When higher education interventions don’t work

Sharing examples of what doesn’t work through large-scale, experimental testing is hugely valuable

I am currently supporting a higher education project in Tunisia and came across an interesting World Bank study considered to be the first of its kind. Final year undergraduates were given the opportunity to graduate with a business plan instead of following the standard curriculum, and were offered 120 hours of training that included ‘most of the components that are considered best-practice for entrepreneurship education’. The optional entrepreneurship track started in 2009/10 and has been running since.

In the first published analysis, short-term impacts were studied:

the entrepreneurship track was effective in increasing self-employment among applicants, but that the effects are small in absolute terms. In addition, the employment rate among participants remains unchanged, pointing to a partial substitution from wage employment to self-employment. The evidence shows that the program fostered business skills, expanded networks, and affected a range of behavioural skills. Participation in the entrepreneurship track also heightened graduates’ optimism toward the future shortly after the Tunisian revolution.

A second paper, published in 2019, examined the medium-term impact using the same cohort:

The medium-term results show that the impacts of entrepreneurship education were short-lived. There are no sustained impacts on self-employment or employment outcomes four years after graduation. There are no lasting effects on latent entrepreneurship either, and the short-term increase in optimism also receded… the lack of medium-term impact holds across sub-groups based on gender, family wealth, skills or social capital.

There are several possible lessons to draw, beyond the clear difficulty of achieving lasting impact. The first is that integrating enterprise education alongside existing curricula, rather than a separate stream, could be an effective alternative. The second, as the second study suggests, is that other limitations are a bigger constraint than the nature of the training, especially accessing capital (there is evidence from Nigeria, cited in the paper, of monetary grants having long-term, positive impact). The third is the importance of continued coaching, training and mentoring beyond the initial period of study.

But what really struck me was how relatively unusual it is to come across randomised control trials of interventions in higher education (and especially published studies of those where the intervention did not work). Fields such as medicine abound with multi-year trials. Primary education has also seen its fair share – this years Nobel prize recognised the work of Abhijit Banerjee, Esther Duflo and Michael Kremer popularised in the excellent book, Poor Economics. There are plenty of large-scale evaluations and analyses of higher education, particularly around student outcomes, but I struggle to think of large-scale, experimental interventions. My hunch is that, as recognition of the role of higher education in development and social change has increased, so too will demand for randomised control trials within the field.

Failures wanted

Tunis, Tunisia

The Tunisian study is helpful as it shows that a ‘common sense’ prescription (give students business and entrepreneurial skills instead of writing an academic thesis as part of their degree) to a commonly-perceived problem (unemployable graduates) simply did not work. I’d love to see more such studies. Sharing examples of what doesn’t work through large-scale, rigorous testing can be hugely valuable, albeit with the caveat that the results may not always be generalisable to other contexts. As I see it, several things need to be in place:

  1. The basic parameters of an academic study: a control group who do not participate in the intervention, careful analysis of the context and environment, benchmarking and continued evaluation, etc.
  2. An acceptance that the intervention may fail. This is why the World Bank is perhaps better placed to fund such a study than the Tunisian government, who would be less willing or able to share widely the outcomes if the project failed, or to experiment with public funds.
  3. A longer-term (multi-year) perspective with no expectation of clear answers in the short term.
  4. A process of freely disseminating the findings and sharing what has (not) worked.
  5. A sufficiently big budget to launch and maintain a long-term effort, and to provide the capacity for effective experimentation, iteration and evaluation. The World Bank’s Tunisia Tertiary Education for Employability Project runs for over five years and commits 70 million USD.

Crucially, these conditions combined distinguish a rigorous, experimental study from a public policy intervention. If you know of any other experimental, evidence-based studies in higher education – especially those that have been deemed to have not worked – please let me know by email or in the comments below.

Photos of Tunis, Tunisia from Unsplash. Credits: main image, article image.

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 )

Google photo

You are commenting using your Google 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 )

Connecting to %s