A Reliability Generalization Meta-analysis of the Leyton Obsessional Inventory Child Version Survey Form

Abstract

BACKGROUND: The Leyton Obsessional Inventory Child Version Survey Form (LOI-CV-SF) is a scale created in order to detect obsessive-compulsive symptoms and their effects on the daily lives of children and adolescents from ages 10 to 18 and for the screening of non-clinic infant population. With the purpose of estimating the reliability of the test scores, and how it varies in relation to the characteristics of the studies, a meta-analysis of reliability generalization was carried out. METHODS: A meta-analysis was accomplished. An exhaustive research allowed to select 13 studies that reported some reliability estimate of the test scores and, through the KR-21 formula, this number was increased to 43 estimations of internal consistency reliability. Taking these coefficients, an estimation of the average reliability and analyses of the variables that could be affecting the heterogeneity of reliability coefficients was accomplished. RESULTS: On average, the reliability by internal consistency of the symptom scale scores was 0.79 (95%CI:0.76 and 0.82), with minimum and maximum values of 0.52 and 0.97, respectively. A large heterogeneity was found (I2 = 96%).The predictive model identified the standard deviation of tests scores as the most relevant variable. The analysis of moderator variables revealed that, mainly, the standard deviation of the test scores is the most statistically related to the reliability, presenting a positive relationship with it. CONCLUSIONS: The results of the LOI-CV-SF showed a satisfactory average reliability for research purposes, but not for clinical practice purposes.