptions of discrimination are not closely tied to the social structure (table 3). (1) The better educated are more likely to see discrimination against migrants, women, and the poor. (2) Men are less likely to see discrimination against women. (3) Those who have benefited from recent economic change see less discrimination against the poor – possibly projecting their favorable situation onto others. (4) Those who see society as relatively equalitarian also tend to see little discrimination against the poor. In all, it is perceptions of discrimination against the poor that has the most links with other variables in the inequality module. Table 3. Who perceives discrimination against (1)migrants, (2) women, or (3)the poor. Standardized partial regression coefficients. Australia, 1997. N=249. ------------------------------------------------------- DMIGPRI# DWOM# DPOOR# ------------------------------------------------------- MALE -.11 -.24* .03 AGE -.04 -.01 -.01 EDUC .26* .20* .21* OCCSTAT .01 .01 -.06 LNFAMINC -.06 -.05 -.09 EYOU_NOW -.13 -.10 -.28* PYRNOW .08 .02 -.13* ------------------------------------------------------- * p<.05 Regression analyses also that (table 3): (1) perceptions of discrimination against migrants has no statistically significant effect on inequality’s “usual suspects”: perceptions of class conflict; perceived type of society; position of low status occupations in society; the legitimate pay of low status occupations; or the legitimate pay of elite occupations. (2) perceptions of discrimination against women have no statistically significant effect on these variables either. (3) perceptions of discrimination against those born into poor families has substantial, statistically significant effects. Those who see more discrimination also see more class conflict; they see Australia as being a less equalitarian type of society; they see low status occupations as positioned lower in society; and they prefer greater pay for low status occupations. Table 3. Consequences of perceptions of discrimination. Effects on (1) class conflict; (2) perceived type of society; (3) position of low status occupations in society; (4) the legitimate pay of low status occupations; and (5) legitimate pay of elite occupations. Standardized partial regression coefficients. Australia, 1997. N=249. -------------------------------------------------------------- CLSCNF# PYRNOW D#LOW O#LOW O#HIGH -------------------------------------------------------------- DMIGPRI# .03 .14 .01 .10 .12 DWOM# .03 .03 -.01 -.04 -.03 DPOOR# .23* -.20* -.19* .16* -.06 MALE -.08 .02 -.03 -.00 .15* AGE -.10 -.08 -.23* .01 .39* EDUC -.06 .07 -.07 -.01 .07 OCCSTAT -.01 .18* -.05 -.03 -.03 LNFAMINC -.08 .00 -.09 .05 .25* EYOU_NOW -.07 .05 -.02 .06 .18* -------------------------------------------------------------- * p<.05 ###--------------- PERCEIVED DISCRIMINATION: RECOMMENDED ITEMS We recommend asking a single item on perceived discrimination against migrants, one on perceived discrimination against women, and two questions on perceived discrimination against those born to poor families. The questions (exact wording was given earlier): “In getting the things one needs to lead a good life in Australia, do these groups receive equal treatment with others, or do they suffer discrimination and unfair treatment...” V321 (DMIGALL): a. Migrants -- compared to others are they treated... (better/worse) V322 (DWOMALL): b. Are women treated... (better/worse) V323 (DPORED): c. How about children of poor families, do they suffer discrimination in access to education? (Yes/no) V324 (DPORJOB): d. Do people who grew up in poor families suffer discrimination in getting good jobs? (yes/no) ###-----Concept: Perceived discrimination against minorities The single recommended item DMIGALL is correlated r=.7 with the multiple item scale and has very similar correlations with other variables (table 2; compare columns 1 & 2). So it is an adequate substitute for a longer scale. ###-----Concept: Perceived discrimination against women The single recommended item, dWomAll, is correlated r=.8 with the full multiple item scale and has similar correlations with other variables (table 2; compare columns 4 & 5). So it is an adequate substitute for a longer scale. ###-----Concept: Perceived discrimination against those from poor families The two recommended items (dPorJob and dPorEd) are highly correlated and have reasonably similar correlations with other variables (table 4, columns 1 & 2). To other items (dPorPay