The Journal of Social Psychology, 1980, 110, 29-37.


University of New South Wales



Previous work with student samples has suggested that authoritarianism may not have the same significance in relation to racism in South Africa as it does elsewhere. It has been proposed instead that both racism and authoritarianism are simply social norms in South Africa. The present study examined these hypotheses on a random sample of 100 residents of Johannesburg. When compared with the results of similar surveys in Australia, the South Africans were found in fact not to be particularly authoritarian or racially prejudiced. Prejudice and F scale score correlated .59, but a scale of authoritarianism in personality inventory format predicted prejudice not at all (r = -.07). It was concluded that the F scale was primarily a measure of social conservatism and that South African institutional racism could best be understood as a response to perceived threat.


The racial problems of Southern Africa have been much in the news in recent years and seem destined to remain so. It seems therefore highly relevant for social psychologists to inquire whether they might not be able to contribute something to an understanding and perhaps to the solution of those problems.

Although acknowledged to be beset with methodological problems, the work of Adorno et al. (1) in The Authoritarian Personality is still influential among social scientists in shaping an understanding of the phenomena of racial attitudes. The basic theories of the work seem to provide a good fit with many familiar and everyday situations.

The situation prevailing in the Republic of South Africa would seem in fact to be one of these familiar situations. We see there a political regime that is at once very authoritarian and very racially discriminatory. It looks like good confirmation of the California theory that authoritarianism gives rise to racism.

So to reason, however, would be to commit the familiar fallacy of inferring causation from correlation. As with all correlations we must allow the possibility of a common third factor influencing our two correlated variables. Such a third factor has in fact commonly been proposed by a variety of writers from Pettigrew (13) through Colman and Lambley (3) to Orpen (11). In broad terms this proposed third factor is, of course, the social situation itself. It has been held that in the South African "sociocultural" climate, both negative racial attitudes and authoritarianism are normative and cannot as such be important as reflectors of personality differences. This does not, of course, give any suggestion as to how such attitudes became normative.

The experimental findings in support of this "third factor" view are extensive. A common feature of them all, however, is that they have been based on sample studies of South African university students (3, 5, 8, 9, 10, 13, 27). Generally, also, they do find at least some correlation between the California F scale and measures of racial attitudes.

The degree to which student attitudes are representative of the society as a whole is unknown. Because there is very little public funding for the support of students at university in South Africa, students there might in fact be an even more atypically elite group than most. What is true of them might have very little to do with what is true in the country as a whole. General population sampling is clearly needed before any questions about the role of personal authoritarianism among South Africans can even begin to be answered.

In the study below, then, an attempt is made to use a sample of the general South African public to examine both the relationship between racial attitudes and authoritarianism and the absolute levels of these two variables. The obvious hypotheses derivable from the existing literature are that even in South Africa, more prejudiced people will be more authoritarian and that both authoritarianism and racial prejudice will show exceptionally high mean scores.


An important innovation in the present study was the use of separate scales to measure authoritarian personality and authoritarian attitudes. The Ray (19) "Directiveness" scale was used for the former and the Ray (15) "Balanced F" (BF) scale was used for the latter. Both were used in 14-item short forms because of the pressures toward brevity normally experienced in doorstep sampling.

The distinction between attitude and personality was one that was deliberately obscured in Adorno et al. (1), but it is nonetheless a traditional and useful one. The reason why it has generally been obscured was the desire by Adorno et al. to use attitudes as a covert measure of personality. Normally, personality would have been measured by some form of behavior inventory. Although the distinction between attitude scales and behavior inventories is not necessarily in every case a clear one, there is a very great difference between asking a person "Do you think hard work is a good thing?" (attitude) and asking him, "Do. you work hard" (personality). Questions of the latter type are generally used in personality inventories precisely because they are better predictors of actual behavior than questions of the former type. They involve fewer intervening assumptions and it is precisely the intervening assumptions that make the relationship between attitudes and actions problematical (2). Since the F scale has generally failed as a predictor of actual authoritarian behavior [see the evidence reviewed in Ray (19)], the reversion to scales in behavior inventory format is an obvious step. The personality scale in the present research has shown good predictions of "the desire or tendency to impose one's own will on others" (as assessed by peer ratings). See Ray (19).

