Methods: Secondary data were used from the Transatlantic Trends: Immigration, 2008 national survey, which employed multi-stage random sampling. U.S. data (n=1,000) were analyzed, and participants were randomly split in half (Study 1: n=498; Study 2: n=502). In each sample, the mean age was approximately 52 years, and roughly one-third identified as politically independent.
Thirty-five survey items were determined to have face validity by tapping into attitudinal dimensions of perceptions of immigrants/immigration. Eight items were omitted due to dichotomous response options, lack of clarity, or being posed to half the sample. Thus, a total of 27 items (=.86) with 4-point Likert response options were suitable for inclusion.
In Study 1, an exploratory factor analysis using principal axis factoring and direct oblimin rotation was conducted with the 27 items in SPSS version 22.0 to explore the latent constructs underlying participants’ attitudes toward immigrants/immigration. In Study 2, a confirmatory factor analysis (CFA) was conducted using the maximum likelihood estimator in Mplus version 7.3 to examine if the specified model relating constructs to items on attitudes toward immigrants/immigration was consistent with the observed data.
Results: In Study 1, the scree plot and parallel analysis suggested four factors underlie the 27 items. Fifteen items were omitted due to factor loadings <.40. A correlated 4-factor solution with the remaining 12 items was the best fit to the data and accounted for 46.37% of the variance. Three items loaded onto each of the four factors and were interpreted as admiration, welcoming acceptance, blatant rejection, and fear of immigrants.
In Study 2, a CFA was run relating the four correlated constructs to the 12 measures. Results indicated good model fit (χ2(66) = 113.73, p < .00; RMSEA: .05 [C.I. .04, .07]; CFI: .95; SRMR: .04). Local fit indicators suggested all items loaded moderately to strongly, with loadings ranging from .37 to .78 (p < .001) on specified factors.
Conclusions and Implications: A correlated 4-factor model was determined to be the best fit to the data. As a preliminarily validated multidimensional instrument of attitudes toward immigrants/immigration, this scale has the potential to fill a major need in the intergroup relations literature. It has utility for research on the origins and consequences of anti-immigrant/immigration sentiment and the design and execution of native-immigrant intergroup interventions.