NSDUH Frequently Asked Questions

NSDUH samples all people living in the United States who are:

• Age 12 or older,
• Not living in institutions such as prisons or nursing homes,
• Not military personnel on active duty,
• Not experiencing homelessness with no fixed address, and
• Able to answer the questionnaire in either English or Spanish.

More information on data collection is available from the annual Data Collection Final Report.

NSDUH is conducted using stratified sampling with multiple levels of selection. Households are chosen geographically. That means that a particular area is randomly selected, and then households are selected within that area. Basic information is first collected from selected households, and then 0, 1, or 2 household members are chosen to complete the full survey. Respondents are chosen so that about a quarter are age 12 to 17, another quarter are age 18 to 25, 15 percent are age 26 to 34, 20 percent age 35 to 49, and 15 percent age 50 or older.

The geographic selection process is complex. The United States is first broken up into 750 sub-state regions that do not cross state lines and have roughly equal populations. From each of these regions, census tracts or collections of tracts are chosen using systematic sampling. Tracts are sorted by socioeconomic status, urban area, percent of the population that is White, and percent of the population that is Hispanic before they are systematically selected. This ensures a broad variety of census tract types are chosen.

Within each census tract or tracts, a “segment” of adjacent block groups or blocks are randomly selected. Segments are formed to contain a minimum of 150 or 250 dwelling units (DUs) in urban areas, and 100 or 200 DUs in rural areas, depending on the state. Eight segments are chosen for interviewing in each of the 750 regions every year.

More detail on sampling procedures, including changes made over the past two decades, can be found in the annual NSDUH Sample Design Report.

The Federal Government has conducted the survey in various forms since 1971, but there have been major changes over the years. For example, before 1990 the survey was only administered every 2 to 3 years and had a very small sample size compared to later iterations. In 1999 the survey shifted from paper-and-pencil data collection to computer-assisted interviewing, and the sample design expanded to include all 50 states and the District of Columbia. In 2002, the name of the survey changed from the National Household Survey on Drug Abuse to the National Survey on Drug Use and Health.

Due to the recent addition of online data collection, 2021 and 2020 data cannot be compared to previous years or each other. There is also a hard break in comparability between 2002 and previous years. Many indicators are not comparable before and after 2008 and 2015 as well. Please review survey documentation for your years of interest before conducting your own multi-year trend analysis.

For historical research, the concatenated public use file (PUF) includes all public-use variables between 2002 and 2019 that can be used in trends.

Generally, the NSDUH sample size is around 68,000 individuals, although disruptions in 2020 reduced the number of households that could be interviewed. See the table below for exact numbers of completed interviews in the past few years.

*In 2020, field operations were interrupted due to the Coronavirus 2019 pandemic, and responses were collected almost exclusively in Quarters 1 and 4

NSDUH uses a stratified sampling design, with the United States divided into 750 sampling regions. While samples are taken from every region each year, many counties (especially sparsely populated counties) are not included in every year.

SAMHSA also uses small area estimation (SAE) methods, in which substate-level NSDUH data are combined with other data from smaller geographies to produce some indicators at the substate level. Some of these substate regions consist of a single county. There are approximately 400 substate regions across the United States. The definitions and maps of the regions are available on the SAMHSA website, along with the estimates for the regions.

Some high-level NSDUH results are available from a recent infographic report.

If you are looking for common, national statistics, try the NSDUH Annual National Report.

For basic Information on survey methodology, try the Methodological Summary and Definitions report. More detailed methodology reports are available from the Methodological Research Books for each year.

For detailed national statistics, try the annual NSDUH Detailed Tables.

To create simple custom tables online, try the Data Analysis System (DAS).

For basic state and sub-state statistics, try the annual State or Substate Specific Tables, or regional Behavioral Health Barometer reports. The tables can also be viewed with this interactive tool.

To access microdata for detailed analyses, see if the public use files (PUFs) meet your needs.

Finally, if you need to access the restricted version of the microdata, you can apply for access to restricted use NSDUH data through the interagency Standard Application Process (SAP) portal.

