Human Relations, Volume 29, Number 4, 1976, pp. 307-325


J. J. Ray*

University of New South Wales

It is pointed out that there is little evidence that authoritarian attitudes and authoritarian behavior are associated. An attempt is made to construct a new scale which will predict authoritarian behavior. Against a peer-rating validity criterion, the new scale (in behavior inventory format) correlated .54, compared to a correlation with the same criterion by an attitude scale of .19.

There is no doubt that the prototype of the authoritarian whom Adorno, Frenkel-Brunswik, Levinson & Sanford (1950) had in mind was the German Nazi. Although carried out mostly in California, the work of these Jewish authors was directed explicitly toward finding an explanation for the rise of German Nazism. I will for my purposes, therefore, take it as given that by an authoritarian person we mean someone prone to behave as the Nazis did -- in an aggressive, domineering, and destructive way toward other people.

The question here is whether the group of people so defined is coterminous with the group of people who hold attitudes identified by Adorno, et al., (1950) in their famous 'F scale' as "Pre-Fascist." Do the attitudes and the behavior go together? It will be submitted that they do not.

Initially, any challenge to the association between attitudes and behavior in this field seems gratuitous. There is no mistaking what the Nazis did and there is equally no mistaking that the ideology they held was such as Adorno, et al. (1950) called "authoritarian" (See Eysenck, 1954). As large a swallow as Nazi Germany might have been, however, one swallow does not make a summer, and people such as Shils (1954) have pointed out that authoritarian political regimes may often have egalitarian and humane official ideologies (e.g., Soviet Russia and Maoist China). Official ideology is a poor guide to action.

The real issue, therefore, is not on the national but on the individual plane. The work by Adorno, et al. and their many successors was, after all, done in what we might loosely term Western democracies. Do people living in such societies tend to behave in an authoritarian way if they hold authoritarian attitudes? There is already some evidence that they do not.

For a start, in the supposedly related field of racial prejudice, it has been shown by several authors from LaPiere (1934) to Ray (1971a) that attitudes and behavior often do not go together. People who acknowledge prejudiced attitudes may or may not behave in a discriminatory way toward members of other ethnic groups. The same is true of people who deny prejudiced attitudes. Overall, there is no relationship. In fact the curious phenomenon that some people who do discriminate against Jews will say, "Some of my best friends are Jews," has given rise to the equally curious myth that this is an anti-Semitic cliche. What genuinely pro-Semitic people (and there are such) are supposed to say we are never really told.

When we turn to work done using attitude scales, we find that most studies assume the basic validity of the scales concerned rather than set out to test it. In spite of the trenchant methodological criticisms leveled at the work of Adorno, et al. (1950) by Christie & Jahoda (1954) and others, the F scale that stemmed from that study is still widely and uncritically used in the belief that it does measure what it purports to measure. It is assumed to tap propensity for authoritarian behavior.

One study which did directly test this assumed validity was that by Titus (1968). He found that high scorers on this scale were not characterized by authoritarian behavior. Regardless of how valid the scale may be as a measure of attitudes, it is not valid at all as a predictor of behavior. What Titus did find, however, was a slight relationship between authoritarian attitudes (F scale score) and submissive behavior.

Similar findings of "incongruity" between authoritarian attitudes and authoritarian behavior were reported by Hollander (1954) and in the extensive review of the literature on the F scale published jointly by Titus & Hollander (1957) they repeated "that there may be real differences between acceptance of the authoritarian ideology and authoritarian behavior as it is traditionally conceived" (p. 56).

Studies with other measures of authoritarian attitudes have tended also to give such results. Using a new "attitude to authority" (AA) scale (see Appendix III), Ray (1971b) found that school students with pro authority attitudes were rated by their teachers to be "submissive" (r = .36) but not "aggressive" (r = .05); nor were they seen as students who "liked to push others around" (r = .07).

In a study with university students, Ray (1972a) used yet another index of authoritarian attitudes (the `A' scale) in finding that high scorers were not seen by their peers as "liking to push others around" or as being "characterized by fixed opinions." To the contrary, they were seen as flexible ("always modifies his behavior to suit the circumstances of the situation") and submissive ("tends to follow instructions without critical thought"). Yet the scale concerned contained items glorifying Mussolini, war, punctuality, discipline, and the Army: Unlike the F scale, both the AA and A scales were completely counterbalanced against acquiescent set. Moving outside peer ratings to an experimental criterion, high scorers on the A scale were found in the same study to be characterized by conceptual diversity rather than black-and-white thinking (r = .40). The criterion was a modified form of Kelly's repertory grid technique.

In an attempt to salvage from these results something of the conventional beliefs about authoritarianism, one might be tempted to seize on the correlations with submissiveness and identify submissiveness alone as sufficient to explain the phenomena of German Nazism. This would be a substantial watering down of the original concept, as Adorno, et al. (1950) believed authoritarians to be rigid and aggressive as well, but certainly submissiveness can explain a lot of things if the leader is demonic (or is believed with hindsight to be demonic). At the Nuremberg war trials most German officers gave as excuse for their behavior the explanation that they were "just following orders" ("Befehl ist Befehl"). Submission in the destruction of others could then be a most fundamental form of authoritarianism, and evidence that the F and other scales are sensitive to submissive behavior could be important predictive validation for those scales.

Unfortunately, it is doubtful if even this much can be salvaged from the conventional conception of authoritarianism. The correlation between authoritarian attitudes and behavioral submissiveness is adventitious rather than strong, and Ray (1972b) reports yet another study with students wherein no authoritarian behavioral trait -- not even submissiveness -- was found to go with high AA scale scores.

