Is the NSDUH Survey Legitimate or a Scam?

The National Survey on Drug Use and Health (NSDUH) is a legitimate, federally funded survey conducted every year by the Substance Abuse and Mental Health Services Administration (SAMHSA), an agency within the U.S. Department of Health and Human Services. It is one of the largest and longest-running surveys of substance use, mental health, and treatment access in the country, with data going back to 1971. If you have been contacted by someone claiming to represent the NSDUH, or you have seen references to the survey and wondered whether it is real, the short answer is that it is a well-established government data-collection effort with strict confidentiality protections enforced by federal law. That said, there are real questions worth asking about how the survey works, how it protects your privacy, and how reliable its findings actually are.

What the NSDUH Is and Who Runs It

SAMHSA commissions the NSDUH to track patterns of drug and alcohol use, mental health conditions, and access to treatment across the civilian, non-institutionalized U.S. population aged 12 and older. The survey has been conducted annually since 1971, originally under the name “National Household Survey on Drug Abuse” before being renamed in 2002. It is the primary source the federal government uses to produce national and state-level estimates of substance use prevalence, and its data feeds directly into funding decisions, policy design, and public health planning.

The survey’s sample sizes are large. In a typical year, tens of thousands of individuals complete the NSDUH, drawn from a scientifically designed random sample of households across all 50 states and the District of Columbia. One study using NSDUH data from 2002 to 2012, for instance, drew on a sample of nearly 98,000 adolescents to examine substance use and teen pregnancy.

How the Survey Is Conducted

The NSDUH is primarily an in-person survey. A trained field interviewer visits your home after your household has been randomly selected. This is not a phone call from an unknown number or a link sent by email. The interviewer carries official credentials and explains the survey’s purpose and your rights before you agree to participate. Participation is voluntary.

The interview itself uses a combination of methods. Less sensitive questions, such as basic demographics, are conducted with the interviewer present using computer-assisted personal interviewing. For sensitive topics like drug use, alcohol consumption, and mental health symptoms, the survey switches to audio computer-assisted self-interviewing, known as ACASI. During ACASI sections, you wear headphones to hear the questions read aloud, tap your answers directly into a laptop, and the interviewer stays out of view of the screen. This setup is specifically designed to reduce embarrassment and encourage honest responses.

Field experiments conducted during the development of this approach found that ACASI increases reporting of drug use, particularly among young people. It also helps respondents who are less comfortable reading, since all questions are read aloud through the headphones. The system even lets respondents correct their own inconsistencies when the software flags them, without involving the interviewer.¹

Privacy Protections and Federal Law

One of the most common concerns people have when a stranger shows up at their door asking about drug use is whether their answers could be used against them. The NSDUH operates under the Confidential Information Protection and Statistical Efficiency Act of 2002 (CIPSEA), which legally prohibits the use of your responses for anything other than statistical purposes. Your individual answers cannot be shared with law enforcement, immigration authorities, insurers, employers, or anyone else. Researchers who work with NSDUH data only see de-identified datasets with no names, addresses, or other information that could be traced back to a specific person.²

These are not just promises. CIPSEA is a federal statute, and violations carry criminal penalties for government employees and contractors who breach confidentiality. The survey also operates under additional protections from SAMHSA’s own data-collection regulations. If a field interviewer asks you about illegal drug use, your admission cannot be subpoenaed by a court or disclosed to any outside party.

How to Tell If a Contact Is Actually From the NSDUH

Because the NSDUH involves strangers showing up at homes and asking sensitive questions, it is reasonable to wonder if a particular contact is genuine. Legitimate NSDUH field interviewers carry government-issued photo identification and can provide contact information for their supervisor and for SAMHSA itself so you can verify the visit. They will never ask for your Social Security number, bank account details, passwords, or any financial information. The survey does not collect that kind of data.

If someone contacts you by phone, email, or text claiming to represent the NSDUH and asks you to click a link or provide personal financial information, that is not the real survey. The NSDUH’s primary mode of contact is in person, though the survey did begin incorporating web-based data collection during and after the COVID-19 pandemic. If you have doubts, you can call SAMHSA directly to confirm whether your household was selected.

