Women's Studies Int. Quart., 1978, Vol. 1, pp. 225-232
Pergamon Press Ltd.
Printed in Great Britain
FEMINISM AND RESEARCH* ALISON KELLY
Sociology Department, University of Manchester, U.K.
(Received and accepted May 1978) Abstract---This paper is concerned with the relationship between feminism and research. It examines the questions 'What is feminist research?' and 'How should feminism enter the research process ?'. Feminist research is defined as research that is undertaken for feminist reasons. Feminism should enter into the choice of research topic and the interpretation of research results. Feminist interpretation is particularly crucial, but is often neglected. However feminism should not impose a new orthodoxy on research methods. Feminist research can employ a variety of methodologies. The argument is illustrated by reference to a research project on girls and science. FEMINISM AND RESEARCH
I have recently completed a research project on girls and science (Kelly, 1978). Having worked on this project for a couple o f years, I realized that I didn't much like the results. This realization forced me to lift m y head from the pages o f computer print-out and start thinking a b o u t some of the wider issues. In particular I started to think about what is meant by feminist research. M y ideas on this are still by no means fully worked out, but I want to put them on paper as a contribution to the development of the concept o f feminist research. I want to begin with some negative observations. It is m u c h easier to say what feminist research is not than what it is. Clearly not all research on women, or all research by w o m e n is feminist. F o r example, I d o n ' t think anyone would claim that Corinne H u t t ' s (1972) work on sex differences, or Katherine Dalton's (1969) work on menstrual cycles were feminist. Both authors were studying women, making women the focus o f their investigations, but this was insufficient to make their work feminist. The question o f exactly h o w these researches are non-feminist is something I want to return to later. It m a y be useful to a p p r o a c h a definition o f feminist research by analyzing feminism. Feminism is, to use Myrdal's (1969) terminology, a valuation. It expresses our ideas o f h o w the world ought to be. Could feminist research, then, be research which provides support for our view o f how the world ought to be ? On this definition research results which suggest that men and w o m e n are equal (although not necessarily identical) or researches which re-evaluate w o m e n ' s social position might be considered feminist. But this a p p r o a c h has its dangers. It has a tendency to be post hoc--having completed the research we then decide whether or n o t the results accord with a feminist view o f how the world should be, in order to decide whether or not the research was feminist. Such a post hoc classification is useless for any practical purposes, and certainly does not accord with researchers' own impressions as to whether or not they are doing feminist research. A n alternative formulation, that feminist research is research which supports the aims o f feminism, but that this is to be decided before, rather than after, the completion o f the research project, implies that the * Paper presented to the BERA seminar on Women, Education and Research, Loughborough University, 15th April 1978. 225
results are known in advance. This is clearly unacceptable. Research, by definition, has to explore the unknown, and so it is meaningless to perform a priori classifications of research on the basis of the results that will be obtained. In politics or social work it is permissable to work towards a desired outcome, but not in research. Myrdal, a socialist researcher, says of his own work that 'my interest in truth seeking was as strong as my interest in social reform' (1969, p. 7) and ! think some element of this feeling is necessary for all research (although many researchers are more cautious in their use of the word 'truth'). The fundamental indeterminacy of outcome of research, means that feminist research cannot be defined as research which supports feminist beliefs. There is a place for polemic and committed literature, but that is not research. So feminist research must be distinguished from feminism. Yet feminist research is clearly value-laden. It includes a valuation term (feminist) in its description. It makes no attempt to claim complete objectivity, although I have argued that some element of 'truthseeking' must remain. How do feminist values enter into feminist research ? The problem of values in social research has been extensively debated by social scientists, from Weber onwards. 1 find Myrdal's little book 'Objectivity in Social Research' (1969) the most useful discussion, and I want now to try to apply some of his arguments to feminist research. Myrdal holds that all research is value-laden, although this is only occasionally admitted. The questions researchers ask illustrate their valuations (i.e. their ideas of how the world ought to be). In Myrdal's words: 'There is an inescapable a priori element in all scientific work. Questions must be asked before answers can be given. The questions are all expressions of our interest in the world; they are at bottom valuations. Valuations are thus necessarily involved already at the stage when we observe facts and carry on theoretical analysis, and not only at the stage when we draw political inferences from facts and valuations' (p. 9). By this token, research is feminist if it expresses a feminist interest in the world, i.e. if the reason for being interested in a particular topic is feminist. Feminist research is research that is undertaken for feminist reasons. Here is a clear demarcation. It enables us to specify why the work of authors such as Hurt and Dalton, although concerned with women, is not feminist; it was not undertaken for feminist reasons. The idea that questions embody valuations is not new. It is present in Mitchell and Oakley (1976, p. 14) who say 'Feminism is a method of analysis as well as a discovery of new material. It asks new questions as well as coming up with new answers'. Similarly, Dale Spender (1976) says 'Feminist research m u s t . . . [question] the whole basis of knowledge in the c u l t u r e ' . . . 