Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
The analysis of self-reports will be severely biased if they are subject to reporting heterogeneity. Moreover, there are several types of such heterogeneity, which have all shown to be widespread in the literature. We consider two predominant types of reporting heterogeneity: differential item functioning and middle inflation bias. We consider and extend approaches for adjusting for each type of reporting heterogeneity in isolation and propose models that allow for both types in combination. Monte Carlo experiments favor more complex models (that allow for reporting heterogeneity), even when the underlying data generating process is of a simpler form. The results suggest that failure to account for these nuances will lead to erroneous inference concerning the analysis of self-reported data. We apply these new methods to the important area of self-reported health outcomes.