Single-sex education: good or bad?

Single-sex education sounds like a rather archaic practice to me but there are people out there who claim it may have great benefits.

Booth, Cardona-Sosa and Nolen (2013) conducted an experiment at a co-educational university in the UK to test the effects of single-sex classes.

The researchers divided students (about 1000) into all-female, all-male and co-ed groups of roughly 30 people who had a one-hour class in these groups each week. The experiment started in the freshman year of an economics program and ended at the end of the year. Grades and behavior in the subsequent sophomore year were also documented.

So what were the researchers looking in their experiments? What effects should single-sex education have? The authors identified four conjectures to test.

Higher pass rates for all-female classes. This hypothesis says that females will have higher pass rates if they are placed in an all-female class. One reason for this might be that females do not do well in male-oriented fields (like economics) because of social pressure. Therefore, putting them in all-female classes will eliminate these stereotypes and they’ll have better grades. (This seems to implicitly assume that only males put social pressure on females who are successful in male-oriented fields, other females do not.)

Result. Indeed all-female classes had better pass rates than females in co-ed classes. Their pass rates were 6.7 percentage points higher. Controlling for the teacher, the difference increases to 7.6 percentage points. And those students who are at the greatest risk of failing (the 15th quantile in this study) had 22% higher scores if they were in an all-female class.

More technical courses chosen. This hypothesis says that females from all-female classes will tend to choose more technical courses in their second year; an effect that has been found in Austrian high schools for instance.

Result. This is negative. The authors did not find any evidence for this. Females from all-female classes were not more likely to take technical courses in their second year.

Long-term effects of single-sex education. Single-sex education might give more confidence to females by destroying the stereotype and psychological threats. It might also create new behavioral habits. All this can lead to the perpetuation of its effects into the sophomore year for instance.

Result. Indeed, females who attended single-sex classes in their freshman year had 10% higher scores in their sophomore year, when everything was back to normal again. It seems the positive effects of single-sex education stuck, at least for a year.

Different distribution of abilities. There is a theory out there that says that if the variance of abilities in a class is smaller, then the teacher can just teach to the median student and that will benefit the whole class. This will lead to better time utilization and better grades. If all-female classes had lower variance in say IQ, then this theory could potentially explain the positive effects of single-sex education (and it’d have nothing to do with gender or stereotypes).

Result. The authors had the students complete IQ tests at the beginning of the year. They found that the standard deviation of all-male classes was 2.86, co-ed 3.00 and all-female 3.21. While statistically these values are equivalent, even if we treat them as different all-female variability is actually higher than co-ed or all-male. Therefore, the distribution of abilities had nothing to do with the results.

Some more issues remain:

Did the teachers teach differently (read: better) in all-female classes? The answer appears to be no. The teachers were evaluated by students and all-female classes did not find their teachers significantly better than did co-ed or all-male groups.

Did single-sex education merely induce behavioral changes that can be achieved in other ways? Indeed, single-sex education changed students’ behavior. Students in all-female classes were 9.3 percentage points more likely to complete optional coursework, and 6 percentage points more likely to attend classes.

Out of all students, those who completed optional coursework were 43 percentage points more likely to pass the exams. Thus roughly 3.9 percentage points (=.43*.093) of the 6.7 percentage point difference in passing rates between all-female and co-ed classes can be explained by this behavioral change.

Similarly, attendance led to 28 percentage point higher passing rates, and thus it can explain 1.7 percentage points (=.28*.06) of the passing rate difference.

Altogether, these two behavioral changes can explain 5.6 percentage points of the 6.7 percentage point difference. This estimate, however, are likely to be biased upwards because more hard-working/intelligent students are more likely to both pass and attend classes/do optional coursework.

Do behavioral changes explain second-year differences? The answer is no. In the second-year, attendance and optional coursework differences are small between the two groups, yet as seen above performance differences remain.

We can conclude therefore that single-sex education does have some positive effects on females while leaving males’ performance unchanged. It appears a great deal of the differences might come from the behavioral changes that single-sex education induces. But single-sex education still has significant direct and long-term effects, which appear to be independent of behavioral changes.

Note that this study was conducted in a university where the students only had one hour of single-sex education per week. So the experiment is not an endorsment of full-out single-sex education, but it rather emphasizes this concept’s benefits in a more limited setting.

As a final remark, let me note that in my opinion the effects of single-sex education might be larger in primary and secondary education when children and their brains are much more easily sculpted by the environment. In summary, it seems that sometimes practices thought to be “archaic” can still be quite beneficial.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s