After the annual release of the ABS Education and Work data last November articles appeared suggesting that female university enrolments might have been hit particularly hard by COVID-19.

I could think of a couple of plausible mechanisms. With children sent home from school and childcare restricted women might have given up study, at least temporarily, to look after their kids. The difficulty of doing required clinical placements and teaching rounds during COVID-19 workplace disruptions might have triggered deferrals, which would probably affect women more than men due to their their large majorities in health and education courses.

On the other hand, the quoted fall in female enrolments – 86,000 – was struggling to pass my ‘does it look right?’ test. And the source, Education and Work, which is conducted each May, has a history of rogue results. It is a sample survey of Australian residents rather than being derived directly from enrolment data. The further users drill down into Education and Work sub-categories – gender, type of enrolment, age group etc – the less reliable it gets (the ABS is upfront about this, and publishes relative standard errors).

Last November I used the TableBuilder version of Education and Work (expensive paywall; university staff can use it) to exclude international students. That caused the female decline in enrolments to go way and became a small increase, although with a narrowing of the gender gap. In 2019 Education and Work reported 1.5 female students for every 1 male student, which declined to 1.42 to 1 in 2020.

While I still had some doubts about the international student numbers, the gender differences could be partly explained by the timing of travel bans. Chinese students were caught by the initial China travel ban on 1 February, and that is a majority female market. The majority male Indian students arrived as usual, as the full travel ban was not imposed until 20 March, producing the gender difference for international students. We know now that Chinese students have been surprisingly willing to study online from their homes in China, but that is not picked up by a survey of Australian residents.

In a report released last night my former Grattan Institute colleagues focus the gender claim on women aged 25-44, who are most likely to have childcare responsibilities.

I used TableBuilder to narrow the analysis down to women with children aged 14 or less compared to women without children in that age bracket. And that does produce a result consistent with one of the original hypotheses – enrolments by women with young children fell by 19,000, while enrolments by women without young children grew by 28,000, bringing us to the overall increase in domestic female enrolments.

Given the known problems with Education and Work I’d want to see the overall student number patterns corroborated with enrolment data, but a plausible theory and consistent data does at least make it credible that COVID-19 affected a category of female enrolments.