Bulletin of the Psychoanalytic Research Society, Volume IV, Number 1, Spring, 1995

In Defense of Empirical Research: A Feminist Response to a Feminist Critique

Gwendolyn L. Gerber, PhD
John Jay College of Criminal Justice, CUNY

Recently, some feminist scholars have described empirical research as a "masculine-typed" approach to science, and have called for women to use a more "feminine-typed" approach involving intuition (Carlson, 1972; Keller, 1985). These scholars want to include more qualitative data and methods as an accepted part of the scientific enterprise. By qualitative data, they mean the "softer" kinds of data that can be gleaned from interviews and case studies, and by qualitative methods they mean an approach to the data which seeks to find meaning and generate new conceptualizations without any preconceived hypotheses.

I'd like to set aside the problem that results from dividing the scientific world into "feminine-typed" and "masculine-typed" methods of study. It perpetuates the traditional division of the world into feminine and masculine spheres. Instead of women being confined to hearth and home, and men to work outside the family, this proposal would restrict the sexes to different kinds of science. And it is this sextyped division of the world that has led to the exclusion of women, and of men, from so many activities in the past.

Instead, I'd like to focus on the proposal that more qualitative kinds of data and methods be accepted within science and within psychology. I would agree that there is value in accepting the results from qualitative data as a legitimate part of the science of psychology. It has always seemed to me, as a clinical psychologist, that psychiatry has had an advantage in being able to develop theories and hypotheses using interview and case study data, sometimes (of necessity) with small numbers of subjects. Using these more qualitative approaches, psychiatry has been able to make creative contributions to our understanding of human behavior that would not have been possible within the limits of the experimental method.

There is a danger, however, in the use of qualitative methods, particularly in the area of sex and gender research. This is because so many of our assumptions about sex and gender appear to us as obvious and immediately apparent. It is extremely difficult, if not impossible, for us to step outside the ways we've learned to perceive experience, and the constructs we use to structure and conceptualize our experience. Without being aware of our motivations, we often strive to confirm our preconceived ideas because they give structure and meaning to the world in which we live.

So when we study phenomena like sex and gender, empirical researchwith its carefully controlled experimentation and institutionalized "objectivity"--is most crucial. And what we have found, when these experimental methods have been used, is that sometimes our most deeplyheld beliefs have been disconfirmed.

I'd like to focus on research involving the so-called "gender-stereotyped personality traits", because this is the area of research with which I've been most involved. I know that within this area of study, the scientific method has paved the way for crucial breakthroughs in our understanding and knowledge.

The term "genderstereotyped traits" refers to personality characteristics that most people think are typical of men or women. People believe that men are high in characteristics that enhance the self, such as self-assertion, decisiveness and independence. They also believe that women are high in characteristics that involve connectedness and concern for other people, such as warmth, consideration and accommodation to others (Bem, 1974; Spence, Helmreich & Stapp, 1975).

A variety of different scales have been developed to measure the genderstereotyped traits. In the earlier scales, the items were bipolar--one pole was masculine and the other pole was feminine. So, when people rated their own personalities on these scales they could either rate themselves toward the pole that was masculine or toward the pole that was feminine. Because of the way the scales were constructed, there was no way that subjects could rate themselves as having both types of traits.

A major breakthrough occurred when masculinity and femininity were conceptualized as separate dimensions that could be separately rated and measured (Bem, 1974; Spence et al., 1975). On the new scales that were developed as a consequence of this reconceptualization, individuals could rate themselves as having the personality traits associated with femininity as well as those traits associated with masculinity. In other words, individuals could describe themselves as warm and accommodating ("feminine-typed" traits) as well as assertive and independent ("masculine-typed" traits).

The use of quantitative and statistical methods were crucial in the development of these scales and in their further refinement. As the scales were employed in numerous empirical studies, the results of these studies showed that gender-stereotyped personality traits need have little to do with biological sex. Even more important, the research showed that these traits were not necessarily linked with gender identity, and therefore could not be used as an index of gender identity (Spence & Sawin, 1985).

When researchers realized that the gender-stereotyped traits were separate not only from biological sex, but also from gender identity, the names of the scales assessing these traits were changed to reflect this new understanding. Instead of referring to the scales as "masculinity" and "femininity", they were called "expressiveness" and "instrumentality", or "communion" and "agency", depending on the focus of the research. The terms that I'll be using here are communion and agency.

