12.2 a There is no one correct answer to this question as the number of variables will depend on the method used to code these descriptive data. If you choose the multiple-response method, five variables will be generated. If the multiple- dichotomy method is used, the number of variables will depend on the number of different responses. b Your first priority is to decide on the level of detail of your intended analyses. Your coding scheme should, if possible, be based on an existing coding scheme. If this is of insufficient detail then it should be designed to be compatible to allow comparisons. To design the coding scheme you need to take the responses from the first 50–100 cases and establish broad groupings. These can be subdivided into increasingly specific subgroups until the detail is sufficient for the intended analysis. Codes can then be allocated to these subgroups. If you ensure that similar responses receive adjacent codes, this will make any subsequent grouping easier. The actual responses that correspond to each code should be noted in a codebook. Codes should be allocated to data on the data collection form in the ‘For office use’ box. These codes need to include missing data, such as when four or fewer ‘things’ have been mentioned.