Even more to the point, examinations have been made of the relation between performance in the Milgram experiment and authoritarian attitudes:

The tradition of research in the social psychological literature that has most directly to do with fascist or Nazi-type behavior is, of course, the Milgram experiment (Milgram, 1963, 1965). The experiment consists of some authoritative figure (usually a member of the Psychology Department staff) instructing the student to press a button which causes a subject (really a stooge) to be given increasing levels of electric shock -- to the point where the subject cries out, screams, and groans with (apparent) pain. Many students continue unsuspiciously to obey instructions even to the point where the imminent death of the subject appears likely. Obviously, this experiment is as near as we can reasonably expect to get in the laboratory to what took place in the concentration camps of Dachau and Belsen. There are some results to show (Kilham, 1971) that dogmatism and politico-economic liberalism do not predict whether or not a subject will acquiesce in or rebel against this situation. Do any of our authoritarianism scales provide a better prediction of how much people will submit to instructions of this sort?

The results to be summarized below in answer to this are derived from a Milgram-type study carried out for another purpose (Sherell de Florance, 1972). There were four experimental treatments given to a total of 47 undergraduate University of Sydney psychology students who took part in the experiment as a course requirement. All subjects received and answered the AA (Appendix III), BF (Appendix II), and Acceptance of Aggression (Shostrom, 1964) scales. The latter scale is designed to measure the handling of aggression as an aspect of good adjustment. Its inclusion was designed to provide some test of whether the authoritarian was in fact a person unable to handle aggression adaptively. The BF scale mentioned was a balanced version of the original F scale (see Ray, 1972c).

The experimental treatments are of interest in another context and need not be elaborated on here. Suffice it to say that one of the treatments was the classical Milgram paradigm and only one of the remaining treatments showed any deviation from the mean score obtained with this paradigm. There are two things, therefore, that can affect a person's score (the score being how far he will go in shocking the subject): 1) The experimental treatment; 2) Individual differences between subjects. We are here interested in the latter. For our purposes the effect of the experimental treatment was controlled statistically so that the only variations left were due to differences between the individuals themselves. This was done simply by dividing each person's score by the mean score of his treatment group.

The results were as follows: The correlation between experimental score and the two authoritarianism scales was BF .04 and AA .09. The correlation between scores on the Shostrom scale and the same two scales was BF .15 and AA .28. Only the last of these four correlations is significant at the < .05 level.

"Debriefing" revealed that all subjects were "taken in" by the experimental manipulation and did not suspect the real purpose of the study. See also Milgram (1963, 1965).

Sherell de Florance (1972) has then shown once again that attitudes conventionally called authoritarian do not predict behavior conventionally called authoritarian. Authoritarians did not obey more in the Milgram situation. AA and BF scores may sometimes predict submissiveness in everyday life but they do not predict the utilization of an authoritative position in a domineering, aggressive, or destructive way. Students of conventionally liberal or radical views are as prone to engage in quasi-concentration-camp behavior ("S.S." behavior) as are students acceptant of authority and authoritarian institutions. (See also Kilham, 1971). It was additionally shown that the particular type of psychological problem imputed to the authoritarian by Adorno, et al. (1950) -- adaptive handling of aggression -- is, if anything, the reverse of the truth. High scorers on the AA scale were significantly better at handling aggression, though there was no relation with BF scale scores.

It may be noted that the findings in this study seem to run counter to the findings by Elms and Milgram (1966) -- who found a relation significant at the < .003 level between obedience in the Milgram situation and scores on the F scale. The sample taken by these authors was, however, rather heterogeneous as to education and when the influence of education was removed statistically the relationship dropped to insignificance. Of potentially even greater importance, however, was the fact that these authors, like so many others, used an unbalanced form of the F scale and the results are hence equally well interpretable as the effects of acquiescence (a variable of obviously great relevance in the Milgram situation) rather than as the effects of authoritarianism. Elms and Milgram did attempt to cope with this possibility by obtaining an independent measure of acquiescence from the MMPI and partialling out this effect. This procedure, however, is vitiated by Martin's (1964) finding that acquiescence in personality scales is independent of acquiescence in attitude scales. Hence the effect of attitude-scale acquiescence was not removed by the authors' procedure, and their work provides no grounds for questioning the Sherell de Florance study. It is perhaps most instructive of all that Elms & Milgram (1966) found a pattern of personality in their "obeyers" (as distinct from their "defiants") which is almost the exact opposite to what Adorno, et al. (1950) postulated for the authoritarian. The following is a quote (pp. 284 and 285) from the "Results" section of the Elms & Milgram paper:

"How close were you to your father when you were a child?" Obedient Ss reported less close than defiants, on a five-point scale from "extremely close" to "extremely distant": (defiant Chi-squared = 1.95; obedient Chi-squared = 2.76; p < .05).

"How were you usually punished?" Several obedient SS reported extremely mild or no punishment at all, and the bulk of the others reported the standard spanking. Defiant Ss more frequently reported physical or emotional deprivation, with several reporting intense physical punishment..

Clearly, this report also undermines the neo-Freudian California account of authoritarianism. It is the rebel, not the acceptor of authority, who has had the harsh upbringing and who is venerative of his father.

