Following the procedure of Vyas and Kumaranayake (2006), education, household income, and subjective class identification
were used to construct a principal component analysis (PCA)-based SES score. Notably, the three interval variables
were standardized and entered into the PCA.4 Since only the associated eigenvalue (accounting for 58% of the variation in
the original data) of the first extracted component was greater than one (k = 1.741), the results indicated that a single component
could be extracted with PCA. Accordingly, using the factor scores from the first principal component as weights (see
Table 1), a new variable was constructed for each respondent, which was regarded as the respondent’s SES score, and the higher
the SES score, the higher the implied SES of that respondent.