Last week I finally got to welcome Professor Sir Roger Jowell to the Agency to present the third lecture in my Chief Scientist’s lecture series. I say ‘finally’ as we had to postpone his original lecture from February because it coincided with the snow storm that all but closed the capital – and of course gave many of my colleagues and their walking boots some excellent exercise as they battled their way in to Aviation House.
The tales of the ‘great snow day’ as it is known in London, and chuckled over in those parts of the country more used to heavy snow, would no doubt provide meat for social scientists, so it is appropriate that more than 100 of us in London, Aberdeen and Belfast heard Roger talk about the work of social science in measuring changes in attitudes and behaviour. It has been known for some natural scientists to be sceptical about the value of measuring what people say and do, but any suggestion that social scientists are ‘fluffy’ was soundly countered by his identification of a range of disciplines where the social and natural sciences overlap – including psychology, epidemiology, maths and, of course, statistics.
Roger was very complimentary about the work of the Agency – which he described as having had a virtuous effect on habits and behaviours in relation to food. Furthermore, he praised our commitment to the use of social science evidence in policy making in the long term. He explained that by considering research into behaviours and attitudes in tandem, social science can help to explain the complexity of people’s interactions – and it is crucial that we further understand both what these are and why they are held or performed in order for the Agency to help effect positive changes in behaviour. Although, of course, an individual’s attitude is shaped by a whole range of factors, including expectations, values, beliefs and opinions.
We spent some time considering bias, and how this can be corrected for and mitigated. The audience was also given a brief introduction to the pros and cons of different types of data collection. Interestingly, recent developments suggest that a telephone approach followed by providing the prospective participant with a choice of how they would prefer to receive the survey is gaining popularity, because it is more cost effective – while it is likely that there are further developments to be made, this is an interesting consideration.
Roger went on to explain three of the different types of change that social science research can help to unravel: cohort (generational) effects such as the decrease in religiosity in the past 20 years; life-cycle effects, such as sympathy towards the provision of a more generous state pension because the population as a whole is ageing; and period effects, such as how attitudes to abortion have become more permissive with time.
Finally, Roger shared some amusing anecdotes about his work on the development of the national sex survey to illustrate different types of internal and external validation. Survey developers expected an under-reporting of the number of visits to a sexual transmitted disease clinic in a specific timeframe – comparing reported answers with clinic data where researchers were able to determine that there was not significant under-reporting (external validation). Conversely, researchers were concerned about over-reporting when questioning individuals about the number of times that they had sex. In this case, data provided by men and women matched closely when reporting for the last week but this was not when the recall period was extended to a month or three months (internal validation).
Now, how tall was that snowman you told me you saw? And how far did you have to walk…?