Psychometric properties of the General Risk Propensity Scale (GRiPS) in a Brazilian sample

Abstract

Risk propensity is the probability of individuals engaging in activities in which there is variability in possible outcomes. In a recent effort to measure general risk-taking propensity, Zhang, Highhouse and Nye (2018) developed and evaluated the General Risk Propensity Scale (GRiPS) using a total sample of 1,523 participants across five studies. The authors found construct validity evidence via the internal structure of item responses and convergence with other self-report measures of risk propensity. They also found that the scale predicted work, academic and life-related outcomes over and above other risk propensity instruments and the big five. Thus, our goal in this study was to investigate the psychometric properties of the GRiPS in a Brazilian sample (n=404). We use factor analyses, TRI analyses and measurement invariance analyses and found good psychometric indices of this instrument to Brazilian samples.

Validity and reliability of the Arabic versions of the General Risk Propensity Scale (GRiPS) and the Risk Proneness Short Scale (R-1)

“…In terms of factorial validity, the EFA-to-CFA approach to explore the best-tting model of the GRiSP suggested that the original one-factor was replicated in this study, except for two items that were removed (item 4 "I would take a risk even if it meant I might get hurt", and item 5 "Taking risks is an important part of my life"). In contrast to our ndings, Porfírio et al [29] were able to replicate the unidimensional model with all eight items loading on a single latent variable in a Brazilian sample of adults from the general population. These variations in factor structure might be explained by cultural variations in people's perceptions of, and attitudes toward, engaging in risky behaviors [31].…”

“…These ndings indicate that both scales measure the RTP construct in the same way across males and females, and that inferences based on gender differences on the variances or means of GRiPS and R-1 latent scores are likely not confounded by measurement bias [60, 61]. Consistently, measurement invariance of the GRiPS was demonstrated in the Brazilian sample [29]. Establishing this psychometric property is of high relevance, given that gender comparisons represent a key research interest in the area of risk-propensity [62].…”

“…In research practice, the GRiPS was reported to be simple-to-administer, easy-to-understand, convenient to score and interpret [25][26][27][28]. The GRiPS has also been validated in Portuguese and yielded adequate psychometric properties, including measurement invariance across gender, in a Brazilian sample aged 16 to 75 years [29]. The second measure selected for validation in this study is the R-1, which is composed of a single item (i.e.…”