The mark on the 100 mm scale was transformed into a figure between 0 and 100, the higher the figure, the more the respondent agreed. The figures were then categorised into five groups, each representing 20 mm of the scale. Respondents scoring 81–100 were considered to "agree strongly" to the statement and those scoring 0–20 to "disagree strongly". The scores were thus assigned to represent attitudes towards gender.

Associations between respondent gender and degree of agreement to each statement were analysed by Pearson's chi-2-test. P-values <0.05 were considered statistically significant. Multivariate logistic regression analyses were used to adjust for speciality, age, years in profession, and academic degree. A 95% confidence interval (CI) was used. For regression analyses the VAS markings were dichotomised >50/≤50 (agreeing/disagreeing). SPSS 8.0 was used.

Of the sample 35 % did not answer the questionnaire. Is there reason to suppose that they differed from respondents in any other way than what is shown in table 3? Did only persons interested in gender answer? It does not seem so. Low ratings existed among male as well as female respondents and there were some very questioning open-ended remarks.

Our study took place at Umeå University in northern Sweden, one of six universities with medical schools in Sweden. The education and curricula do not differ in any particular way from other medical universities in Sweden. Moreover, Sweden is a country known for a long time for its official policy and ideology which very actively encourages equality between women and men in all sectors of society. Thus we have no grounds to believe that the gender differences found in our study would be less pronounced in other medical schools in Sweden or in the western world. Still one must be cautious when trying to generalise our results, the sample referring to only one medical school.

We wanted to assess gender attitudes of physicians engaged in education. We found no gender-attitude questionnaire used before so we created one where we used statements about the importance of gender to agree or disagree with as an instrument. It might be argued that doctors' own assessments on the scales do not disclose awareness of gender issues and that our statements do not reveal or characterise gender attitudes. Other methods, for example observations or open-ended interviews might have been more reliable. Nevertheless we argue that our statements represent attitudes towards gender. We consider finding gender important as a prerequisite for introducing gender issues into medical curricula. We also believe that whether teachers agree or disagree to the importance of gender in different aspects of their work represent gender attitudes, since it affects the way they meet and understand male and female patients, students and colleagues and accordingly investigations, and treatments as well as teaching and research. We find it less likely that a person who is interested in gender issues, aware of the role of gender, and who recognises gender as a determinant of health would mark low on our scales. If you score gender of low importance you are probably not aware of the gender order that affects the health of women differently than the health of men and that permeates into professional as well as private relations. Several comments on the open-ended answers support these assumptions. For example, one low-scoring man wrote: "I am solely a professional – neutral and genderless."

Differences between women and men were greatest at the ends of the scale (figure 1). It might be argued that this reflects different ways for men and women to express themselves, women putting more emphasis on it when they agree and men when they disagree. Still, there were significant differences also when dichotomising in the middle of the scale (table 4) indicating that the differences between men and women were not linguistic.

Since gender issues are sometimes regarded as a controversial subject it might be that some respondents exaggerated the importance of gender to answer politically correct. If so, it was probably the male respondents who might feel uncomfortable and tempted to overstate their agreement. We therefore believe that the gender differences found underestimate, rather than overestimate true differences.

Both female and male physicians assigned more importance to gender in doctor-patient encounters than in the teacher-student dyad. This probably reflects that there has been more discussion about gender in consultation research 6 than in research about medical teaching and tutoring. Still, it was surprising that the teaching physicians rated the importance of their own gender in tutoring lower than the other statements. In Sweden as in most countries in the west, there has been an ongoing discussion on how to increase the number of women among the academic teachers. The rationale behind the efforts made has been explained foremost in terms of numbers and justice on behalf of the female teachers. Because of the low ratings on importance of gender in clinical tutoring we suggest that more attention should be given to the role of gender in the teaching situation. For example, many women teachers stressed the significance of same-gender role models in their open-ended answers in comments like "Role models are important. I missed female role models myself."

