March 2, 2008
The New Gender Gap in Education
The New Gender Gap in Education-Becker
Until the mid 1960's, female high school graduates were less likely than male graduates to go to college, and female college students were far more likely to drop out than were male students. The direct reason for this difference was that many younger women married and then dropped out of school - mainly to start having families. Perhaps a more basic reason for this gender difference in education was that women did not participate in the labor force so much in those days, and hence many women did not believe a college education was useful.
All this has changed radically since 1970. Female high school graduates are now no less likely to enter college than are male graduates, and a much larger fraction of girls than boys finish high school. These two facts imply that considerably more women enter college than men. The fraction of college students who are female is further increased by the greater propensity of women who enter college to finish and graduate. About 57% of American college students are women, and they constitute about 60% of those who graduates. Similar trends toward making women a majority of college students apply to many European countries, and some Asian countries as well.
What explains this reversal from under representation of women in college to over representation ((see the related discussion in my blog entry for July 17, 2006)? One important cause is that marriage and child-rearing exert a much weaker pull out of school for women than in the past since women marry and start families at much later ages than 40 years ago. This increase in age at marriage is related to the decline in birth rates, and to the increased time that women want to spend working rather than caring for children and running households.
A college education is more attractive to women who spend greater time in the labor force since going to college significantly raises earnings of women as well as men. The financial attractiveness of a college education has grown sharply for both sexes since the 1970's because of the large rise in the earnings premium from a college education. The average hourly earnings of college-educated persons grew from about 40 % higher than that of high school graduates in 1980 to about 80% higher in recent years. This trend toward a much higher college education premium is also found in many other countries as well as the United States.
Although quantitative evidence on non-earnings benefits are more limited, the advantages of a college education in improving health, raising children, managing financial assets, responding to adversity, and in other areas of life have also grown along with the growth in the college earnings premium. This implies a widening advantage of a college education even to women who spend a significant portion of their time raising children and managing a household. In addition, the propensity of college-educated women to be married has increased a lot relative to the marital rates of women with less education, so that graduating college no longer significantly reduces a woman's chances of marriage.
Since these forces pushing women toward a college education have been strong during the past several decades, it is no surprise that a much larger fraction of young women now enter and complete college than a half century ago. This does not, however, fully explain why women are more likely than men to be in college since most of these forces have been just as powerful for men, and college-educated men still spend a larger fraction of their time working in the labor force than do college-educated women.
An important reason why women not only closed the education gap with men, but also changed the direction of that gap, relates, I believe, to the better performance of women in school. The average grades of women at every education level exceed the average grades of men, while the variation around the average is larger for men. Persons with low grades find school unpleasant since their teachers criticize them, and they come to believe that they are failures. Since many more boys than girls in high school have low grades because both average grades are lower and the variance in grades is greater for boys, more boys than girls find high school unpleasant and drop out before graduating. Dropouts truncate the grade distributions of graduates at the lower end, so that average grades of boys who graduate high school are closer to the average grades of girls who graduate than are the averages for all boys and girls in high school.
The same process operates at the college level. Men have much lower grades in college and find the experience less pleasant, so they drop out of college in much larger numbers than women, and are much less likely to graduate. That many more men than in the past continue on to college after high school indicates that they are aware of the rise in financial and other benefits from college. That they drop out of college in large numbers presumably indicates that they are either discouraged by their low grades, or they just do not like being students.
Why women at all ages do better in school than men is not so easily understood. It is unlikely that women do better mainly because they expect to remain in school longer- this is causation from remaining in school longer to better grades- since women had better average grades than men even when they were more likely to drop out of school. One line of explanation argues that women are more diligent students, less rebellious, and more docile students. Whatever the explanation for the remarkable shift in college attendance rates of men and women during the past 40 years, this shift is likely to have major implications for future changes in the gender gap in average earnings, the fraction of heads of business that are women, and other measures of gender differences in achievement.
The New Gender Gap in Education--Posner's Comment
It is no surprise that female enrollment in college has increased over the last half century. The later age of marriage and childbearing and the greatly increased job opportunities of women explain the trend. Another factor, stressed by Becker in his pathbreaking economic analysis of the family, is increased emphasis on quality rather than quantity of children; parental education is an important factor in the quality of children.