and dPoorAll also have reasonable measurement properties but, for lack of space, we do not recommend them strongly (columns 3 & 4). Table 4. Correlations among items measuring perceived discrimination against those born to poor families. Australia, 1997. N=240. ---------------------------------------------------- DPORJOB DPORED DPORPAY DPOORALL ---------------------------------------------------- DPORJOB 1.00 DPORED .60 1.00 DPORPAY .71 .51 1.00 DPOORALL .49 .56 .39 1.00 MALE .00 -.07 .07 -.02 AGE .03 -.04 .01 .01 EDUC .14 .05 .15 .02 LNFAMINC -.07 -.15 -.13 -.11 SUB#NOW -.05 -.20 -.04 -.18 EYOU_NOW -.20 -.31 -.21 -.28 EALL_NOW -.19 -.28 -.13 -.25 PYRNOW -.15 -.11 -.11 -.12 D#LOW -.19 -.13 -.12 -.14 D#HIGH -.04 -.12 -.06 -.03 CLSCNF# .23 .29 .16 .23 O#LOW .16 .08 .15 .16 O#HIGH -.06 -.10 -.11 -.07 DV_PREF -.22 -.29 -.16 -.34 TRIGHT -.23 -.26 -.19 -.21 B#AUTHR -.22 -.15 -.17 -.17 B#FAM .13 .23 .17 .13 ---------------------------------------------------- ###---------------------------------------------------------------------- ###---------------------------------------------------------------------- ###-------------------- PERCEPTIONS: RELATED TOPICS ###--- Perception: Meritocracy (Skill, effort, intelligence rewarded): The perception that rewards are distributed on the basis of merit is in part the opposite of discrimination: achievement based on effort and merit precludes inequality based on ethnicity, gender, or family background. We measure perceived meritocracy with two items from the International Social Justice Project (isSkill, isEffort), part of a longer battery of four items they used (page 48, question 2). The two items load on the same factor, although with just two items this is not strong evidence (table 5). The scale: isSkil# "Atrib of succes: Skill, effort (IJSP) Scale = MEAN( ISSKILL ,ISEFFORT ). Table 5. Other perceptions of what it takes to get ahead. Factor loadings, varimax rotation. Australia 1997; n=249. ---------------------------------------------------------------- Factor 1 Factor 2 Factor 3 Factor 4 ---------------------------------------------------------------- (Perception: Meritocracy (skill, effort, intelligence rewarded = isSkil#) ISSKILL -.09 .09 -.09 .83* ISEFFORT -.11 .08 -.11 .81* ISNEED -.03 .30 -.09 .60* ISEQOPP -.27 .34 -.22 .43 (Perception: Poor able to help themselves) CANRAISE -.07 .84* -.06 .13 IMPRVLOT .03 .83* -.07 .16 DECIDHOW -.11 .69* .04 .20 NOTTRY .19 .66* -.20 -.04 NOTRAISE -.16 .57* -.08 .10 (Perception: Crime, dishonesty needed to get ahead. ORGCRMBZ .87* -.02 .19 -.08 CROOKED .77* -.05 .24 -.07 CORRUPT .82* -.05 .30 -.09 MAFDEAL .87* .01 .09 -.09 ORGCRMEN .84* .01 .15 -.09 DISHONST .63* -.19 .48 -.09 CONSCIEN .55* -.14 .46 -.14 TYCONADV .35 -.10 .49 -.19 GAORGCRM .43 -.02 .67* -.17 GACHEAT .36 -.07 .79* -.11 GATAKEAD .12 -.15 .82* -.06 ----------------------------------------------------------------- ###--- Concept: Perception -- Poor are Able to Help Themselves. Whether the poor are poor because of things beyond their own control – for example, discrimination -- or because of things under their own voluntary control is an important issue. It has important implications for free market economics: if poverty is largely voluntary, then changing economic incentives (for example reducing welfare benefits, or reducing tax ‘poverty traps’) can be expected to have an impact. But if, on the other hand, poverty is due to factors beyond the control of the poor themselves, then changing their incentives is irrelevant, save on pure welfare/charity grounds. This is also a politically divisive issue, with the right usually taking one view and the left another. We developed a number of items on this topic for the Developmental Questionnaire. Three form a clear factor (table 3). The scale: Vol#Poor "Perception: poor able to help themselves Scale= MEAN(CANRAISE ,IMPRVLOT ,DECIDHOW ). ###--- Concept: Perception -- Crime, dishonesty needed to get ahead. Particularly in Eastern Europe, crime and dishonest are widely thought important in getting ahead in business. In the Developmental Questionnaire we included a large number of items on this topic, with the results shown in table 3. The first 5 items form a strong scale; but the last 3, although similar conceptually, are empirically distinctive, forming a separate dimension; three other items straddle the two dimensions. We constructed a scale from items in the first, and clearest, dimension: is$Crime "Perception of success: Crime, dishonest needed Scale = MEAN(ORGCRMBZ , CORRUPT ,MAFDEAL ,CROOKED ,ORGCRMEN ). ###----------------------- RESULTS: ANALYSIS (table 6) PERCEIVED EQUALITY OF OPPORTUNITY has a small impact on norms about the legitimate earnings of high status occupations, with those who see equal opportunity willing to pay the elite well (table 6, row 1, column 5). But perceived equality of opportunity has little or no effect on other relevant inequality variables (columns 1 to 4). Opinion on whether or not POVERTY IS VOLUNTARY has little or no effect on any usual inequality variables (table 6, row 2). However the PERCEPTION THAT CRIME IS REQUIRED TO GET AHEAD does matter (table 6, row 3). Those who see criminal links as important: (1) see more class conflict (col 1); (2) see society as more hierarchical, less equalitarian (col 2); (3) would be content with slightly lower incomes themselves (col 3); and are opposed to high pay for the elite (col 5). Table 6. Consequences of perceptions of discrimination. Effects on (1) class conflict; (2) perceived type of society; (3) how much respondent thinks they, themselves, ought to earn;(4) the legitimate pay of low status occupations; and (5) legitimate pay of elite occupations. Standardized partial regression coefficients. Australia, 1997. N=249. -------------------------------------------------------------- CLSCNF# PYRNOW O$YOU O#LOW O#HIGH (1) (2) (3) (4) (5) -------------------------------------------------------------- ISSKIL# -.12 .13 .03 -.01 .13* -- equal opportunity VOL#POOR .10 .02 .00 -.07 -.10 -- poverty is voluntary IS$CRIME .20* -.17* -.13* .02 -.19* -- crime important DPOOR# .22* -.08 .02 .16* .00 -- discrim against poor MALE -.09 -.01 .32* .00 .15* AGE -.11 -.08 .23* .00 .39* EDUC -.05 .11 .19* -.01 .08 OCCSTAT .02 .17* .15* -.04 -.06 LNFAMINC -.02 -.06 .31* .05 .19* EYOU_NOW -.04 .01 .06 .07 .15* -------------------------------------------------------------- * P<.05 ### ----------------------------- RECOMMENDED ITEMS (Moderate priority) ###-----Concept: Perception—country is a meritocracy (ie skill, effort, & intelligence are rewarded) 2 questions. -- RECOMMEND TO INCLUDE IN MODULE, WITH MODERATE PRIORITY We recommend that two items, (ISSKILL =p48q2d) and (ISEFFORT= p48q2b) be included in the questionnaire, with “moderate” priority. The items are from the International Social Justice Project. Mateju reports that his analysis of ISJP data shows they are very useful in understanding differences between nations. While they do not do much in the Australian pretest, they are conceptually interesting and the prospect of cross-national diffeences is attractive. (Exact question wording was given earlier.) The frequencies (table 7): - - - - - - - - - - - - - - - - - - - - - - - - - - - Table 7. Perception of meritocracy. Australia, 1997. - - - - - - - - - - - - - - - - - - - - - - - - - - - ISSKILL p48q2d In Australia: rewarded for intelligence & skill Value Label Value Frequency Percent No!! 0 8 3 No 25 34 14 ?? 50 46 19 Yes 75 151 61 Yes!! 100 7 3 . 3 1 ------- ------- Total 249 100 Mean 62 Std dev 22 - - - - - - - - - - - - - - - - - - - - - - - - - - - ISEFFORT p48q2b In Australia: rewarded for effort Value Label Value Frequency Percent No!! 0 5 2 No 25 43 17 ?? 50 49 20 Yes 75 144 58 Yes!! 100 4 2 . 4 2 ------- ------- Total 249 100 Mean 60 Std dev 22 - - - - - - - - - - - - - - - - - - - - - - - - - - - ###--------------Digression: Alternate Measures of Meritocracy In the 1992 ISSP inequality module, we attempted to measure a similar concept with a series of “attribution” questions: AAMBIT (V7 p48q1d) Getting ahead: having ambition ASMART (V8 p48q1e) Getting ahead: natural ability ASTRIVE (V9 p48q1f) Getting ahead: hard work Surprisingly – and distressingly – answers to these attribution questions are virtually unrelated to the two IJSP questions IsSkill and IsEffort (table 8). Moreover, the two sets of questions have very different correlations with relevant inequality variables: IsSkill and IsEffort have substantial correlations with many other relevant variables, while the old “attribution” questions have no substantial correlations. It is thus clear that the two sets of items give entirely different pictures of the world. They cannot both be right. The internal evidence of high correlations with other relevant variables suggests the two IJSP items are at least measuring something relevant. These worrying results are a strong argument against using the old “attribution” questions (which in any case are scheduled to be dropped from the