ABSTRACT
Resumen El uso de Internet está masificado en casi todo el mundo. Existe evidencia de que su uso excesivo tiene consecuencias negativas, por lo que es necesario contar con instrumentos que evalúen su uso. El Cuestionario de Experiencias Relacionadas con Internet (CERI) es un instrumento de dos factores que evalúan conflictos (intrapersonales e interpersonales) asociados a su uso excesivo. Estudios realizados en distintas partes del mundo han utilizado este instrumento sin analizar sus propiedades psicométricas con la suficiente profundidad y han privilegiado el uso del puntaje total en lugar del puntaje de los dos factores. Por esta razón, en este artículo se evalúan las propiedades psicométricas del instrumento en población chilena. Se realizaron dos estudios en los que participaron un total de 626 estudiantes universitarios (180 en el estudio 1 y 446 en el estudio 2) que respondieron el CERI y un cuestionario de abuso del teléfono móvil. En el estudio 1 se realizó un análisis factorial exploratorio que sugirió seleccionar la estructura unifactorial en lugar de la estructura bifactorial original, aunque ambas presentaban adecuada validez convergente (relaciones directas con el abuso del teléfono móvil). En el estudio 2 se realizó un análisis factorial confirmatorio y ambas estructuras factoriales presentaron buen ajuste, pero la consistencia interna del factor único fue mayor que la de los dos factores originales. Es necesario continuar evaluando la estructura factorial del CERI en orden a determinar cuál es la mejor alternativa factorial, mientras tanto la estructura unifactorial parece entregar mayores garantías psicométricas.
Abstract A large proportion of the world's population has access to the Internet. On the one hand this has brought undeniable benefits to society by facilitating access to information. But on the other hand, the excessive use of the Internet has also brought with it a series of new problems ranging from physical injuries (for example: tendinitis), psychological pathologies (for example: Internet addiction) to new forms of crime (for example: online sexual abuse). For this reason, it is necessary to have reliable and valid instruments that evaluate the use and abuse of the Internet. The Internet Related Experiences Questionnaire (CERI) has two factors that evaluate conflicts (intrapersonal factor, with 6 items and interpersonal factor, with 4 items) associated with excessive use of the Internet. Studies in different parts of the world (Spain, Chile and other Latin American countries) have used this instrument without analyzing its psychometric properties with sufficient depth. These studies have mostly used the total CERI score (10 items) instead of the two-factor structure, but they have not justified this choice. For this reason, this paper intends to analyze the factorial structure of the instrument, its convergent validity and its internal consistency in the Chilean population. We conducted two researches in which participated a total of 626 university students (180 in the study 1, and 446 in the study 2). The participants responded to the CERI and a questionnaire of abuse of the mobile phone (CERM). In the study 1, the results of exploratory factor analyzes suggest to use a unifactorial structure which included the 10 items, instead of the original bifactorial structure. In particular, in the bifactorial structure item 1 was not included in any factor (factorial loads less than .2), two items (items 5 and 10) presented similar factor loads in both factors and other two items (items 8 and 9) were included in the factor that it did not correspond. The single factor was greater internal consistency than that of the original intrapersonal and interpersonal factors (Cronbach´s alpha of .78, .68 and .56, respectively). Although both alternatives have adequate convergent validity (direct relations with mobile phone abuse). In study 2, a confirmatory factorial analysis was carried out in order to compare the unifactorial structure obtained in study 1 with the bifactorial structure of the original study. The results show that both factorial structures presented good fit, but again the internal consistency of the single factor was greater than that of the original intrapersonal factor and interpersonal factor (Cronbach´s alpha of .74, .61 and .51 respectively). The low reliability of interpersonal and intrapersonal factors may be due to the few number of items grouped in them (4 and 6 respectively). It is necessary to continue evaluating the factorial structure of the CERI in order to determine which is the best factorial alternative or if it is necessary to increase the number of items of interpersonal and intrapersonal factors. Although the results are not conclusive, for the moment, it seems advisable to use the one factor structure as it seems to give further psychometric guarantees.