One of the Foundation’s main initiatives currently under way is the Q-Step programme. The project aims to improve quantitative social science training at undergraduate level across 15 universities in the UK. We asked Sharon about the background to how and why Q-Step came about. ‘I’m always at pains to say that neither I nor the Foundation, think that all social science needs to be quantitative. But it had become clear over the years that there was a shortfall in good funding applications coming through to us that looked at the wide range of statistical evidence - and we thought that reflected a capacity issue more generally.’
She continues by explaining her personal view as to why the programme is aimed directly at social science undergraduates. ‘I was trained as a statistically adept sociologist and one of my personal worries was that a range of important problems could seemingly only be addressed by economists or statisticians, who might not be thinking of some of the broader questions that other social science disciplines would ask.’
Q-Step was originally going to be a Nuffield solo project, but in the period of its development, the potential of the project began to widen. ‘The worry was that leaving this intervention to doctoral level was too late as people had chosen their thesis topics and were hardwired into a particular pathway. The ESRC had initially agreed to put a relatively small amount of money in, as their remit is postgraduate and later, but when they began looking at their own grant applications and the international benchmarking studies they had done for sociology, geography and so on, the lack of quantitative skills had become a recurring theme. So, with the support of BIS, they took a bolder step and helped shape the programme too, including a clearer vision about pathways to postgraduate work.’
The ESRC broke with their usual focus on postgraduate level as it became evident how an investment in quantitative training at undergraduate level would bear fruit further down the academic line. As Sharon says, ‘this all came together in a way that unlocked significant resources from the ESRC and HEFCE who also had an interest in statistics and maths, including outside traditional STEM subjects. So this meant Q-Step went from being a £5 million initiative to what is now a £19.5 million programme over five years.’
Such a significant investment means the project aims to create a step change in the structure of social science degrees. ‘This isn’t just about producing a cohort of undergraduates with these skills. It is about changing the notion of what skills, tools and training undergraduates will need, both in general and for those pursuing advanced pathways. Quantitative skills should open up big and interesting questions and we wanted to ensure they aren’t just seen as a set of techniques. The point of understanding what the statistical method does in conjunction with what the problem is potentially opens up an exciting set of issues.’
During our conversation it became clear that Sharon thinks that improving quantitative skills ultimately needs to start earlier than the undergraduate years. As Sharon put it, ‘This is a problem that is worse in Britain because we specialise so early and relatively narrowly at A-level. Most other countries have a much broader secondary school curriculum. I suspect this will remain an issue - and after all, the debate goes back more than 100 years when the stereotype was that you either did the classics or you did science.’
Sharon appreciates the progress being made with the introduction of core maths into the school curriculum, but she also cautions that the changes ‘are not going to be a stable resting place. If we are going to have 50% of the age cohort going to university, there is inevitably going to be pressure to broaden the curriculum. The UK is an outlier here - in most other countries at least half of the age cohort takes maths all the way through secondary level. In the UK it is much lower at around 20%, with the Scots doing a bit better as they have Highers. I suspect core maths may still not be enough.’
This demonstrates how ambitious these efforts are to improve the UK’s capacity in quantitative skills, but Sharon is realistic about what this will mean. ‘We have to work with the world as it is, but this is partly about getting a debate going about whether all students could benefit from learning some quantitative skills. It’s not to say that everything should be quantitative, but it does help you understand a lot of complex social problems if you can understand the statistics behind them - as well as research design issues like social selection effects, and thinking analytically about causal pathways.’
Using statistics to explain the complex problems facing the country is something that is going to be used (or abused as the case may be) in next year’s upcoming general election. Sharon reiterated that the Foundation is committed to advocating that strong evidence is useful for an informed public debate. However, there are limits to how far this can go in an election cycle.
At the 2010 election, the Foundation funded an Institute for Fiscal Studies (IFS) study of the main parties’ spending plans. The IFS will be doing this again for the 2015 election with funding from Nuffield, but Sharon notes that economic analysis may be one of the more straightforward areas of political policy analysis.
The coalition have been cutting social spending in various ways over the course of the parliament and what effect this has had is not as easy to calculate. ‘You can look at the spare room subsidy as the government calls it - or the bedroom tax as others call it - and the fact that they call it different names alerts you to the fact that it is highly political. You can look at how many have downsized, what is stopping them from downsizing and at least try to get some facts on the table. You can then say “here is what we think the data show” - it doesn’t mean that this gives you a straight answer as to what an appropriate policy should be, even if you agreed about the aims.’
However as Sharon says, the interpretation of social policy data can change depending on the political perspective. ‘In any of these areas you will have data or evidence that can be read in more than one way. Sometimes you can say for certain that if you do x, then y will follow. But in other areas it’s not so clear cut. So the politicians will cherry pick, and then Full Fact will ask did they cherry pick? Of course, in some cases the UK Statistics Authority has a role in ensuring that the statistics are represented fairly. I’m not naïve enough to think that everybody is reading the detail of this, but it at least helps draw the attention to the fact that some evidence and data are more robust than others.’
The political cycle is a poor time period in which to really assess the success or failure of social policy. With this in mind, we asked Sharon what assets the UK needed to evaluate these questions in the years and decades ahead?
‘The last few years have seen an increase in the stability of the statistical data infrastructure for a lot of these social policy issues. But I do worry that there is still not enough money going in to large scale evaluations of some public policies. In terms of the long term data infrastructure, for the first time in a long time we have had a new birth cohort and all the birth cohorts on infrastructure funding or the possibility of that. The question for the post-election period is whether this should continue as part of the national data infrastructure with capital funding.’
Sharon did note that the ONS complement of datasets had been under pressure, but that the important ones have remained intact with the possible exception of the wealth and assets survey. ‘That is one we are concerned about - it is important that it continues because it will actually become an appreciating asset over time.’
The UK has a wealth of data being generated that paints a vibrant picture on the social state of the nation, as Sharon appreciates. ‘One of Britain’s strengths is that we have a very strong model for what that infrastructure should be in social science statistics. The ESRC’s very welcome investment in the administrative data centres should mean cross-sectional studies over time can now be done using this data. But all of this depends on having more human capacity for people to do this analysis, and that is likely to be a growing bottleneck unless we take steps to address it. I’m proud of Q-Step and I’m very hopeful about it, but I think it is only a small part of what we need.’