Starting largely in 2021, NSDUH became a fully mixed-mode survey, with many responses from both web and in-person administration. Repeated analyses showed that the web responses are not comparable to the in-person responses and that the comparability is not consistent in a way that we can fully account for. Because of this, data from 2021 should not be compared to estimates from 2019 or earlier.

Estimates from 2021 should also not be compared to estimates from 2020 or even from just quarter 4 of 2020. The reasons are a bit complex. First, the 2020 NSDUH is missing two quarters of data. Tests of data from before 2020 show that estimates based on just quarters 1 and 4 are not comparable to estimates based on the entire year. This indicates that 2020 estimates should not be compared to 2021.

Moreover, while the methodology for Quarter 4 of 2020 is mostly the same as for the 2021 NSDUH, data in Quarter 4 of 2020 mostly consists of web interviews. In 2021, the proportion of in-person interviews increased over the year as county-level restrictions related to COVID-19 were gradually lifted, and field interviewers were able to follow up in more areas. These mostly single-mode estimates from quarter 4 should not be compared to the multi-mode estimates from 2021.

A full description of the analyses can be found in chapter 6 of the 2021 Methodological Summary and Definitions report.

Because of the coronavirus disease 2019 (COVID-19) pandemic, major changes were made to the methods used in data collection in 2020. There is no way to separate out the true changes in behavior from the changes due to the new methodology.

The main methodological changes were:

• Almost no data collection from mid-March through September 2020,
• Introduction of web data collection in October 2020 with very limited in-person data collection and
• Additions to the questionnaire beginning in October 2020.

The 2020 NSDUH is missing two quarters of data. Tests of data from before 2020 show that estimates based on just quarters 1 and 4 are not comparable to estimates based on the entire year. This indicates that 2020 estimates should not be compared to previous years. Repeated analyses have shown that web responses are not comparable to in-person responses, and that the comparability is not consistent in a way that we can fully account for. For these reasons, it is not recommended to compare any estimates from 2020 to estimates from other survey years.

A full description of the analyses that were conducted can be found in chapter 6 of the 2021 Methodological Summary and Definitions report.

Substance use disorder (SUD) was assessed by DSM-IV criteria in 2019 and earlier but was assessed by DSM-5 criteria for all of 2020—both quarter 1 and quarter 4. Although there were major changes to the survey over the course of that year, questions assessing SUD were the same for both quarters. This continues in 2021 and beyond.

Measures of Major Depressive Episodes (MDEs) are based on DSM-5 and have been since 2017.

Any mental illness (AMI) and serious mental illness (SMI) are currently based on a model that is calibrated to DSM-IV, but plans are underway to change to DSM-5 in the next few years.

Estimates for pain reliever use prior to 2015 are not comparable to estimates in 2015 and later. NSDUH underwent a redesign in 2015, and questions about pain reliever use were updated.

More details on these changes are available from the 2014 and 2015 Redesign Changes report.

Due to changes in measures and methodology for mental illness, mental illness estimates from 2008 through 2019 should not be compared to prior years. Because of overall methodological changes, data from 2020 or 2021 should not be compared to prior years or each other.

Detailed information on changes specific to mental illness estimates can be found in section 3.4.7 of the 2019 Methodological Summary and Definitions document.

A full analysis of 2020 and 2021 comparability can be found in chapter 6 of the 2021 Methodological Summary and Definitions report.

In 2020, the criteria for a substance use disorder (SUD) in NSDUH was updated from DSM-IV to DSM-5. These changes altered who could be considered to have an SUD.

No. Regions are defined for NSDUH by state governments and differ between years. Definitions of the regions for each set (both maps and descriptions) can be found in the following reports:

2016-2018 NSDUH Substate Region Definitions

2014-2016 NSDUH Substate Region Definitions

2012-2014 NSDUH Substate Region Definitions

2020 estimates cannot be pooled or combined with other years of data. This is because of major changes in survey mode (i.e., web versus in-person data collection) caused by the Coronavirus 2019 pandemic. In 2019, data collection was exclusively in person, but in 2020, most respondents in Quarter 4 answered via the newly introduced web survey. People answer differently in person and via the web, both because different people answer the survey and because they think about the questions differently in different contexts. This is called a mode effect.