'Male definitions have been responsible for deciding what is good knowledge (anything to do with power . . . ) and what is poor knowledge (who needs to know about housework--it's not even proper work); what is relevant knowledge and deserves to be pursued (maternal deprivation) and what is irrelevant knowledge and not worthy of further study (paternal deprivation). The cultural definition of knowledge, which for us is a man-made one, is all pervasive in decreeing what knowledge is important, how it will be gathered and how it will be used'. Both Mitchell and Oakley and Spender emphasize the newness of the questions asked in feminist research. To the extent that feminism provides new reasons for asking questions, and so generates new questions, this is the same point as Myrdal's. In many, perhaps most,
Feminism and Research
instances of feminist research, the questions are indeed new. But this is not necessarily so, and it is worthwhile to consider the extent to which the same question can be asked for feminist and non-feminist reasons, and the way in which this affects the course of the research. This brings me back to the problem with which I began, my own research on girls and science. I am interested in the theoretical question of why girls do worse than boys in school science, and in the practical issue of finding ways to improve girls' achievement in science. I think this is feminist research, because my reasons for being interested in the problem were feminist. I set these out in my research report as follows: 'As a feminist my c o n c e r n . . , is to increase opportunities for women, and to avoid the reduction of options implicit in rejection of, or failure at, science. Girls who cannot or will not learn science are cut off at an early age from a wide range of careers, hobbies and interests. By conforming to a feminine stereotype that excludes science they are channelled towards traditional women's occupations and the low pay and low status which frequently accompanies such occupations. Girls who succeed in science have a wider choice than those who fail, so the feminist seeks to reduce failure . . . . These concerns are with power and autonomy. Science is changing society, [and] the feminist position is that such an important process should not be controlled exclusively by men. To avoid this situation substantial numbers of women should become professional scientists and technologists. And all women should have sufficient scientific understanding to be able to take a full and informed part in current debates on such issues as environmental pollution, energy resources and genetic engineering. At a practical level science is important for a feeling of competence and environmental mastery. A woman in an industrial society is surrounded by machinery and labour-saving devices; often she does not understand or control this equipment sufficiently even to carry out simple repairs. Such a situation encourages a passive acceptance of an incomprehensible male-structured social order'. (Kelly, 1978, p. 2). Thus the question as to why girls do worse than boys in school science can lead to feminist research. However it is by no means a new question. It has been asked for at least 20 years (Arregger, 1958). Nor is it always a feminist question. Sometimes it is studied simply through intellectual curiosity (see e.g. Ormerod, 1975; Smithers, 1978). More often it is approached in terms of 'manpower planning' (sic)--the country needs more engineers or scientists, there aren't enough boys coming forward for training, so how can we motivate girls to fill these positions? Comments such as 'With the number of male engineering apprentices failing the [Engineering Industry Training Board] saw women as an untapped source' (New Scientist, 1978) abound. Feminist research may have a tendency to be more action oriented than other research. But neither the theoretical question as to the origin of sex differences in science achievement nor the concern to remedy girls' under-achievement in science are specifically feminist. Is there anything else, besides motives, which distinguishes feminist from non-feminist research? My answe'r would be yes. The question 'What is feminist research ?' can be reformulated as 'At what points does feminism enter the research process ?', and I want to examine that second question now. The research process can be crudely divided into three stages: (1) choosing the research topic and formulating hypotheses (2) carrying out the research and obtaining the results (3) interpreting the results
I have already argued that feminism enters into the first stage, that of choosing the research topic and formulating hypotheses. I want now to examine the extent to which feminism can and should influence the other two stages. This may be an appropriate point to mention that I am talking primarily about positivist research (i.e. traditional 'scientific' research). My discussion may not be relevant to other types of research, such as literary or historical research, or even classroom interaction studies. I feel that some of the same arguments do apply, but I really don't have enough experience in these areas to say. To illustrate this discussion I want to return to my own research on girls and science. I started the research with three main hypotheses as to why girls achieve less well than boys in science. These were (a) there are cultural pressures against girls achieving well in science (b) science is presented in schools in a way more suited to boys than to girls (c) boys have better attitudes than girls towards science. All these were, I