Copyright Sociological Research Online, 1999


Alan Aldridge (1999) 'Prediction in Sociology: Prospects for a Devalued Activity'
Sociological Research Online, vol. 4, no. 3, <>

To cite articles published in Sociological Research Online, please reference the above information and include paragraph numbers if necessary

Received: 9/9/1999      Accepted: 28/9/1999      Published: 30/9/1999


The nineteenth-century founders of sociology were in no doubt that sociology is a predictive discipline. This was a key component of its constitution as a science, in contrast to religious mythology and metaphysics. Comte's formula 'Savoir pour prévoir et prévoir pour pouvoir' neatly captures the mood and mission of the sociological enterprise at its foundation. As the twentieth century closes, prediction has become almost a taboo word, connoting an embarrassing affiliation to vulgar positivism, scientism and technocracy. This article argues that many of our current fears about prediction are exaggerated or misplaced. If sociology is to regain its standing in the public domain we need to reclaim prediction as a core element in the sociological project.

Positivism; Prediction; Self-fulfilling Prophecy

From Totem to Taboo

'Savoir pour prévoir et prévoir pour pouvoir' - roughly, know in order to foresee and foresee in order to act. Auguste Comte's once famous, now notorious, formula underlined the crucial role played by prediction in the mission of positivist sociology as the supreme science of society.

Comte's vision of sociology has not fared well. His works are seldom consulted by contemporary sociologists or taught with any conviction to our students, who learn that 'positivist' is one of the gravest professional insults we can utter. His law of three stages serves as a paradigm case of the futile attempt to write a meta-narrative of history. His religion of humanity and positivist catechism, so embarrassing that few of us can bear to read about them, are dismissed as the sad products of his doomed love for Clotilde de Vaux. His positivism, far from being scientific, is discredited as misconceived and dogmatic. Even his incalculable impact on his contemporaries is forgotten. Comte's own work is now seen as irretrievably dated, but it inspired the French Impressionists, whose canvasses retain all the freshness of a thoroughly modern vision of the world.

According to Wright Mills, prediction enjoyed a pre-eminent position in the social sciences even as late as the 1950s. He claimed that:

'Among the slogans used by a variety of schools of social science, none is so frequent as: "The purpose of social science is the prediction and control of human behaviour"' (Mills1970: 127).

Mills argued with characteristic vigour that to make prediction a central element of the sociological project is to capitulate to an anti-democratic, 'bureaucratic ethos'. We squander our birthright of moral choice, not even for a mess of pottage but for an empty technocratic slogan. His overriding fear was of a growing duopoly, a stranglehold on the sociological imagination exerted by grand theory and abstracted empiricism. The high-flown conceptualizations of grand theory grant ideological legitimacy to the bureaucratic state, while abstracted empiricism supplies it with a technology of social control.

The rise of futurology in the 1960s appeared to bear out Wright Mills's fears. Long-range forecasting and scenario-building by countless commissions and think-tanks achieved an unprecedented ascendancy in social theorizing. The year 2,000 acted as a powerful symbol of the future (Kahn and Wiener 1967), in ironic contrast to the time of writing in 1999, when millennialist concerns have shrunk to more limited preoccupations: exaggerated fears of millenarian religious movements, hypochondria about the millennium bug, and mistaken anxiety about a global shortage of champagne.

A striking feature of futurology, rightly emphasized by Kumar (1978: 185-92), was its technocratic, jargon-ridden language. Kumar argued that perhaps the most shocking thing about the future was its dismal prose. It is not a trivial point. Here was a ruthless language fitted to the task of thinking the unthinkable.

While we may well concur with Mills's critique of technocracy, it is surely difficult to maintain that the slogan, 'the purpose of social science is the prediction and control of human behaviour', continues to be the dominant perspective in sociology. Quite the contrary: prediction has become virtually a taboo subject, erased from most of the dictionaries of sociology, the introductory textbooks, and the advanced overviews and position papers.

As an example, consider the testimony of Rodney Stark. Stark is one of a network of American sociologists - including Bainbridge, Chaves, Finke, and Iannaccone - who have developed what Stephen Warner (1993) has termed 'the new paradigm' in the sociology of religion. Stark's own approach is a Popperian, hypothetico-deductive, rational-choice theory. Together with Bainbridge, he has advanced a theory of religion derived deductively from a limited set of formal axioms (Stark and Bainbridge 1987). Their theory unashamedly yields predictions. One of these concerns the Church of Jesus Christ of Latter-day Saints, popularly known as the Mormons.

