While most survey research on immigration attitudes routinely excludes respondents with an immigrant background, empirical justifications for doing so are rarely provided. Yet it is crucial to examine whether respondents with an immigrant background can be included in the analysis, as excluding them with no further consideration would ignore the opinion of a considerable share of multicultural societies. With multigroup analyses on Swiss data we illustrated how to evaluate whether both the measurement and the prediction of immigration attitudes are invariant across natives and individuals with an immigrant background. The slight revealed differences did not justify the exclusion of respondents with an immigrant background. The paper concludes by discussing ways of addressing these differences in measurement and prediction.