The use of a balanced form of the F scale to measure authoritarian attitudes may seem unnecessary in view of the observations by Rorer (24), but such a conclusion would ignore the work of Peabody (12). Peabody made a strong case that the F scale is particularly ambiguously worded and although Rorer (24) showed that there was no such thing as a general trait of acquiescence or yeasaying; even he did acknowledge that acquiescence could be a problem with ambiguously worded scales. Since one can probably never know in advance how ambiguous one's scale will appear to each given group of respondents, one might indeed conclude that balanced scales should still continue to be used on all occasions.

Unlike previous balanced F scales in which the negative items were selected by a priori methods, the negative items for the present scale were selected by empirical means and the average correlation between the two halves of the scale in a variety of Australian studies has been around -.6 (before reversals).

The scale of attitude to blacks was a South African adaptation by Heaven and Moerdyk (6) of the Ray (18) "Attitude to Aborigines" scale. They apparently found that the things said of Australian Aborigines were similar to those said of the South African Bantu and simply replaced the word "Aborigine" wherever it appeared by the word "Black." They also gathered evidence which they interpreted as validating the so-modified scale. For the purposes of the present study, the scale was augmented by four items having special reference to South Africa. Melamed (7) found a .48 correlation between antiblack attitudes and antiblack behavior among a group of South African students; hence living in a racist culture does not necessarily wipe out individual difference in this respect.

A questionnaire incorporating the three scales mentioned above was administered to a random cluster sample of 100 people living in the Johannesburg Greater Metropolitan area. Cluster sampling is the method almost universally used by public opinion polls where it generally gives very accurate results. One problem with commercial surveys is the problem of responses faked by the interviewer to save him/her effort. This is normally overcome by running a percentage follow-up of all interviews by a second person. To do this, however, names or phone numbers have to be taken and this breaches any guarantee of anonymity. As race can be expected to be a sensitive matter with South Africans, a guarantee of anonymity was thought to be especially necessary on the present occasion. The solution to the problem of obtaining genuine interviews was therefore for the author to go to South Africa and carry out all interviews personally. This took place in December, 1978.

As is usual in cluster sampling, geographical starting points rather than people were chosen randomly. The points were chosen off the master map of the local street atlas by a "pin and blindfold" method. Each point so chosen was then visited and the 10 people living nearest to it were interviewed.

The N of 100 was chosen because correlations explaining as little as 4% of the variance will be shown as significant with this sample size. Smaller effects were not thought important.

Neither Johannesburg or any other South African city is, of course, "typical" of South Africa. It is, however, South Africa's largest city and does as such itself account for a significant fraction of South Africa's white population (one out of four million whites live in the Witwatersrand area). As other studies with the present scale of authoritarian personality have recently been carried out in the largest cities of England, Scotland, and Australia (21), the Johannesburg sample also enabled potentially valuable cross-cultural comparisons to be made.

The mean age of the sample was 38.68 years; the mean education was between four and six years of high school; the male/female ratio was 52/48 and the ratio of manual to nonmanual occupations was 26/74. As Afrikaners are normally bilingual, all testing was done in English. However, there was one Afrikaner who was not sufficiently fluent for interviewing plus approximately 12 people (recent immigrants) who spoke Southern European languages only. None of these could, therefore, be included in the sample.


The reliabilities ("alpha") of the three scales were as follows: BF (authoritarian attitudes) .65, Authoritarian personality .75, Attitude to Blacks .79. Without the four items added for the present study only, the Attitude to Blacks scale showed a reliability of .73. The correlation between the two halves of the BF scale was -.21.

