The role of theory in social work research: a further contribution to the debate.
Both authors agree on the need for more research evaluating the effectiveness of social work interventions in order to improve services to users. Thyer points out that this need has been recognized for decades, yet we are still far from developing a strong evaluative tradition. Without such studies, the drive towards making practice more evidence-based is inevitably undermined. How can we exhort practitioners to base their decisions on evidence when there are such big gaps in our knowledge? Thyer raises a very pertinent question in asking why it is proving so difficult to do evaluative research when it seems so widely accepted that it should be done. His answer is that researchers have the wrong priority and are preoccupied with testing theories at the expense of appraising practice. In support of this claim, he quotes an analysis of the content of articles published in thirteen major social work journals (Thyer, 2001b, p. 12). This categorized 47% as empirical research and, of those, 49% were explanatory studies (testing a theory aimed at explaining a phenomenon). Only 3% reported credible outcome studies of the results of social work practice. Thyer (2001b) also criticizes PhD programs for requiring students to formulate a theoretical framework for their study even when the study has not been guided by any explicit theory:
An otherwise sound piece of clinical or program evaluation may be forced to rest uneasily on a Procrustean bed of theory-testing research, sometimes being distorted beyond recognition. (p. 14)
The comments on how students artificially add theory onto their work resonate with my experience in the United Kingdom. Student social workers are required to do detailed case studies of their practice, making explicit the theories that guided their assessment and intervention. For many, if not most, the theoretical exposition is fabricated after the work has been done and merely to satisfy course requirements. The teaching of formal theories in college is often so cursory that, unless the practice supervisor uses a clear theoretical framework, the student is unable to apply any one theory in an explicit and systematic manner in their direct work with clients. Their academic tutors and practice supervisors often know about and collude with this cynical use of theory. This is pernicious because it promotes a culture that encourages social workers to see theories as irrelevant to their practice and as merely some kind of game played by academics. This anti-intellectual stance has coexisted with the drive to develop a scientific base throughout social work's history and is possibly the greatest obstacle to widespread adoption of evidence-based practice.
The anti-theoretical stance is also misleading because it implies that social workers can practice without using theory of any form. Here the disagreement between Thyer and Gomory comes to the fore. Thyer quite rightly says that practitioners can and do function without explicitly using formal theories but Gomory (2001a) is also right in saying that they inevitably use implicit theoretical assumptions as they make sense of a client's problems and try to help (pp. 30-34). Their disagreement is based on very different definitions of theory. They both explain the particular way in which they are using the term and, rather than adjudicate on which usage is "right" in some absolute sense, it seems more fruitful to appraise the substantive points they are making with the differing meanings of the concept of theory.
Thyer (2001b, p. 16) explicitly uses "theory" to refer to formal theories that explain and predict various aspects of human behavior. He cites psychoanalytic theory and Kohlberg's theory of moral development as examples. This use of "theory" is frequently found in the social work literature. Gomory (2001a), on the other hand, uses "theory" in a much broader way that is more familiar to philosophers: to refer to any assertion that goes beyond the evidence given to us through our senses (pp. 32-34).
In Thyer's sense, the research tradition at present can be faulted for concentrating too much on testing formal theories, which are, in fact, rarely used in any systematic way by practitioners. I agree with Thyer's argument that more effort should be made to evaluate practice as it is carried out, that is, without assuming it must involve testing some formal theory. Thyer's example (2001b, p. 18) of evaluating a counseling center where no specific theoretical orientation was being used seems to me to be a very common circumstance. It would be nonsense to impose a theory that was not, in practice, influencing the decisions and actions of the workers. However, I would disagree with his argument that such research can or should be theory-free in Gomory's sense of the term. Practitioners may not be using formal theories but they are using theoretical assumptions in their practice wisdom. They are guided in their understanding of how the client's problems arose and what responses might be helpful by some speculations about the causes of human actions. This background thinking is, for the most part, implicit, and effort is needed by researchers to try and help them formulate it, even in very general terms. It is, of course, possible to evaluate practice without trying to explicate the social workers' reasoning.
