Sunday, October 09, 2005

The Journal of Social Psychology, 1985, 125(2), 271-272

DEFECTIVE VALIDITY in the Altemeyer Authoritarianism Scale


School of Sociology, University of New South Wales, Australia

FEW BOOKS HAVE WON WIDER ACCLAIM than Right-Wing Authoritarianism by Altemeyer (1981), a particularly convincing critique of The Authoritarian Personality (Adorno, Frenkel-Brunswik, Levinson, & Sanford, 1950). It contains a new RWA scale developed by Altemeyer to replace the F scale and used by him to carry out a number of new studies of the sources of authoritarianism. What he establishes by its use is, however, disappointingly sparse: a number of things (such as high scorers having high-scoring parents) that could be equally well explained by saying that the scale measures nothing more than conservatism. In fact, a perusal of the scale's items positively suggests such a conclusion. They cover a range of issues very familiar in conservatism scales; what makes them particularly authoritarian is certainly not immediately evident. Positive validation of the scale as measuring authoritarianism as well as conservatism is, therefore, clearly called for.

Two other scales exist that were designed to measure solely authoritarianism and solely conservatism. If the RWA scale measures a mixture of these two attributes, then it should show moderate correlations with both. The first is the Ray Directiveness scale (Ray, 1976), a measure of authoritarian personality. It has repeatedly been shown both to give strong predictions of authoritarian behavior and not to predict overall conservatism (Ray & Lovejoy, 1983). The second is the Ray Conservatism scale (Ray, 1982), balanced against both acquiescence and authoritarianism; that is, it contains an equal number of pro-authoritary and anti-authority conservatism items as well as pro-authority and anti-authority "leftist" items. The scale, thus, measures conservatism because any tendency to measure authoritarianism is experimentally controlled for. The scale exists in 68-, 22-, and 14-item forms (Ray & Heaven, 1984), but only the 68-item form has exact balancing against authoritarianism. Because 68 items are too many for most surveys, a new 24-item form having complete balancing was devised for the present work on the basis of the item analyses of the full scale. Only items showing high item-total correlations were chosen; they included all those in the 14-item form.

The 24-item Conservatism scale, the 24-item Altemeyer scale, the 14-item Directiveness scale, and a social desirability scale were included in a questionnaire administered to a random cluster sample of 84 people interviewed door to-door in the Australian city of Brisbane. The reliabilities of the 24- and 14-item forms of the Conservatism scale were .82 and .79 (alpha). The Altemeyer scale had an alpha of .89 and a correlation between its positive and negative halves of -.58 (before reverse scoring). With strongly agree to a conservative item scoring 5 points and strongly disagree scoring 1 point, its mean was 75.62 with a SD of 15.56.

The Altemeyer scale correlated .76 and .81 with the two forms of the Conservatism scale and -.024 with the authoritarianism scale. The authoritarianism scale correlated -.045 and -.039 with the two conservatism scales. Clearly, the Altemeyer scale is just another conservatism scale.


Adorno, T. W., Frenkel-Brunswik, E., Levinson, D. J., & Sanford, R. N. (1950). The authoritarian personality. New York: Harper.

Altemeyer, R. A. (1981). Right-wing authoritarianism. Winnipeg: University of Manitoba Press.

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

Ray, J.J. (1982) Authoritarianism/libertarianism as the second dimension of social attitudes. Journal of Social Psychology, 117, 33-44.

Ray, J.J. & Heaven, P.C. L. (1984) Conservatism and authoritarianism among urban Afrikaners. Journal of Social Psychology, 122, 163-170.

Ray, J.J. & Lovejoy, F.H. (1983). The behavioral validity of some recent measures of authoritarianism. Journal of Social Psychology, 120, 91-99.


I should have mentioned above that there is another Canadian study that is everything which Altemeyer's work is not -- the study by Sutherland & Tanenbaum (1980). This was a remarkably rigorous study that used a large Canadian general population sample and applied to it scales that distinguished carefully between the various supposed "components" of authoritarianism. It may be noted from their Table III that high and low scorers of their measure of "General Obedience" (excerpted from the F scale) were virtually identical in political party orientation -- both being on average very much at the political centre in fact.

And in accordance with that, in a later book Altemeyer (1988, p. 239) reports that Right Wing Authoritarians as detected by his scale, "show little preference in general for any political party"! In other words, according to the RWA scale, half of Right-Wing authoritarians vote for Leftist political parties! So how can they be Rightist if they vote for Leftist parties? The scale therefore fails its most basic validity test. I was too generous above. The RWA scale is not even Right-wing, let alone Right-wing authoritarian. In what sense are the statements in the scale "right-wing" if right-wingers are no more likely to endorse them than Leftists are? Altemeyer is like a character in "Alice in Wonderland" where words can mean anything that he says they mean.

Even Altemeyer however seems eventually to have become perturbed about the meaning of his scale after the decline and fall of Communist regimes in the Soviet Union and Eastern Europe enabled use of his RWA scale there. Studies in the East such as those by Altemeyer & Kamenshikov (1991), McFarland, Ageyev and Abalakina-Paap (1992) and Hamilton, Sanders & McKearney (1995) showed that high RWA scores were associated with support for Communism!! So an alleged "Rightist" scale went from being non-political to being a measure of Leftism! If you took it at face-value, it showed Communists were Rightists! -- the absurdity of which I was not slow to point out at the time (Ray, 1992).