Our findings that female faculty consider gender more important than male faculty are in agreement with the work of others. In a Canadian survey of a psychiatry faculty about teaching on gender-related issues female physicians rated personal interest in the subject higher than male physicians did 21. Women also rated the importance of future teaching on these topics higher and the adequacy of current teaching lower than men did. A study in the US, at Stanford University School of Medicine, showed that the women teachers thought that gender discrimination and gender insensitivity was more of a problem than the men did 22. In a Swedish interview study about the doctor-nurse relationship female physicians believed that the gender of the physician plays a role for this relationship while male physicians were much more unsure 23. These findings are not surprising – but important and worth reflecting on. The gender differences found in assessments in our study probably reflect more than attitudes, such as differences in experiences and working conditions. One woman explained why she finds gender important in contact with colleagues and staff in this way: "Wow – many years of experience and a few years of awareness!"

This study, like most medical gender research, focuses on differences between women and men. However, applying a gender perspective is not restricted to revealing differences and/or dispossessed groups. A gender perspective includes that differences and inequity have to be understood, described and analysed in relation to the social construction of gender and the sociocultural contexts that create opportunities for gender-associated experiences and preconceptions to appear over and over again.

The fact that women physicians in this study had a higher response rate and that women more than men found gender important, can be interpreted as a reflection of sociocultural norms and expectations about women and men. Men have historically and culturally been perceived as norm for human beings while women have been conceived of as aberrant. Medical schools and medical research have reflected and reinforced such traditional stereotype and dichotomous views of women and men. In 1970 health professionals associated a healthy adult with a healthy male; a healthy female was described quite differently 24. In an investigation among medical students in the 1990s, they equate adults with men and see women as "not adults" or "others" 25. In line with this, our study illustrates that although both men and women faculty have a gender, it is foremost the women who are aware of it.

Gender is not a homogenous analytical category. Gender theory underlines that other contextual hierarchical categories, such as class, ethnicity and age, interfere and interact with gender to mediate personal agency, available choices and power 26. In this study age was not shown to interfere with physicians' attitudes towards the importance of gender (table 4). Class and ethnicity were not inquired about. Senior physicians were all considered part of upper middle class. As for ethnicity, members of the medical school faculty in this study are quite homogenous (Caucasian).

When focusing on gender differences there is a risk of reinforcing gender-related dichotomies and producing individual gender bias. What is true on a population level is not necessarily so on an individual level. Not all men in this study found gender less important. Eight percent of them "agreed strongly" to all five statements, thus agreeing to a higher degree than many individual women did. In spite of this variation between individual men and between individual women, gender differences on the group level and their consequences have to be addressed. Finding a way to do so, without the risk of reconstructing gender dichotomies at the same time, is a delicate matter in research as in everyday life. Our results should not (but might wrongly) be taken to mean that every individual male physician found gender of little importance.

The attitudes, interest and knowledge of faculty are crucial factors when implementing a gender perspective in medical education. Experiences show that a strong, clear commitment from the faculty leaders is required to prevent backlash 22. Offering same-gender seminars followed by mixed-gender seminars to discuss their own reactions and feelings has been shown to be a useful way to create a climate where teachers of both genders feel engaged and committed 16.

Our study showed that there was a substantial number of male teachers who assessed the importance of gender very low. This implies that men were more "blind" to gender than were women. Open-ended remarks suggest that several of them were not only "blind" but very negative. One low-scoring man put it like this: "I hope health care professionals stop thinking about gender and start dedicating themselves to helping poor women as well as poor men who suffer in the health care system and in society!" Is it possible to convince these "blind" and negative men to join the new initiatives or at least to pursue them to accept that gender issues are important? On the other hand, there were both men and women among those "agreeing strongly" to all the statements, and those teachers are a promising resource. We believe that it might be of certain weight to recruit the men interested in gender issues to the implementation work. If men participate to a larger extent in the discussions and execution of gender issues, i.e., if the question of gender in education is not left to women alone to solve, the result will be that gender stands out as an academic area instead of a field for women teachers struggling for better conditions. Instead of being solely a question of numbers and equality, however important that might be, gender issues will also be more easily identified as a question of competence and knowledge. This means that a large responsibility is put on the men who are interested in becoming involved and joining the work. However, it has to be underlined that more research about the implementation of a gender perspective is needed.