In a series of articles, Stark (1990, 1996, 1998) has documented the rise of the LDS Church from six founding members in upstate New York in 1830 to some ten million worldwide today. He has also projected the Church's growth rates into the next millennium, arguing that by the year 2080 there will be, on a rather conservative estimate, sixty million Mormons in the world. If the Church maintains its current rate of growth, the figure will be two hundred and sixty million.

These projections have been greeted with joy by the LDS Church and alarm by some of its critics. The response of Stark's colleagues has been revealing. He reports that he has been given the benefit of 'an amazing amount of counselling concerning the pitfalls of straight-line projections' (Stark 1996: 175). Stark's experience suggests that, at the turn of the second millennium, making predictions is not a way for sociologists to enhance their reputation among their fellow practitioners. It is a contrarian strategy.

In the case of the sociology of religion, Stark and others operating with the new paradigm have repudiated the dominant European orthodoxy, that is, the secularization thesis (Aldridge, 1999 forthcoming). The contrast is quite marked. Stark is engaged in prediction, while secularization theorists are more like prophets. The prophet does not foretell but forth-tell. Whereas religious prophets call down God's judgement on society, secularization prophets call down society's judgement on religion. I suggest that more predictions and rather fewer prophecies is the route for us to take.

Misplaced Fear # 1: Self-Fulfilling and Self-Negating Prophecies

Runciman's recent stimulating text, The Social Animal, (Runciman 1999) confronts the issue of prediction. Runciman repudiates the notion of sociology as a predictive science. I wish to suggest that although his arguments are cogent, we should not - given the current state of sociology - be too concerned about some of the familiar problems.

First, there is the issue of self-fulfilling and self-negating prophecies. Runciman gives a trenchant example of the latter. He knows, as we all do, that if he holds out his hand to greet the stranger Joe Soap, the chances are very high that Joe will shake his hand, as cultural conventions prescribe. But if Runciman makes the mistake of predicting that this will happen, offering $100 as a bet on the outcome, my tactic, as Runciman knows, is pretty obvious. I take Joe Soap aside, and offer him say $50 not to participate in the familiar ritual. Joe and Alan each gain $50, and Gary loses $100: fine sociology, lousy gambling.

The example may appear trivial, but as Runciman says it highlights the truth that social life involves human purposes and goals, and self-conscious decisions to pursue them.

The classic sociological source on the problem of self-fulfilling and self-negating prophecy is Merton (1957). He offers a fictionalized account of the collapse of a bank following widespread but unfounded rumours of its insolvency, commenting: 'The parable tells us that public definitions of a situation ( prophecies or predictions) become an integral part of the situation and thus affect subsequent developments' (Merton 1957: 423).

Self-fulfilling prophecies play a major part in ethnic and 'racial' conflict - and much of Merton's article is devoted to exposing to ridicule the tissue of self-serving self-fulfilling prophecies on which white racists have based their bogus theories of white supremacy. In keeping with his subject matter, Merton's article is an angry, morally highly charged piece of writing. He urges us to reject fatalistic acceptance of the power of self-fulfilling prophecy. In the case of banking, he points out that Roosevelt's New Deal, with its enactment of banking legislation and establishment of the Federal Deposit Insurance Corporation, produced a drastic reduction in the annual rate of bank insolvency. The implication is that social action can similarly deal with the cancer of racism. Significantly, Merton was not concerned with the more parochial problems that self-fulfilling prophecy presents to sociologists.

At one level, as Giddens argues (1979: 244), the problem of self-fulfilling and self-negating prophecy can be seen as purely practical. The solution is to insulate the prediction from the audiences which might respond to it. Runciman's (hypothetical!) error was to declare his bet in the hearing of a self-interested colleague in search of a fast buck. I suggest, however, that this is not a pressing problem for sociologists in the UK at the present time. We do not face a vast audience eager to exploit our every word. On the contrary, for most of us most of the time the problem is that we do not have such an audience at all.

Misplaced Fear # 2: Lucky Guesses

To be worthy of the name, as Runciman cogently points out, a prediction must be more than just a lucky guess. We are interested not in lucky guesses but in rationally grounded predictions whose merits can be discussed before the event (Rescher 1995).

