Orthorexia nervosa is characterized by a preoccupation to eat healthily. However, reliability and validity of some of the existing measures of orthorexic symptomatology are questionable. Therefore, the aim of the current study was to examine internal reliability of and intercorrelations between four of the most popular self-report scales for measuring orthorexia nervosa: Bratman's Orthorexia Test (BOT), the ORTO-15, the Eating Habits Questionnaire (EHQ), and the Düsseldorf Orthorexia Scale (DOS). Five-hundred and eleven adults (63% female) completed all four instruments. Model fit of the originally proposed factor structures of the BOT, DOS, and EHQ was good but was unacceptable for the ORTO-15. Similarly, internal reliability was good for the BOT, EHQ, and DOS, but was unacceptable for the ORTO-15. The BOT, EHQ, and DOS were highly correlated with each other while correlations with the ORTO-15 were of medium size. A subsequent exploratory item analysis suggested that the poor psychometric properties of the ORTO-15 are largely due to the originally proposed scoring procedure. In conclusion, the BOT, EHQ, and DOS are internally reliable instruments that seem to measure the same construct-orthorexic eating behavior. In line with previous suggestions, we conclude that the ORTO-15 cannot be recommended for the measurement of orthorexia nervosa, at least not when the originally proposed scoring procedure is used.
Keywords: Item analysis; Orthorexia; Psychometrics; Reliability; Self-report measures; Validity.
Copyright © 2019 Elsevier Ltd. All rights reserved.