Measurement Invariance of the Satisfaction With Life Scale Across 26 Countries

Abstract

The Satisfaction With Life Scale (SWLS) is a commonly used life satisfaction scale. Cross-cultural researchers use SWLS to compare mean scores of life satisfaction across countries. Despite the wide use of SWLS in cross-cultural studies, measurement invariance of SWLS has rarely been investigated, and previous studies showed inconsistent findings. Therefore, we examined the measurement invariance of SWLS with samples collected from 26 countries. To test measurement invariance, we utilized three measurement invariance techniques: (a) multigroup confirmatory factor analysis (MG-CFA), (b) multilevel confirmatory factor analysis (ML-CFA), and (c) alignment optimization methods. The three methods demonstrated that configural and metric invariances of life satisfaction held across 26 countries, whereas scalar invariance did not. With partial invariance testing, we identified that the intercepts of Items 2, 4, and 5 were noninvariant. Based on two invariant intercepts, factor means of countries were compared. Chile showed the highest factor mean; Spain and Bulgaria showed the lowest. The findings enhance our understanding of life satisfaction across countries, and they provide researchers and practitioners with practical guidance on how to conduct measurement invariance testing across countries.

Examining the cross‐cultural validity of the positive affect and negative affect schedule between an Asian (Singaporean) sample and a Western (American) sample

“…Based on Milfont and Fischer's () guidelines, we first conducted a CFA on each sample independently to ascertain that a two‐factor structure (i.e., PA and NA each indicated by their respective 10 items) fits both data sets well. Following current measurement invariance studies (e.g., Jang et al., ), model fit was assessed based on three goodness‐of‐fit statistics: chi‐square significance test, comparative fit index (CFI), and root mean square error of approximation (RMSEA). For CFI, values between .90 and .94 indicate acceptable fit whereas values of .95 or greater indicate good fit (Hu & Bentler, ).…”

“…We then conducted a test of partial metric invariance by constraining each factor loading one at a time (instead of constraining all factor loadings simultaneously) to determine which specific items were exhibiting metric noninvariance between our samples (Byrne, Shavelson, & Muthén, ). Following previous measurement‐invariance studies (e.g., Jang et al., ), the items with the least residual variance when loaded onto their respective latent factors were selected as reference items and constrained to be equal between groups. The items “interested” and “afraid” were selected as reference items for the latent PA and NA factors, respectively.…”

Testing the Measurement Invariance of Nativism*

“…Strict invariance provides evidence that the mean differences across groups are driven by real group differences and not by error variance. While researchers prefer strict invariance as the ideal threshold for construct validity, they consider scalar invariance empirically appropriate to compare factor or observed means because of the rigidness required to meet the strict invariance requirements (Davidov et al., ; Jang et al., ; Meredith, ).…”

“…The model fit is interpreted based on three indices: chi‐square, comparative fit index (CFI), and root mean squared error of approximation (RMSEA). Because chi‐square values are sensitive to sample size and the number of groups included in the analysis (Bentler and Bonett, ; Jang et al., ), we primarily rely on CFI and RMSEA. By convention, the CFI ≥ 0.95 and the RMSEA ≤ 0.05 indicate acceptable model fits (Hu and Bentler, ).…”

“…Within the MG‐CFA context, however, previous research advocates for a more liberal RMSEA cutoff based on simulation results showing the number of groups affects the RMSEA cutoff (Rutkowski and Svetina, ). In this study, because we are investigating metric invariance across 19 countries, we use a 0.10 level threshold following the suggestions of prior studies (Jang et al., ; Rutkowski and Svetina, ).…”