Rosenberg-P self-esteem scale (EAR-P) in university students: New psychometric evidence of its internal structure

Abstract

Self-esteem is a psychological term that since its emergence in 1890 with William James has been widely studied, the most used concept being that of Rosenberg (1965) who defines it as that positive or negative attitude towards oneself. In addition, this construct has been studied in different areas of psychology, including the clinical field, where its relationship with mental health has been demonstrated. One of the most widely used instruments for measuring this construct is the Rosenberg Self-Esteem Scale (EAR), made up of 10 items that measure the positive or negative assessment of oneself. The EAR has been adapted in different countries worldwide where psychometric studies were carried out in adolescents, university students, adults and older adults. Despite its wide use, the EAR has shown variations in terms of factorial structure. In this way, different positions are held on whether the scale is two-dimensional or one-dimensional. Part of this problem is due to the negative wording of the items that favors the appearance of a new dimension. In addition, other studies found the presence of a method effect associated with the negative items that has been examined by applying statistical procedures. To this end, the Rosenberg-P Scale (EAR-P) was proposed as a new version of only positive items that made it possible to overcome the method effect associated with the negative wording of the items and provided a one-dimensional measurement of the self-esteem construct as proposed by Rosenberg. However, the EAR-P presented limitations given that no evidence was shown to identify the relationship between self-esteem and other similar constructs, and it is also added that the sample could have been larger in order to obtain better results of psychometric properties. The objective of the study was to evaluate the validity based on the internal structure from a Classical Test Theory (CTT) and Item Response Theory (IRT) approach, obtain evidence of validity based on the relationship with other variables and estimate reliability. of the Rosenberg Self-Esteem Scale-P (EAR-P). For this, 797 Peruvian university students were evaluated, of which 235 were men (29.49%) and 562 women (70.51%) with an age range of 18 to 35 years (M = 22.55, SD = 3.32). The results show that the unidimensional structure of the EAR-P presents adequate adjustment indices (CFI=.98, TLI=.97, SRMR=.05), the parameters of difficulty and discrimination showed favorable values and an acceptable value was found in internal consistency (ɷ = .96; α =.97). In addition, the EAR-P was found to be negatively related to anxiety depression (-.45), anxiety (-.33), and stress (-.30). It is also positively related to life satisfaction (.69). The findings of the one-dimensional model were adequate. However, it was found that one of the indices did not provide a good fit (RMSEA=.164). One of the possible explanations is related to the small degrees of freedom identified in the model (df=35), because small values in the degrees of freedom cause inappropriate adjustments in the RMSEA. In addition, it has been found that its operation is altered when there are few response categories, specifically when there are 4, as is the case of the EAR. Faced with the limitations of the RMSEA, support has been given to other adjustment indices such as the SRMR and CFI, which are robust to being influenced in their values when the model shows small degrees of freedom. Likewise, the SRMR is considered as an index that allows to adequately evaluate the fit of the model regardless of the estimation method used. Therefore, it is concluded that the EAR-P presents adequate psychometric properties of validity and reliability, this allows to have a useful and pertinent instrument for the measurement of self-esteem in the Peruvian university students.