think feminist hypotheses. They contain feminist valuations. I was hoping to show that, all other things being equal (such as the number of science courses taken, the pupil's liking for science, and the cultural pressures on children) girls would do as well as boys in science. In other words I wanted to show that the sexes were fundamentally equal with respect to science achievement. Unfortunately I couldn't. The results are set out in detail elsewhere (Kelly, 1978), but basically what I found was that the gap between the sexes in science achievement stayed more or less constant under all the conditions that I studied. As you may imagine, I wasn't very happy with these results. I began to wonder what, as feminists, we should do with results that seem to conflict with our value commitments. Where had I gone wrong ? Stage 1 of the research process (choosing the research topic and formulating hypotheses) had embodied feminist ideas. Had stage 2 (carrying out the research and obtaining the results) been non-feminist ? Or would stage 3 (interpreting the results) save the situation ? One reaction to results that we don't like is to attack the research methods as invalid. We can examine the assumptions and techniques in detail, and, since the perfect research project has yet to be performed, we can certainly find faults. But this takes specialist knowledge. If it is to be convincing it has to be done carefully and in detail, by people familiar with the topic. As such it is an elitist critique, one which is inaccessible to the bulk of feminists who may encounter the results in the press or the pub. This strategy is clearly essential in some situations (the recent scare report that high achieving women grow body hair like men is a good example (Owen, 1977)). But it is not essentially productive. Every study can be criticised, and this destructive reaction will not serve to develop feminist thought. A more radical version of the same approach is being advocated by some feminists. They argue that the traditional ways of doing research are all masculine ways, and that feminists must develop new research techniques. Objectivity and rationality are masculine traits, whereas subjectivity and personal experience are characteristic of feminism. Jessie Bernard (1973) speaks of the 'machismo' element in positivist research and says that 'a masculine bias has been embedded in the [methodology]'. She argues instead for a communal approach, operating by way of 'naturalistic observation, sensitivity to qualitative patterning, and greater personal participation by the investigator'. To label these contrasting approaches masculine and feminine seems to me a dangerous approach, liable to reify and crystalise
Feminism and Research
the differences between gender roles. I accept that subjectivity and personal testimony deserve a better reception than they have previously had in the research community, but this should constitute a widening of the range of possible methodologies rather than a replacement methodology. I accept also that different methodologies are appropriate for different problems--that questionnaires tap individualistic, conscious motivations whereas participant observation is a better technique for group feelings or unconscious motivation. But feminist questions can concern individuals o r groups, conscious o r unconscious motivations. I don't want to see one type of research technique branded as feminist, and others rejected, so that the field becomes narrow and sectarian. Just as research to pre-determined results was rejected because of an interest in 'truth-seeking', so research to pre-determined methodologies should be rejected. Worthwhile knowledge can be generated from many varied research strategies, and we should avoid the imposition of new orthodoxies on liberated thought. In addition to the cramping effect there is also a logical problem with the approach which rejects masculine research practices. If we believe that under patriarchy men have appropriated the best for themselves, then feminists should be seeking to emulate at least some of these attributes. Clearly we can also argue that power corrupts, and that men have been corrupted by patriarchy. But we must beware of rejecting human attributes when we reject masculine attributes. Rationality and objectivity, the ability to reflect on our actions, are important human attributes, which can be seen as the fullest development of our intellectual capabilities, and we should not lightly disown them. An alternative reaction to unpalatable results is to attack the research as socially irresponsible. We can argue that results antipathetic to feminism undermine the dignity of women, and that research funds should not be allocated to these studies. This type of response is clearly evident in some of the Race/IQ debate, and traces of it can be detected in some feminist responses to biological work on sex differences (Sherman, 1976; Griffiths and Saraga, 1977). But I find this type of censorship emotionally repulsive--what about the truth-seeking element in research ? Moreover I think it is counter-productive. Any attempt to suppress research findings is an expression of weakness, on admission that one is frightened of the results. Far better to face the problem and argue it out. The wilful suppression of knowledge seems more likely to lead to an eventual backlash than to solve the problem posed by the distasteful results. The drift of this argument is that I reject the idea that feminism should enter into the second stage of the research process, that of carrying out the research and obtaining the results. I think feminist research should be a catholic activity, embracing a wide range of methodologies and accepting the results that are produced. However I believe that feminism is a crucial component of the third stage of research, that of interpreting the results. This element of the research process is frequently neglected. But results do not speak for themselves, they have to be interpreted. It may be necessary to delve quite deeply into the topic to make the relevant feminist interpretation, when, as in my research, the obvious hypotheses don't work. But I believe that it is possible, and that re-interpretation of the existing body of research results from a feminist viewpoint is long overdue. This is an area which feminist theoreticians could profitably explore. This point can be illustrated by considering one of the thorniest problems for feminists, that of biologically based sex differences in behaviour. Again, I was forced to think about this because of the way my own research turned out. I did include a biological pre-disposition among my original hypotheses on the origins of sex differences in science achievement, but this was purely formal since I had no data which were in any way relevant to it. However
