Table 1.
Comparison of Ad Hoc Analyses of Question Characteristicsa
| Knauper et al. 1997 | Holbrook, Cho & Johnson 2006 | Yan and Tourangeau 2008 |
| Question length (number of words) [Understanding] | Question length (number of words) | Number of words per clause [Complexity of the question] |
| Question complexity (for example, embedded or inverted sentences) [Understanding] | Reading level (Flesch-Kincaid Grade Level) | Number of clauses [Complexity of the question] |
| Instructions to respondent (for example, “not counting any money or assets…”) [Understanding] | Qualified judgment (e.g., specified time frame or excluding items) | |
| Ambiguous terms (e.g., “Do you have less than one drink…”) [Understanding] | Level of abstraction (very abstract, somewhat abstract, concrete) | |
| Introductory phrase (e.g., “The next question might not be easy…”) [Understanding] | ||
| Whether or not question asked for a retrospective report [Retrieval/judgment] | ||
| Attitude/opinion [Complexity of the question/question type] | ||
| Behavioral frequency report [Retrieval/judgment]Quantity report [Retrieval/judgment] | Factual/behavioral [Complexity of the question/question type] | |
| Demographic [Complexity of the question/question type] | ||
| Whether a response scale is provided [Response formatting] | Response format (requesting numeric values, verbal labels, yes/no) | Nature of the response categories (not a scale, fully labeled scale, end-point-only-labeled scale, rating scale, frequency scale) [Complexity of the response options] |
| Number of response categories [Complexity of the response options] |
a
Phrases in square brackets indicate the authors’ label for the concept under which the specific characteristic appears. The dependent variables in the three analyses are, respectively, “don't know” answers, behaviors that suggest comprehension or mapping difficulties, and response times.