The fact that women tend on average to get better grades in college helps to explain their lower dropout rate, but this is nothing new; even in the era when women dropped out of college to marry and have children, they had higher grades than men. That women are better students than men is pretty much a constant--and a puzzle.
When one observes members of one group outperforming another in a competitive environment in which, therefore, substitution of inputs is possible, a possible explanation is discrimination against the members of the superior group. If a college wants to have the the best students it can attract, and the women attending the college have better grades than the men who attend it, something is wrong--the school could increase the quality of its student body by admitting more women and fewer men. That it does not do so may be because it values other gender-dependent factors--for example, female students may prefer a lower ratio of female to male students than a purely meritocratic admissions policy would produce, and this preference may influence the college's admissions decisions. But this is unlikely to be a good explanation for the superior female academic performance today. The incentive to discriminate against female college applicants was much stronger in the old days, yet the female-male performance gap has not (so far as I can discover) diminished.
Women might outperform men academically because they worked harder, and they might work harder because they had more to gain from completing college successfully and doing so with high grades. But as Becker points out, since male participation in the labor force continues (and probably will continue) to exceed that of women, and since there is a large wage premium for college graduates, men actually have more to gain from completing college than women do. Yet not only do they drop out at a higher rate; but male college enrollment has not increased nearly as rapidly as female college enrollment has. Women are not just catching up with men on the educational front; they are becoming better educated than men.
So there are two puzzles: why women get better grades than men, and why men have a lower elasticity of response to the effect of education on earnings than women do. At this stage of our knowledge, the answers to these questions must be highly speculative; what follows, then, is guesswork.
The first question is, though, I think, a little easier than the second. From the standpoint of most teachers, right up to and including the level of teachers of college undergraduates, the ideal student is well behaved, unaggressive, docile, patient, meticulous, and empathetic in the sense of intuiting the response to the teacher that is most likely to please the teacher. Those are traits less characteristic of boys than of girls. Moreover, there is more variance in IQ among boys than girls--to exaggerate, more morons and more geniuses--and both the morons and the geniuses are difficult for most teachers, the morons for obvious reasons, the geniuses because they are easily bored in a class geared to the comprehension of the average student. So girls are easier to teach, and so are "rewarded" (not deliberately) with higher average grades.
Nothing in the suggested answers to the first question, however, can explain why males should be less responsive to the growing value of a college education than females. One possibility is that there is nothing more that men can do to improve their academic performance, given genetic limitations. Notice the curious fact that the more men in the lower tail of the male IQ distribution drop out at some stage in their academic career, the higher the average grades of the men who remain school should be; the "genius" tail pulls up the average, while the "moron" tail, being depleted because of dropouts, pulls it down less than it would if the students in that tail did not drop out disproportionately and thus cease to figure in the determination of grades. Maybe the "genius" tail, because of the publicity that its members attract, has obscured the fact that women may on average be more intelligent, or at least have innately a suite of qualities more supportive of academic perfornance, than men. The key is "innately." If aggressiveness and other psychological or cognitive qualities that inhibit male academic performance are innate, men may have maxed out long ago, while women did not reach their peak then because of factors extraneous to ability, such as lack of demand for women in high-skilled jobs, until recently.
Another possibility is that the decline of the conventional "patriarchal" family since the 1960s has been harder on boys than on girls. Because of rampant divorce and illegitimacy, a boy's biological father is less likely to be a continuous presence during the boy's formative years, and this is only one factor in what appears to be a decline in the disciplining of children. If docility is as I have suggested a factor in academic performance, a decline in discipline is more likely to harm the academic performance of boys than of girls because the former need more discipline to instill docility in them. It is difficult to test this hypothesis empirically, however, because grade inflation bedevils any effort to use changes in average grades over time as a measure of the trend in academic performance.
But, to repeat, these suggested answer to the puzzle of the gender education gap are highly speculative--a stimulus (I hope) to further thought, not the end of the inquiry.