A comparison of fully labeled and top-labeled grid question formats

Placing items in a grid requires respondents to engage in the difficult response task of connecting information in rows and columns and therefore has been shown to reduce response quality. In web surveys, dynamic programming features can be used to help respondents with this task, thereby reducing some of the negative effects of the grid format, but no such features are available in mail surveys. In this chapter, we compare responses to survey questions, data quality indicators, and respondent processing (i.e. eye-tracking data) across traditional grid designs and fully labeled grid designs where the response option labels are repeated in every row of the grid, thereby reducing the need for vertical processing. We find few differences in responses or data quality, but respondents to the fully labeled grid did appear to process more efficiently.