Political Psychology, Vol. 6, No. 1, 1985, pp. 141-146.
RANDOM SAMPLING MIGHT NOT BE IMPOSSIBLE AFTER ALL
J.J. Ray (1)
Treatises on sampling generally seem to assume that a random sample has been obtained. In real-life sampling, however, this seems never to be so - due to refusals to co-operate on the part of some of those drawn. Most of our statistics are therefore based on mythical assumptions. A way out of this problem would seem to emerge from a recent paper by Heaven (1983) which reports that in a door-to-door survey of white South Africans there were 106 completed interviews out of 110 planned. The present author has also carried out a random doorstep survey in another part of South Africa and obtained similar results. Refusals to cooperate are also very rare in India. It seems that with a little more effort the first truly random sample of a significant human population may be gathered from one of these two societies. Since white South African society is in almost all ways very similar to other Western societies, South Africa may be an important future venue for research where the theory to be tested demands rigorous accuracy in sampling.
KEY WORDS: sampling; South Africa; India; statistics; random; survey.
To judge by the research they report in their journals, psychologists seldom even attempt to gather random samples. Available groups of college students and white rats seem to generate at least 95% of the data analyzed. This tends to suggest that most psychologists feel the discipline to be still at a programmatic stage. There is more interest in exploring what a psychology of human beings could be like than in actually creating such a psychology. We are still at the stage of developing and trying out our research methods rather than being at the stage of giving them serious application. In saying this, I am assuming that everyone knows that relationships observed among college students need not tell us anything about relationships in the population at large (Ray, 1981). The same, of course, applies a fortiori to white rat subjects.
Nonetheless, it seems that at some stage we must face the need actually to start doing serious research into what really is going on out there among the people at large. Those of us who have already started making tentative steps in that direction, however, will be aware of one enormous problem: If we are to produce generalizable results, we need to study either the whole population or a representative sample of it. Yet both these options seem no options at all. Statisticians can tell us many marvelous things about the representative properties of marbles drawn out of a barrel, but one possibility they never seem to deal with is what we do when some of the marbles fight back and say (in effect), "Hey, I don't want to get out of this barrel! Go and draw some other marble!" That, regrettably, is what a lot of people tend to do when you try to sample them. All working statisticians and survey-takers have developed pragmatic ways of trying to compensate for such fractiousness but, whatever one does, the result can still never be a random sample. In fact, the present author hereby offers a small but valuable prize to anyone who can point him to a single truly random sample that has ever been gathered of any ordinary community. Samples gathered for government agencies are sometimes reported as if they were truly random but if one talks to actual field workers, it is not hard to ascertain how many "shortcuts" are in fact customarily necessary for such samples to be generated. Academic authors such as Martin and Westie (1952) or Scott (1979) report that even after extensive call-backs on "not at homes" and the like, non-completion rates for planned interviews can easily be 25% or more. The present author is sufficiently eccentric to have been doing much of his own doorknocking for over 10 years now and guarantees that no one who has ever done similarly will ever be in any doubt about how those marbles fight back. If a man just says: "Not interested" and slams the door in your face, what do you do then? No matter how carefully you prepare the ground for your visit, someone always will be that difficult.
Even government compulsion simply leads to evasion. As has previously been reported (Ray, 1974), army conscripts would seem to be a suitably captive group of subjects, but as has also been reported, such subject groups do seem to give rise to a remarkable lot of zig-zag response sets and the like. As the old anti-Nazi song said: "Die Gedanken sind frei". You cannot compel people to tell you their real thoughts. You must rely on voluntary co-operation.
Even quota sampling is nothing like a solution. For just one thing, how can you know that you have got the quotas right? If you rely on official census data, you are assuming that the census data are correct. Anyone who has ever seen former refugees from Central Europe asked to fill out a census form will be sure that there at least is one category that is sure to be under-represented. Certain life experiences can make people very reluctant to appear on any list of names -- let alone a government list.
Perhaps the one comfort in this situation is that the problem we face in obtaining generalizable psychological data is itself a psychological one. We are faced with a "physician, heal thyself" situation. If we have so far any claim to a useful understanding of human motivation, surely we should be able to manipulate people into co-operating with our data-collecting objectives. The first test of our ability to apply our psychology lies in our own backyard.
No doubt, therefore, many solutions to the problem of non-co-operation among our targeted subjects will in time be offered. Perhaps the simplest example of such a solution might simply be to look for a culture which is very polite and cooperative. If we can find a society in which refusals of cooperation run against the prevailing social norms, most of our work will have been done for us in advance. Such cultures do exist. It could even be argued that highly co-operative cultures have been the norm throughout human history. Individualism is a modern luxury.
Unfortunately, however, it is people in the context of modern Western culture that we are most likely to be concerned to study. Is there a compromise? Is there a society that is at once modern and also characterized by village courtesy? No doubt several possibilities will spring to the mind of various readers, but two in particular will be considered here. The first, perhaps surprisingly, is India. Aside from those involved in the tourist industry, Indians are extraordinarily "nice" people by Western standards. They have exactly the patience and courtesy needed to answer a questionnaire. Additionally, India is in many ways a very modern nation. Aside from that most dismal cachet of modernity -- its own atom bomb -- India has also orbited its own earth satellites and an extraordinarily high proportion of its population are skilled in the everyday use of a foreign language (English). And, after all, the first man to synthesize life (Khorana) did come from the Indian subcontinent. Despite the poverty, Indian cities (unlike Chinese ones) have virtually the same diversity as Western ones. There are numerous fairly autonomous Indian academic journals in the social sciences, even though the printing may be poor, the bindings flimsy and the English "quaint." Despite being poor and brown-skinned, urban Indians at least are very much like us.