The evaluation of a counseling center that Thyer quotes seems to be of this type. But this "black box" research can only tell you whether the intervention had the desired effect; it has severe practical limitations. The administrators at the counseling center studied by Vonk and Thyer (1999) may well have been pleased to be told that the outcomes for the students looked very good but how are they to use these findings? Since the counselors used many diverse approaches, does it mean that they are of equal value? Can the administrators cost out the different styles of work and only offer the cheapest? Does it imply, perhaps, that the diverse approaches were irrelevant to the positive outcome? How can others read this study and use the findings to run another effective counseling center? The major drawback, from a practical point of view, is that we do not know what that intervention comprised and so cannot identify the causally important components. This is necessary for others to be able to replicate the effective response. They cannot copy any intervention in exact detail (at the very least, the place and personnel will differ) but need to know which factors are relevant to the successful outcome. Since the major reason for conducting evaluative research is to provide evidence to guide other practitioners, this flaw in black box research is fundamental.
Moreover, this black box type of research may be free of formal theories but it is far from neutral since it inevitably draws on some implicit theoretical assumptions and values. To amplify Gomory's points on his version of theory, practitioners, in making assessments and planning interventions, are going beyond what they directly observe and are speculating, at some level, about the causes of their clients' problems and what actions by themselves might lead to a positive outcome.
This broader definition of theory is central to one of the fundamental arguments against the positivist philosophy of science. To positivists, who accepted this broad conception of "theoretical," all statements were theoretical except those that were reports of sensory experience. The latter, "observation statements," had a special significance in their account of science in that their truth or falsity could be established by experience alone, thus providing an independent, empirical test of theories (Carnap, 1953, p. 367). As Gomory (2001a, p. 32) succinctly explains, this doctrine of positivism has been subjected to overwhelming criticism. We cannot approach the world as neutral observers but always have some preconceptions or interests that influence what we notice (O'Hear, 1989, p. 14). Moreover, when we report on that observation we use language that goes beyond the limited data received through our senses and draws on theoretical assumptions (Popper, 1959, chapter 5). Therefore, observation statements cannot be regarded as neutral and infallible reports of sensory experience but are, to some degree, theory-laden.
The practical implications for social work of this critique of the positivist account of observation are, first, that it challenges a traditional view in social work that practitioners can make an absolute distinction between the activities of observing and theorizing. An early version of this view is expressed by Mary Richmond in her influential textbook, Social Diagnosis, in which she tells social workers that, `in social study, you open your eyes and look, in diagnosis, you close them and think' (1917, p. 347). The second implication is that researchers cannot be neutral, and this challenges Thyer's claim that it is possible to evaluate practice in a theory-free way. Researchers inevitably bring some preconceptions to the research process that affect, at the very least, what questions are asked, how it is established that the client has received an intervention, and how the outcome is specified and measured.
Thyer believes that some types of intervention are common sense and not based on any theory. He uses as an example a service, Habitat for Humanity, that pairs poor people up with community volunteers and helps them build their own homes. He suggests that there would be "guffaws if we asked them about the HFH psychosocial theory regarding the causes of poverty" (2001a, p. 54). To me, there seems to be a clear assumption about the causes of poverty in this service in that it is believed that re-housing people will help them to a better economic state. In British history, it is easy to find people who would argue against this, claiming that poverty is the result of weaknesses within the individual, not society; re-housing them will not remove these flaws and so will not lead to any long-term improvement in their lot. Testing the efficacy of this service, therefore, will have significant value in testing the rival ideas on the causes of poverty.
To return to the problem posed by Thyer--why is there so little evaluative research--perhaps the answer lies in the complexity of conducting useful evaluations. Since social workers so rarely base their practice on an explicit, systematic use of a formal theory, the researcher needs to do a lot of groundwork in trying to identify what ideas are guiding their practice so that the findings can be used and replicated by others.
