NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Academy of Sciences (US), National Academy of Engineering (US), and Institute of Medicine (US) Committee on Maximizing the Potential of Women in Academic Science and Engineering. Biological, Social, and Organizational Components of Success for Women in Academic Science and Engineering. Washington (DC): National Academies Press (US); 2006.

Cover of Biological, Social, and Organizational Components of Success for Women in Academic Science and Engineering

Biological, Social, and Organizational Components of Success for Women in Academic Science and Engineering.

Show details

Social Influences on Science and Engineering Career Decisions1

Yu Xie

University of Michigan

Abstract

Our study on the career processes and outcomes of women in science has four major components. First, rather than focusing on specific segments of a science/engineering (S/E) career, we studied the entirety of a career trajectory. Second, we analyzed seventeen large, nationally representative datasets. Third, we tried to be as objective and “value-free” as possible and to emphasize empirical evidence. Finally, we based the book on a life-course approach, a combination of special methodological perspectives which recognize the following phenomena:

  1. Interactive effects across multiple levels, such as the individual level, the family level, and the school level. Individuals do not live or work in isolation from one another.
  2. Interactive effects across multiple domains, such as education, family, and work. What we do in one domain of our lives affects what we do in other domains.
  3. Individual-level variations in career tracks resulting from differences among individuals, even those with the same demographic characteristics.
  4. The cumulative nature of the life course. What happened before affects what happens now, and what is happening now affects what comes next. This is also called “path-dependency.” Because of path dependency, small differences at particular points in time can deflect trajectories and subsequently generate large differences in career outcomes.

The life course approach places a high demand on data. Ideally, we would like to have longitudinal data over the entire career of many scientists and non-scientists. We looked very hard but were not able to find a perfect data set. Lacking such a data set, we were still able to carry out our study by piecing together many datasets to paint a composite picture of gender differences in science careers, a method which is called synthetic cohort in demography. Figure 2-12 shows the data sets that we used to look at different segments of science/engineering careers.

FIGURE 2-12. Synthetic cohort life course, career processes, and outcomes examined, and data sources.

FIGURE 2-12

Synthetic cohort life course, career processes, and outcomes examined, and data sources.

Needless to say, our study contains complicated and nuanced analyses. These analyses led us to conclude that women’s severe underrepresentation in science and engineering is an extremely complex social phenomenon that defies any attempt at simplistic explanations. Due to the complex and multi-faceted nature of women scientists’ career processes and outcomes, especially how these processes and outcomes affect, and are affected by, other life course events such as marriage and childbearing, we were uncomfortable recommending concrete policy interventions intended to increase women’s representation in science and engineering No single explanation or hypothesis testing should or could substitute for the richness of the empirical results from these analyses, though we did consider and reject several widely accepted hypotheses, as the following discussion shows.

The “Critical Filter” Hypothesis

One longstanding hypothesis in the literature is that women are less likely to pursue science/engineering careers because they are handicapped by deficits in high school mathematics training. In a classic statement of this position, Sells (1980) claims that “[a] student’s level of high-school mathematics achievement acts as a critical filter for undergraduate college admission for blacks and limits choices of an undergraduate major for women in general once they are admitted to college.” This hypothesis is appealing for its simplicity and the clear remedy it implies.

From our research, we find that the gender gap in average mathematics achievement is small and has been declining since the 1960s (see Table 2-5). The numbers in Table 2-1 are mean gender differences in math achievement scores (in standard deviation units). The declining trend shown in Table 2-5 casts doubt on the interpretation that the gender gap in math achievement reflects innate, perhaps biological, differences between the sexes. We also find that the gender gap in representation among top achievers remains significant (see Table 2-6). This finding was cited by Harvard President Larry Summers in his remarks at an NBER conference on January 14, 2005, which made international news. However, President Summers failed to cite the following finding: gender differences in neither average nor high achievement in mathematics explain gender differences in the likelihood of majoring in science/engineering fields.2

TABLE 2-5. Standardized Mean Gender Difference of Math Achievement Scores Among High School Seniors by Cohort.