Nor is this proposal mere theory. In a recent random population survey of the city of Bombay (Ray, 1982) the only non-co-operation reported by the interviewers was from the highly Westernized area around the skyscrapers of Nariman Point. Generally speaking, non-co-operation is something Indian survey organizations hardly have to think about. In a city less Westernized than Bombay, the holy grail of 100% cooperation might conceivably be reached.
Even so, however, the fact that India is very poor, very religious and still much influenced by what is to us a very alien culture must restrict the generalizability of Indian data. Can we therefore find a culture that is as co-operative as the Indian one but which is also affluent, predominantly Protestant in its traditions and thoroughly in touch with all that goes on in the modern world?
A recent paper by Heaven (1983) suggests that there may be: white South Africans. Heaven drew 110 people to be sampled and, obtained completed questionnaires from 106 of them. A less than 4% refusal rate! I have heard other psychologists tending to dismiss this remarkable report with some skepticism so I wish to add to it. I also carried out a random doorstep survey in South Africa (Ray, 1980). Although not previously reported, I had only one refusal in the entire survey. The refusal was an old lady who was too afraid to open her door at night. Had I been using true random sampling (I was in fact using cluster sampling), I could have called back in daylight, completed the interview, and obtained the first ever (to my knowledge) random sample of a significant human population!
At the time, I put my remarkable success as an interviewer (I did all the interviews myself) down to the fact that my introductory words were, on all occasions, "Hello, I'm from Australia." As my accent is clearly not South African and as most South Africans have probably considered Australia as a possible refuge should things in South Africa "go bad", this should have been a rather attention-getting greeting. It seems from Heaven's (1983) report, however, that I might have been congratulating myself unduly on my own cleverness.
White South Africans are co-operative even to South African interviewers. Heaven's interviews were carried out by Heaven and his students in the Afrikaner city of Bloemfontein and used uniformly the Afrikaans language. My interviews were carried out in the predominantly English origin city of Johannesburg. I found that people would ask me in and answer my questions even while they were having their Sunday dinner!
The reasons behind this exceptional courtesy can at this stage only be speculated upon but several hypotheses spring to mind. Heaven himself (personal communication) put his remarkable success down to the fact that "Afrikaners trust their own" but my similar findings among non-Afrikaner whites suggest that a broader explanation is needed. There may in fact be several factors at work. South Africans do have a tradition of hospitality of which they are very proud, for instance. The most critical reason, however, may lie in the "outside enemies" effect. National leaders in any country (Argentina being the most notorious recent example) are prone to use actual or threatened conflict with other countries to unify their population under their rule. Whites in South Africa find themselves in a similar position. There is the ever-present threat of their being swamped by the vast black majority. Anybody who is not a black is therefore a friend.
Execrated though they may be by most social scientists, therefore, white South Africans may have a unique value for social scientists in particular. Ironically, they may be the only advanced Western society where attributes and responses likely to be influenced by volunteer artifact (and what trait or response could confidently be excluded?) can be studied without such artifact present. South Africa could become our only source of true random samples. If this ever penetrates the consciousness of the discipline as a whole, tickets to South Africa may become as important items for funding purposes as experimental equipment now is.
The major doubt we might have concerning the use of South African data stems, of course, from their unusual social system ("apartheid"). As those who have been there know, however, it is remarkable how "apart" the black and white communities do live. The everyday life of a South African white is virtually indistinguishable from that of anyone else in the Western world. The major difference is the availability of domestic help for white housewives. Even so the washing will generally be done in a washing machine and white mothers tend to think that only they can look after their own children.
Perhaps, however, the sort of people who support a system as iniquitous as apartheid have to be different? As South African whites are of the same racial and cultural origin as many other Westerners, this seems unlikely. South African whites have even been shown not to be particularly prejudiced against blacks when their attitudes are compared with attitudes in other parts of the Western world (Ray, I980). While this is probably the best instance yet of the classical attitude/behavior discrepancy, it does have to be explained. A reasonable explanation is to see apartheid as simply a situation-specific response to perceived threat. If whites from the land of George Wallace, the Confederacy, and James Earl Ray were outnumbered 5 to 1 by blacks, what might they do? At least for some purposes, then, it seems that South African whites could indeed be a very useful model for Western man generally.
Heaven, P. C. L. (1983). Authoritarianism or acquiescence? South African findings. J. Social Psychol. 119: 11-15.
Martin, J. G., and Westie, F. R. (1952). The tolerant personality. Ant. Sociological Rev. 24: 521-528.
Ray, J.J. (1974) Are the workers authoritarian, conservative or both? Ch. 43 in Ray, J.J. (Ed.) Conservatism as heresy Sydney: A.N.Z. Book Co.
Ray, J.J. (1980) Racism and authoritarianism among white South Africans. Journal of Social Psychology, 110, 29-37.
Ray, J.J. (1981) Is the ideal sample a non-sample? Bulletin of the British Psychological Society 34, 128-129.
Ray, J.J. (1982) Authoritarianism and achievement motivation in India. J. Social Psychology 117, 171-182.
Scott, W. A. (1979). Correlates of political ideology. In Scott, W. A., Osgood, U. W., Peterson, C., and Scott, R. (eds.), Cognitive Structure: Theory and measurement of individual difference, Ch. 14, V. H. Winston, Washington, D.C.
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 or 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)