2. The Incredible Shrinking Woman
Caroline Clarke Hayes
"The primary aims for this chapter are to present data from national data sets on the representation of women in multiple STEM fields over time (1966-2006), so as to characterize the extent and depth of this phenomenon, and propose a possible explanation for the recent decline in the proportion of undergraduate women in computer science. Additionally, the chapter will outline continuing research questions important for better understanding the phenomenon and discuss several approaches for change."
The data will be used to show the growth of computer science as an educational discipline, the changing representation of women at multiple levels of educational achievement (undergraduate, graduate, and faculty), and the changing representation of women in the computing workforce, illustrating parallels between computing education and the workforce. There is a wealth of existing literature out there on the factors that make it difficult for women in traditionally male-dominated STEM field, including a dearth of women role models, stereotypes of women and their skills being at odds with stereotypes of people in these fields, implicit gender bias, and an often unconscious reluctance from people in positions of power to include women in their networks or to include them as professional allies.
These factors do not explain why fields such as medicine, psychology, and the biological sciences have seen increases in representation of women, while computer science has seen a decline for the past twenty years. Existing hypotheses for the small proportion of women in computer science fail to explain why it has followed a trend so different from other STEM fields. For example, it is often suggested that women do not enter into computer science because they do not like math, but since 1985 between 45 and 50% of bachelor’s degrees in math have been awarded to women. Equal numbers do not necessitate equal status in all STEM fields, for even as representation of women has increased over the past 40 years, regardless of whether a field is male-dominated or not, there are proportionately fewer women holding management and leadership positions. Women also continue to be promoted more slowly and paid less than their male counterparts for comparable achievements.
Evaluation bias is the tendency to judge people's talents more positively when they match the stereotype for a discipline, can differentially impact the rates at which men and women are hired, retained or promoted, and in fields like computer science importunately favor men. The number of women in computer science grew steadily between 1967 and 1984, when it peaked at 37% of computer science bachelor’s degrees being awarded to women. The number has since fallen to 20% in 2006, raising a number of questions, first among them, is the situation in computer science unusual relative to other disciplines, and if so, how unusual?
Between 1972 and 1984, the proportion of women earning degrees in computer science increased more rapidly than any of the 21 STEM disciplines tracked by the National Science Foundation, but after 1984 it was the only STEM field to see prolonged decrease in participation of women. The rise may have been a result of the 1970s feminist movement which encouraged women to pursue careers in male-dominated fields, and computer science was a relatively new discipline without a set of gendered rules at that time. Yet no other STEM field has seen the prolonged decline in women’s participation at the undergraduate level that computer science has, even in cases such as electrical engineering and industrial science, which experienced temporary dips in enrollment but then bounced back.
The next question to examine then is whether this trend carries over from education to the workforce, and data from the Bureau of Labor Statistics would indicate that it does, with all computing job categories seeing rapid increases through the 1980s and even mid-1990s followed by rapid falloffs continuing to 2006. There appears to be a two-year lag between the peak of women earning degrees in computer science and women employed in the computer workforce, which may suggest that there is some transitional period between graduation and being hired, but since only 60% of computing professionals report having a four-year degree, it would seem that these aren't the same people that graduated two years earlier. Alternatively, the data could indicate that both undergraduate and professional populations are shaped by the same cultural biases.
The computing workforce grew rapidly and steadily from 1971 on, not experiencing the same fluctuations as education, and not all computing professionals came from a computer science education background (there would be no other way for the workforce to grow while degrees stagnate), indicating that social factors are independently shaping both who studies computer science and who chooses computer jobs.
The overall fluctuations in proportions of undergraduates earning bachelor’s degrees in computer science is not uncommon among established fields. The rapid growth before 1986 is somewhat, but this is reflective of the "start-up" status of the field. This is of importance because one result of "professionalization" is the creation and solidification of internal identities (how people within the discipline see themselves) and external ones (how the field is viewed by others, including stereotypes.)
"With few exceptions, all disciplines follow a similar pattern in which the percentage of women shrinks (and the percentage of men grows) as rank increases."
The percentage of women earning doctoral degrees in computer science has risen steadily from 1966 to 2006, same as all fields, at first following the same trends as the physical sciences and math, but after mid-1980s it starts to look more like engineering. This may be because computer science programs were originally housed within mathematics or physical science departments, often within liberal arts colleges, and were then rehoused in the same buildings as engineering after being formalized as their own departments, possibly leaving a large cultural impact on computer science as a discipline.