questionnaire). But they also create some doubts about the IJSP questions. Table 8. Correlations, Australia 1997 (n=249). -------------------------------------------------------------- ISSKILL ISEFFORT ASMART ASTRIVE AAMBIT -------------------------------------------------------------- ISSKILL 1.00 ISEFFORT .57 1.00 ASMART .09 .09 1.00 ASTRIVE .07 .08 .38 1.00 AAMBIT -.10 -.07 .31 .50 1.00 EDUC .08 .11 -.02 -.18 -.06 OCCSTAT .22 .11 .13 -.09 -.13 LNFAMINC .20 .21 -.02 -.03 .02 SUB#NOW .31 .18 -.00 -.09 -.03 EYOU_NOW .25 .15 .08 .06 .06 DV_FREE .21 .16 .04 .06 .13 O#LOW -.04 -.03 -.03 .11 .02 O#HIGH .21 .19 .12 -.06 -.04 ------------------------------------------------------------- ###--- CONCEPT: PERCEPTION -- CRIME, DISHONESTY NEEDED TO GET AHEAD (2 QUESTIONS) -- RECOMMEND TO INCLUDE IN MODULE, WITH MODERATE PRIORITY Varibles (exact wording was given earlier): v327 (ORGCRMBZ p49q5f) Bz success: link to organized crime v328 (CROOKED p49q5d1) Bz success: crooked dealing The frequencies are(table 9) show that few Australians think outright crime is necessary, but many more think that some crooked dealing may be common. - - - - - - - - - - - - - - - - - - - - - - - - - - - Table 9. Perception – crime, dishonesty needed to get ahead. Australia, 1997. - - - - - - - - - - - - - - - - - - - - - - - - - - - ORGCRMBZ p49q5f Bz success: link to organized cri Value Label Value Frequency Percent No!! 0 79 32 No 25 107 43 ?? 50 44 18 Yes 75 10 4 Yes!! 100 4 2 . 5 2 ------- ------- Total 249 100 Mean 25 Std dev 23 - - - - - - - - - - - - - - - - - - - - - - - - - - - CROOKED p49q5d1 Bz success: crooked dealing Value Label Value Frequency Percent No!! 0 27 11 No 25 70 28 ?? 50 68 27 Yes 75 61 25 Yes!! 100 19 8 . 4 2 ------- ------- Total 249 100 Mean 47 Std dev 28 - - - - - - - - - - - - - - - - - - - - - - - - - - - ### ------------------------------ CONCLUSION We recommend that these scales have a moderate priority (not high and now low)for inclusion in the 1999 ISSP Inequality module. This presumes that the 2 family background questions which are discussed elsewhere (see “background-2”) are in the final questionnaire – if not, they would have to be added here to allow analyses of actual discrimination. -------------------------------------------------------------------------------- -------------------------------------------------------------------------------- (this separates two files) -------------------------------------------------------------------------------- -------------------------------------------------------------------------------- To: ISSP From: Inequality Drafting Group Subject: ISSP INEQUALITY MODULE: “Econ-3” Evaluation of economic system. (file: econ3.doc) ------------------------------------------------------------------------ POSSIBLE NEW QUESTIONS FOR THE ISSP INEQUALITY MODULE: EVALUATION OF THE ECONOMIC SYSTEM AND CHANGES OVER TIME IN IT ------------------------------------------------------------------------ Australia (convenor) J. Kelley M.D.R. Evans Czech Republic P. Mateju Norway K. Knudsen K. Skjak Poland B. Cichomski P. Morawski Slovenia B. Malnar ###---------------- INTRODUCTION & BACKGROUND This is an argument for 4 new questions for the inequality module (to add to the ‘core’ questions that have already been circulated). ---The Economic System of the Past. The first question was a straightforward one asking whether the past economic system –from 1945 to 1985 -- was beneficial to “you & your family” (variable v313 below). In these times of historic change to a more market oriented economy in both East and West, we need to get a more precise picture of how people around the world experience processes of social/economic changes and how they react to them. In the May meeting in the Netherlands, we all (Drafting Group and ISSP members generally) liked the proposed question on whether respondent benefited from the old, pre-reform system, and its clone about the country as a whole (v315). ---Self-interest versus Altruism. The reason for also asking about the “nation as a whole” as well as about “you & your family” is that an important theoretical issue in sociology and political science is whether people shape their politics more by their perception of their own self-interest (as economists assume) or, alternatively, shape their politics by their perceptions of WHAT IS GOOD FOR THE NATION AS A WHOLE -- i.e. altruistically. Surprisingly, the weight of evidence is in favor of altruism in the politics of the USA and several other