Again, I suggest that this fear is somewhat misplaced. We have reached the point where we shy away from making any predictions at all. On the predictive front, we are not just risk-averse, which is rational, but risk-evading, which amounts to a refusal to engage fully in the social-scientific project. Who would listen to a weather forecast if all it did was tell us what the weather had been? How, then, can we expect the public to listen to us? If we insisting on waiting until we have a fully worked-out theory, we shall wait for ever.

It is worthwhile to reflect on the example of meteorology. First, weather forecasts are notoriously unreliable. We know that the weather is hard to predict: climatic conditions are mathematically chaotic, so small differences in initial conditions rapidly escalate into widely divergent outcomes.

Second, meteorologists do more than forecast the weather. They also tell us what the weather has been and what it currently is, with an explanation of why it is so. These narratives are an important element in what meteorologists do. Prediction is necessary but not sufficient.

Third, even though their predictions frequently prove inaccurate, their retain their audience. For all the problems of chaos theory, when the weather forecast predicts snow for the UK it is rational to put Wellington boots and a snow shovel in your car. We know we are dealing with probabilities, not certainties. If the snow does not fall, the meteorologists will offer an interesting explanation of why it missed us. They do not risk the sack for being wrong, even spectacularly so. On the contrary, a famously wrong prediction can turn you into a celebrity.

Prediction and Practical Consciousness

I am not arguing for a return either to technocratic futurology or to Comte's vision of sociologists as positivist scientists, as though nearly two centuries of sociological thought counted for nothing. We know now that these projects were na´ve. I have focussed so far on exaggerated fears, but, clearly, there are profound problems with prediction.

One problem has already been mentioned: chaos theory. Minute differences in initial conditions, differences which lie below the threshold of detectability, can produce vast differences in outcomes. Chaos theory is a salutary reminder that long-range forecasting is extremely hazardous.

A second problem, emphasized by Popper and reiterated by Runciman, is the necessary unpredictability of future knowledge. We cannot claim to know things which we have not yet discovered. This is the rock on which many predictions have foundered, for example Comte's assertion that we would never be able to determine the chemical composition of the stars.

It is worth adding that we should be foolish to hope that the future might one day be fully predictable. A literally surprise-free world would be intolerable, as in George Eliot's story The Lifted Veil, whose protagonist has the self-destructive ability to foresee in complete detail his own future. Were such a capacity possible, it would make Weberian 'salvation anxiety' over predestination seem no more than a minor ache.

Prediction should not be seen as something confined to the caste of scientists. It is an everyday activity, a key component of practical consciousness and necessary to evolutionary survival. Calling upon sociologists to make their predictions explicit is not to propel us back into a Comtean world of philosopher-kings remote from the common people. Giddens is quite right: orthodox 'scientific' sociology had an 'oversimple revelatory model of social science', in which sociologists produced startling insights which supposedly left the common people in awe (Giddens 1979: 248). Even so, recognizing the limits to prediction and abandoning prediction as a grand projet, does not mean that no predictions should be attempted. By repressing the predictive impulse, we are not displaying our humanistic credentials but being untrue to the life-world we share with our informants.

A Concluding Prediction

Some years ago, Giddens (1987: 275-96) wrote of 'the perils of punditry'. Despite his warning, I propose not simply to advocate prediction, but to make one. It is this:

Within twenty years, virtually all of the defined benefit occupational pension schemes in the UK will have disappeared. This will be true of USS, the Universities Superannuation Scheme which covers the vast majority of academic and academic-related staff in the pre-1992 universities.

I believe this prediction will prove to be true. It is also, clearly, testable. I doubt that it will be either self-fulfilling or self-negating, for the reasons given earlier. I also expect it to come true long before twenty years are up. I must make clear that I have no 'inside knowledge' of USS or any other defined benefit occupational scheme, nor am I suggesting that any of them is insolvent or improperly managed. The point is a more general one, related to broader patterns of socio-cultural change which sociologists have explored in great depth. For that reason, I would argue that the prediction is more than a lucky guess, since there are rational grounds for believing it - grounds which testify to the richness of sociology as a discipline.

