One in Five Research Participants Underreported Drug Use When Checked Against Hair Tests | Research

The takeaway

About 21% of participants denied using a substance that their hair test detected, highlighting the limits of relying on self-report alone in research.

Cannabis researchers, clinicians assessing substance use, survey methodology researchers

21.3% of participants denied using a substance that their hair test detected

What the researchers found

21.3% of 75 participants underreported use of at least one substance (negative self-report but positive hair test). For cannabis specifically, self-reported frequency and quantity predicted THC detection in hair, but years of use and time since last use did not.

Why it matters

Cannabis research depends heavily on self-reported use, but this study shows that more than one in five people may deny use that biological testing confirms. This has direct implications for how we interpret survey-based cannabis research.

The numbers in context

21.3% underreported at least one substance; moderate-to-large effect sizes observed for MDMA and cocaine discordance; cannabis use frequency and quantity predicted THC hair detection; years of use and time since last use did not predict detection

How the study worked

75 adults completed substance use questionnaires and provided hair samples analyzed by liquid chromatography-mass spectrometry for drug detection over the prior 3 months. Underreporting was defined as denying use while testing positive.

What this study cannot tell us

Small sample (75 participants), hair testing has its own limitations (can miss very light use, affected by hair treatments), no data on why participants underreported, single timepoint

How to read the evidence

Small cross-sectional study comparing two measurement methods; provides useful methodological insights but limited sample size

When this study was published

Published 2025

The bigger picture

If roughly one-fifth of research participants underreport substance use, studies relying solely on self-report may systematically underestimate cannabis use prevalence and misclassify participants, potentially skewing findings across the field.

Questions still open

Would underreporting rates differ in clinical vs. research settings? Do legal status and social stigma affect underreporting rates for cannabis specifically? Could combined self-report and biological measures become standard in cannabis research?

Common questions

How accurate are self-reports of cannabis use in research?

In this study, about 1 in 5 participants denied using a substance that their hair test detected, suggesting self-reports alone may underestimate actual use.

Does how often someone uses cannabis predict whether a hair test will find THC?

Yes. Self-reported frequency and quantity of cannabis use predicted THC detection in hair, but total years of use and time since last use did not.

Read the original research

Substance use assessment: comparing self-reports with objective data in a research setting.

Frontiers in public health, 13, 1628519

View on PubMed ↗Publisher’s paper ↗

Citation

Binkowska, Alicja Anna; Pałczyński, Piotr; Jakubowska, Natalia; Czarny, Jakub; Raczkowski, Michał; Brzezicka, Aneta. (2025). Substance use assessment: comparing self-reports with objective data in a research setting.. Frontiers in public health, 13, 1628519. https://doi.org/10.3389/fpubh.2025.1628519