APPENDIX 1
THE SAMPLES

In both inquiries the aim was to obtain a representative sample of the adult civilian population of Great Britain. The methods of sampling used were not however the same, and in the March inquiry an adult was defined as a person who has left school and in the October inquiry as a person aged sixteen or over.

The method used to select the areas in which interviews were to be made was the same in both inquiries. The total number of interviews to be made was first divided between the twelve Civil Defence regions on the basis of the Registrar General’s population figures. Then within each region the correct proportion of interviews was allotted to urban and rural administrative districts. The rural interviews were allotted to one or more districts, chosen at random, and the urban interviews were distributed by grouping the towns of the region by size, allocating the correct proportion to each group and then choosing a town or towns to represent that group.

In the inquiry made in October names and addresses were drawn at equal intervals from the adult portion of the Maintenance Registers in the districts selected. The people so chosen were then visited at their homes and interviewed. If the interviewer could not make contact with the person selected after at least two re-calls, a substitute was taken from a shorter list of names drawn up in a similar way. Evidence collected suggested that the representativeness of the sample was not affected by substitution.

In the inquiry made in March the method of quota sampling was used. Representative numbers of men and women in different broad occupation groups were selected in each of the Civil Defence regions, the selection of individuals within the groups being made by the investigator so far as possible by an improvised method based on some random principle e.g. selecting workers at regular intervals from a list at a factory, or housewives at houses selected at regular intervals in a street. Workers were for the most part interviewed at their places of work and housewives and the retired and unoccupied at their homes.

In the March sample 2,000 interviews were set and 2,014 were obtained. In the October sample 3260 were set and 3,137 obtained.

Some analyses of the two samples are shown below:

It will be seen that the older age groups are somewhat under-represented in the March sample in which the method of quota sampling was used. There is however good agreement between the age distribution of the random sample in October and that of the population.