Statistical skills are in short supply in social research, according to both the Economic and Social Research Council and the European Social Fund ("Quantitative is qualitative" and "Europe, know thyself: social science solutions to the biggest problems", 15 April).
The poverty of statistical research in the final social work/social policy research assessment exercise was raised by Paul Wiles, the chair of the incoming research excellence framework panel, at a meeting of the Academy of Social Sciences (AcSS) in London on 14 January.
This weakness has been around for years, certainly since I chaired a multidisciplinary NHS group for social care research in 2003. Most practitioner researchers could never access relevant support by chartered statisticians.
Various public bodies have sometimes offered one-off sessions on statistical methods for fledgling researchers, but this occasional input is unsustainable unless their supervisors have a good grasp of quantitative analysis.
In the short term, there is only one solution. In each higher education institution, academics within the social sciences (especially PhD supervisors) must volunteer to learn new analytic skills. This sort of professional development is not alien to university life - for example, this grey-haired professor keeps taking instruction in IT skills.
Some disciplines, such as economics and psychology, have a stronger statistical tradition, and may have to seed experience into other disciplines locally.
One positive lesson from the AcSS is that social scientists in general can communicate well across our diverse traditions. In healthcare, I have researched successful mutual-aid groups (sorry - I mean successful "communities of learning"): perhaps the ESRC or ESF could promote sustained mutual learning for statistics across higher education?
Woody Caan AcSS, Professor of public health, Anglia Ruskin University.