The mode effect is not the same across different groups. As a result, comparing averages that include 2019 and 2020 data does not produce valid results. The differences are inconsistent across measures, so they cannot be accounted for with weighting.

Because of changes to the prescription drug use questions in 2015 and context effects, we also caution against pooling 2015 data with prior years, and pooled estimates were not published.

Many (but not all) NSDUH estimates from 2022 and 2023 are comparable with estimates from 2021 as long as updated 2021 estimates presented in the 2022 Detailed Tables are used. When the NSDUH moved from in-person-only survey administration to a multi-mode approach (with both in-person and web modes), it was determined that the mode of administration affected the resulting estimates. The proportion of interviews that were completed in person was lower in 2021 than it was in 2022, likely due to the COVID-19 pandemic. In response to this, estimates for 2021 were updated using adjusted weights to account for changes in the proportions of web vs. in-person interviews. The resulting updated 2021 estimates set the contributions of respondents for each mode to the same proportions as the 2022 estimates, thus removing the potential for bias in estimates of change due to shifts from the web to in-person interviewing. For more details, see section 3.3.3 of the 2022 NSDUH Methodological Summary and Definitions.

SAMHSA has no plans to re-release the 2021 National Report and Detailed Tables, and the 2021 NSDUH National Report and 2021 Detailed Tables do not present this updated data. Therefore, these documents should not be used to compare estimates between 2021 and 2022. The updated 2021 estimates are available in the 2022 Detailed Tables. Consequently, estimates for 2021 listed in the 2022 Detailed Tables may differ slightly from these previously released estimates.

NSDUH published data does not differentiate between clinical and recreational marijuana use. However, NSDUH does have questions asking if any or all of the respondent’s marijuana or cannabis use was recommended by a doctor or other health professional. These results can be accessed in the public use file (PUF) and through the Data Analysis System (DAS), using the variables MJANYMEDYR and MJALLMEDYR starting with the 2022 data year and MEDMJYR and MEDMJALL for earlier data. NSDUH does not distinguish between marijuana use permitted by state laws and use prohibited by state laws.

In the public use file (PUF) and single-year NSDUH data in the Data Analysis System (DAS), no continuous age variables are available. The categorical variables for age in the PUF are CATAGE, CATAG2, CATAG3, CATAG6, and CATAG7. DAS also allows users to collapse categories using the “recode” function. Please see the PUF Codebook for more details about the age variables.

Multi-year files in DAS have a number of additional age variables, which can be found under the “Background: Age” variable category.

The only race/ethnicity variable that is available in the Public Use File (PUF) or single-year NSDUH data in the Data Analysis System (DAS) is NEWRACE2, which includes the following racial/ethnic categories:

• Non-Hispanic White
• Non-Hispanic Black or African American
• Non-Hispanic Native American or Alaska Native
• Non-Hispanic Native Hawaiian or other Pacific Islander
• Non-Hispanic Asian
• Non-Hispanic more than one race
• Hispanic

With the multi-year files in DAS, there are 5 other available variables. These can be recoded to collapse categories if desired. The table below lists these variables:

SAMHSA has several resources to find statistics on substance use, substance use disorders, mental health, and treatment, including:

• Data Analysis System (DAS): https://datatools.samhsa.gov/
• Overview page that describes NSDUH reports: https://www.samhsa.gov/data/data-we-collect/nsduh-national-survey-drug-use-and-health
• Key Substance Use and Mental Health Indicators in the United States – Results from the 2022 National Survey on Drug Use and Health Annual Report: https://www.samhsa.gov/data/report/2022-nsduh-annual-national-report

• Detailed tables: https://www.samhsa.gov/data/report/2022-nsduh-detailed-tables  

There are race/ethnicity variables in the recoded demographics section of the public-use file (PUF), but they do not include the detailed categories for Asians and Hispanics. All the detailed categories for Hispanics have been collapsed to "Hispanic" for the PUFs, which includes Cuban, Mexican, Puerto Rican, etc. You will need to access the restricted files in RDAS for the detailed categories. To access RDAS, select Analyze > Restricted Data from the toolbar.