In interpreting this repeated failure of attitudes and behavior to correlate, we must initially acknowledge that there is, no doubt, some pressure for the person to justify the way that he himself behaves. Countering this, however, there are also pressures to disavow such behavior (See Kelman, 1974). In the case of the authoritarian, this pressure could well be the extreme aversion that the exercise by others of authority on him evokes. A dominant person is unusually resentful when he has to take orders. This more than cancels out (in deciding what he would advocate for the society in general) the satisfaction that he himself gets from dominating others.

After this point, then, it must seem that henceforth any unqualified use of the term "authoritarian" will be ambiguous. Do we mean a person who is authoritarian in attitudes or a person who is authoritarian in behavior? For Adorno, et al. this ambiguity did not exist. They assumed the two to be identical without proof. The nearest thing they had to such proof was that Hitler's Germany embodied both authoritarian ideology and authoritarian behavior. As mentioned above, however, with Stalinist Russia we had the phenomenon of egalitarian ideology combined with authoritarian behavior. Had their consideration of historical data been more careful, Adorno et al. would have been forced to conclude what has been here concluded from psychological data -- that authoritarian behavior sometimes has associated with it authoritarian attitudes and sometimes egalitarian attitudes. There is no overall association.

[t might seem at this point that the question posed in the title to this paper has been answered: "No, authoritarians do not have authoritarian attitudes." The practical problem that this raises for us, however, is that we are left totally without a predictor of behavioral authoritarianism. We appear to have been taken right back to first base. We are no further advanced than when the California authors first began their studies. There is also the problem that there are at least two ways in which this predictive failure can be interpreted: we could see it as simply a methodological failure, or as substantive information about the true state of affairs. The first interpretation leaves the California hypotheses quite unimpeached. The second interpretation seriously undermines and confutes those same hypotheses.

If we see the results obtained in the literature so far as simply inconclusive, we might do so on the grounds that, after all, a failure of attitudes and behavior to correlate is no new thing. Attitude scales in many fields have been shown to lack predictive validity (But see Kelman, 1974). The attitude scales are just not good enough and the California hypotheses have so far simply not been adequately tested. This line of argument is, of course, a potentially hazardous one in that, carried to an extreme, it would never allow any hypothesis to be rejected. The above summary does include results from four separate authoritarianism scales. When do we stop suspecting that our instruments might be at fault and accept the results obtained with them at face value?

Nonetheless there does seem something wrong with there being absolutely no relationship between self-report and actual behavior, and doubts about the validity of one's instruments will never be really laid at rest until a demonstrably valid instrument is actually produced. It is believed therefore that there are grounds for at least one more attempt to devise an authoritarianism scale that does predict behavior. Obviously, however, any such attempt will have to employ radically different techniques from the attitude-scale approach that has been found above to be so uniformly unsuccessful.

A very simple alternative to the attitude-scale approach that does not appear so far to have been tried in this field is the personality scale, or behavior inventory approach. This is perhaps particularly surprising when we consider the widespread success such scales have had in the clinical field (e.g., Eysenck & Eysenck, 1969). The difference between the two sorts of scale, then, is that, whereas the attitude scale asks the subject about something outside himself (the army, Mussolini etc.), the behavior inventory asks the subject directly about his own behavior. The difference between the two is perhaps best seen in the response format typically attached to the items: where the attitude-scale responses are from "strongly agree" to "strongly disagree," the personality scale is generally responded to simply "yes" or "no." One can readily see why a person who agrees that "hard work is a good thing" (attitude item) might still say "no" when asked "do you work hard?" (personality scale item). It is the behavior inventory item, then, that we would on quite a priori grounds expect to be the best predictor of actual behavior. It is simply a matter of fewer intervening assumptions being required.

The big drawback with the personality-scale approach, of course, is its overtness. Adorno, et al. (1950) originally favored the attitude-scale approach precisely because it was more covert. Events, however, would seem to have shown that they were too covert by far -- so covert, in fact, as to lose all predictive validity. There may then be simply no choice but to use the more overt approach. One certainly appears to have nothing to lose.

For the purposes of constructing such a scale, it is, of course, necessary to start out with at least the preliminary conceptual exercise of formulating a definition of the construct to be scaled. The new scale should also, if possible, have a distinctive name.

In formulating a definition of the construct to be scaled, there is the considerable difficulty that the word "authoritarian" is used to cover a multitude of sins. Masling (1954) found that it was simply the social scientist's all-purpose pejorative adjective -- more a swear-word than a description. There is also the problem that the word's popularizers (Adorno, et al.) might be conceived as having something very like proprietary rights to it -- it means the constellation of traits that they described and nothing else. The word was, however, in existence before Adorno, et al. came along and in popular usage it seems to be most commonly elicited as a description for the behavior of someone who seems to like bossing other people around. Whether such people are also characterized by rigidity, intolerance of ambiguity, conservatism, and all the rest is really an empirical hypothesis rather than part of the definition. Certainly Shils (1954) has made a strong case for not restricting its application to describing people on the ideological right.

Bearing in mind the description of an authoritarian person suggested in the first paragraph of this paper as "someone who behaves in an aggressive, domineering, and destructive way toward other people," we might then formulate the irreducible minimum meaning of authoritarianism as being "the desire or tendency to impose one's own will on others." This will be the formal definition of the construct to be measured by our new scale. In accordance with this definition, the title of the new scale will be the "directiveness" scale. This title makes it very clear that the view of Adorno, et al. that an authoritarian person is one who is at once submissive (to superiors) and dominant (to inferiors) is regarded as a matter for proof rather than as something that may be assumed. While such ambivalent behavior may be seen as typical of a military situation, it does not follow that there is a personality type which matches this situation in some sense. It could well be that there are only two types of people -- dominants and submissives. The dominants seek officer or NCO positions, while the submissives do not. Dominants may submit on the occasions where they have to, not because they like it, but rather because it is a necessary evil that must be tolerated if they are to get their chance of doing some dominating.

