Information restrictions have received significant traction as a policy and legislative tool to fight employment discrimination. These policies forbid employers from requesting potentially prejudicial information like criminal records, salary history, or credit scores until the final stage of hiring. The assumption is that this information would disproportionally remove minority and female candidates from contention in the hiring process.

Yet, studies in the wake of these policy changes have suggested that racial and gender disparities remain under these conditions. Drawing upon social science literature, I suggest that restricting limited pieces of information about applicants, as commonly touted, may be an overoptimistic solution for mitigating bias. Employment restrictions rely on the assumption that people will not think about what is not in front of them, but information restrictions may inadvertently make the forbidden information more salient. Instead, employers may seek out information because it is withheld from them.

To test how an information restriction impacts employer interest, this Article leverages two original empirical studies with over six hundred respondents, who serve as employers in an economically realistic scenario. I find exploratory evidence that restrictions on information about candidates’ protected characteristics results in respondents’ increased interest in those protected characteristics.

But simply providing more information to employers does not rectify the bases on which employers continue to discriminate. Rather, I argue that regulators, employers, and courts ought to define the distributional consequences at stake, the criteria they rely on, and the rules of disparate impact in a data-driven world.