The NSDUH 2-year restricted data datasets are generally sequential – 2006-2007, 2008-2009, 2010-2011, 2012-2013, 2014-2015, and 2016-2017. NSDUH underwent a partial redesign in 2015.

There are several measures that “broke trends” that year, so estimates from 2015 and later are no longer comparable to their 2014 and earlier counterparts. Because of this, an extra dataset was created for the restricted data for 2015-2016, but there is no 2-year dataset for 2013-2014.

After the 2015 redesign, the datasets include 2015-2016, 2016-2017, 2017-2018, and 2018-2019. There was another trend break in 2020, and no restricted dataset was created for that year. The sequential nature of the datasets begins again with 2021-2022.

No NSDUH data is available from before 2002 in DAS, restricted or otherwise.        

SAMHSA has several resources to get information on treatment rates from the National Survey of Drug Use and Health (NSDUH): There are sections on adult and youth mental health treatments as well as substance use treatment in the most recent annual national report. A few state and regional variables on treatment are available in the small area estimation tables. Sections 5-7 of the NSDUH detailed tables also include treatment information.

If none of these sources meet your needs, custom NSDUH tables can be created in the Data Analysis System.     

NSDUH geographic information is not available in the public-use data file, but some information for states and regions can be found in multi-year data. This is available in both published tables and interactive data tools.

Selected estimates for states are available from the NSDUH state reports. These are generally based on models of data concatenated over two years. These data are also available in an interactive data tool, which can be used both to visualize and download the information by year and state.

Additionally, there are Behavioral Health Barometer reports that summarize key mental health and substance use indicators by state and compare them to regional and national averages.

The Data Analysis System (DAS) also allows users to analyze restricted NSDUH datasets, concatenated over multiple years, to use some geographic identifiers in simple crosstabulations. The table below lists some of the geographic variables that are available in RDAS.

NSDUH does not include geographic indicators like county or state in public-use datasets. Full restricted single-year NSDUH data files are available only through the Research Data Center (RDC). If you need to access detailed geographic data, you can apply for access to restricted-use NSDUH data through the interagency Standard Application Process (SAP) portal.

NSDUH data for counties with a population of 100,000 or greater are available through the Research Data Center (RDC). However, because of our geographic sampling methods, some counties may not have data in a particular year. Data for some larger counties is also available in several more easily accessible forms.

A few basic estimates for some large counties are available from the NSDUH substate reports. These are based on models of data concatenated over three years. Smaller counties are clustered to produce substate regional estimates.

These data are also available in an interactive data tool, which can be used both to visualize and download the information by year and substate region.

The  Data Analysis System (DAS) allows users to analyze restricted NSDUH datasets, concatenated over multiple years, to use some geographic identifiers for simple crosstabulations. While state-level information is available from the 2-year NSDUH DAS files, only the 10-year files include substate-level indicators. The table below lists some geographic variables available in DAS.

NSDUH does not include geographic indicators like county or state in public-use datasets. Full restricted single-year NSDUH data files are available only through the RDC. If you need this detailed geographic data, you can apply for access to restricted-use NSDUH data through the interagency Standard Application Process (SAP) portal.

NSDUH began reporting on national opioid misuse beginning in 2015. Opioid misuse estimates from 2020 should not be compared to earlier estimates due to the methodological changes in the survey. Similarly, 2021 numbers are not comparable to 2020 or earlier for similar reasons.

The 2019 annual report on Key Substance Use and Mental Health Indicators includes national trends in opioid misuse from 2015 to 2019 (figure 25, page 25) and opioid use disorder (figure 44, page 40) by age. In the 2019 detailed tables, a number of trend tables of opioid use and opioid use disorder from 2015 to 2019 are available in section 7.

Additionally, the 2023 Companion Infographic Report compares national estimates of opioid and Fentanyl misuse, including illegally made fentanyl (IMF), between 2021, 2022, and 2023 (page 8). Note that IMF is sometimes present in products that are sold as heroin or in counterfeit prescription drugs. People who use IMF are often not aware they are doing so. Because of this, estimates of IMF use are likely underestimates. The 2023 Detailed tables also include information on opioid use disorder for 2022 and 2023. 

