This blog was kindly contributed by Rhys Williams. Rhys is an Economic Consultant at London Economics (@LE_Education) specialising in the Education, Labour Markets and Competition sectors. This blog post is based on a Working Paper, available here. Rhys can be found on Twitter @rys_williams.
The financial effects of the COVID-19 crisis have already had a significant impact on the UK higher education sector. With a decline in student intake predicted for the 2020/21 cohort, the sector is likely to be further impacted for years to come. To maintain operations, substantial savings will need to be made across the entire sector.
One method under consideration is to reduce the costs of teaching staff through increasing the use of zero-hour, variable hour and per-session contracts. Teaching staff employed on such contracts are considered casualised workers and are often cheaper to employ and easier to dismiss than permanent teaching staff.
Existing evidence has documented the increasing prevalence of casual contracts over time; however, against the potential cost savings associated with casualised contracts, other than the direct impact on employees, are there wider consequences for higher education institutions from loading teaching on casualised staff?
To address this, using survey information merged for the first time, we looked at the relationship between casualisation and student satisfaction.
What data did we use?
The existing literature provides scant evidence relating to the effects of casualised staff on student satisfaction. To address this research gap, we combined data from the 2014/15 Student Academic Experience Survey produced by HEPI and the Higher Education Academy (which is now part of Advance HE) with the University and College Union’s bespoke survey of higher education staff. This enabled us to create a dataset for student satisfaction and the proportion of teaching delivered by staff on casualised contracts.
This dataset combines information on the level of satisfaction that students report of their time in higher education, ranging from very satisfied to not at all satisfied (4-point scale) alongside the estimate of the proportion of weekly undergraduate teaching time delivered by teaching staff on casual contracts.
The casual teaching load was estimated from the responses provided by academics to the survey. Both datasets relate to the 2014/15 academic year (the focus of our analysis) and the analysis was undertaken at university-subject pair level (the fundamental unit of analysis). To ensure that the results were not dependent on just a few responses in the UCU survey, we restricted the sample to university-subject pairs with at least 5 responses related to the proportion of teaching information (with robustness checks on minimum 3 responses yielding similar findings). Overall, there were 8,145 university-subject pairs, covering 118 different universities and 15 subject areas, with a sufficient number of responses to allow for subsequent analysis.
What was the modelling approach?
To undertake the econometric analysis, we used ordered logistic regressions and ran a number of different model specifications. All the models control for a variety of factors which might affect both student satisfaction and/or the extent of staff casualisation.
These control variables include the total number of students and staff at the institution, the proportion of staff on fixed-term contracts and region.
We also attempted to incorporate university and subject fixed effects but found these factors are statistically insignificant (both individually and jointly) and so we prefer models which do not include these fixed effects.
What was the main result?
Increasing the proportion of teaching delivered by casual staff led to a reduction in student satisfaction in UK universities.
Controlling for a range of other factors, our key finding is that a five percentage point increase in the proportion of teaching delivered by staff employed on a casual contract leads to a 0.58 percentage point reduction in the probability that a student will be satisfied with the university experience.
What are the implications?
This suggests that students are less satisfied when there is a higher proportion of teaching delivered by staff on casual contracts, and that universities might be risking a reduction in student satisfaction levels by loading up teaching on staff with casualised contracts. This does not seem like a large impact. However, it is important to remember how clustered student satisfaction scores are and how student satisfaction scores (such as the National Student Survey) play a key role in third-party rankings – and ultimately prospective student demand.
What else did we find?
Somewhat surprisingly, once we controlled for these other university characteristics, we did not find an effect of a university’s overall financial position on student satisfaction. However, this might be a result of financial information only being available at the university-level rather than the subject-university level considered in our study.
We also found no statistically significant effect of membership of the Russell Group on student satisfaction once the total number of staff and students at the institution had been controlled for. This suggests the effect of being a Russell Group university is in some respects a proxy for university size.
Going back to the main result, our dataset does not allow us to explore directly why increasing staff casualisation results in reduced satisfaction, but there are a number of possible hypotheses for why this might be.
- Firstly, students, most of whom paid around £9,000 in tuition fees each year (in England in the year in question), may expect to be taught by faculty members employed on permanent contracts.
- It might also be the case that casual teaching staff have significantly less time available to dedicate to preparing class materials, providing feedback and assessment and delivering out-of-class contact time, all of which are significant determinants of overall student satisfaction levels.
- A third potential explanation is that casual staff may have less experience and receive inadequate training from their institution, leading to lower levels of quality which impact student satisfaction.
These are just a few potential explanations which, unfortunately, we are unable to currently test for.
What is clear is that universities face a difficult balancing act, between cutting costs while maintaining high levels of student satisfaction. We provide evidence that the recent approach of generating cost reductions in universities through the increased dependency on casualised staff leads to lower student satisfaction. As this approach may have knock-on effects in terms of reputational damage and declines in future student demand, it might be a false economy.
Essentially, cutting costs at the expense of student satisfaction only offers short-term financial relief, but does not offer a long-term strategy for institutional sustainability.
Acknowledgements: This blog is based on research that was in partial fulfilment of a PGC qualification in 2019-20. The Working Paper is available here. I would like to thank Matt Waddup at the University and College Union and Nick Hillman from the Higher Education Policy Institute for providing access to the data in the first instance. I am grateful to Gavan Conlon and Maike Halterbeck for providing the idea and all their help and guidance throughout, and also to Alice Battiston, Rohit Ladher, Edgar Liberis and Deborah Williams for helpful comments and suggestions.
Update: The paper is now available here.