Reduction in variation in response function for Social Sciences variables: Job Satisfaction

In Chapter 11 Bon tried to solve the problem of variation in response functions by what can be called a 'post data collection correction procedure'. However, the author was forced to admit that it was unclear whether one is correcting variation in response functions or variation in opinion.
Saris et al. (Chapter 12) suggested that it is better to prevent variation in response functions in the data collection by restriction of the respondents' freedom to answer by providing two references points on the response scale. The success of that psychophysical experiment stimulated this new study concerning the efficiency of the suggested procedure for a social sciences' variables.
The topic we have chosen for this purpose is job satisfaction, which is, normally seen as a private experience (Andrews and Withey 1974) and therefore should be measured by means of subjective indicators (Andrews and Withey 1976).
Diener (1984) mentions that over 700 studies about psychological Well-being have appeared in the last 25 years, so this seems a rather important topic. On the other hand, the judgements asked are subjective evaluations. Subsequently the problems discussed in this book are very likely to occur for this measurement instrument. Thus, an experiment for this topic to reduce the expected variation in response functions makes sense
First of all, the design of the experiment and the data collection will be discussed. Toen the model used for the analysis will be specified. After that the results follow and a discussion of these results.