How Evidence Tests a Theory
Turning to the other issue I wish to address, the way that evidence tests a theory, I again find myself agreeing and disagreeing with both authors to some extent. Without getting involved in an extensive philosophical discussion, I want to offer an alternative view because both authors' positions seem unnecessarily narrow and so are restricting the role of scientific method in social work.
I want to begin with Gomory's Popperian account of scientific method and deal with two issues: the importance of falsifications and the role of inductive support (both of which involve proposing a third definition of "theory").
Karl Popper is renowned for expressing so eloquently and persuasively that science progresses by trying to falsify theories; scientists derive predictions from their theories and then conduct tests to see if they are confirmed or disproved (Popper, 1959). A theory that is compatible with any state of affairs whatsoever and, so, is not falsifiable is not scientific. This element of Popper's philosophy of science is not seriously contested by anyone. However, neither is this element peculiar to Popper; it features in every philosophy of science (although not in most people's common sense understanding of science). What is perhaps peculiar to Popper is his ability to present the idea in such a vivid and lucid way that it has been readily understood and accepted by non-philosophers.
Although simple, the notion of trying to falsify one's beliefs is radical because it differs so much from the way that we intuitively reason about the world. Francis Bacon, the seventeenth-century philosopher, first drew attention to the bias in the way people test their ideas: they notice the evidence that confirms their current beliefs and tend to ignore or dismiss evidence that challenges them. Once the human mind has formed an opinion
It draws all things else to support and agree with it. And though there be a greater number and weight of instances to be found on the other side, yet these it either neglects or despises, or else by some distinction sets aside and rejects. (Urbach, 1987, p. xivi)
Modern psychologists have confirmed this observation (Kahneman, Slovic, & Tversky, 1990, Nisbett & Ross, 1980). Scientific method is, therefore, significantly different from everyday, intuitive reasoning in that it systematically and deliberately tries to expose beliefs to severe testing. The point is highly relevant to social work since it would suggest that practitioners are likely to notice a biased range of (confirming) evidence so that they have unwarranted confidence in their judgements, an error that is apparent in child protection work where practitioners continue to believe children are safe despite growing evidence that they are suffering abuse (Munro, 1999). Social workers need to make a deliberate effort to be more critical and aware of counterevidence to the theoretical assumptions on which they are basing their decisions and actions (see Munro, 2002).
As I mentioned above, although trying to falsify theories is a standard element of philosophies of science, its importance is not widely understood by the general public. They tend to have what is called a "naive inductivist" view of science, two core features of which are a belief that scientists can study the world from a neutral position and that they then build up theories from these observations, theories whose justification rests on being developed in this way.
Gomory puts very clearly the reasons why this naive inductivist view is wrong. The case against neutral, assumption-free observation was made in the previous section. The case against theories being reliable because of the way they are created rests on the problem of induction. They cannot be built up in a mechanical, infallible manner from observations because they inevitably go beyond what has been observed and this entails inductive reasoning. David Hume (1739/1964), an eighteenth-century philosopher, first exposed the problem, demonstrating that inductive reasoning is fallible: However many instances of a correlation we may have observed, it is logically possible that the next instance will be different.
The fallacy that theories are derived from observation alone has been particularly influential in social work in contributing to the mistaken view that science in the social sciences should limit itself to observable behavior and cannot study the processes of the mind. Thyer, to some extent, puts forward this viewpoint when he stresses the importance of studying correlations--interventions and outcomes--without theorizing about why this correlation occurs (2001b, p. 21). In contrast to this view, science proceeds by making conjectures about the, often unobservable, processes behind the observed correlations, as in Newton-Smith (1981): "The discovery of correlations between observables, far from being the end of science, is but its beginning. Science begins when, having noted correlations, we seek an explanation of why they obtain" (p. 211).
