Professor Susan Golombok:

Yes. The numbers are quite small. It really comes back to the point I made about the need for replication. Of course, there are different ways of studying families. Our approach at the Centre for Family Research, Cambridge, is to do very in-depth longitudinal studies and obtain very in-depth data using many different methods and informants in the family. Other researchers might survey larger numbers of people. We need all these approaches.

Given the in-depth nature and reliability of the measures we used, we would certainly have picked up any moderate differences had they existed. You cannot really make blanket statements about what constitutes a large or small sample; however, for family studies that are as in-depth as ours, you would usually and ideally aim for about 50 families in each group. As it happened with the surrogacy ones, we managed to recruit only 42 within our sampling timeframe. However, the 42 were very representative because we were fortunate in that we were able to collaborate with the UK’s Office for National Statistics, which is where parental orders are registered when the intended parents become the legal parents of the child. On our behalf, the Office for National Statistics approached the families and we had a very high response rate. In some ways, the representativeness of the sample is more important than the number because, with volunteers, you might get only families with an axe to grind or families who are doing particularly well. Those who are not doing well might not want to take part in research. Representativeness is one big plus of the study but I completely acknowledge it is but one study.

The results have been positive but it would be helpful to have more research focusing on larger samples, though then the methods would necessarily be less in-depth.