The hypothesis that people who like bossing others around also like being dominated by their own superiors is an interesting one, but it is one in need of proof rather than something that may be embodied in a definition. Indeed it is a hypothesis which the results of Ray (1971b) clearly contraindicate.


The new scale (see Appendix I) was constructed in two stages. In the first, 14 new items were written by the author to center around the formal definition given above. These were embodied in a larger questionnaire and administered to a first-year Sociology class (N = 117) at the University of New South Wales. The balanced 14-item scale showed a reliability ("alpha") of .61 and all items but one showed significant correlations with the scale total -- even after correction for overlap. Its correlation with an "attitude to Apartheid" scale (reliability .73) also included in the battery (see Appendix V) was a nonsignificant .050. This suggests that authoritarians in the present sense are not racists.

The new scale was shown to be unrelated to a whole range of phenomena that Adorno, et at. (1950) would have led us to believe are authoritarianism-related: It was not related to history of participation in antiwar or anti-Apartheid demonstrations, to membership in conservative political clubs, or to national voting preference. It was not related to religious belief, religious background, moralism, or attitude to morality: (Details of the moralism and attitude to morality scales are given in chapter 53 of Ray (1974)). Most encouragingly, it was not related to social desirability response set as measured by a short form (reliability .69) of the Marlow-Crowne Social Desirability scale.


As it stood so far, the scale represented a rather negative achievement. It showed that self-reported authoritarian people were not what Adorno, et al. had said they were. What was now needed was some positive validation. The two sorts of validation planned for were of the criterion groups and concurrent types. Both types were, however, somewhat modified to fit the needs of the particular project. As the first stage of this, the theory and findings up to date were presented to a second-year Sociology class. At the conclusion of this, the formal definition of authoritarianism given earlier was presented and explained. The students as a class (N = 250) were then invited to suggest items that might reflect the construct embodied in the definition. A total of 16 new items was collected in this way. Half were devised as positive and half as negative. The upshot was that there now appeared to be two directiveness scales of different origins -- a 14-item scale devised by the author and a 16-item scale devised by students. Their correlation would obviously give some suggestion whether both scales had succeeded in embodying items to express the same construct:

The two scales were included in a questionnaire which also contained the same social desirability and attitude to authority (AA) scales (see Appendix III) that were mentioned previously. Each student in the class was then asked to take two questionnaires each and administer them to the two people he knew whose behavior was to his mind most extreme in low authoritarianism and high authoritarianism. It was thought this way to form two criterion groups against which to validate the new scale: If the high-authoritarian group got higher scores on the new scale than did the other group, this would represent a conventional indication of validity for the scale. A slight problem that arose in this connection, however, was that any one student might not know two people who were in absolute terms at the extremes of authoritarianism. One student's "high" authoritarian might be in absolute terms only medium in authoritarianism -- even though he was higher than the same student's "low" authoritarian. To overcome this as far as possible, a refinement of the usual criterion groups procedure was introduced: instead of just assigning a person into one of the two categories ("high" and "low"), the students rated each subject in terms of five categories of authoritarianism and in terms of three categories of submissiveness. Instead of examining mean score differences, the validation procedure then becomes one of correlating the ratings with the scale scores. The questionnaire was administered anonymously, and each student was provided with a university envelope in which to seal the questionnaire as soon as it was filled out by the subject. This created some difficulty in matching up the rating of the person with that person's questionnaire, but this was overcame either by preinserting the completed rating sheet in the envelope or by pinning the completed rating sheet to the outside of the envelope.

Of the questionnaires received in this way, a total of 282 were available for analysis. Checks were made to exclude incomplete questionnaires and other questionnaires not filled out in accordance with the administration instructions.

The two directiveness scales were found to correlate .52. This was, then, persuasive concurrent validation for the two scales. Their reliabitities were .64 for the scale described in Study I and .48 for the student-written scale. The correlations of the two scales with the predictive validity criterion (authoritarianism ratings) were .53 and .41 respectively. We may say, then, that two scales written around the same definition have separately been shown to have strong predictive validity and the two scales themselves correlate highly. This leaves little doubt that both scales do in fact measure what they purport to measure.

The next step was obviously to combine the two scales into a single measure. This was done and the usual item-analysis checks carried out. The balanced 30-item scale so produced was found to have a reliability of .71. However, 4 of its items (all from those contributed by the students) were found to have a corrected item-total correlation of less than .1. These were excluded and a second item-analysis carried out. The result was a balanced 26-item scale with a reliability of .74 with only 1 item failing to correlate significantly with the scale total (after overlap correction). See Appendix I.

The correlates of the new scale are, then, of considerable interest. Its correlation with the prime predictive validity criterion was a very high .54. Subjects who described themselves as authoritarian by use of the 26-item directiveness scale were also highly likely to be seen as behaving in an authoritarian way by the students who rated them. By contrast, the attitude-to-authority scale also included in the questionnaire correlated only .19 with rated authoritarianism. This figure is also significant but obviously of a very different order from the .54 recorded with the personality scale. Interestingly, the correlation with submissiveness previously observed on some occasions with this scale did not appear here. The submissiveness rating and the AA scale correlated only .01. Overall, then, the previous observations of at best marginal behavioral validity for attitude scales of authoritarianism stand confirmed in the present study also.

