A sample of forty married people was decided upon (twenty wives and their twenty husbands). Although this was an arbitrary number, it was felt that many more would be unnecessary unless the aim had been to make the sample very much bigger in order to be able to generalise.

The sample characteristics decided in advance were that all should be legally married and not previously divorced; living with their spouses; and residing in Aberdeenshire or the city of Aberdeen. Other research and general knowledge of the area do not show those living in the north-east of Scotland as being markedly different from those living in other parts of provincial Britain.

A four-cell quota sample was chosen, so that there would be some variation, but not too widespread a variation, within the sample. The quotas set were that, first, half the sample should involve marriages in which the husband was at present in a manual occupation, and half should involve those in which he was in a non-manual occupation (there were to be no unemployed or retired husbands though they could be away from work through illness or on a training course). Second, half the sample should involve people married for five years or less, and half people married for fifteen years or more. Again this was in order to permit comparison and also to restrict the extent of variation by length of marriage.