Design, Evaluation, and Analysis of Questionnaires for Survey Research, 2nd Edition

PART IV

APPLICATIONS IN SOCIAL SCIENCE RESEARCH

In Chapter 12, we saw that the quality of the questions can be rather different across methods and across countries. This means that the effect of variables measured with different methods or in different countries cannot be compared without correction for measurement error. We also saw that knowledge of the quality of questions involved in MTMM experiments is only very limited, and therefore, we looked at the possibility of developing an algorithm to predict the quality of any question in any possible language. In the last chapter, we showed that such an algorithm has been developed on the basis of the knowledge we have collected with the MTMM experiments. Surely, the present algorithm is not as general as covering all possible methods and languages, but it is much more general and better than the earlier algorithm incorporated in the program SQP 1.0. Now, the algorithm is covering more methods and many more languages.

This algorithm has been incorporated in a new program, SQP 2.0, which will be introduced in Chapter 13. In this chapter, the program and the possibilities for improvement of questions will be discussed.

In Chapter 14, we will illustrate how the information available in SQP 2.0 can be used to estimate the quality of concepts-by-postulation.

In Chapter 15, the fundamental issue of correction for measurement errors in survey analysis will be discussed. In this case, we will pay attention to the correction for measurement ...