Premium
Psychometric Properties of the Dutch Version of the Washington Psychosocial Seizure Inventory
Author(s) -
Swinkels Wilhelmina A. M.,
Kuyk Jarl,
Van Dyck Richard,
Spinhoven Philip
Publication year - 2004
Publication title -
epilepsia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.687
H-Index - 191
eISSN - 1528-1167
pISSN - 0013-9580
DOI - 10.1111/j.0013-9580.2004.55203.x
Subject(s) - psychology , psychosocial , exploratory factor analysis , confirmatory factor analysis , psychometrics , reliability (semiconductor) , epilepsy , factor analysis , clinical psychology , scale (ratio) , explained variation , variance (accounting) , construct validity , psychiatry , statistics , structural equation modeling , mathematics , power (physics) , physics , accounting , quantum mechanics , business
Summary: Purpose: The psychometric properties of the Dutch version of the Washington Psychosocial Seizure Inventory (WPSI) were investigated. Methods: The dimensional structure, reliability, and validity of the WPSI scales were assessed in 218 patients with epilepsy. The association with relevant patient and epilepsy characteristics also was studied. Results: An exploratory seven‐factor principal components analysis was compared with a confirmatory factor analysis, and a loss of 3.36% of total variance was found. To increase the homogeneity of the scales, 30 items with low factor loadings or highest factor loading on the wrong clinical scale were excluded. This resulted in a higher amount of explained variance and improvement of the factor loadings of the remaining items. The reliability and validity of the WPSI scales was satisfactory to good. No clinically relevant associations were found between the WPSI scales and patient and epilepsy characteristics. Conclusions: The factor structure, reliability, and validity of the Dutch translation of the WPSI are good and comparable to the original English version of Dodrill. The results demonstrate that shortening the WPSI makes the inventory more reliable and the clinical scales less dependent on each other.