TABLE 2-5

Standardized Mean Gender Difference of Math Achievement Scores Among High School Seniors by Cohort.

TABLE 2-6. Female-to-Male Ratio of the Odds of Achieving in the Top 5% of the Distribution of Math Achievement Test Scores Among High School Seniors by Cohort.

TABLE 2-6

Female-to-Male Ratio of the Odds of Achieving in the Top 5% of the Distribution of Math Achievement Test Scores Among High School Seniors by Cohort.

The Pipeline Paradigm

A dominant perspective in the literature on women in science is the “pipeline” paradigm. According to this paradigm, the process of becoming a scientist can be conceptualized as a pipeline, called the “science pipeline,” which is essentially a developmental process. Change in the developmental process along the life course is unidirectional—leaving science versus staying in science.

However, we find career processes to be fluid and dynamic. Exit, entry, and reentry are real possibilities. Many persons, especially women, become scientists through complicated processes rather than by just staying in the pipeline. Also, we show that participation gaps are greatest at the transition from high school to college. This is illustrated in Figure 2-13.

FIGURE 2-13. Sex-specific probabilities for selected pathways to an S/E baccalaureate.

FIGURE 2-13

Sex-specific probabilities for selected pathways to an S/E baccalaureate.

In Figure 2-13, we observe that in the senior year of high school, women are much less likely than men to plan a science/engineering major in college. In addition, women experience a much larger attrition from the science/engineering educational trajectory than men do at the transition from high school to college. In the later college years, however, we find women and men to have similar transition rates to attaining degrees in science and engineering.

The “Productivity Puzzle”

In an influential paper, Cole and Zuckerman (1984) state that “women published slightly more than half (57%) as many papers as men.” They found that the gender gap had persisted for many decades at this level and could not find any explanations for it. Out of despair, they called this gender difference the “productivity puzzle.” Later, Long (1992), after considering possible explanations, reaffirms this characterization with the observation that “none of these explanations has been very successful.”

We analyzed data from four nationally representative surveys of faculty in postsecondary institutions in 1969, 1973, 1988, and 1993.3 Two major findings emerged from our work about the puzzle. First, sex differences in research productivity declined sharply between the 1960s and the 1990s, even without any controls. This is shown in Figure 2-14. Women scientists’ research productivity has improved because their overall structural positions, such as institutional affiliation, have improved. This improvement in women’s productivity relative to men’s suggests that the large gender gap observed for earlier decades should not be attributed to innate biological differences between men and women.

FIGURE 2-14. Trends in female-male ratio of publication rate.

FIGURE 2-14

Trends in female-male ratio of publication rate.

Second, most of the observed sex differences in research productivity can be attributed to sex differences in background characteristics, employment positions and resources, and marital status. This is shown in Table 2-7. The first line of Table 2-7 reproduces the observed trends presented earlier in Figure 2-14. In lower lines, we included statistical adjustments for the fact that women and men differ in relevant characteristics, such as rank, year from a bachelor’s degree to PhD, and institutional affiliation. Thus, even in the earlier decades, the observed sex differences in productivity can be explained once these relevant attributes are controlled for.

TABLE 2-7. Estimated Female-to-Male Ratio of Publication.

TABLE 2-7

Estimated Female-to-Male Ratio of Publication.

Family Life and Women Scientists’ Careers

A common theme is the importance of considering the family in studies of women in science. In particular, we find that it is not marriage per se that hampers women’s career development. Married women appear to be disadvantaged only if they have children. For example, we show that, relative to their male counterparts, married women with children are less likely to pursue careers in science and engineering after the completion of science/engineering education4 less likely to be in the labor force or employed, less likely to be promoted,5 and less likely to be geographically mobile.6 Although some of the gender differences are attributable to the advantages that marriage and parenthood bestow upon men, they clearly suggest that being married and having children create career barriers that are unique to women—as opposed to men—scientists.

