2026 Volume 6 Pages 1-11
Background: Cultural differences in questionnaire scores have been reported in the assessment of eating disorders. The Eating Disorder Inventory-3 (EDI-3) includes both the Eating Disorder Risk and Psychological Scales, as well as Response Style (RS) Indicators. While RS Indicators were originally introduced to assess the validity of responses, they may also reflect psychological tendencies and culturally influenced self-presentation. Previous cross-cultural studies using the EDI-2 have suggested that cultural context may influence symptom expression and psychological characteristics. However, few studies have reported EDI-3 RS Indicators in samples outside the EDI-3 standardisation samples. This study examined RS Indicators and their relation to subscale scores in Japanese patients and non-clinical participants compared with the U.S. adult patient sample.
Subjects and Methods: Participants included 114 adult Japanese patients with eating disorders and 183 Japanese university students as non-clinical participants. All EDI-3 subscales and RS Indicators were analysed. Subscale scores in Japanese patients were compared with summary statistics from the published U.S. adult patient sample, whereas comparisons between Japanese patients and non-clinical participants were conducted using raw data. RS Indicators were evaluated by deriving category proportions from U.S. adult patient frequency tables and comparing them with those of Japanese patients.
Results: Subscale patterns were generally consistent with previous studies. Compared with the U.S. adult patient sample, Japanese patients showed higher scores on most subscales except for Drive for Thinness, Body Dissatisfaction, Perfectionism, and Asceticism. The RS Indicators showed marked differences. Japanese patients were more frequently classified as Atypical or Very Atypical. For the Infrequency (IF) Scale, endorsements in the extreme symptomatic direction, which were infrequent in the U.S. adult patient sample, occurred more frequently in Japanese patients, indicating that the assumption of low-frequency responses was not supported. In 10 IF items with less than 2% endorsement in the U.S. standardisation samples, response rates in Japanese patients ranged from 6.1% to 24.6%.
Discussion: Patterns in the Negative Impression (NI) and Inconsistency (IN) Scales, as well as the IF Scale, suggest that EDI-3 RS Indicators capture culturally influenced differences, similar to cross-cultural variation in subscale scores. These differences between Japanese patients and U.S. adult patients, which were also observed to some extent in Japanese non-clinical participants, appear to reflect self-presentational tendencies shaped by culture, including modesty and negative self-evaluation, rather than invalid response. The low endorsement rates of IF items may indicate response biases influenced by cultural factors. They should, therefore, be considered when interpreting results of the IF Scale.
Conclusion: Japanese patients showed distinctive response patterns in both EDI-3 subscale scores and RS Indicators. These patterns exhibited culturally influenced styles of self-evaluation rather than simple response invalidity, highlighting the need to consider culture and response styles when interpreting scores.