Whereas extreme behaviorists such as J. B. Watson (1924) have claimed that a scientific social science cannot study the mind because it is not directly perceptible, the natural sciences are full of examples of theorizing about imperceptible processes and entities, such as neutrons and quarks. In this respect, science is not significantly different from the type of speculation implicit in social workers' practice wisdom as, reflecting on their experience, they develop more complex ideas about why clients behave as they do.
For most scientists and philosophers of science, induction is seen as necessary and the implication of Hume's argument is that the products of induction--scientific theories--are fallible. We can never say with certainty that they are true but we can rigorously test them so that our confidence in them grows; we consider they have a higher probability after subjecting them to severe tests.
The distinctive feature of Popper's philosophy, and the one for which he is most criticized, is found in his efforts to develop a scientific method that does not involve any inductive reasoning. In this enterprise, he is attempting to prescribe not describe science since scientists themselves do talk of theories being more or less probable, of being supported by the evidence. Popper claims that scientists can avoid induction by seeking only to use deduction. If they deduce observations that should obtain if the theory is true, then test them by experiments, they will either confirm or refute the prediction. Typically, scientists take confirmation as providing inductive support but Popper (1959) says they should only respond to the instances when the prediction is falsified: "The method of falsification presupposes no inductive inference, but only the tautological transformations of deductive logic whose validity is not in dispute" (p. 42). This has the logical implication of falsifying the hypotheses. To repeat Gomory's example, the hypothesis "all swans are white" cannot be proved inductively, however many white swans we find, but it can be falsified deductively by one single instance of a non-white swan.
There are three main criticisms of this philosophy. First, it assumes that one certain feature of science is that an observation statement that conflicts with a prediction from a theory can falsify that theory beyond dispute. This assumes that the truth or falsity of the observation statement can be determined infallibly but Popper accepts (and indeed has been a major proponent of) the view that all observation statements are, to some degree, theory-laden and therefore fallible. If it is logically possible that the observation statement is wrong, then it is logically possible that the theory, although apparently falsified by it, is true.
The second criticism is based on a more complex account of a scientific theory. It rests on the Duhem-Quine thesis. Both Duhem (1905) and Quine (1953) pointed out that scientific theories are complex structures and, at most, a falsification shows that there is a fault somewhere within the set of premises from which the observation was deduced. It does not, in itself, tell you where the fault lies and therefore which hypothesis should be rejected. Even in the simple example of the white swans given earlier, there must be at least two premises to choose between: "all swans are white" and "this is a swan." But science rarely deals with such simple hypotheses. Lakatos (1970) and Kuhn (1970) have both written extensively on this, Lakatos using the concept of "scientific research programs" and Kuhn the similar concept of "paradigms" to capture the many layers in a scientific theory.
The practical import of this criticism becomes apparent if we apply it to theories known to social workers, for instance psychoanalytic theory and behavioral theory. If either of these are tested and falsified, do or should people reject the entire theoretical approach? History tells you that people certainly do not give up--particularly in the case of psychoanalytic theory where the evidence on the effectiveness of the associated therapies is very negative. But they are not being irrational in holding onto the core assumptions. It would be irrational to claim that all is well despite the counterevidence but there are many options for changing lower-level hypotheses while protecting the central elements. Consequently, as Quine (1953) emphasized, "Any statement can be held to be true come what may, if we make drastic enough adjustments elsewhere in the system.... Conversely, by the same token, no statement is immune to revision" (p. 43).
Popper (1963) famously challenged psychoanalysts to state in advance what evidence would make them abandon psychoanalytic theory:
Criteria of refutation have to be laid down beforehand: it must be agreed which observable situations, if actually observed, mean that the theory is refuted.... What kind of clinical responses would refute to the satisfaction of the analyst not merely a particular diagnosis but psychoanalysis itself? (p. 38)
Unfortunately for Popper, this challenge is just as hard for natural scientists to meet. Lakatos (1970), Popper's colleague at the London School of Economics, points out that it is equally true of Newtonians: "But what kind of observation would refute to the satisfaction of the Newtonian not merely a particular version but Newtonian theory itself?" (p. 101).