Online survey fraud is a growing problem in public health research more broadly. One study documented how a Facebook-recruited health survey received nearly 2,600 screening attempts and 950 completed responses within a single hour of launching an advertisement that had reached only about 125 real people, a volume that was overwhelmingly fraudulent.³ The NSDUH’s reliance on in-person household visits with verified interviewers is itself a guard against this kind of fraud, because it is much harder to fake a face-to-face visit with official credentials than to fabricate an online form.

What Happens If Minors Are Selected

The NSDUH surveys people as young as 12, which raises understandable questions about how the process works for minors. Before an adolescent can participate, the field interviewer must obtain explicit parental or guardian permission. The parent signs a consent form, and then the adolescent provides their own assent. Once the interview begins, ACASI ensures the teenager’s specific answers remain private even from the parent in the next room.

This protocol is thorough but has consequences for the data. Research comparing NSDUH youth substance-use estimates with those from school-based surveys like Monitoring the Future (MTF) and the Youth Risk Behavior Survey (YRBS) has found that NSDUH estimates tend to be lower. One reason is that the parental-consent requirement may discourage participation by adolescents with higher rates of substance use, since those teenagers (or their parents) may be less willing to engage with a government survey. Adolescents who are frequently absent from school or who have dropped out are also harder to reach through a household survey.⁴ SAMHSA has published reports directly comparing youth estimates across the NSDUH, MTF, and YRBS to help researchers understand where the numbers diverge and why.⁵

How Accurate Are the Numbers

The NSDUH’s data is based entirely on self-report, which is both its strength and its most criticized limitation. Even with ACASI and confidentiality guarantees, some people underreport their drug use. Researchers who study NSDUH data described it as providing “a relatively accurate source of information” in part because ACASI “protects individuals’ privacy and confidentiality and is expected to increase honesty in reporting.”⁶ But “relatively accurate” is not the same as perfectly accurate, and the gap between self-report and biological reality has been studied extensively.

A clinical validation study compared the NSDUH’s substance-use-disorder questions against structured diagnostic interviews (the gold standard for psychiatric diagnosis) in a sample of nearly 300 adults and adolescents. The study found “fair to moderate” agreement overall, with somewhat better agreement for dependence diagnoses than for abuse diagnoses, and better agreement among adults than adolescents.⁷ “Fair to moderate” is a candid assessment rather than a glowing one, and it means the NSDUH’s diagnostic questions are useful for population-level estimates but imperfect for any individual case.

When researchers have tried to estimate how much underreporting affects the NSDUH’s headline numbers, the adjustments can be substantial. A simulation study estimated that the NSDUH’s self-reported cannabis prevalence of about 10.5% in 2018 would rise to roughly 16% after correcting for underreporting, and self-reported cocaine prevalence of 0.8% would jump to somewhere between 2.7% and 5.5%.⁸ The estimated self-report sensitivity varied dramatically by substance: around 78% for cannabis (meaning most users reported their use) but as low as 14% to 22% for cocaine (meaning the vast majority of cocaine users did not report it).⁹

These are not flaws unique to the NSDUH. Every self-report survey of illegal behavior faces the same problem. The NSDUH’s advantage is that it uses a consistent methodology year after year, which means even if the absolute numbers are undercounts, the trends over time are meaningful. If the NSDUH shows cannabis use doubling over a decade, the real increase may be somewhat different in magnitude, but the direction and general scale of the trend are reliable.

What Biomarker Studies Show About Self-Report

Some studies have gone beyond surveys and compared what people say to what biological tests detect. The results are sobering and help put the NSDUH’s limitations in context, even though these studies were not conducted within the NSDUH itself.

In a New York City health survey that paired self-reported drug use with urine testing, the self-report numbers were actually higher than what the urine tests found: about 14% of participants reported past-year use of heroin, cocaine, or prescription drug misuse, while urine tests detected drug metabolites in only about 4%.¹⁰ That might seem backwards, but urine tests have a narrow detection window of just a few days for most substances, while self-report covers an entire year. A urine test on a single day cannot capture whether someone used cocaine six months ago. This is a reminder that biological testing and self-report are measuring different things and are not straightforward to compare.

A study of young adult opioid users compared urine testing and hair testing against self-reported drug use and found that each method had different strengths. Urine testing confirmed a higher proportion of self-reported heroin and marijuana use than hair testing did, while hair testing was better at detecting reported cocaine use.¹¹ No single method catches everything. When researchers in another study compared self-reported substance use against urine tests in both a research setting and a primary health care clinic, the agreement was only fair for opioids and slight for cannabis and methadone, with self-report sensitivity for some substances falling as low as 10% to 20%.¹²

The takeaway is not that the NSDUH’s data is unreliable, but that self-report data across all survey contexts tends to undercount drug use. The NSDUH’s ACASI design is one of the best available methods for minimizing that undercount in a large national sample, even though it cannot eliminate it entirely.

