Ethnic and Racial Studies, Volume 6 Number 3 July 1983, pp. 363-364.

(With two post-publication addenda following the original article)

T. Hanf, H. Weiland and G. Vierdag, SOUTH AFRICA: THE PROSPECTS OF PEACEFUL CHANGE, London. Rex Collings, 1981, 492 pp., 15.

John J. Ray

University of New South Wales, Australia

In times when emotional responses to South Africa are so common, it is refreshing to find social scientists making their contribution to a solution of the South African problem in the way that they best know how: by doing research into how the problem came about, why it persists and what future developments are likely. If such research helps improve our understanding of why South Africa is as it is, then there is surely at least the possibility that some contribution to a workable solution to the problems in South Africa might one day be found without the preliminary necessity of the much-feared bloodbath. Given the occasional boast by South African whites that they have 20 million hostages to guarantee their safety from outside attack, there can be little doubt who would suffer most from such a bloodbath.

I therefore approached this book by Hanf and his colleagues (Hanf alone holds the copyright) in a very positive frame of mind. As I first heard of Hanf's research when I was myself in South Africa doing a survey of white ethnic attitudes, you might say that I was rather biased to say good things about Hanf's type of enterprise. As my approach to the problem is primarily psychological and as Hanf's is sociological, we should have complemented one another nicely.

I am afraid that I will have to say straight away, however, that in my view the work is so methodologically flawed as to be probably worthless. To explain this, I have to mention something that most psychologists know but many sociologists (including Hanf) appear not to know: that to examine people's attitudes by the use of single questions can be seriously misleading. There are two major problems: People are very prone to vacillate in the answers they give (as can be observed when you ask the same people the same question on two different occasions a month apart. The answers given on the two occasions often correlate quite poorly) and they very often take the question to mean something quite different from what it is intended to mean. Psychologists tackle these well-known problems by asking not one question on any one issue but rather a whole set of questions (often called a 'scale') that canvass all the aspects of a particular issue. Whether a person is 'for' or 'against' something is then judged not just by his stand on one question but rather by his average stand on many versions of what is ultimately the same question. This careful procedure does generally tend to produce information that is more reliable in every sense.

The great pitfall (and also the great strength) in the psychologist's procedure is that questions that the researcher has designed in the view that they all reflect aspects of the one underlying issue can very easily turn out to be responded to by the people surveyed as if they had nothing in common at all. The answers fail to intercorrelate. This does mean 'back to the drawing board' for the researcher but it also protects the researcher from a mismatch between his perceptions of what questions mean and the people's perceptions of what a question means. The researcher's 'scale' will be very unlikely to be 'reliable' (i.e. the answers will be unlikely to correlate to a satisfactory degree) until the researcher has succeeded in designing questions that are perceived by the respondents in the way that he intended the respondents to perceive them. Standard tests for 'scale reliability' (such as Cronbach's `alpha') are widely known and used. A set of questions with an `alpha' of .5 would generally be considered unsatisfactory while an alpha of .8 would be considered quite good (cf. Ray, 1980).

As it happens, Hanf does go some way towards the careful procedure of the psychologist. In many cases, he does include in the survey on which the book is based several questions on the one underlying issue. As far as I can see, however, Hanf gives no data that would enable us to judge the reliability of his 'scales'. As it is almost a byword in psychology that scales for which the reliability is not given almost invariably turn out to be very unreliable (e.g. Ray, 1971), this was a grave disappointment.

I therefore wrote immediately to Hanf to ask him if he had any reliability information and asked him for a copy of his raw data (in any form suitable to him) so that I could calculate the reliability statistics myself in case he was not interested to do so. As my correspondence with Hanf had been promptly answered before publication of the English-language version of the book, I expected this letter to receive a prompt response also. Over a year and several follow-up letters later, I have received no response. I can only conclude that Hanf now knows that his scales were unreliable (i.e. that they did not mean to the people surveyed what Hanf intended them to mean) and is unwilling to co-operate with any investigation that might prove this. That certainly seems to me to be a very disreputable stratagem for a scientist or scholar to adopt.


RAY, J.J. (1971) Correspondence: Regarding the Lynn n-Ach test. Bulletin British Psychological Society, 24, 352.

RAY, J.J. (1980) Racism and authoritarianism among white South Africans. Journal of Social Psychology, 110, 29-37.


1. One of Hanf's research-assistants did reply to this article, revealing that at least some of my letters had been received, but claiming that Hanf HAD replied to my various letters. Strange that the post-office stopped delivering his letters as soon as I asked for a copy of his data!

2. In the above comments on the use of single questions, I expressed the orthodox view among psychologists about their limited usefulness. And my own approach to attitude research has been overwhelmingly orthodox in that regard. I should make it clear, however, that I believe there to be some circumstances where single questions can yield useful information -- e.g. where political issues well-known to the public are being canvassed. I myself have used single questions as a secondary source of data in Ray (1979) and have reported research findings regarding their usefulness in Ray (1974 & 1975).


Ray, J.J. (1974) Are trait self-ratings as valid as multi-item scales? A study of achievement motivation. Australian Psychologist 9, 44-49.

Ray, J.J. (1975) Public opinion polls and attitude measurement. Current Affairs Bulletin 52, 24-30.

Ray, J.J. (1979) Does authoritarianism of personality go with conservatism? Australian Journal of Psychology 31, 9-14.

Go to Index page for this site

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