nations. This conclusion is based on analyses that explicitly measure BOTH self-interest and altruism, and then use both to simultaneously predict people’s politics. That requires that both self- interest and altruistic questions be asked, hence two questions (but in very different parts of the questionnaire to avoid bias). ---The Present Economic System. We also planned to add two parallel questions on the CURRENT economic system, but had not then decided on their wording. In the developmental questionnaire, we adopted a straightforward extension of the original ‘past’ questions to cover the current situation. Again, this was for both “you and your family” and separately for the “nation as a whole”. Here are the four questions, which seem to have worked out very well: ###------------ Concept: Impact of economic change on RESPONDENT & THEIR FAMILY. V313: Do you think the kind of economic system [COUNTRY] had between 1945 and 1985 or so -- [with most businesses private but some owned by the government, strong trade unions, and high tariff protection for manufacturing] -- brought you and your family... Only benefits More benefits than harm As much benefit as harm More harm than benefits Only harm V314: And how about Australia’s economic system now -- with almost all businesses owned by private enterprise, less government regulation of business, and less tariff protection for manufacturing – does it bring you and your family... Only benefits More benefits than harm As much benefit as harm More harm than benefits Only harm [NOTE: The descriptive phrases in the two questions -- "[with most businesses private but some owned by the government, strong trade unions, and high tariff protection for manufacturing]" and “[almost all businesses owned by private enterprise, less government regulation of business, and less tariff protection for manufacturing]” – are to be varied according to the conditions of the country; the version shown here would be suitable for many Western European nations.] ###------------ Concept: Impact of economic change on NATION AS A WHOLE (ALTRUISM). V315. Do you think the kind of economic system [COUNTRY] had between 1945 and 1985 or so -- [with most businesses owned by private enterprise, some owned by the government, strong trade unions, and high tariff protection for manufacturing] -- brought to the majority of [COUNTRY] people... Only benefits More benefits than harm As much benefit as harm More harm than benefits Only harm V316. Australia’s economic system now -- [almost all businesses owned by private enterprise, less government regulation of business, and less tariff protection for manufacturing] – does it bring the majority of [COUNTRY, eg Australians]... Only benefits More benefits than harm As much benefit as harm More harm than benefits Only harm [NOTE 1: The descriptive phrases in the two questions -- "[with most businesses private but some owned by the government, strong trade unions, and high tariff protection for manufacturing]" and “[almost all businesses owned by private enterprise, less government regulation of business, and less tariff protection for manufacturing]” – are to be varied according to the conditions of the country and should be the same as the corresponding phrases in V313 and V314. NOTE 2: These questions should be placed in the questionnaire far from V313 and V314.] ###------------------ Results for “YOU & YOUR FAMILY”: Description Here are results from the Developmental pretest in Australia: - - - - - - - - - - - - - - - - - - - - - - - - - - - V313: p55q6a YOU: economic sys 1945-1985: benefit YOU & YOUR FAMILY? Value Label Value Frequency Percent Only harm 0 0 0 More harm 25 14 6 equal 50 99 40 More benefits 75 125 50 Only benefits 100 4 2 . 7 3 ------- ------- Total 249 100 Mean 62 Std dev 16 - - - - - - - - - - - - - - - - - - - - - - - - - - - v314: p55q6b YOU: economic sys NOW: benefit YOU & YOUR FAMILY? Value Label Value Frequency Percent Only harm 0 3 1 More harm 25 81 33 equal 50 110 44 More benefits 75 47 19 Only benefits 100 1 0 . 