People Who Underreported Marijuana Use Also Underreported Tobacco Use in Drug Surveys | Research

The takeaway

Comparing self-reported tobacco use with biological testing in 627 survey respondents showed that underreporting of marijuana strongly predicted underreporting of tobacco, suggesting survey data on drug use may be less accurate than assumed.

Read this if you want to understand how reliable self-reported drug use data is in research studies.

Self-report tobacco accuracy fell below the 90% threshold, and marijuana underreporting predicted tobacco underreporting

What the researchers found

Researchers compared self-reported tobacco use with saliva cotinine testing in 627 respondents from an epidemiological drug-use survey. While outright underreporting of tobacco was relatively rare, self-report sensitivity was below the 90% level established in prior reviews, even after adjusting for passive smoke exposure.

A key finding was that underreporting of marijuana use showed a strong association with underreporting of tobacco use. Race and ethnicity also predicted tobacco underreporting.

The study used Audio Computer-Assisted Self-Interview (ACASI) to minimize social desirability bias, and also tested hair, urine, and oral fluid for illicit drugs including marijuana, cocaine, heroin, and amphetamines.

Why it matters

If people who underreport marijuana use also underreport tobacco use, it suggests that drug survey data may systematically underestimate use in certain populations. This has implications for interpreting all cannabis research that relies on self-reported use data.

The numbers in context

627 respondents via multistage sampling. Self-report sensitivity was below the 90% threshold from prior reviews. Marijuana underreporting strongly predicted tobacco underreporting. Biological tests included saliva (cotinine), hair, urine, and oral fluid.

How the study worked

Cross-sectional epidemiological study using multistage sampling of 627 respondents. Participants completed household surveys via Audio Computer-Assisted Self-Interview (ACASI) and provided saliva samples for cotinine testing. Hair, urine, and oral fluid were tested for illicit substances. Self-report sensitivity was calculated by comparing reported use against biological confirmation.

What this study cannot tell us

The study cannot determine why respondents underreported. The single geographic sample may not generalize to other populations. Passive cotinine exposure complicates the comparison between self-report and biological testing. The study focused on tobacco reporting validity with cannabis as a predictor, not the reverse.

How to read the evidence

Cross-sectional survey with biological validation. Provides useful methodological insight but is limited to a single sample and geographic area.

When this study was published

Published in 2005. Survey methodology has continued to evolve, though the fundamental challenge of self-report accuracy in drug research remains.

The bigger picture

This study touches on a fundamental challenge in cannabis research: most studies rely on self-reported use data. If underreporting is systematic and correlated across substances, it could bias findings about the health effects of cannabis, tobacco, and other drugs.

Questions still open

How much does underreporting of cannabis use affect the conclusions of health research that relies on self-reported data? Would different survey methods reduce underreporting?

Common questions

How accurate are drug surveys?

This study found self-reported tobacco use accuracy was below 90% when compared against biological testing, and that patterns of underreporting were correlated across different substances. This suggests drug survey data may systematically underestimate use in some populations.

Why would someone underreport both marijuana and tobacco?

The study did not determine specific reasons, but suggests that factors driving underreporting (such as social stigma or desire to present oneself favorably) may apply broadly across substances rather than being substance-specific.

Read the original research

Tobacco-reporting validity in an epidemiological drug-use survey.

Addictive behaviors, 30(1), 175-81

View on PubMed ↗

Citation

Fendrich, Michael; Mackesy-Amiti, Mary Ellen; Johnson, Timothy P; Hubbell, Amy; Wislar, Joseph S. (2005). Tobacco-reporting validity in an epidemiological drug-use survey.. Addictive behaviors, 30(1), 175-81.