Writing up and publishing findings
There are four intertwined issues here: (1) where a research team had gathered data on both sexes; (2) did they publish all these data?; (3) how authors set about describing male and female subjects in their texts; and (4) which results they highlighted in their publications, and which of their respondents they believed and therefore reported as
‘credible’ to their readers. Feminist sociologists discovered that researchers had left data on women unpublished (actually in many cases unanalysed) while reporting on males only. One famous example is the social mobility study done in Britain in 1949 by David Glass. He gathered data on men and women, but only analysed and reported on the data on men in his publication (Glass, 1954). The data, on men and women, were all destroyed, preventing further analysis. Feminists found that males and females were described very differently in their reports. One example here is Lortie’s (1968) study of American school teachers. He gathered data on men and women, but described the women in stereotyped terms, blaming them for not only their lowly grades in the occupation, but also for the low status of the occupation in American society. It transpired that results pertaining to males were highlighted in publications while those on females were glossed over or downplayed. Nash (1977), for example, followed a co-educational class from primary to secondary school, but chose to write a journal article on the boys and not one on the girls.
There was another form of sexism prevalent. Re-analyses of original studies revealed that authors had believed male respondents’ accounts of social phenomena and disbelieved female respondents – reporting them as ‘misguided’, or ‘deluded’. Here Irene Jones’s scrutiny of the Murdock and Phelps data revealed a blatant example. Simply put, when the adolescent boys and girls studied by Murdock and Phelps told them different things, the researchers reported the boys’ version as the facts. The girls’ views were dismissed as ‘claims’, offered to the reader as delusions. The boys overwhelmingly told Murdock and Phelps that they spent their leisure time in all-male friendship groups. Half the girls said they spent their time in mixed groups. Murdock and Phelps believed the boys’ accounts and called the girls ‘hangers on’ who make ‘claims’ to belong to what are actually all-male groups. This was, as Jones pointed out, simply bad social science. Murdock and Phelps had actually discovered something interesting, and then ignored it. A finding that adolescent boys and girls see the same phenomenon differently is interesting. A stereotyped report that boys’ gangs are hindered by girls hanging around is not interesting. Lesley Smith (1978) subsequently replicated, in a small qualitative study, what Murdock and Phelps had found and missed.
One response to all these types of critique was to call for a ‘cleansing’ of methods: to replace sexist methods with non-sexist ones. Eichler (1988) published an influential book on this topic. The other response was to develop feminist methods, and these are the subject of the rest of the chapter.
However feminist methods are defined, and as the rest of the chapter shows, they are a contested territory, there are some commonly held
beliefs about what empirical work done since 1970 should look like. If it is large-scale, quantitative work, the investigator should do the following:
• either pose a non-sexist research question or one in which a potential sex difference is the question;
• should sample both sexes or explain and justify sampling only one;
• should design research instruments to test for sex differences rather than assume them in the design;
• analyse the data objectively;
• write them up so both sexes are portrayed as rational actors in their settings, and publish the data in a gender-neutral manner so stereotypes are not reinforced.
This, in short, means obeying the rules of positivist research, being objective, and not imposing one’s own values.
In qualitative, especially interpretivist research, avoiding sexism involves employing a tough-minded reflexivity. A minimalist manifesto for non-sexist research would include the following: Good researchers need to do several things. First, collect and report data on gender in the field setting; second, pay equal attention to all the informants in the setting, whether they are male or female (see L. S. Smith, 1978); third, collect data on how the actors in a field setting understand and view gender; fourth, gather data on how those beliefs are enacted (e. g. in speech, or in non-verbal behaviour); fifth, examine the relation between gender and power in the field setting; and all the time the researcher needs to make his or her own beliefs about gender problematic (Delamont and Atkinson, 1995: Chapter 9).
In the light of all these criticisms of the existing published research in sociology, it is not surprising that in the early 1980s feminists began to develop specific feminist methods.