Estimates of opioid misuse and opioid use disorder in the past year by state in 2021-2022 are available in the Interactive NSDUH State Estimates tool, as well as trends in pain reliever misuse and pain reliever use disorder for states and substates. These data can also be found in the form of tables and maps in the state estimates and substate region estimates.

Immigration status information is only available in the restricted data. For simple cross-tabulations, you can use the multi-year files in the Data Analysis System (DAS), which allows users to analyze restricted NSDUH datasets concatenated over multiple years. IRIMMBORNUS is the best variable for immigration status; it indicates whether the respondent was born in the United States or not. Those born in US territories, including Puerto Rico, self-identified as to whether they were born in the United States or not.

For more detailed analysis, you can apply for access to restricted use NSDUH data through the interagency Standard Application Process (SAP) portal. However, disclosure controls may limit the analyses available.

The following information is available by state for the years 2013 to the present:

• marijuana use in the past year,
• marijuana use in the past month,
• perceptions of the great risk of smoking marijuana once a month,
• first use of marijuana, and -illicit drug use disorder in the past year.         

The NSDUH data collection page has links to the Annual, State, and Substate Reports by year.

The NSDUH reports are located on the SAMHSA archive: https://www.samhsa.gov/data/data-we-collect/nsduh-national-survey-drug-use-and-health.

Three annual reports for the 2017 NSDUH have been released and are available at: https://www.samhsa.gov/data/nsduh/reports-detailed-tables-2017-NSDUH

• 2017 Key Substance Use and Mental Health Indicators Report
• 2017 NSDUH Detailed Tables
• 2017 Methodological Summary and Definitions Report

SAMHSA’s NSDUH Reports page includes annual, state, and substate reports in addition to detailed tables, data reviews, and methodology reports. The NSDUH Reports page is available at: https://www.samhsa.gov/data/data-we-collect/nsduh-national-survey-drug-use-and-health

The National Survey on Drug Use and Health (NSDUH) collects highly sensitive information from individual respondents, including data on substance use and mental health concerns. Ensuring respondent confidentiality is crucial to encouraging participation in the survey, and responses are protected under the Confidential Information Protection and Statistical Efficiency Act (CIPSEA).

To ensure that no individual respondent can be identified from his or her responses, NSDUH public-use files (PUFs) have been treated with a number of disclosure-avoidance methods. For NSDUH, these include, but are not limited to:

• The public-use data do not include any geographic identifiers, including state.
• The public-use data include only collapsed age categories and do not include continuous age variables (including age at first use of substances).
• Some variables such as income have been recoded to courser levels than those found in the restricted data.

Other demographic variables (i.e., immigrant status) have also been removed to protect the confidentiality of respondents.

When addressing suppression, sample size, variance, and design effects are all considered. Sometimes, that means that estimates for larger groups are suppressed while estimates for smaller groups are available.

The Statistical Inference Report outlines procedures for when suppression is applied. Chapter 9 explains these procedures, and they are summarized in Table 9.1. Sample codes for identifying unreliable output in SUDAAN®, Stata®, SAS®, R, and SPSS® are available in Appendix A. A shorter explanation is also available in the annual Methodological Summary and Definitions (MSD), chapter 3.2.

This Statistical Inference Report outlines procedures for identifying unreliable estimates. Chapter 9 explains these procedures, and they are summarized in Table 9.1 (pg. 66). Estimates that do not meet the criteria in the guidelines should not be reported or used. Sample codes for identifying unreliable output in SUDAAN®, Stata®, SAS®, R, and SPSS® are available in Appendix A (Exhibits A.2.3, A.3.3, A.4.3, A.5.3, and A.6.3, respectively).

A shorter explanation is also available in the annual Methodological Summary and Definitions (MSD), chapter 3.2.