Table 2-8 presents the female-to-male odds ratio of post-baccalaureate career paths by family status. There are five destinations for graduates with a bachelor’s degree in science and engineering: (1) out of work and school altogether, (2) graduate school in science/engineering, (3) graduate school in nonscience/ engineering, (4) work in science/engineering, and (5) work in nonscience/engineering. For the five outcomes, we made four contrasts and found that in all four, married women with children are disadvantaged in terms of science/engineering careers. Column 1 shows that married women with children are less likely than men to either work or attend graduate school. In column 2, we see that they are less likely than men to be in graduate school rather than working. Furthermore, married women with children are less likely than men to be in science/engineering, either in work (column 3) or in graduate school (column 4). Similarly, we also find married women with children disadvantaged in terms of other labor force outcomes.7

TABLE 2-8. Female-to-Male Odds Ratio of Post-Baccalaureate Career Paths by Family Status.

TABLE 2-8

Female-to-Male Odds Ratio of Post-Baccalaureate Career Paths by Family Status.

Summary

While the conventional wisdom often draws on casual analyses of non-representative data, our tentative conclusions are based on very good data and careful analyses. Table 2-9 shows the contrast between conventional wisdom and our findings.

TABLE 2-9. Comparison Between Conventional Thinking and Our Findings.

TABLE 2-9

Comparison Between Conventional Thinking and Our Findings.

There appear to be two types of simplistic explanations. At one extreme, some observers claim that gender differences in science are all due to innate biological differences between men and women. At the other extreme, some scholars are tempted to make a sweeping claim that all gender differences are due to discrimination against women in school and at work. Our research shows that both positions are wrong. Otherwise, it would not be possible to explain either the rapid improvement of women’s position in science, which cannot be attributed to change in biological differences between the sexes, or the interaction between gender and parental status, which suggests that factors outside educational and work settings play an important role.

Women’s underrepresentation in science/engineering has deep social, cultural, and economic roots that will not be transformed by a few isolated policy interventions or programs. Increasing women’s representation in science/engineering requires many social, cultural, and economic changes that are large-scale and interdependent. After spending ten years searching for explanations, our research indicates we should stop looking for simple explanations and easy fixes, as attractive as they may be to us as human beings. Instead, we should look at the actual social processes that generate gender differences in science, and base policy interventions on empirical knowledge about these processes. Finally, while there may be policy changes that could address some of the complex reasons for women’s underrepresentation, we should not expect any individual policy change to bring about gender equity in science overnight.

References

  1. Cole JR, Zuckerman H. The productivity puzzle: Persistence and change in patterns of publication of men and women scientists. Advances in Motivation and Achievement. 1984;2:217–258.
  2. Long JS. Measures of Sex Differences in Scientific Productivity. Social Forces. 1992;71:159–178.
  3. Sells LW. The mathematics filter and the education of women and minorities. In: Fox L, Brody L, Tobin D, editors. Women and the Mathematical Mystique. Baltimore, MD: Johns Hopkins University Press; 1980.
  4. Xie Y, Shauman KA. Sex differences in research productivity revisited: New evidence about an old puzzle. American Sociological Review. 1998;63:847–870.
  5. Xie Y, Shauman KA. Women in Science: Career Processes and Outcomes. Cambridge, MA: Harvard University Press; 2003.

Footnotes

1

This presentation is based on the book Yu Xie co-authored with Kimberlee Shauman entitled, Women in Science: Career Processes and Outcomes, published by Harvard University Press in 2003.

2

See Xie and Shauman (2003). Ibid, Chapters 3 and 4.

3

See also Y Xie and KA Shauman (1998). Sex differences in research productivity revisited: new evidence about an old puzzle. American Sociological Review 63:847-870.

4

Xie and Shauman (2003). Ibid, Chapters 5 and 6.

5

Xie and Shauman (2003). Ibid, Chapter 7.

6

Xie and Shauman (2003). Ibid, Chapter 8.

7

Xie and Shauman (2003). Ibid, Chapter 7.

Copyright © 2006, National Academy of Sciences.
Bookshelf ID: NBK23778

Views

  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (1.7M)

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...