The purpose of this study was to evaluate the measurement invariance for measures that assess the three domains that make up SWB (i.e., life satisfaction, positive affect, and negative affect). We hypothesized that there would be limited measurement invariance across racial groups on measures of SWB based on the premise established by CRT that youth of color experience life differently than their White peers.
Methods: Data and sample: This study used data from the first wave of the Add Health study. The sample was restricted to youth aged 15 to 17 who self-identified as Black or White, which resulted in a sample comprised of 5,521 participants: 717 Black males, 766 Black females, 1972 White males, and 2066 White females
Measures: Subjective well-being was assessed using a composite of three subscales that assessed participants’ life satisfaction, positive affect, and negative affect. A scale that assessed participants’ perceived bias was also constructed in an effort to contextualize participant responses within a CRT framework.
Data Analyses: To assess configural, metric, and structural invariance in the observed measures, the analyses were handled using R software (v. 3.5.1) mIRT (v. 1.29) package for all estimations. We compared the measurement structures of the empirical forms to the hypothetical model structure. Evaluations of measurement invariance were conducted using the likelihood ratio statistic while using the SRMSR statistic to interpret local group differences in fit. All models with a significance was set at α = .05 using the limited information Goodness of Fit statistic.
Results: Our results indicated that configural invariance failed to be rejected between the groups (M2(32) = 44.45, p = .07) suggesting that the measurement structure and loading patterns replicated across racial and gender groups. Next, the equivalence of factor loadings across groups, was tested and rejected (M2293 = 493.21, p < .0001), demonstrating a failure of both metric and scalar invariance. This pattern indicates that while the interpretative meaning of the items holds between groups (configural invariance), the relationship between individual items and their respective latent measure spaces are non-equivalent.
Conclusions and Implications: This study provides partial evidence that racial/ethnic differences cannot be ignored when assessing SWB in youth. Critical race theory provides a context for us to understand observed differences in SWB as an outcome. It is important to note that the findings presented here are likely symptomatic of a much deeper issue. That is, race and racism not only influence how constructs are conceptualized and measured, but how findings are interpreted.