it has been suggested to me that by weakening the credibility of my other three hypotheses I have strengthened the credibility of a biological hypothesis. I see no reason to accept this argument--there are a multitude of possbile hypotheses, and weakening the original suggestions does not strengthen the biological hypothesis as against any of the other possibilities. In my own interpretation of the results I favour a cognitive-developmental model of socialization which views science as a masculine subject, followed by a feedback loop in schools (Kelly, 1978). Nevertheless there are some indications that som ecognitive abilities are biologically based (Maccoby and Jacklin, 1974), and I want to consider how we might develop a feminist interpretation of such research. Many feminists seem to react almost instinctively against the very idea of a biological basis to sexually dimorphic behaviour. And they have good reason. It is undeniable that biological theories have been opportunistically produced and used to justify women's oppression. Griffiths and Saraga (1977) document this with historical examples. For example discussing research on brains they say 'The earliest ideas suggested that the frontal cortex was the seat of all intellectual skills, and it will not be a great surprise to us to discover that observations demonstrated that the frontal lobe was in fact larger in men than in women! However towards the end of the century, further research suggested that it was the parietal lobes n o t the frontal lobe that was the location of intellectual functions. One of the workers in this field, Patrick, wrote in 1895 ' . . . the frontal region is not as has been supposed smaller in women but rather larger relatively . . . but the parietal lobe is somewhat smaller, [furthermore] a preponderance of the frontal region does not imply intellectual superiority . . . the parietal region is really the most important". Griffiths and Saraga go on to argue that all research on the biological basis of sex differences in intellectual ability is politically suspect and should be abandoned. They hold that the present revival of interest in biological bases for sex differences in behaviour is an opportunistic response to the rise of the Women's Movement; and that such research is invalid because it makes simplistic assumptions such as extrapolating from animals to humans without taking account of the social characteristics of humans, extrapolating from biological to social roles, separating nature and nurture, and assuming one way causation from genes or hormones to behaviour. Many of these points are certainly valid, although some of them are in turn simplistic and can be easily answered. There is a continuing debate, but this is a matter for experts and need not concern us here. My point is that, rather than rejecting the work, and demanding that it be abandoned (which is unlikely to happen) we should develop a feminist analysis of the results. Many researchers find these problems fascinating (for whatever reason) and I am loath to try and censor their activities (truth-seeking again). But if supporting evidence for a biological basis to sex differences in behaviour continues to accumulate it is necessary to be forewarned with a feminist interpretation of these results. This approach has the added advantage that it provides arguments which non-specialists can use. When confronted with biological arguments I think there are three main lines that feminist replies can take. The first is to stress the large overlap between the sexes in all measured characteristics. Many research reports concentrate on the differences rather than the similarities between the sexes. Means for males and females and the difference between means, are far more often reported than standard deviations or measures of overlap. But group differences cannot be applied to individuals in any meaningful way. As Maccoby and
Feminism and Research
Jacklin (1974, p. 367) point out, even if the relevant skills were completely determined by the genetic distribution of spatial ability there would be 'more than enough women to fill our engineering schools if women's talents were developed through the requisite early training and interests'. This brings us to the second line of argument. Biological factors provide only a predisposition to acquire certain skills, and do not imply that these skills are unlearned or unlearnable. Biology is not destiny, and societies have the option, through schooling and socialisation, of providing additional training in the areas where each individual is weakest, so as to produce citizens with well rounded personalities and competences. This is already happening with reading. Boys have less verbal ability (on average) than girls, possibly for biological reasons. Remedial reading classes are provided in most primary schools, not for the less able group (boys) but for individuals (of both sexes) who have trouble reading. The majority of the pupils in these classes are boys. Reading is considered such an important skill in this society, that it is worth making