The correlation between the two types of scale is also of no little interest --representing as it does the clearest answer yet to the question posed in the title of this paper. The coefficient found was .05 -- which is nonsignificant. When a behaviorally valid authoritarianism scale has been devised, we find it to have no relation to authority-approving attitudes. Some authoritarians approve of authority and some do not.

A question often overlooked in scale-validation studies is the question of the reliability of the criterion. If the criterion is itself unreliable, any correlation with it is of little importance. There was a means provided in the present project of running some check on this. This check is based on the fact that a second rating of the subject was made by each rater -- the submissiveness rating. When students were being instructed on how to complete the two ratings, they were specifically told not to assume that authoritarianism and submissiveness were two sides of the same coin. On each rating sheet appeared separate definitions of the two constructs which enabled the two attributes to be considered independently. These were respectively "By 'authoritarian' we mean: Tries to or would like to impose his will on others" and "By 'submissive' we mean: Tends to accept direction from others." Thus, if a student had rated a person as "highly authoritarian" (for instance), he was still to consider the possibility that the person might also be "highly submissive." The correlation between the two attributes actually observed in the study would therefore reflect the empirical relation between the two attributes rather than the requirements of some a priori definition.

The correlation actually found between the two ratings was -.45. This very decisively upsets the contention by Adorno, et al. (1950) that the two attributes are positively related. Submissiveness is the opposite of authoritarianism, not its concomitant.

This being so, the two ratings can be treated as reflecting two aspects of a single underlying phenomenon. In Likert scaling terms, they form a 2-item balanced scale. It is therefore reasonable to apply to these two peer ratings, the same statistical procedures that might be applied to two self-ratings. Applying the Spearman-Brown formula, we find that the reliability of the "scale" is .61 -- an eminently satisfactory result for a scale composed of only 2 items. The correlation of our 26-item directiveness scale and the rating scale was .44. Two quite separate and reliable sources of information have therefore been found to give convergent results.

Other correlations found with the 26-item directiveness scale that are of interest are as follows: Authoritarians (high scorers on the directiveness scale) are significantly more likely to be males (r = .28), are significantly better educated (r = .23) and are of significantly higher occupational status (r = .16). They may be of any age (r = -.03) or of any political persuasion (r = .04). All these relationships also hold for the original 14-item version of the directiveness scale. The 14-item and 26-item versions themselves correlated .90.

Another contrast of interest between the attitude and personality scales was in the relationship with social desirability. The attitude scale correlated positively (r = .26) while the personality scale correlated negatively (r = -.24). This means that saying you approve of authority was seen by the respondents as socially desirable but saying that you yourself were authoritarian in behavior was seen as socially undesirable. The latter correlation is not, however, high enough to affect seriously the validity of the scale. There was no significant relationship between either of the ratings and social desirability scale score.

The social desirability correlations throw an interesting sidelight on the composition of the sample. Designed as it was for what was basically a criterion groups study, there was of course no pretense of the sample being random. Students were, however, told that non-student respondents were preferable and this seems to have been largely accomplished. On the criterion of occupation alone, roughly 40% of the sample worked in manual occupations. When we come to the social desirability correlations, however, the non-student character of the sample is more vividly in evidence. One thing that is anathema to sociology students in particular and humanities students in general is authoritarianism. Approval of authority is extremely socially undesirable. Yet the correlations show that the people who responded to the questionnaire saw it as desirable. Very few of them can have been students.


The raison d'etre of the original authoritarianism studies was, of course, the supposed relationship between it and ethnocentrism or racism. This has been briefly touched on in Study I using a student sample only. In this third Study an attempt is made to examine the relationship using a community sample. This will also afford an opportunity of checking the directiveness scale's reliability on a community sample. Scales constructed on more educated samples do sometimes collapse when applied to more broadly based samples.

This study formed part of a project to investigate attitudes towards Aboriginals --Australia's indigenous colored minority. Since Aboriginals form only 1% (roughly) of the population and tend to live clustered together in only one or two depressed areas, very many Australians have had no contact with them whatever and have only a hazy notion of what they are like. Attitudes held toward them would then be of only the most superficial kind. For this reason, the present study was confined to an area where there was in fact frequent contact between the white and Aboriginal populations. This was the Sydney suburb of La Perouse -- where an Aboriginal reservation is located.

Sampling proceeded by moving outwards from the reservation and endeavoring to interview someone from every household in each block. Total N was 68.

The reliability observed for the directiveness scale was .72 -- quite comparable with what had been previously obtained. The mean score on the scale was also slightly (but non-significantly) lower for this sample. For the sample of Study II the mean was 55.41 (S.D. 8.07) while for this sample it was 52.52 (SD 7.57). One difference between this sample and the previous one which might have some role in explaining the difference in mean scores is the fact that in this sample roughly 70% worked in manual occupations. Given this opposite polarization in the two samples, it is doubly reassuring that the scale underwent no reliability collapse.

As well as the directiveness scale the questionnaire also contained an attitude-to-Aborigines scale (see Appendix IV). This was a balanced 10-item scale with a reliability of .62. The correlation between the two scales was .01. This confirms that, when authoritarianism is validly measured, it bears no relationship to, and hence cannot be used to explain, racialism or racial prejudice.


Where the present study would appear to differ from previous ones is in its failure to-accept the orthodoxy that the much flawed "California F scale" is a valid measure of authoritarianism. Once this view is cast off, the way is opened to a whole new view of the field.

