Validity and Reliability of a Self-Report Likert-Type Measure of Nightmare Frequency

Abstract

Background and Objective: Nightmare frequency is typically measured using interval and ordinal, but rarely Likert- type, scales. The current study aimed to examine the validity and reliability of a Likert-type measure of retrospective nightmare frequency.

Materials and Methods: Participants included a convenience sample of undergraduate students. They completed a recent- ly developed Likert-type nightmare frequency item, an established Nightmare Frequency Scale, an estimate of nightmare frequency over the past week, and measures of hypothetically related variables including neuroticism, psychological dis- tress, and trauma symptoms. Convergent and discriminant validity was evaluated through relationships between the Likert- type nightmare frequency item with other nightmare measures and hypothetically related variables. Reliability was esti- mated using a correction for attenuation approach designed to provide a conservative estimate of single-item reliability.

Results: The sample included 233 (119 female, 51.1%; 114 male, 48.9%) university students (mean age: 19.05 ± 1.91). The average convergent validity coefficient between the Likert-type nightmare frequency item and other nightmare measures (r = 0.647) was significantly different (z = 5.45, P < 0.001) from its average correlation coefficient with hypo- thetically related measures (r = 0.256). The single-item reliability coefficient was 0.722.

Conclusion: The reliability and convergent and discriminant validity of the Likert-type nightmare frequency item were supported.