The proliferation of team-authored academic work has led to the proliferation of two kinds of authorship misconduct: ghost authorship, in which contributors are not listed as authors and honorary authorship, in which non-contributors are listed as authors. Drawing on data from a survey of 2,222 social scientists from around the globe, we study the prevalence of authorship misconduct in the social sciences. Our results show that ghost and honorary authorship occur frequently here and may be driven by social scientists' misconceptions about authorship criteria. Our results show that they frequently deviate from a common point of authorship reference (the ICMJE authorship criteria). On the one hand, they tend to award authorship more broadly to more junior scholars, while on the other hand, they may withhold authorship from senior scholars if those are engaged in collaborations with junior scholars. Authorship misattribution, even if it is based on a misunderstanding of authorship criteria rather than egregious misconduct, alters academic rankings and may constitute a threat to the integrity of science. Based on our findings, we call for journals to implement contribution disclosures and to define authorship criteria more explicitly to guide and inform researchers as to what constitutes authorship in the social sciences. Our results also hold implications for research institutions, universities, and publishers to move beyond authorship-based citation and publication rankings in hiring and tenure processes and instead to focus explicitly on contributions in team-authored publications.