Suicide researchers commonly use a variety of assessment methods (e.g., surveys and interviews) to enroll participants into studies and assign them to study conditions. However, prior studies suggest that different assessment methods and items may yield different responses from participants. This study examines potential inconsistencies in participants' reports of suicidal ideation (SI) and suicide attempt (SA) across commonly used assessment methods: phone screen interview, in-person interview, self-report survey, and confidential exit survey. To test the reliability of the effects, we replicated the study across two samples. In both samples, we observed a notable degree of inconsistent reporting. Importantly, the highest endorsement rates for SI/SA were on a confidential exit survey. Follow-up assessments and analyses did not provide strong support for the roles of purposeful inaccuracy, random responding, or differences in participant experiences/conceptualizations of SI. Although the reasons for such inconsistencies remain inconclusive, results suggest that classification of suicidal/control participants that uses multiple items to capture a single construct, that uses a Graded Scale to capture a broad spectrum of thoughts and behaviors, and that takes into account consistency of responding across such items may provide clearer and more homogenous groups and is recommended for future study. (PsycInfo Database Record (c) 2021 APA, all rights reserved).