This article has been published in the November issue of Physics World.
The Research Assessment Exercise (RAE),which has been used to measure the quality of UK research for more than 20 years, has now morphed into the equally clunky sounding Research Excellence Framework (REF). Like its predecessor, the REF is a means for “judging” research in university departments and “scoring” them in ways that have significant consequences for future funding.
The amount of money each university department received in the RAE depended on how many researchers were submitted to the exercise and the quality of their work. The RAE has, however, raised a storm of protest around equality issues. In 2004 the Association of University Teachers, which was then the main trade union representing academic staff, said the 2001 RAE was blighted by “institutionalized sexism”. That view was backed by an analysis carried out by the Higher Education Funding Council for England (HEFCE), which revealed that around 64% of men, but only 46% of women, were submitted to the RAE.
The disadvantages for women were obvious given the propensity for institutions to associate inclusion in the RAE as a criterion for promotion. In the 2001 exercise no allowance was made for time lost on maternity or other long-term leave, leading to an apparent lack of productivity from some women who had been pregnant and taken maternity leave during the census period, hence their exclusion from the exercise.
For the next RAE in 2008 a dedicated attempt was made to eliminate this unfair bias. The rules allowed women to submit, subject to a case being made, fewer than the standard four “outputs” (papers, book etc) upon which quality was to be judged. Thus, a woman who had one or more periods of maternity leave could specify how long she had been on leave and justify a reduction in numbers of publications entered. This could be argued as amounting to “special pleading”. Furthermore, getting it wrong, so that the argument for reduction was not accepted, meant that the missing output counted as unclassified and not scored, so it remained a risk for an institution to permit someone to do this. Having sat on the 2008 RAE physics panel, I recall very few people submitted fewer than four outputs in physics.
It is clear that the newly configured REF is now heading the same way. A consultation period to allow people to comment on draft criteria that the panels will use to judge researchers’ quality has just closed, so we will not know until early 2012 what changes the funding councils may make in light of these responses. But maternity issues have again stirred up strong feelings. Given the sensitivities raised in the previous exercises, it is perhaps surprising that the initial recommendations regarding maternity leave in the REF consultation document appear illjudged and somewhat arbitrary.
With the results of the REF to be published in December 2014, HEFCE has laudably tried to remove the need for special pleading by introducing an explicit entitlement for women taking maternity leave to reduce the number of outputs they need to submit. It is not obvious there is any simple allowance made for men taking additional paternity leave, which itself raises equality issues. Bizarrely, however, the draft rules say that women will only automatically qualify for the reduction in outputs from four to three if they have taken a minimum of 14 months off work (with further reductions possible for longer periods off but increasing time off required pro rata for part-time working). The problem with this is that a woman who has taken only a full year off after childbirth – as is typically the case – will see no allowance made at all under this recommendation.
The logic of this 14 month cut-off is not obvious. Indeed, I rather suspect it is unclear to HEFCE too. In the consultation document it says: “In discussions with REF panels, a possible alternative approach was identified taking account of pregnancy and maternity: that staff who had periods of maternity leave during the assessment period may reduce the number of outputs by one for each discrete period of maternity leave, without penalty in the assessment.” This statement makes much more sense and removes the special pleading “stigma” from the last RAE. It takes into account that being pregnant and giving birth do impact very substantially on one’s productivity, albeit temporarily.
To be fair to HEFCE there is still the possibility of using special pleading. They say “complex circumstances” may arise – and not just around pregnancy but other identified situations such as long-term ill-health. It means that an institution can write up the details of “constraints related to pregnancy or maternity” for an external panel to judge whether these count as sufficiently complex to justify a reduction in outputs.
This is better than nothing but again raises the spectre of a missing output being graded as unclassified if the panel rejects the statement. I am not sure how sympathetic this panel might be to repeated statements along the lines of “my brain went to mush during pregnancy and subsequently stayed mush for 18+ months because I was sleep-deprived and focused on my new baby”, which is probably a true reflection of how I myself felt.
As my former mentor, the Cambridge condensed-matter theorist Sir Sam Edwards, once wonderfully and encouragingly said to me at a crucial stage of my career: “Intelligent women should have babies”. So, let us hope HEFCE sees the light and opts for the administratively simple and logical step of saying one pregnancy equals one fewer output needed. It is due to release its final criteria for REF in early 2012 and I urge it to adopt this approach.
Since this article was written and went to proof there have been developments. Clearly HEFCE received many responses to their Consultation Document about this matter. I, for instance, wrote along the same lines as the above in my capacity as chair of the Athena Forum. Every person I spoke to who had any interest in these matters seemed to be involved in some way or another in submitting a response to the consultation. An article in the THE also made many of the same points as this (at that time unpublished) piece for Physics World. It is therefore very satisfying that HEFCE announced last week that on this front they have made an early decision stating
An overwhelming majority of respondents to the consultation supported the proposal that researchers may reduce the number of outputs in a submission by one, for each period of maternity leave taken during the REF period. In light of the response, the funding bodies have decided that this approach will be implemented across all panels.
So, victory for common sense on this front. Or, as one of my correspondents put it, ‘Hurrah!’