This Statistical Inference Report outlines procedures for calculating statistical significance and confidence intervals in chapters 7 and 8. Sample codes in SUDAAN®, Stata®, SAS®, and R are available in Appendix A. A shorter explanation is also available in the annual Methodological Summary and Definitions (MSD), chapter 3.2.

The annual Methodological Summary and Definitions (MSD) report includes:

• A basic description of the survey,
• A discussion of statistical measures and measurement,
• Special topics for the given year,
• An overview of alternate sources of drug use and mental health data, and
• A glossary of key terms

Other methodological reports can be found as part of the annual Methodological Resource Book (MRB) collection. The MRB generally includes:

• The NSDUH questionnaire, including screener specifications, show cards, and pill images;
• The Field Interviewer Manual;
• The Statistical Inference Report, with guidelines and sample code for analyzing NSDUH data;
• The Sample Design Report, which gives details about the sampling design;
• The Sample Experience Report, which details how the sampling design was implemented in the field;
• Reports on the sampling weight calibration;
• The Editing and Imputation Report; and
• The Data Collection Final Report.

Reports are released as they are produced, so more recent MRBs may not yet include all reports.

NSDUH public use files (PUFs) are available in SAS, SPSS, Stata, ASCII, Delimited, and R formats. Setup code is also available in SAS, SPSS, and Stata formats to read in the file and apply formatting information.

There are race/ethnicity variables in the recoded demographics section of the public-use file (PUF), but they do not include the detailed categories for Asian and Hispanic. For instance, all the detailed categories for Hispanics have been collapsed to "Hispanic" for the PUFs, which includes Cuban, Mexican, Puerto Rican, etc. You will need to access the restricted files in the Data Analysis System (DAS) for the detailed categories.

Starting in quarter 4 of 2020, NSDUH included several questions directly or indirectly related to the Coronavirus 2019 (COVID-19) pandemic. These questions continued to be asked in the 2021, 2022, and 2023 NSDUH questionnaires. Other questions were altered in response to the pandemic.

One questionnaire change due to COVID-19 is the addition of virtual treatment to several questions. NSDUH added the option of virtual treatment -- over the phone, by email, or through video calling -- to several questions on treatment type.

Additionally, people who answered yes to questions about suicidal thoughts and behavior were newly asked the follow-up question, “Was this because of the COVID-19 pandemic?”

A special “Covid-19” section with eight questions was also added to the questionnaire. This module included questions about the effects of the pandemic on mental health, alcohol consumption, drug use, financial worries, homelessness, and access to mental health, substance use, and other medical treatment. Details on question-wording can be found in the COVID-19 module of the questionnaire.

Starting with the 2022 NSDUH, questions were added to the emerging issues section of the questionnaire to assess the use of illegally made fentanyl (IMF). New measures were produced starting in 2022 that included the use of IMF. These new measures included the use and misuse of fentanyl (i.e., pharmaceutical fentanyl or IMF) and the use and misuse of opioids, including IMF. See sections 3.4.1 and 4.4 of the 2023 NSDUH Methodological Summary and Definitions for details on data collection and management for these measures, as well as a discussion of which summary measures of opioids include IMF and which do not.

Estimates of IMF use are available throughout the 2023 Detailed Tables. In particular, Tables 1.69 and 1.107 present estimates for any fentanyl use (i.e., any prescription fentanyl use or IMF use), fentanyl misuse (i.e., prescription fentanyl misuse or IMF use), and IMF use in the past year by demographic, geographic and socioeconomic characteristics. It should be noted that because people who used IMF may have been unaware that they used it, caution must be taken when interpreting estimates of IMF use; these estimates are most likely an underestimate of true IMF use.

In addition, NSDUH provides current and historical data on the use and misuse of prescription fentanyl. Data for 2020 and 2021 are not comparable to previous years or to each other.

Information on prescription fentanyl use and misuse can be found in the Data Analysis System (DAS) or in the following detailed tables:

*Due to the introduction of web response in 2020, these data are not comparable to previous years.
**Due to question changes related to fentanyl, as well as the introduction of mixed-mode data collection in 2021, these data are not comparable to previous years.

NSDUH does not collect information on carfentanyl xylazine, or ayahuasca use.