The third criticism of Popper's philosophy was well expressed by Thyer (2001a, p. 56). Popper portrays an unreal picture of science, of scientists who are only interested in knowledge for its own sake. This is far from the truth. Science is about developing knowledge that can be used. Therefore, scientists must be concerned about the future reliability of their theories, not just their past performance, making inductive reasoning essential. Social workers use theories as guides to future actions. By encouraging evidence-based practice, people are not recommending that social workers should use an intervention because there is no evidence of its failure so far, but because there is some evidence that it has the desired effect. As the American philosopher Hilary Putnam (1981) puts it
Since the application of scientific laws does involve the anticipation of future successes, Popper is not right in maintaining that induction is unnecessary. Even if scientists do not actively anticipate the future (and of course they do), men who apply scientific laws and theories do so. (p. 335)
The limitations of both naive inductivism and naive falsificationism highlights the complexity of appraising theories. There is no simple algorithm, or mechanical process, that allows us to judge with certainty how reliable a theory is. There is a large philosophical literature examining what criteria are or should be used by scientists. There are two aspects relevant to social work on which all agree. As Popper stressed, a theory is more severely tested by, and so (to all but Popperians) receives more support from, a prediction that is highly improbable given existing background knowledge without the theory. Secondly, responding to a falsification by modifying some premise in the theory is not the end of the story. Scientists then test that modification by deriving new predictions and checking whether they are true or false. This can usefully be applied in social work since, as Thyer critically comments (2001b, p. 20), it is easy to find examples of social workers dismissing counterevidence on flimsy grounds which they do not test further. Suppose, for instance, that a child protection worker who thinks a mother is non-abusive dismisses an allegation of abuse from a child on the grounds that the child is in conflict with his or her mother and so is trying to get her into trouble. The worker should not rest easy with this explanation but try to test it further by looking for more evidence both for and against this hypothesized conflict.
There are three major implications for evidence-based practice that we can draw out of these philosophical debates. First, they reveal the impossibility of adopting a neutral stance in theorizing. People inevitably have some assumptions as they try to make sense of the world, to understand the causes of human behavior, or to look for ways of changing behavior. Instead of pretending to be value-or theory-free, we should rather try to make explicit the base from which we are approaching social work practice. As Thyer points out, practitioners make little systematic use of formal theories, so testing these theories, while having some value, leaves out a large part of social work practice. However, to evaluate this less formal method of practice requires researchers to try to formulate practitioners' implicit reasoning, and this is a considerable challenge. It is essential, however, if we are to be able to use the findings of evaluative studies to guide future practice.
The more optimistic point to draw out of this discussion is that, in refuting the positivist distinction between theoretical and observation statements, behavioral statements no longer have a privileged logical status. Psychological statements are not second class but share with behavioral ones the problem of having sufficiently clear rules of use so that there is high inter-user agreement about their truth or falsity. Whereas there seemed an absolute schism between social workers' intuitive practice wisdom and the strict version of behaviorism that avoided all reference to mental processes, there is a strong resemblance between the hypothesizing found in the early stages of developing scientific theories and in practice wisdom. Therefore, the enterprise of drawing out the assumptions guiding social work practice, though difficult, seems achievable. Scientific methods can build on, not reject, the current preference for intuitive reasoning.
The final point relates to the issue of evaluating a theory. Neither induction nor falsificationism provides an absolute answer. However well supported a theory is at present, it is logically possible that future observations will conflict with it. When a theory is falsified, all that can be inferred is that there is a flaw somewhere within the total set of premises. Deciding which premise to alter or whether to abandon the theory altogether is not a mechanical process.
