Values of Men and Women, Part 2

Last week I blogged about differences between the career-related values of men and women. I used data from the 2003 National Survey of College Graduates, which was conducted by the Census Bureau on behalf of the National Science Foundation. Respondents were asked about the importance of various factors (which may as well be called values) that they consider when they think about a job.

Last week I looked at which values male and female respondents most often identified as very important. This week, I’m looking at what the different college major choices of men and women indicate about their value differences.

Of course, the values that the survey covers are not necessarily the only considerations that men and women bear in mind when they make the choice of a major. Some of these values, such as salary and job location, are expressed in job-related terms and therefore can influence the choice of a major only insofar as students consider the relationship between their prospective major and their future job. On the other hand, the job-major relationship is a very significant (often paramount) influence on many students’ major decisions, so it’s understandable that men’s and women’s feelings about careers would be related to their preferences for certain majors. Furthermore, some of these values, especially intellectual challenge, apply equally well to a major as to a job. Therefore, it seems likely that any male-female differences choosing between these values will be reflected in their choices of majors.

I thought it would be useful to look at the correlation between (a) the percentage of graduates of a major who say a value is very important and (b) the percentage of male and female graduates of a major. In other words, I’m asking which values tend to be rated highest by graduates of the majors with the most female graduates and lowest by graduates of the majors wit the least female graduates (and the same question for men). This is another way of getting at the question of which values characterize each sex. It may even be more meaningful than the results I looked at last week, which were based solely on professed preferences. This time, I’m also looking at behaviors--the college majors that were chosen and completed.

Here are the results for women, in descending order. Keep in mind that a score of 1.0 would mean a perfect correlation.

• Contribution to Society: 0.60
• Benefits: 0.17
• Location: 0.12
• Security: 0.11
• Responsibility: 0.05
• Challenge: –0.01
• Independence: –0.20
• Advancement: –0.53
• Salary: –0.59

Now, compare this to the very different order I found last week, looking only at expressed opinions:

• Benefits: 64%
• Security: 62%
• Challenge: 59%
• Independence: 59%
• Location: 59%
• Salary: 54%
• Contribution to Society: 53%
• Responsibility: 43%
• Advancement: 37%

Next, here are the correlations for men. These are actually the same as for the women but in the opposite order. That is, the minus signs change to plus signs and vice versa.

• Salary: 0.59
• Advancement: 0.53
• Independence: 0.20
• Challenge: 0.01
• Responsibility: –0.05
• Security: –0.11
• Location: –0.12
• Benefits: –0.17
• Contribution to Society: –0.60

Finally here are last week’s very different results for men, in descending order:

• Benefits: 65%
• Security: 61%
• Independence: 59%
• Salary: 58%
• Challenge: 55%
• Location: 48%
• Advancement: 45%
• Responsibility: 44%
• Contribution to Society: 38%

What strikes me is that when the two sexes are compared, the rankings I compiled last week (the percentages saying a value is very important) are a lot more similar than the rankings I compiled this week (the correlations). Specifically, the values associated with the majors that the men and women chose (and completed) adhere much more closely to the stereotypes of women as nurturers and men as strivers.

What explains these different findings? Last week, I created only two pools of graduates, one male and one female. Taking the averages of the two pools, as I did last week, washed out a lot of the differences that were present among subpopulations within each pool. But I was able to tease out some of these hidden differences by breaking up these large pools into smaller sets based on their past behaviors--the majors that they chose and completed. In addition, using correlations allowed me to detect the values profiles of those graduates who had gravitated toward majors that were dominated by one sex or the other. These grads may have been a minority--their opinions are hard to detect when you look at overall averages--but they show that the sex-stereotypical values profiles remain a reality for a significant group of people.

The larger lesson to take away is that the simple percentages one sees in many survey results (for example, what percentage of voters is currently backing a certain candidate) can disguise some information that would valuable to know about subpopulations in the sample.