After that, Altemeyer more or less gave up his original claim and engaged in a bit of historical revisionism. He said (Altemeyer, 1996, p. 218) that when he "began talking about right-wing authoritarianism, I was (brazenly) inventing a new sense, a social psychological sense that denotes submission to the perceived established authorities in one's life". It is true that he did originally define what he was measuring in something like that way (in detail, he defined it as a combination of three elements: submissiveness to established authority, adherence to social conventions and general aggressiveness) but what was new, unusual or "brazen" about such a conceptualization defies imagination. The concept of submission to established authority was, for instance, part of the old Adorno et al (1950) work. What WAS brazen was Altemeyer's claim that what he was measuring was characteristic of the political Right. But it is precisely the "Right-wing" claim that he now seems to have dropped and the RWA scale is now said to measure simply submission to authority.

Even that claim, however, seems ambitious. In a general population survey, Heaven (1984) found that the peer-rated behaviours that the RWA scale significantly predicted were submissiveness (r = 0.22) and authoritarianism (0.20) but the very low level of the correlations may be noted. More importantly, however, there is evidence showing that there is no such thing as a consistent or overall attitude to authority -- not even to conventional authority (Ray, 1972; Ray & Lovejoy, 1990). People are discriminating about what authority they will accept and when they will accept it. So "acceptance of conventional authority" is now clearly a "unicorn" concept -- i.e. there turns out to be no reality there to correspond the words. But anybody who talked to committed U.S. conservatives about the U.S. Supreme Court in recent years would soon get an idea of how little respect conservatives have for THAT major example of conventional authority! James Lindgren has also drawn together some U.S. public opinion poll data showing that respect for authority among the public at large is anything but monolithic.

It may also be noted that, despite all the evidence to the contrary and Altemeyer's own backdown, the RWA scale still seems to be referred to by all its users as measuring something "Right-wing". As I have pointed out at some length elsewhere (Ray, 1987) psychologists hold to their prejudices so rigidly that they rarely let little things like evidence disturb them.

So what DOES the RWA scale measure?

There is, however, one shred of justification for psychologists continuing to refer to the RWA scale as measuring something "Right-wing". Like the F scale which was its inspiration, the RWA scale seems to have little to do with voting behaviour or, indeed, any sort of behaviour, but it does correlate well with various other measures of conservative attitudes -- as seen above. It does appear to measure some sort of conservatism, if not a politically relevant type of conservatism. So what could that possibly be?

The most probable answer is that, like the F scale, it simply measures old-fashioned attitudes. Conservatives undoubtedly have some respect for the past so they would tend to find some sense or plausibility in attitudes from the past and would say so. But the evidence is that they no more allow such thinking to influence their political decisions than Leftists do.

Another possibility is that vote is little influenced by ideology -- a view which has been put forward by Lipset (1959), among others. So maybe the RWA scale simply demonstrates that. Undoubtedly, there is a large grain of truth in that. Ideology clearly is only one factor in vote. Economic self-interest is another obvious factor. But to declare ideology irrelevant would surely be to declare irrelevant just about the whole of politics -- which does depend heavily on ideological statements of one sort or another! Fortunately, we do not have to do that. As I have shown in my own work, an empirically constructed scale of conservative attitudes can correlate up to .50 with vote and by using several scales, a multiple R of up to .70 can be obtained. So ideology may not be the whole story but it is still highly relevant -- if measured in a relevant way. Altemeyer's way was not relevant.

A third possibility has to do with the tone of the RWA items. They are generally expressed in a combative, aggressive, hostile way. It is clear that Leftists are angry and hate-filled people (witness how they behave when they gain unrestricted power -- as in Communist regimes) so that aspect of the RWA items could resonate with Leftists and attract assent from them -- while conservatives assent more to the basic content of the item. But which ever way you look at it, Altemeyer's scale is simply irrelevant in telling us anything about contemporary politics.

Altemeyer is of course aware of the validity problems with his scale but seems to be largely in denial about it. He has made various attempt to demonstrate validity for it and chief of these attempts seems to be how students and others perform in playing a game of Altemeyer's own devising. I will simply quote what another writer says about the game concerned:

"Altemeyer's favorite proof of right-wing turpitude comes from something he designed called the "Global Change Game." Altemeyer does not explain the game in detail, but, essentially, participants control various regions of the globe and then make decisions (e.g., wage war, allocate "resources," restrain population growth) about what their respective regions will do. Apparently, when only RWAs played the game, "after 40 years, not counting nuclear war, 2.1 billion people had died."

Frightening, no? Only until one reads that the 2.1 billion figure was calculated "according to a complicated formulae used in the game to take into account the consequences of war, long-term unemployment, malnutrition and poor medical infrastructures." In other words, the results of any game simply reflect the designers' assumptions as to how the world really works. Altemeyer takes it for granted, for example, that foreign aid from wealthy countries reduces suffering in poor countries, notwithstanding the contrary theory that foreign aid makes matters worse by entrenching kleptocracies and rewarding government failure. Hence, the hapless high RWAs who don't see the world the way Altemeyer does necessarily fail when they play the game. The Global Change Game, in short, proves only that Altemeyer's political views differ from those of conservatives. As he is hardly reticent about making this point to begin with, it is unclear why he needed a "sophisticated simulation" to prove it.


Altemeyer did however have still more to contribute in his role as the clown of political psychology. He went on to develop a scale of Left-Wing Authoritarianism -- the LWA scale. When he tested it on over two thousand people however, he could not find one single high-scorer on it! The LWA scale did not detect a single Left-wing authoritarian! Again he himself proved that his scale was not valid -- unless of course one is so totally one-eyed as to accept that there ARE no Left-wing authoritarians. If you are as good at waving magic wands as Altemeyer is, you might perhaps be able to claim that no such thing as Communism has ever existed, I guess.


No comments: