Psychometric properties of the Inventory of Study Habits CASM-85: a multicenter study with Peruvian high school students

• Autores: Figueroa-Quiñones, Joel, Pomahuacre Carhuayal, Juan Walter, Chávez, Yuri, Gómez, Heber, Portocarrero, Roxana, Gamero-Vega, Giulianna, Calderón-Giraldo, Ada, López, Elizabeth, Castro, Martín, Bazo-Alvarez, Juan Carlos
• Formato: artículo
• Fecha de Publicación: 2019
• Institución: Universidad de San Martín de Porres
• Repositorio: Liberabit
• Lenguaje: español
• OAI Identifier: oai:ojs2.ojs3.revistaliberabit.com:article/244
• Enlace del recurso: http://www.revistaliberabit.com/index.php/Liberabit/article/view/244
• Nivel de acceso: acceso abierto
• Materia: hábitos de estudio; CASM-85; propiedades psicométricas; modelos MIMIC

Objective: To assess the internal structure, measurement invariance, criterion validity and reliability of the CASM-85 in Peruvian high school students. Materials and methods: Two thousand seventy-five (2075) students of both sexes, aged 11 to 18 years, from public and private high schools of Arequipa, Chimbote, Lima and Pucallpa were evaluated. An item-item and item-rest correlation analysis for each dimension of the CASM- 85, a confirmatory factorial analysis and MIMIC (Multiple Indicator, Multiple Cause) models were used to evaluate the measurement invariance across the groups according to their sex, age, school and city. Reliability was estimated using alpha and omega coefficients. The criterion validity between the CASM-85 and academic performance was determined using Pearson’s r and regression models. Results: The heterogeneous items were removed, obtaining a 45-item short form compared to the original 53-item form. In the original form, CFI and TLI were < .9 in all the cities, whereas in the short form, CFI and TLI were < .9 only in Chimbote. SRMR and RMSEA were ≤ .10 and ≤ .09 in the original and short forms, respectively. The MIMIC models confirmed the measurement invariance in all variables, with the exception of the original form between cities (|ΔCFI| = .09, |ΔTLI| = .08). Reliability was moderate for both the original (Ω ≥ .68; α ≥ .81) and short forms (Ω ≥ .77; α ≥ .84). Pearson’s r showed a positive and weak correlation in the original (r = .17, p < .001) and short (r = .18, p < .001) forms. Conclusion: The CASM-85 short form (45 items) has shown better psychometric properties than its original full form.