ABSTRACT
The pilot study is a trial run on a small number of ‘guinea pigs’ who should be as like the true respondents as possible in age, intelligence, and social class. Most pre-testing effort should be put into areas thought likely to present problems and areas most crucial to the purposes of survey. Government surveys generally achieve a good response rate. It is usually found that ‘personalization’ of the letter of introduction is of no benefit, nor is use of a printed letterhead; neither the colour of paper nor the day received seem to make any difference. Organizing a survey is a non-trivial exercise. Several methods of data collection are possible and a pilot study will be necessary to check that the operation runs smoothly. At all stages of a survey, but particularly analysis of the data, considerable cooperation is required among: the sponsor, the interviewers and coders, the data preparation team, the statistician, and the computer programmer.