Philosophers of science focus on the problem of deciding between two rival theories but, in social work, although this does arise to some extent, we are more often choosing between theories that are complementary rather than conflicting. One intervention may focus, for example, on improving an abusive mother's parenting skills, while another may be trying to reduce her social isolation. The effectiveness of one does not rule out the value of the other. Most social problems are thought to have multiple causes. Deciding which one to tackle is partly guided by knowledge of what works but is also strongly influenced by economic and resource issues and moral or political judgements about where, in the system, change should occur. However extensive an evidence base we develop, decisions on action will always be strongly influenced by values.
Carnap, R. (1953). Testability and meaning. In H. Feigl & M. Brodbeck (Eds.), Readings in the philosophy of science. New York: Appleton-Century-Crofts.
Duhem, P. (1905). The aim and structure of physical theory. Princeton, NJ: Princeton University Press.
Gomory, T. (2001a). A fallibilistic response to Thyer's theory of theory-free empirical research in social work practice. Journal of Social Work Education, 37, 26-50.
Gomory, T. (2001b). Critical rationalism (Gomory's blurry theory) or positivism (Thyer's theoretical myopia): Which is the prescription for social work research? Journal of Social Work Education, 37, 67-78.
Hume, D. (1964). A treatise on human nature. Oxford, England: Clarendon Press. (Original work published 1739)
Kahneman, D., Slovic, P. & Tversky, A. (Eds.) (1990). Judgement under uncertainty: Heuristics and biases. Cambridge, England: Cambridge University Press.
Kuhn, T. (1970). The structure of scientific revolutions (2nd ed.). Chicago: Chicago University Press.
Lakatos, I. (1970). Falsification and the methodology of scientific research programmes. In I. Lakatos & A. Musgrave (Eds.), Criticism and the growth of knowledge (pp. 91-196). Cambridge, England: Cambridge University Press.
Munro, E. (1998). Understanding social work: An empirical approach. London: Continuum Press.
Munro, E. (1999). Common errors of reasoning in child protection work. Child Abuse and Neglect, 23, 745-758.
Munro, E. (2002). Effective child protection. London: Sage.
Newton-Smith, W. H. (1981). The rationality of science. London: Routledge/Kegan Paul.
Nisbett, R., & Ross, L. (1980). Human inference: Strategies and shortcomings of social judgements. Englewood Cliffs, NJ: Prentice-Hall.
O'Hear, A. (1989). An introduction to the philosophy of science. Oxford, England: Clarendon Press.
Popper, K. (1959). The logic of scientific discovery. London: Hutchison.
Popper, K. (1963). Conjectures and refutations. London: Routledge/Kegan Paul.
Putnam, H. (1981). Reason, truth and history. Cambridge, England: Cambridge University Press.
Quine, W. V. O. (1953). From a logical point of view. Cambridge, MA: Harvard University Press.
Richmond, M. (1917). Social diagnosis. New York: Russell Sage Foundation.
Thyer, B. A. (2001a). Research on social work practice does not benefit from blurry theory: A response to Tomi Gomory. Journal of Social Work Education, 37, 51-66.
Thyer, B. A. (2001b). What is the role of theory in social work practice? Journal of Social Work Education, 37, 9-25.
Urbach, P. (1987). Francis Bacon's philosophy of science. La Salle, IL: Open Court.
Vonk, E. M., & Thyer, B. A. (1999). Evaluating the effectiveness of short-term treatment at a university counseling center. Journal of Clinical Psychology, 55, 1095-1106.
Watson, J. B. (1924). Psychology from the standpoint of a behaviorist. Philadelphia: Lippincot.
Eileen Munro is lecturer in social policy, Department of Social Policy, London School of Economics.
Address correspondence to Eileen Munro, Department of Social Policy, London School of Economics, Houghton St., London WC2A 2AE, UK; email: E.Munro@lse.ac.uk.
|Printer friendly Cite/link Email Feedback|
|Publication:||Journal of Social Work Education|
|Date:||Sep 22, 2002|
|Previous Article:||Bridging the gap between micro and macro practice: large scale change and a unified model of narrative-deconstructive practice.|
|Next Article:||Popper, positivism, and practice research: a response to Munro.|