Analysis of Rank Scales in Mass Surveys

Abstract

The problem of measurement in sociology is always relevant. This is especially noticeable in design a questionnaire for survey. One of the types of scales that are trying to avoid in paper surveys is ranked scale. In electronic surveys, the use of ranking scales is much simpler. The question of interpretation of responses to such scales remains actual. In the paper there is analysis of a set of 14 of entrant’s preferences of the choice of specialty when joining to HEI. At first, this set is considered as 14 ordinal questions, then – as one ranked question with 14 alternatives, which respondents had to set ascending importance. Then, based on responses to the ordinal questions, a rating of preferences was constructed. For analysis of the rank question, two methods were applied: the Schulze method and the Borda count. For each method, ratings were constructed and the principles were underlying the analysis in each of these approaches were explained. The results were written in a consolidated table and a comparison of these results was performed: as an analysis of the ordinal scale questions with the Schulze and Borda methods, and the methods themselves. At the end of the paper recommendations for the application of these methods in social studies were formulated.