Comparison of single- and multiple-item nightmare frequency measures

Abstract

Nightmare frequency research usually relies on single-item measures. The current study compared some single- and multiple-item measures of nightmare frequency. Three single-item measures representing different response formats (binary, ordinal, and nominal) were compared with both an existing and newly devised multiple-item nightmare frequency scale, the Nightmare Frequency Index (NFI). To collect psychometric information on the new scale, 276 university students completed the NFI. A subsample of 146 participants completed additional measures of nightmare frequency and constructs previously related to nightmares. The NFI had adequate psychometric properties. Generally, nightmare frequency measures demonstrated convergent and discriminant validity. A higher-order nightmare frequency factor was found with all nightmare frequency measures contributing. Multiple-item measures had slightly higher average correlations than a binary single-item measure, but not ordinal or nominal single-item measures. The results are discussed in the context of choosing frequency measures for research