Background: Understanding how policymakers define research and differentiate it from other sources of data is critical for scientists to improve how they conduct and communicate research to policy audiences. Yet, few studies have explicitly asked policymakers -- particularly state legislators in the USA -- how they define research evidence. Methods: We sought to fill this gap via in-depth interviews with 168 policymakers from two Midwestern states; 32 of whom were nominated by their colleagues as exemplar research users. Findings were triangulated via interviews with experienced key informants from both states. In-depth interviews were the preferred methodology for our research question, as they offered legislators the chance to describe research in their own words and elaborate on examples when needed. Findings: For many legislators, definitions of research largely aligned with how the scientific community might define research; both Republicans and Democrats defined research as peer-reviewed studies with specific qualities that distinguish research evidence from other types of information. However, some legislators defined research with a broader lens, including different types of information (for example, anecdotes) and qualities of information (for example, accessibility, relevance, credibility, and unbiased) as part of their definition. Discussion and conclusions: Researchers may better engage policy audiences by referring to the types and qualities legislators mentioned because policymakers prefer evidence from rigorous studies to those that are poorly executed or politically motivated. Legislators called this 'bogus' research, 'party' research or 'pseudoscience'. Researchers can signal their credibility by being transparent regarding funding sources and reasons or motivation for conducting studies.