an effort to teach it to all pupils, irrespective of their genetic pre-disposition. Boys are not devalued because they predominate in these classes. Why not similar classes for pupils who have difficulty with the spatial skills underlying maths and science--clases which would cater mainly, but not exclusively, for girls ? We must avoid the naturalistic fallacy--what is natural is not necessarily right or inevitable. Even if, in most primitive societies, women do most of the domestic tasks, this does not determine what should happen in our society. We are not a primitive society, we can change these patterns by our technology and our conscious decisions. After all, the 'natural' infant mortality rate is very high, but no one argues that this is right or unalterable. The third line of reply to biological arguments rests on the differential evaluation of male and female attributes. This is a point that will be familiar to most feminists. It is put extremely well by Jessie Bernard (quoted in Spender, 1976) 'How does it happen that so much is made of the fact that the blood of males has more androgen than the blood of females, but nothing is made of the fact that it also has more uric acid ? And how does it happen that the net effect of the vast corpus of research leads to the conclusion that men are superior to women on all the variables that are valued highly in our society: namely muscular or kinetic strength, competitiveness, power, need for achievement and autonomy ? . . . These are the variables that interest men. These are the variables they judge one another by. These are the variables that are rewarded in our society. These are the variables we need to know about in order to operate successfully in our society. These are the prestige variables !' So we need to attack the prestige variables. When confronted with a putative biologically based sex difference, we should ask what is so important about it ? Why is it considered a desirable attribute? Are there any compensating virtues possessed by the opposite sex? For example, when told that men are less emotional than women and so more fitted to posts of responsibility, we can point out that women are less aggressive than men, and less likely to have heart attacks than men, and so more fitted to posts of responsibility. This is a more constructive line of argument than disputing the fact that in present society women are more emotional than men. This discussion of biological arguments is included merely for illustration. The main point is that feminist interpretation of existing research results is possible, and that we do not have to reject results which at first sight appear unpalatable. Interpretation is an
important c o m p o n e n t o f the research process, and feminist interpretation is an important c o m p o n e n t o f feminist research. In s u m m a r y then, I have attempted to anwer the question ' W h a t is feminist research ?' by examining the points at which feminism should enter the research process. I think that research is feminist if the reasons for undertaking it are feminist. Feminism should enter the research process at the initial stage o f choosing a research problem. This frequently but not always leads to new questions being posed. Feminism should enter the research process again at the final point o f interpreting the results. Feminist analysis can and should be applied to any research results. This is a crucial ingredient o f feminist research, but one that has received little theoretical discussion and is often neglected in practice. But I think that feminism has little place in determining the research methods used or the results o b t a i n e d - - a n y research m e t h o d o l o g y can be feminist in the right hands. A n d feminist research has to be clearly distinguished from feminist literature or politics. In literature and politics a case is argued, a point o f view is put across. But research must be more impartial, its results determined by where the material leads, rather than by what the a u t h o r desires. Feminist research has an important part to play in the liberation o f women, but this can only be realized by a careful combination o f ideas from feminism and from research practice. REFERENCES
Arregger, C. E. 1958. The Recruitment of women as physicists--the problem in the schools. Bulletin of the Institute of Physics 9, 289-294. Bernard, J. 1973. My four revolutions: an autobiographical history of the ASA. In Huber, J. (ed.) Changing Women in a Changing Society. University of Chicago Press. Dalton, K. 1969. The Menstrual Cycle. Penguin. Griffiths, D. and Saraga, E. 1977. Sex differences in a sexist society. Paper presented at Sex-Role Stereotyping Conference, Cardiff University. Hutt, C. 1972. Males and Females. Penguin. Kelly, A. 1978. Girls and Science: An International Study of Sex Differences in Science Achievement. Almqvist & Wiksell. Maccoby, E. E. and Jacklin, C. N. 1974. The Psychology of Sex Differences. Stanford University Press. Mitchell, J. and Oakley, A. 1976. The Rights and Wrongs of Women. Penguin. Myrdal, G. 1969. Objectivity in Social Research. Duckworth. New Scientist. 1978. Short of men, engineering firms turn to women, 77, p. 811. Ormerod, M. B. 1975. Single sex and co-education: an analysis of pupils' science preferences and choices and their attitudes to other aspects of science under these two systems. Paper presented at Girls and Science Education Conference, Chelsea College. Owen, L. 1977. The myth of the hairy lady. The Guardian, p. 11, 28 July. Sherman, J. 1976. Effects of biological factors on sex-related differences in mathematics ach!evements. Paper from Women's Research Institute of Wisconsin. Spender, D. 1976. Man made knowledge: problems of research in feminism. Introduction to workshop at Women's Studies Conference, University of Manchester. Smithers, A. G. 1978. Girls in science. Paper presented to the Findlay Society, University of Manchester. March 1978.