After reading a TechCrunch article on how Stanford is closing the gender gap in computer science I got curious. How is University of Ljubljana doing?
A friendly professor Neza Mramor-Kosta from Faculty of Computer Science helped me to obtain some raw data. It seems no one has requested such detailed information before. Luckily faculty administration was very helpful. Today’s post is trying to analyze some of that data.
Let’s start with the total student body. It has about 1500 students each year. When combining everything together (professional degree, B.Sc., M.Sc. and PhD. for all CS and CS+Math programs we get this graph of women as part of the student body:
This gives us a top-level view. Now if we want to look at shorter term trends we might want to look at first academic year (professional degree + B.Sc.) in CS and CS+Math:
What seems to have happened is that in 2009 the positive trend has been broken. What could have caused that? Let’s compare this with the second academic year data:
Indeed, something changed in 2009!
Ratio of women from then on stays the same from first to the next academic year which was not the case before. Have professors been discriminating female students before 2009 and then stopped? Did economic crisis make women more diligent students, but not men?
Probably not. I have the following set of hypothesis:
- Virtual enrollment skews the data. Being enrolled provides state sponsored student benefits. Assuming virtual enrollment to be gender-neutral, this has moved the percentage of women up. I am not sure if virtual enrollment somehow got under control in 2009? Or how it affects the 1st year data.
- Bologna reform CS program started in 2009. This led to a lesser number of women enrolling into CS (maybe because other programs were more alluring)
- New Bologna program was better at teaching women students and led to higher percentage of them progressing to the next year.
However 2009 was the first year of Bologna reform freshmen programs, so how could they already have effect on the second academic year data for 2009? That’s simple – since there was no more old first year program in 2009, so the rules for progress of 2008 cohort got relaxed. To be sure, let’s plot men and women progression from the first to the second academic year. The data is unfortunately not very precise, since we are mixing first time and second time students of the same academic year.
Wow, Faculty of Computer Sciences has been seriously tightening up their criteria for progression – going from 80% to about 50%. The good news is that women seem to be getting ahead in the last few years – their progression matches that of men. The data for 2011 could be a reason for optimism. This might mean that the new type of teaching fits them better.
Prof. Mramor-Kosta commented on the draft of this post that relationship between lectures, homework and exam has changed dramatically with Bologna reform – students are more involved during the year and almost all serious students pass the final exams. Interestingly the downward trend is very obvious for men, while there’s a higher variability for women.
Now let’s look at data about graduations (B.Sc. and professional degree) for each year:
Well this data isn’t very encouraging, basically there is no trend. And since we’ve seen above that the percentage of women in second academic year is increasing constantly, this is really worrisome. It means women drop in second year or later.
To be sure, let’s take a look at percentages of women in the first year, the second year and the percentage of graduates. We’ll use 2009-2012 data since it seems to be less biased :
Yes, this confirms what we’re seeing before – women aren’t giving up computer science after a year, they drop somewhere else. I wonder if this the case just for computer science or is this the norm in higher education in Slovenia.
Prof. Mramor-Kosta observed that some pre-Bologna students who have dropped out have decided to finally get the degree since it won’t be available any more. These are older back-to-schoolers and women are less represented in the group due to having families or for other reasons. So the situation with new wave of regular students might actually be different than these charts show.
Naturally there’s more data to be crunched. How is Math+CS doing versus pure CS? what’s happening with postdoc studies? What really happened between 2008 and 2009? What if we look just in new Bologna program students?
If you are interested to dive in, here’s the raw data and analysis on Google Docs. There you can also find pure raw data as it was kindly provided by the administration of the Faculty of Computer Science and Informatics at University of Ljubljana.
If you find something interesting in the data, do let me know! I hope to write more posts on this topic not necessarily only about this data set. So please share any thoughts or insights in comments!