The essence of the new outlook is to confine our conception of authoritarianism to its single most basic characteristic -- desire or tendency to impose one's own will on others. This is by contrast with the California view of authoritarianism as a complex of many covarying traits. As mentioned earlier, the reason for rejecting such a complex view is simply its failure at behavioral prediction and the untenability of the assumptions it embodies. Putting the latter point another way, the supposedly covarying traits just do not covary. See for instance Brown (1965) on the relationship between authoritarianism (as measured by the F scale) and rigidity.

The most important failure to covary, however, is the one concentrated on here -- the failure of attitudes to covary with behavior. This Adorno, et al. (1950) did not attempt to prove -- they just assumed it.

Some stress should be laid on just what one takes as the criterion of authoritarian behavior. There are, of course, many studies where various kinds of performance in experimental situations have shown some relation to scores on the California F scale. With most of these, of course, whatever the F scale measures is inextricably confounded with acquiescence. Such results, then, prove little. The contrast in the results obtained by Elms & Milgram (1966) and Sherell de Florance (1972) suggests that control for acquiescence can be sufficient completely to remove an effect formerly thought significant.

More importantly, however, in most of these studies the basic validity of the scale is assumed and it is actually an inference from the presumed nature of authoritarianism that is tested. Confirmation of this inference is not sufficient to validate the scale. It could well be that such an inference would be equally well justified by some other conception of what the F scale measured (such as conservatism).

What is needed, then, is a direct test of whether the F scale measures those central things that its authors explicitly claim it measures. Such tests are rare but, as indicated above, are clearly negative in import. The validation procedure used in Study II above, by contrast, could scarcely be more explicit or more central as a test of whether the scales concerned measured authoritarianism. If authoritarians are not dominant and not submissive (the two traits tested) they are nothing. Take these two traits away and the concept of authoritarianism becomes a very hollow shell indeed.

The method of obtaining validation information is also important. Results derived from behavior in one or two highly contrived and artificial laboratory situations are intrinsically severely limited in their generalizability. By contrast, the peer ratings method used here treats the rater as an accumulating data bank about the subjects' everyday and overall behavior. A rating is not based just on one or two items of information, nor is it based on information obtained in artificial settings. Its generalizability is intrinsically high.

It is hoped then that the present report will help to lay the admittedly, plausible ghost of Adorno, et al. (1950). The continued acceptance of the conclusions of these authors in spite of the mountain of methodological criticism leveled at their work is ample testimony, however, to the axiom that bad theories are driven out not by criticism but only by better theories. There must, however, be some limit to this phenomenon. Surely there must come a point when, no matter how attractive and how plausible a theory is, it will still be buried if the evidence is clearly against it.

It is of course in the spirit of "bad is only driven out by better" that the A, AA, BF, and directiveness scales have been provided. The A scale is an attempt to measure "classical authoritarianism" (i.e., authoritarianism as it was conceived before Adorno, et al. came along); the AA scale is an attempt at a simple overt measure of attitudes to authority in general without the assumptions of the F scale (though it does correlate very highly with the F scale); the BF scale is an attempt to provide a balanced version of the F scale without introducing any new item content, and the directiveness scale would appear to be the first attempt at a valid authoritarianism scale. It is hoped that the availability of these alternative instruments will in the long run enable more informative research on authoritarianism to be done.


ADORNO, T. W., FRENKEL-BRUNSWICK, E., LEVINSON, D. J., & SANFORD, R. N. The authoritarian personality. N.Y.: Harper, 1950.

BROWN, R. Social psychology. Glencoe, Ill.: Free Press, 1965.

CHRISTIE, R., & JAHODA, M. Studies in the method and scope of "The authoritarian personality" Glencoe, Ill.: Free Press, 1954.

ELMS, A. C., & MILGRAM, S. Personality characteristics associated with obedience and defiance toward authoritative command. J. Expt. Res. Pers., 1966, 1, 282-289.

EYSENCK, H. J. The psychology of politics. London: Routledge, 1954.

EYSENCK, H. J., & EYSENCK, S. B. G. Personality structure and measurement. London: Routledge, 1969.

HOLLANDER, E. P. Authoritarianism and leadership choice in a military setting. Journal of Abnormal Social Psychology, 1954, 49, 365-370.

KELMAN, H. C. Attitudes are alive and well and gainfully employed in the sphere of action. American Psychologist, 1974, 29, 310-324.

KILHAM, W. J. Level of destructive obedience as a function of transmittor and executant roles in the Milgram situation. Unpublished BA (hons.) thesis, Department of Psychology, University of Sydney, 1971. ,

LA PIERE, R. T. Attitudes and actions. Social Forces, 1934, 13, 230-237.

MARTIN, J. Acquiescence-Measurement and theory. Brit. J. Soc. Clin. Psychol. 1964, 3, 216-225.

MASLING, M. How neurotic is the authoritarian? Journal of Abnormal Social Psychology, 1954, 49, 316-318.

MILGRAM, S. Behavioural study of obedience. Journal of Abnormal Social Psychology, 1963, 67, 371-378.

MILGRAM, S. Some conditions of obedience and disobedience to authority. Human Relations, 1965, 18, 57-76.

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

RAY, J.J. (1971) An "Attitude to Authority" scale. Australian Psychologist, 6, 31-50.

RAY, J.J. (1972) Non-ethnocentric authoritarianism. Australian & New Zealand Journal of Sociology 8(June), 96-102.

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

RAY, J.J. (1974) Conservatism as heresy Sydney: A.N.Z. Book Co.

SHERELL DE FLORANCE, A. The disobedient victim -- an empathy hypothesis. Unpublished BA (hons.) thesis submitted to the Department of Psychology, University of Sydney, 1972.