7 3 ------- ------- Total 249 100 Mean 46 Std dev 19 - - - - - - - - - - - - - - - - - - - - - - - - - - - These results show that Australians had a generally favorable view of their earlier, pre-reform economic system, with a majority feeling they had benefited, 40% thinking “as much benefit as harm”, and only a handful, 6%, thinking they were harmed. The mean was 62 points out of 100. In sharp – and surprising – contrast, Australians have a much less favorable view of the current reformed, more free-market economic system. About 20% think that they have mainly benefited; 44% see as much benefit as harm; and fully 34% see it as mainly harmful to themselves and their family. The mean is only 46 points out of 100. Zagorski and others have found somewhat similar patterns in post-communist Poland, with important changes in the last few years as well. Who Benefits from Economic Reform? It is informative to see which groups Australians think benefit from economic reform (although we do NOT recommend these questions for the module). In the developmental questionnaire (page 4) we asked the current economic system was “good for” various groups. The answers, scored as mean points out of 100 are: Table: Current economic system “Good for...”. Mean points out of 100. (Australia 1997; n=249) --------------------------------------------- Working class 35 Middle class 49 Elite 83 Consumers 49 Business 48 Future generations 39 ---------------------------------------------- Thus Australians think that the current economic system is harmful to the working class; mixed for the middle class; but very good for the elite. They see mixed results – near the neutral point – for consumers and business, but mostly harm for future generations. Given the substantial support that recent economic reforms have had from intellectual circles and the leaders of both political parties, this is a surprising result. Which of these is important in shaping people’s overall evaluation of the economic system? (see table) Table: What shapes people’s feeling that the present economic system benefits them & their family (V315)? Standardized partial regression coefficients for Australia, 1997 (n=249). ------------------------------------------------------ Variable Beta T Sig T ------------------------------------------------------ Economic system good for: Working class .28 3.79 .00 Future generations .28 4.12 .00 Consumers .16 2.40 .02 Business ns -1.11 .27 Middle class ns 1.14 .26 Elite ns -.65 .52 Education .14 2.38 .02 MALE ns 1.46 .15 AGE ns .43 .67 Occupation ns .28 .78 Income ns -1.30 .20 ------------------------------------------------------ ns - not significantly different from zero, p<.05. This regression analysis suggests that the most important factors in people’s mind are: The impact on the working class. Impact on the middle class is probably not relevant and impact on the elite is quite irrelevant. The future matters a lot: those who think the new economic system will make things better for future generations are much more favorable. Effects on consumers matter, but not so much. Well educated respondents are more favorable. But there is little or no difference between rich and poor, male and female, old and young, or those with good jobs and those with bad ones. ###------------------ Results for ALTRUISM - “AUSTRALIA AS A WHOLE”: descriptive The results for benefit to the nation as a whole are, in broad outline, similar. Australians mainly think the previous pre- reform economic system was good for the nation as a whole, but that the new, reformed system is mostly harmful to the nation: - - - - - - - - - - - - - - - - - - - - - - - - - - - v315: p4q8a AUSTRALIA: Econ system 1945-85 benefit NATION AS A WHOLE Value Label Value Frequency Percent Only harm 0 2 1 More harm 25 21 8 equal 50 80 32 More benefits 75 137 55 Only benefits 100 4 2 . 5 2 ------- ------- Total 249 100 Mean 62 Std dev 18 - - - - - - - - - - - - - - - - - - - - - - - - - - - V316: p4q8b AUSTRALIA: Econ system NOW benefit NATION AS A WHOLE Value Label Value Frequency Percent Only harm 0 5 2 More harm 25 103 41 equal 50 81 33 More benefits 75 54 22 Only benefits 100 2 1 . 4 2 ------- ------- Total 249 100 Mean 44 Std dev 21 - - - - - - - - - - - - - - - - - - - - - - - - - - - ###------------------ ANALYSIS People who, they felt, benefited from the pre-reform economic system of the past generally felt they did a little worse from the present, reformed economy (see table, columns 1 & 2). Similarly, those who thought the pre- reform economy was of benefit to Australia as a whole, tend to feel that reform has made Australia worse off (columns 3 & 4). There is a strong link between what one thinks about the effect of the economy on ‘you & your family’ and what one thinks about its effect on Australia as a whole (r=.5 for pre-reform and r=.6 for the present economy): Table: Correlations (Australia 1997; N=249) --------------------------------------------------- Economy benefit y