Adaptation of the Bergen Instagram addiction scale in Argentina: calibration with item response theory

Abstract

The items of the Bergen Facebook Addiction Scale (Andreassen et al., 2012) were adapted to study the construct in Instagram users. This construct is defined as the tendency to be overly concerned about Instagram, spending so much time and effort on this platform to the point of harming other activities in the individual’s life, interpersonal relationships and psychological well-being. Different Item Response Theory (IRT) models was applied for polytomous items in order to determine which one best explains the responses of those evaluated. The six items were translated into Spanish and evaluated by an expert jury. A total of 754 Instagram users from the Buenos Aires Metropolitan Area took part in the study (65.8% women). An Exploratory Factor Analysis was performed to verify the unidimensional assumption and the assumption of local independence was corroborated. The results referring to the models fit data allows us to provide evidence that the Grades Response Model (GRM) presents the most adequate parametric structure to shape the subject’s responses. All items were adjusted to GRM. Parameters a ranged from 2.02 to 3.45 (Mean = 2.69; SD = 0.49). Parameters b were located within the expected range (-3 to 3) with a minimum θ of -0.51 and a maximum of 2.74. The Ordinal Alpha and Omega were elevated. The Information Function shows a maximum value of 14.26 around a θ of 1.2. This function shows a negative asymmetry, indicating that the scale is more accurate for measuring high levels of the trait.