ABSTRACT
We have already lamented that the large majority of empirical studies do not include the deliberate construction of a variable before performing the statistical analysis. Raw scores, or counts, are treated as measures, and analyses often are conducted at the item level rather than the variable level. For example, the ubiquitous checklist is an item format commonly used in surveys, particularly in the discipline of health education. A researcher who wished to know respondents' perceived barriers to getting involved in determining public health policy might ask them to check all statements that apply to them. Some of the options offered to the respondent might include the following: lack of time, lack of money, involvement not personally gratifying, and so on (Boardley, Fox, & Robinson, 1999).