SHILS, E. Authoritarianism, "Right" and "Left." In R. Christie & M. Jahoda (Eds.) Studies in the method and scope of "The authoritarian personality. Glencoe, Ill.: Free Press, 1954.

SHOSTROM, E. L. An inventory for the measurement of self-actualization. Educational and Psychological Measurement; 1964, 24, 207-218.

TITUS, H. E. F scale validity considered against peer nomination criteria. Psych. Record, 1968, 18, 395-403.

TITUS, H. E., & HOLLANDER, E. P. The California F scale in psychological research: 1950-1955. Psychological Bulletin, 1957, 54, 47-64.


JOHN RAY has written extensively on the measurement of authoritarianism. Most of this is in the journal literature, but several papers are also to be found in his recent book Conservatism as Heresy. He is a psychologist by training but is at present lecturing in Sociology at the University of New South Wales. His second major research interest is achievement motivation. He took his MA from the University of Sydney and his PhD from Macquarie University.


The items of the directiveness scale. alpha = .74. The first 14 items are those written by the author. The remainder were written by students. Response is "Yes" (scored 3), "?" (scored 2) or "No" (scored 1) unless the item is marked "R"-in which case the scoring is 1, 2, and 3 respectively.

1. Are you the sort of person who always likes to get his own way?
2. Do you tend to boss people around?
3. Do you like to have things "just so"?
4. Do you suffer fools gladly? R
5. Do you think one point of view is as good as another? R
6. Are you often critical of the way other people do things?
7. Do you like people to be definite when they say things?
8. Does incompetence irritate you?
9. Do you dislike having to tell others what to do? R
10. If you are told to take charge of some situation does this make you feel uncomfortable? R
11. Would you rather take orders than give them? R
12. Do you dislike standing out from the crowd? R
13. Do you find it difficult to make up your own mind about things? R
14. If anyone is going to be Top Dog would you rather it be you?
15. Do you give in to other people rather easily? R
16. Do you tend to-dominate the conversation?
17. Do you let your wife (or husband, as the case may be) get his/her own way most of the time? R
18. Are you generally a follower rather than a leader? R
19. Do you like to make your own decisions without assistance from others?
20. When you are going out socially, do you always like to have the say about where you will go?
21. Are you a fast driver?
22. Are you argumentative?
23. Do you like being waited on?
24. Would you prefer to hear a lecture rather than give one? R
25. Would you prefer to be a worker rather than a manager? R
26. Do you very often accept advice from other people? R


The BF Scale

The first 14 items are new versions and are scored 1 to 5 for "strongly agree" to "strongly disagree." The last 14 are F scale originals and are scored 5 to 1.

1. The rebellious ideas of young people are often a constructive source of change for the better.
2. Homosexuality between consenting adults may be distasteful but it should not be regarded as a crime.
3. Many of the radical ideas of today will become the accepted beliefs and practices of tomorrow.
4. Familiarity does not breed contempt.
5. People who want to whip or imprison sex criminals are themselves sick.
6. What a youth needs most is to be free to make up his own mind, to be flexible, and to work and fight for what he considers right personally, even though it might not be best for his family and country.
7. Many good people honestly could never bring themselves round to feeling much love, gratitude, or respect for their parents.
8. You need to get out and rub elbows with all kinds of people to get new ideas and broaden your understanding of life.
9. There are times when it is necessary to probe into even the most personal and private matters.
10. It's all right for people to raise questions about even the most personal and private matters.
11. Insults to our honor are not always important enough to bother about.
12. Sex crimes, such as rape and attacks on children, are signs of mental illness; such people belong in hospitals rather than in prison.
13. Most honest people admit to themselves that they have sometimes hated their parents.
14. Disobedience to the government is sometimes justified.
15. If people would talk less and work more, everybody would be better off.
16. The businessman and the manufacturer are much more important to society than the artist and the professor.
17. Every person should have complete faith in some supernatural power whose decisions he obeys without question.
18. No sane, normal, decent person would ever think of hurting a close friend or relative.
19. Young people sometimes get rebellious ideas but as they grow up they ought to get over them and settle down.
20. An insult to our honor should always be punished.
21. Sex crimes, such as rape and attacks on children, deserve more than mere imprisonment; such criminals ought to be publicly whipped or worse.
22. Obedience and respect for authority are the most important virtues children should learn.
23. Most of our social problems would be solved if we could somehow get rid of the immoral, crooked, and feeble-minded people.
24. People can divide into two distinct classes: the weak and the strong.
25. What the youth needs most is strict discipline, rugged determination, and the will to work and fight for family and country.
26. Familiarity breeds contempt.
27. There is hardly anything lower than a person who does not feel great love, gratitude, and respect for his parents.
28. Homosexuals are hardly better than criminals and ought to be severely punished.


Attitude to Authority Scale

All items are scored from 5 to 1. Normed on 96 army conscripts. a = .83. Items marked "R" are reverse scored.

Subset 1.
View of leader: executive versus decision-maker

1. If there is a disagreement about a policy, a leader should be willing to give it up. R
2. A leader should always change his actions to ensure agreement and harmony in the community. R
3. It is important for a leader to get things done even if he must displease people by doing them.
4. A national leader should follow the wishes of the community, even if he thinks the citizens are mistaken.
5. If a leader is himself sure of what is the best thing to do, he must try to do this, even though he has to use some pressure on the people.
6. It is all right for a leader to do something unauthorized, if he is sure it will be for the good of the people in the long run.
7. It is most important to have the participation of everybody in making decisions, regardless of their knowledge of the issues involved. R
6. It's always better' to try to talk people into doing things, rather than give them straight-out orders. R

Subset 2.
Evaluation of authoritarian institutions and other examples of the exercise of authority.