The BF scale correlated .59 with attitude to blacks and -.20 with authoritarian personality. Authoritarian personality correlated -.07 with attitude to blacks. The first two coefficients only were significant at the .05 level. The meaning of the correlations is that authoritarian attitudes and antiblack attitudes were strongly associated, but authoritarian personality and antiblack attitudes were not associated at all. People with authoritarian behavior tendencies also showed a slight tendency to reject authoritarian attitudes.

The mean scores (and SDs) observed were as follows: BF 42.14 (5.50), Authoritarian personality 28.08 (5.79), and Attitude to Blacks (10 original items only) 29.75 (4.88). All three means were almost exactly on the theoretical midpoint of the respective scales (number of items multiplied by the midpoint for each item).

For comparison the scores obtained by identical sampling methods (N = 95) in the Sydney metropolitan area, Australia, may be used (21). The mean score on the same short scale of authoritarian personality was 29.69 (SD 6.16). The short BF mean was 37.42 (8.20). The reliabilities of the scales in the Australian study were as follows: BF .81, Authoritarian personality .78. Note that Ray (21) also shows mean scores on authoritarian personality that do not differ as between Australia, England, and Scotland. On the attitude scale only the differences in authoritarianism between Australia and South Africa are significant (t = 3.2).

The evaluation of the South African mean score on Attitude to Blacks is not quite as easy. In three separate doorstep applications of the same 10-item scale in Australia, the means were as follows: 28.55 (SD = 7.05, N = 88); 28.63 (SD = 7.03, N = 90); and 30.58 (SD = 5.59, N = 68). See Ray (18, 19). Although all three means were obtained again in the Sydney metropolitan area, they were samples of particular suburbs rather than samples of Sydney as a whole. In addition they tended to be working class suburbs. For this reason, then, preferable norms are those obtained from a previously unpublished postal survey of the Australian state of New South Wales. In this study, 500 questionnaires containing the attitude to Aborigines scale were mailed out to persons selected at random from the electoral rolls. The 140 returns formed a sample which was indistinguishable in demographic structure (age, sex, occupation, and education) from contemporaneous samples of the Sydney metropolitan area obtained by random doorstep interviews. Thus, while there was in the sample the inevitable bias towards more co-operative people, the sample was in other ways highly representative. The mean and SD observed on this sample was, then, 29.14 (6.26). Against all criteria, then, white South Africans showed a degree of prejudice that is identical with that shown by Australians. It is clear that, at least with these samples, the view of South Africans as extremely prejudiced people is not supported.

Even the South African mean on the BF scale is in fact not especially high by Australian standards. The BF mean for the Australian population as a whole in the study described in Ray (22) was 41.57 (SD = 7.97, N = 200). This was again a postal survey but was carried out by a commercial polling organization. Again the demographic structure of the sample closely matched that of the overall Australian population. As a result of careful preliminary approaches to the respondents, the response rate in this survey was over 70%. Again, then, we see that the Australian mean was not significantly different from the Johannesburg mean.

These South Africans, then, were no more authoritarian in personality than Australians, Scots, or Englishmen, nor were they more racially prejudiced.


The picture of South Africans that emerges from these results is very much one of average people with average responses. While this should not on one level be a particularly surprising finding, it does, of course, run contrary to much popular opinion. The view of South Africans as not being particularly racially prejudiced may prove particularly difficult to assimilate.

A tempting response might, of course, be to see the results as more revealing about Australia than they are about South Africa. Since Australians have been shown to be no different from the English or the Scots on variables such as authoritarianism of personality (21) and achievement motivation (23), however, this view is not easily supported. Australians do have a record of various racist practices but certainly no more so than do the British or the Americans. Because Australians and South Africans have similar backgrounds in very many ways, in fact, the comparison between the two forms a fairly well controlled "natural" experiment.

If South African racist practices cannot be explained by greater racism in attitudes, how, then, can they be explained? A very obvious explanation flows from such work as that by Sales (25, 26), who shows that authoritarianism in both behavior and attitudes is increased by perceived threat. South Africans are hard on blacks not because they despise them but because they fear them. There are, no doubt, very good grounds for such fear. The accuracy of this explanation is borne out by the now well-documented finding that South Africans suffer from especially high levels of chronic anxiety [see Van Der Spuy and Shamley (28)].

