ABSTRACT
This chapter addresses the influence of both the number and the labeling of response-option categories in a test-construction process to demonstrate how the design of rating scales has a large impact on the quality of the responses elicited and to show how the Rasch model provides an appropriate framework for carrying out such empirical investigations. It examines category frequencies and average measures to identify categories with low frequencies and lack of monotonicity across average measures. When rating scale categories were used by respondents, adjacent categories can be combined and the data reanalyzed. In the Rasch world, Wright and Masters first demonstrated the problems by plotting the estimates for the positively worded items against those for the negatively worded items on a series of questions about attitudes toward drugs. In addition to plotting the pairs of estimates for the two types of items: negatively worded items and positively worded items, potential problems are likely to be flagged by Rasch diagnostics.