Communion refers to personality traits that reflect concern for other people and accommodation to the needs of others (and that, as noted earlier, people generally believe are more typical of women than of men). Agency refers to selfassertive personality traits that promote the enhancement of the self (and that people think are more typical of men than of women).

My own research has sought to explain what leads to the belief that men and women have different personalities. It has shown that these beliefs stem from status and power differences between the sexes. Because men typically have higher status and are generally in more powerful roles than women, they appear to be high in agency or assertiveness; because women typically have lower status and are usually in less powerful roles, they appear to he high in communion or accommodation (Gerber, 1988, 1991). In other words, my research has found that a man or woman who is in a very powerful role is perceived as high in agency or self-assertiveness, whereas a man or women who is in a non-powerful role is perceived as high in communion or accommodation.

In most of our everyday experience, sex and power are confounded. For example, if we think of the complementary roles of business executive and secretary-or even doctor and nurse-traditionally the more powerful roles have been occupied by men while the less powerful roles have been occupied by women. As a consequence, in the "real" world it is difficult or even impossible to disentangle the effects of sex and power and thereby study all the ways in which these variables can interact. It is only within the laboratory, using the techniques of experimental psychology, that it is possible to disentangle these two variables and study their effects.

As we have begun to re-examine what we mean by the social categories of "woman" and "man", we have begun to realize that they can be defined separately from the social roles that women and men traditionally have played. They can be separate from the occupational roles of homemaker and employee, separate from sexual orientation, separate from the choice to marry, separate from the choice to remain a parent, and even separate from the nature of the parenting role that is performed. Men can be "mothers" too!

However, this new understanding leads to a dilemma. We want to have a social definition, in addition to a biological one, of what it means to be a man or woman. Such definitions help us to characterize ourselves as "male" or "female", and help to reinforce our sense of gender identity. So it is increasingly uncomfortable for us to realize that many of the assumptions we've used in the past to anchor ourselves as "male" or "female" can no longer be used.

In reaction to this predicament, some feminist theorists, using qualitative methods of research, have focused on "connectedness with others" as the central defining attribute of femininity, and as central to a woman's sense of self (Belensky, Clinchy, Goldberger & Tarule, 1986; Gilligan, 1982). "Separateness" or "self-enhancement" is seen as the central defining attribute of masculinity, and as central to a man's sense of self. Because of the differing socialization experiences of men and women, these definitions resonate with our experience, and with the culturallydefined ways we've learned to interpret our experience.

These are problems with this conceptualization. When femininity is defined as connectedness or communion and masculinity is defined as self-enhancement or agency, this simply provides a more contemporary rationale for long-standing beliefs about the differences between the sexes. And as we've seen through our objective research methods, a person who is high in communion has little power, while a person who is high in agency has much power. Thus, with the best of intentions, this approach perpetuates traditional relationships between the sexes in which men are socialized to think of themselves as having attributes associated with power, and women are socialized to think of themselves as having attributes associated with powerlessness.

Even more importantly, empirical research does not support the notion that communion is central to a woman's selfconcept, and that agency is central to a man's selfconcept. It is true, as research has shown, that the stereotypes about men and women are very different. When you ask people to describe the "typical man" and the "typical woman", they describe them in very different ways; the man is very high in agency and the woman is very high in communion (Rosenkrantz, Vogel, Bee, Broverman & Broverman, 1968).

However, one of the best-kept secrets in psychology is that actual men and women are not very different from one another in their self-concepts. When men and women describe themselves using the traits of agency and communion, the differences between them are so small that they don't always reach statistical significance (see Gerber, 1989, pp. 4748). Consequently, how can we possibly place characteristics on which men and women see themselves as so similar into a central role in our theories regarding gender? I don't think we can

So we're left with our dilemma: How do we create definitions for the social categories of "woman" and "man" that take into account our current state of knowledge about ourselves and our world? I don't have the answer to this question. I just intend to keep on doing more objective, scientific, quantitative research.


About the author

Back to Spring, 1995 Bulletin

Psychoanalytic Research Society Home Page