Development and validation of the Eating Pathology Symptoms Inventory (EPSI).

Abstract

Many current measures of eating disorder (ED) symptoms have 1 or more serious limitations, such as inconsistent factor structures or poor discriminant validity. The goal of this study was to overcome these limitations through the development of a comprehensive multidimensional measure of eating pathology. An initial pool of 160 items was developed to assess 20 dimensions of eating pathology. The initial item pool was administered to a student sample (N = 433) and community sample (N = 407) to determine the preliminary structure of the measure using exploratory and confirmatory factor analyses. The revised measure was administered to independent samples of patients recruited from specialty ED treatment centers (N = 158), outpatient psychiatric clinics (N = 303), and students (N = 227). Analyses revealed an 8-factor structure characterized by Body Dissatisfaction, Binge Eating, Cognitive Restraint, Excessive Exercise, Restricting, Purging, Muscle Building, and Negative Attitudes Toward Obesity. Scale scores showed excellent convergent and discriminant validity; other analyses demonstrated that the majority of scales were invariant across sex and weight categories. Eating Pathology Symptoms Inventory scale scores had excellent internal consistency (median coefficient alphas ranged from .84-.89) and reliability over a 2- to 4-week period (mean retest r = .73). The current study represents one of the most comprehensive scale development projects ever conducted in the field of EDs and will enhance future basic and treatment research focused on EDs.

Scale development: ten main limitations and recommendations to improve future research practices

“…Given its importance, future studies should to consider alpha evaluation as a central point of measurement reliability, and yet, as much as possible, involve the assessment of internal consistency with other measures of reliability. In the sequence of internal consistency, the following methods were identified by this review: test-retest reliability (analysis of the temporal stability; items are applied on two separate occasions, and the scores could be correlated) (e.g., Forbush et al 2013), item-total/inter-item correlation reliability (analysis of the correlation of each item with the total score of the scale or subscales/analysis of the correlation of each item with another item) (e.g., Rodrigues and Bastos 2012), split-half reliability (the scale is split in half and the first half of the items are compared to the second half ) (e.g., Uzunboylu and Ozdamli 2011), and inter-judge reliability (analysis of the consistency between two different observers when they assess the same measure in the same individual) (e.g., Akter et al 2013;DeVellis 2003;Nunnally 1967).…”

“…The first and the most representative way was related to the sample type. Several studies used homogeneous sampling (e.g., Forbush et al 2013;Morean et al 2012), whereas others used convenience sampling (e.g., Coker et al 2011;Flight et al 2011). Both homogeneous and convenience samples were related to limitations of generalization.…”

Men, Muscles, and Eating Disorders: an Overview of Traditional and Muscularity-Oriented Disordered Eating

“…Measures available for assessing broad ED symptoms among males include those focused on traditional ED symptoms, such as the Eating Disorder Examination-Questionnaire [EDE-Q; 33] as well as newer measures that assess both traditional and muscularity-oriented ED symptoms, such the Eating Disorder Assessment for Men [EDAM; 34] and the Eating Pathology Symptoms Inventory [EPSI; 35]…”

Are common measures of dietary restraint and disinhibited eating reliable and valid in obese persons?

“…Participants were normal weight (N = 510), overweight (N = 201), and obese (N = 101) individuals recruited from the community to participate in a study of "health and eating behaviors" (Forbush et al, 2013). Participants were recruited from fliers posted in local community establishments and on buses.…”