9. There's generally a good reason for every rule and regulation in public service departments.
10. In the Army, soldiers should not obey an order if it is obviously morally wrong. R
11. If the Army allowed more room for individuality it might be a better institution.
12. There is something wrong with anybody who likes to wear military uniform. R
13. When the dictator Mussolini made Italy's trains run on time, that at least was an important thing to achieve.
14. Two years in the Army would do everyone the world of good.
15. The Army is very good for straightening men out and smartening them up.
16. Civilians could learn a lot from the Army.
17. I disagree with what the Army stands for. R
18. You can be sure that Army procedures will be good, because they have been tried and tested.
19. Schoolchildren should have plenty of discipline.

Subset 3.
Freedom versus regulation.

20. People should be guided more by their feelings and less by the rules. R
21. People should be made to be punctual.
22. Efficiency and speed are not as important as letting everyone have their say in making decisions. R
23. There is far too much regimentation of people nowadays. R
24. You know where you're going when you have an order to obey.
25. People should not be expected to conform as much as they are today. R
26. People who say we can't have enough freedom here in Australia don't know what they're talking about.
27. I don't mind if other people decide what I am to do, or advise me how to do it.
28. It would be much better if we could do without politics altogether. R


The attitude to Aborigines scale. There is a 5-point response option to each item from "strongly agree" (S) to "strongly disagree" (1). Items marked "R" are reverse scored.

1. Aborigines generally don't show much inclination to work.
2. Aborigines have been unfairly discriminated against. R
3. Aborigines are not very hygiene-conscious.
4. Drunkenness is one of the greatest problems with Aborigines.
5. Aborigines often get into fights with one another.
6. Given the chance, the Aborigine will work as hard as the white man. R
7. It is only because they haven't had the same chance to get an education that Aborigines can't get work. R
8. Aborigines are a kind and gentle people. R
9. The Aborigines are a rather ugly race.
10. We could learn a lot from the way Aborigines always share with one another everything they've got. R


The attitude to Apartheid scale. There is a 5-point response option for each item. Items marked -R- are reverse scored.

1 . The South African government should definitely abolish its apartheid policy as soon as possible. R
2. Apartheid may not do as much harm as some people think.
3. The South African blacks are every bit as good as the whites and should not be discriminated against. R
4. Some separation of different races is a good thing if conflicts are to be avoided.
5. The South Africans are ruled by a brutal Fascist government. R
6. People here in Australia should do everything they can to smash Apartheid. R
7. If the blacks in South Africa were really being oppressed as much as some people make out, they could always leave the place or use force to get a better deal from the white minority.
8. It is survival of the fittest in this world so if the South African blacks can't help themselves they don't deserve to be helped.
9. It is no good us condemning the South African whites because we in Australia would probably do the same if we were outnumbered 4 to 1 by a less civilized race.
10. The South African attitude toward their blacks is a disgrace to humanity. R


Replication is one of the cornerstones of science. A new research result will normally require replication by later researchers before the truth and accuracy of the observation concerned is generally accepted. If a result is to be replicated, however, careful specification of the original research procedure is important.

In questionnaire research it has been my observation that the results are fairly robust as to questionnaire format. It is the content of the question that matters rather than how the question is presented (But see here and here). It is nonetheless obviously desirable for an attempted replication to follow the original procedure as closely as possible so I have given here samples of how I presented my questionnaires in most of the research I did. On all occasions, respondents were asked to circle a number to indicate their response.

Further to the finding above that dominant people were NOT submissive (contrary to Adorno) it may be of interest to note that Fox et al. (1977) also found -- using a large random U.S. community sample -- that "the command class was less likely than the obey class to acquiesce to authority".

It is also perhaps worth mentioning that Presley (1985) found big differences in attitudes but negligible differences in personality between political "resisters" and others. And one of her personality scales was in fact a short Directiveness scale thematically similar to the one described above but derived from Borgatta & Bohrnstedt. See Ray & Lovejoy (1986) for more on different Directiveness scale versions.

The Introduction to this article is an abridged version of a more comprehensive treatment of the subject here


Fox, W.S., Payne, D.E., Priest, T.B. & Philliber, W.W. (1977) Authority position, legitimacy of authority structure, and acquiescence to authority. Social Forces, 55 (4), 966-973.

Presley, S. L. (1985) Moral judgment and attitudes toward authority of political resisters. J. Research in Personality, 19, 135-151.

Ray, J.J. & Lovejoy, F.H. (1986) A comparison of three scales of directiveness. Journal of Social Psychology 126, 249-250.

Go to Index page for this site

Go to John Ray's "Tongue Tied" blog (Backup here)
Go to John Ray's "Dissecting Leftism" blog (Backup here)
Go to John Ray's "Australian Politics" blog (Backup here)
Go to John Ray's "Gun Watch" blog (Backup here)
Go to John Ray's "Education Watch" blog (Backup here)
Go to John Ray's "Socialized Medicine" blog (Backup here)
Go to John Ray's "Political Correctness Watch" blog (Backup here)
Go to John Ray's "Greenie Watch" blog (Backup here)
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 -- Backup here)

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 (Backup here)
Go to Selected pictures from John Ray's blogs (Backup here)
Go to Another picture page (Best with broadband)