How NSDUH Data Shapes Policy

The practical stakes of the NSDUH’s accuracy are high because the data is used to allocate billions of dollars in federal funding. State-level NSDUH estimates of substance use and mental illness inform how grants from SAMHSA and other agencies are distributed. Research examining opioid-related federal grant funding found significant problems with how state prevalence rates, derived in part from NSDUH data, were used as benchmarks. The study found that roughly one-sixth of federal opioid funds, totaling about $1.5 billion, would have needed to be reallocated to achieve equal severity-adjusted funding across states. Less populous states were typically overfunded relative to their burden, in part because NSDUH estimates for small states have wider margins of error.¹³

This highlights a real tension. The NSDUH is the best available tool for state-level substance-use estimates, but its precision drops in smaller populations. Using those less precise numbers to drive funding decisions can create meaningful mismatches between need and resources. Researchers and policymakers are aware of the issue, and there is ongoing work to improve the statistical models that generate state-level estimates, but it remains an active challenge.

Tracking Long-Term Trends

Where the NSDUH really earns its keep is in tracking how substance use patterns shift over years and decades. Because the survey uses a consistent design and asks the same core questions year after year, it produces one of the most reliable records of changing drug-use behavior in the United States. Researchers analyzing NSDUH data alongside older survey data found that between 2008 and 2022, the rate of people reporting past-year cannabis use roughly doubled, while the total number of days of use reported across the population more than tripled. Perhaps most striking, there was a 15-fold increase since 1992 in the rate of people reporting daily or near-daily cannabis use.¹⁴

These kinds of findings would be essentially impossible without a survey like the NSDUH running consistently for decades. No other data source captures this breadth of substance-use behavior at a national scale with this level of methodological continuity. Emergency room records, treatment admissions, and overdose death counts are all useful, but they only capture the most visible and severe end of the spectrum. The NSDUH fills in the picture for the much larger population of people whose substance use never results in a hospital visit or an arrest.

When the Survey Itself Changes

One complication researchers watch for is what happens when the NSDUH’s methodology changes. The biggest recent disruption was COVID-19. In 2020, in-person household visits became impractical, and SAMHSA shifted to a multimodal approach that included web-based data collection. Any time a survey changes how it collects data, comparisons to previous years become harder to interpret. Did a jump in reported use reflect a genuine increase, or did people answer differently on a web form than they would have in person with ACASI? SAMHSA acknowledges these breaks in the data series and publishes guidance on which year-to-year comparisons are valid and which are not.

The 2002 rename from “National Household Survey on Drug Abuse” to “National Survey on Drug Use and Health” was itself accompanied by methodological changes that created a data break, meaning pre-2002 and post-2002 estimates are not directly comparable without adjustments. Researchers who work with this data routinely account for these discontinuities, but casual users of NSDUH statistics sometimes miss them.

Who the NSDUH Misses

The NSDUH surveys the civilian, non-institutionalized population. That exclusion matters. People who are incarcerated, living in residential treatment facilities, hospitalized, serving on active military duty, or experiencing homelessness without a fixed address are not captured. These are precisely the populations with some of the highest rates of substance use and mental illness. The NSDUH’s portrait of American drug use is therefore a portrait of the housed, non-incarcerated population, which is important to keep in mind when interpreting its findings.

Similarly, the survey’s household-based design means it depends on people being willing to open their doors and spend time answering questions. Response rates for the NSDUH, like most large household surveys, have been declining over the years. Lower response rates increase the risk that the people who agree to participate are systematically different from those who refuse, which could bias the estimates in ways that are hard to measure. SAMHSA uses statistical weighting to adjust for known demographic differences between responders and non-responders, but no weighting scheme can fully correct for unknown sources of bias.

None of these limitations make the NSDUH illegitimate. They make it a tool with a well-understood set of strengths and blind spots, which is the most you can ask of any survey. The researchers who use NSDUH data most frequently are also its most clear-eyed critics, and the survey’s methodology has been continuously refined over its five-decade history in response to exactly these concerns.