Given that the levels of attitudinal authoritarianism and racial prejudice are not as overweeningly high in South Africa as was once thought (e.g., 3), it must come as no surprise that the relationship between the two is also of normal magnitude. Because the authoritarian personality scale, on the other hand, gives no such correlation, we are forced to seek explanations for the correlations somewhat different from those used in the past. Note that Australian and British surveys using the scale of authoritarian personality have also shown it to be unrelated to racial attitudes (19, 20).

If then BF scores and their correlates are not to be explained as the outcome of authoritarian personality, what do they reflect? Ray (17) has set out extensive historical and statistical evidence leading to the view that what the F scale measures is little more than social conservatism, perhaps at most a particularly tough-minded brand of social conservatism. In view of the repeated failure of the F scale to predict basic authoritarian behavior [see the review in Ray (19)], this explanation can be expected to gain steadily in plausibility. Racism, then, must be seen as another aspect of social conservatism; it is an aspect of what the F scale measures rather than being caused by what the F scale measures (16). That conservatives are more likely to acknowledge racist attitudes, however, must not mislead us into assuming that they will be more likely to behave in racist ways (14).

The low correlation observed between the two halves of the BF scale, however, may suggest that the F scale is not valid in South Africa. If two sets of items are, on both theoretical and empirical grounds, conceived as being highly opposite in meaning and yet do not in fact correlate highly negatively, we must surely conclude that the scale concerned is lacking even in construct validity, let alone predictive validity. The items very obviously do not mean the same to South Africans as they do to Australians. Of all the findings from the present study, then, the results from the BF scale would seem to be most in need of cautious interpretation.

It may be of interest that, with the exception of two items advocating harsh treatment for sexual deviants, all positive BF items received an overwhelming number of "agree" responses. The reversed items, by contrast, appeared to be responded to much more thoughtfully and distributed responses much more evenly between "agree" and "disagree." Both halves, however, correlated in the predicted way with attitude to blacks (.52 and -.37). The high level of agreement with original F items fits in well with Hartmann's (4) view of the F scale as measuring "Victorian" values which "hold considerable sway in our society." As he says, "Few F scale items can be regarded as foreign to the dominant value systems of modern Western societies" (4, p. 91).

Overall, then, the present work has counterindicated the view that the high levels of institutional racism in South Africa could be accounted for by maladaptive personality patterns among South Africans. Their institutional racism is then probably best seen as a fear response in the face of a far from imaginary threat. That sanctions and other forms of international pressure will decrease their perception of threat seems unlikely.


1. ADORNO, T. W., FRENKEL-BRUNSWIK, E., LEVINSON, D. J., & SANFORD, R. N. The Authoritarian Personality New York: Harper, 1950.

2. AJZEN, L, & FISHBEIN, M. Attitude-behavior relations: A theoretical analysis and review of empirical research. Psychol. Bull., 1977, 84, 888-918.

3. COLMAN, A. M., & LAMBLEY, P. Authoritarianism and race attitudes in South Africa. J. Soc. Psychol., 1970, 82, 161-164.

4. HARTMANN, P. A perspective on the study of social attitudes. European J. Soc. Psychol., 1977, 7, 85-96.

5. HEAVEN, P. C. L. Do authoritarians hold authoritarian attitudes? The case of South Africa. J. of Psychol., 1977, 95, 169-171.

6. HEAVEN, P. C. L., & MOERDYK, A. Prejudice revisited: A pilot study using Ray's scale. J. Behav. Sci., 1977, 2, 217-220.

7. MELAMED, L. The relationship between actions and attitudes in a South African setting South African J. Psychol., 1970, I, 19-24.

