In this appendix I consider only those independent variables which either required some particular work to be done or which need a more detailed description of their definition. I shall consider the following variables:

1. (1) per capita income (youth unemployment sample only)

2. (2) labour market indicators (youth unemployment sample only)

3. (3) preferences (both samples)

4. (4) social classes (both samples).

Per capita income of the family

In the questionnaire presented to the unemployed young people, subjects were asked to give the monthly income of their family; this was precoded as a level variable of nine possible categories (in thousands of lire these were: 0–150, 151–300, 301–450, 451–600, 601–750, 751–1, 000, 1, 001–1, 500, 1, 501–2, 000, and 2,001 and above).

In the analysis I gave each case the value of the mid-point of the category (for the highest income class I took the arbitrary value of 3,000) and, in order to obtain the per capita (or per child) income of the family, I divided the mid-point value by the number of family members living in the same household at the time of the survey.