The UK experienced a major scandal in the 1980s over the misselling of personal pensions to teachers, miners, nurses, police officers and others who were badly advised to opt out of their superior occupational schemes. Commission-hungry salespeople were widely blamed for the misselling, which still has not been fully resolved. Most of the leading financial institutions were implicated. This fiasco was enthusiastically launched in the 1980s by the then Conservative government. Without wishing to exonerate the guilty parties, including those whose 'professional' advice was truly abysmal and indisputably self-serving, there is one uncomfortable truth which should be confronted: occupational pensions were an extraordinarily good deal. It is not so much that personal pensions were a poor product, as that occupational pensions were and are too good to be true. In a market economy this is not a stable situation. Things that are too good to be true tend to be too good to last.

The salient features of defined benefit occupational pension schemes are three:

I would argue that this pattern of provision is incompatible with contemporary social trends. The evidence of the private sector supports this. Employers are withdrawing their occupational pension schemes in favour of money-purchase personal pensions, where the employee's contributions build up a fund which is used on retirement to purchase an annuity. The trend is one feature of the decline of corporatism. Among the reasons why the same process can be predicted for the public sector are the following:

This article began with the claim that prediction has moved from totem to taboo. Taboos do not necessarily succeed in regulating our impulses, which often prove too strong; taboos are a tribute which 'virtue' pays to 'vice'. For all the proscriptions, predictive activity in sociology is commonplace, as several colleagues have pointed out to me. We do not highlight our predictions, however. They remain implicit in our work: colleagues can discern them, but they are not made explicit to a wider public.

My own work has been no exception. Some years ago I wrote a number of articles about gender in the Church of England. A clear implication in this work was that opponents of the ordination of women to the priesthood would not be able to hold the line for very much longer. Why did I not make this explicit? The unedifying answer is that I thought it would detract from my chances of getting published in refereed journals.

Our efforts at Habermasian rational communication in the public sphere have not been fruitful because hardly anyone is listening. If we took the risk of making explicit predictions, we might stand a higher chance of giving sociology the public profile it deserves. As things stand, any anxiety that our prophecies might prove to be self-fulfilling or self-negating is less a reality than an aspiration.


ALDRIDGE, A. E. (forthcoming) Religion in the Contemporary World: a Sociological Introduction. Cambridge: Polity.

FREIDSON, E. (1994) Professionalism Reborn: Theory, Prophecy, and Policy. Cambridge: Polity.

GIDDENS, A. (1979) Central Problems in Social Theory: Action, Structure and Contradiction in Social Analysis. London: Macmillan.

GIDDENS, A. (1987) Social Theory and Modern Sociology. Cambridge: Polity.

HUTTON, W. (1995) The State We're In. London: Cape.

KAHN, H. and Wiener, A. J. (1967) The Year 2000: a Framework for Speculation on the Next Thirty-three Years. New York: Macmillan.

KEAT. R. and Abercrombie, N. (editors) (1991) Enterprise Culture. London: Routledge.

KUMAR, K. (1978) Prophecy and Progress: the Sociology of Industrial and Post-industrial Society. Harmondsworth: Penguin.

MERTON, R. K. (1957) 'The Self-fulfilling Prophecy', in R. K. Merton, Social Theory and Social Structure (revised edition). New York: Free Press, 421 - 436.

MILLS, C. Wright (1970) The Sociological Imagination. Harmondsworth: Penguin.

RESCHER, N. (1995) 'Prediction', in T. Honderich (editor), The Oxford Companion to Philosophy. Oxford: Oxford University Press.

RUNCIMAN, W. G. (1999) The Social Animal. London: Fontana.

STARK, R. (1990) 'Modernization, Secularization, and Mormon Success', in T. Robbins and D. Anthony (editors), In Gods We Trust. New Brunswick: Transaction, 201 - 218.

STARK, R. (1996) 'So Far, So Good: a Brief Assessment of Mormon Membership Projections', Review of Religious Research, Vol. 38, No. 2, 175 - 178.

STARK, R. (1998) 'The Basis of Mormon Success: a Theoretical Application', in J. T. Duke (editor), Latter-day Saint Social Life: Social Research on the LDS Church and its Members. Provo: Brigham Young University, 29 - 70.

STARK, R. and BAINBRIDGE, W. S. (1987) A Theory of Religion. New Brunswick: Rutgers.

WARNER, R. S. (1993) 'Work in Progress Toward a New Paradigm for the Sociological Study of Religion in the United States', American Journal of Sociology, Vol. 98, No. 5, 1044-1093.

Copyright Sociological Research Online, 1999