8. NIEUWOUDT, J. M., & NEL, E. M. The relationship between ethnic prejudice, authoritarianism and conformity among South African students. In S. J. Morse & C. Orpen, Contemporary South Africa. Cape Town. South Africa: Juta, 1975.

9. ORPEN, C. Authoritarianism and racial attitudes among English-speaking South Africans. J. Soc. Psychol., 1971, 84, 301-302.

10. ORPEN, C. The reference group basis of racial attitudes: An empirical study. South African J. Sociol., 1973, 6, 67-72.

11. ORPEN, C. Authoritarianism revisited: A critical examination of "expressive" theories of prejudice. In S. J. Morse & C. Orpen, Contemporary South Africa. Cape Town, South Africa: Juta, 1975.

12. PEABODY, D. Authoritarianism scales and response bias. Psychol. Bull., 1966, 65, 11-23.

13. PETTIGREW, T. F. Personality and sociocultural factors in intergroup attitudes: A cross-national comparison. J. Conflict Resolut., 1958, 2, 29-42.

14. RAY, J.J. (1971) Ethnocentrism: Attitudes and behaviour. Australian Quarterly, 43, 89-97.

15. RAY, J.J. (1972) A new balanced F scale -- And its relation to social class. Australian Psychologist 7, 155-166.

16. RAY, J.J. (1972) Militarism, authoritarianism, neuroticism and anti-social behavior. Journal of Conflict Resolution 16, 319-340.

7. RAY, J.J. (1973) Conservatism, authoritarianism and related variables: A review and an empirical study. Ch. 2 in: G.D. Wilson (Ed.) The psychology of conservatism London: Academic Press.

18. RAY, J.J. (1974) Are racists ethnocentric? Ch. 46 in Ray, J.J. Conservatism as heresy Sydney: A.N.Z. Book Co.

19. RAY, J.J. (1976) Do authoritarians hold authoritarian attitudes? Human Relations, 29, 307-325.

20. Ray, J.J. (1978) Determinants of racial attitudes. Patterns of Prejudice 12(5), 27-32.

21. RAY, J.J. (1979) Authoritarianism in Australia, England and Scotland. Journal of Social Psychology 108, 271-272.

22. RAY, J.J. (1979) A short balanced F scale. Journal of Social Psychology, 109, 309-310.

23. RAY, J.J. (1979) A quick measure of achievement motivation -- validated in Australia and reliable in Britain and South Africa. Australian Psychologist 14, 337-344.

24. RORER, L. G. The great response-style myth. Psychol. Bull., 1965, 63, 129-156.

25. SALES, S. M. Authoritarianism. Psychol. Today, 1972, 6, 94 ff.

26. SALES, S. M. Threat as a factor in authoritarianism. J. Personal. ~'r Soc. Psychol., 1973, 28, 44-57.

27. VAN DEN BERGHE, P. Race attitudes in Durban, South Africa. J. Soc. Psychol., 1962, 57, 55-72.

28. VAN DER SPUY, H. I. J., & SHAMLEY, D. A. F. The Psychology of Apartheid. Washington, D.C.: University Press of America, 1978.

School of Sociology, The University of New South Wales P. O. Box 1, Kensington, New South Wales, Australia 2033

Go to Index page for this site

Go to John Ray's "Tongue Tied" blog
Go to John Ray's "Dissecting Leftism" blog
Go to John Ray's "Australian Politics" blog
Go to John Ray's "Gun Watch" blog
Go to John Ray's "Education Watch" blog
Go to John Ray's "Socialized Medicine" blog
Go to John Ray's "Political Correctness Watch" blog
Go to John Ray's "Greenie Watch" blog
Go to John Ray's "Leftists as Elitists" blog (Not now regularly updated)
Go to John Ray's "Marx & Engels in their own words" blog (Not now regularly updated)
Go to John Ray's "A scripture blog" (Not now regularly updated)
Go to John Ray's recipe blog (Not now regularly updated)

Go to John Ray's Main academic menu
Go to Menu of recent writings
Go to John Ray's basic home page
Go to John Ray's pictorial Home Page