NSDUH does not currently collect data on people denied treatment. For people who felt that they needed substance abuse treatment but did not get it, 2023 detailed table 5.35 shows the reasons that people indicated they did not receive treatment. Additionally, 2023 detailed tables 6.34-6.36 show the reasons that adults did not receive mental health treatment among those who perceived that they needed it.

NSDUH does not currently collect data by a specific program. Other SAMHSA surveys may have the information you are looking for.

The National Mental Health Services Survey (N-MHSS) surveys public and private mental health specialty facilities. This includes psychiatric hospitals; nonfederal general hospitals with separate psychiatric units; U.S. Department of Veterans Affairs medical centers; residential treatment centers for children and adults; community mental health centers; outpatient, day treatment, or partial hospitalization mental health facilities; and other multi-setting mental health facilities.

The Mental Health Client-Level Data (MH-CLD) gives information on patients of some mental health treatment facilities. It provides information on their mental health diagnoses, mental health treatment services and outcomes, and demographic and substance use characteristics. Only those in facilities that report to individual state administrative data systems are included.

The National Survey of Substance Abuse Treatment Services (N-SSATS) is an annual census of treatment facilities. Information is collected on the location, organization, structure, services, and utilization of substance abuse treatment facilities in the United States. 

No. In order to protect the privacy of our respondents this is not allowed. Moreover, using the original variables will not produce better-quality research. NSDUH has developed sophisticated, scientifically sound procedures for data recoding and imputation. The imputed and/or recoded values are almost always more reliable for analysis than the raw data.

If you have a research project specifically focused on non-response, you can apply for access to restricted-use NSDUH data through the interagency Standard Application Process (SAP) portal.

The weights were adjusted for 2021 to make the estimates comparable to 2022 estimates.

When the NSDUH began including web responses along with in-person responses, it was determined that the mode of survey administration had an effect on the resulting estimates. In 2021, a smaller proportion of interviews were completed in person than in 2022, likely because of the COVID-19 pandemic. To account for mode differences between 2021 and 2022, estimates for 2021 were updated using adjusted weights.

The updated 2021 estimates used weights to set the contributions of respondents for each mode to the same proportions as the 2022 estimates, thus removing the potential for bias in estimates of change. See section 3.3.3 of the 2022 NSDUH Methodological Summary and Definitions for more details.

The 2021 detailed tables use the unadjusted estimates, while the 2022 detailed tables use the adjusted estimates to allow comparisons.

Having an SUD “in the past year” means the 12 months prior to taking the survey, not the calendar year. People who did not have an SUD in the past year may have had one in the past, even if they currently do not have problematic substance usage.   

Individual observations are weighted so that the sample can represent the civilian, non-institutionalized population in the United States. The person-level weights in NSDUH, including for youth, are calibrated by adjusting for sample design, nonresponse, and break-off and matching to known population estimates (or control totals) obtained from the U.S. Census Bureau. Additional pair weights are also created to analyze households with two respondents.

NSDUH collects information from residents of households and noninstitutional group quarters (e.g., shelters, rooming houses, dormitories) and from civilians living on military bases. The survey excludes people experiencing homelessness who are not staying in shelters, military personnel on active duty, and residents of institutional group quarters, such as jails and hospitals. Therefore, our weighted total population tables (detailed tables, section 9) and Census population may not be the same.

More information on weights is available in the Person-Level Sampling Weight Calibration Report and the Questionnaire Dwelling Unit-Level And Person Pair-Level Sampling Weight Calibration report.

In general, states cannot be simply ranked using state estimates alone. To make direct comparisons between states (i.e., to determine whether they are mathematically different from one another), p-value tables should be used. P-value tables also allow comparisons to state and regional averages.

In our online map tool and published maps, the colors (quintiles) that are shown should not be used to directly compare states.  While maps are technically based on a ranked list, these categorizations do not indicate a statistically significant difference between any two particular states.

A quintile is a statistical value that represents one-fifth (20%) of the range of values in a data set. For example, the first map quintile (color) includes the 10 states with the lowest prevalence of a particular outcome; the second includes the next 10, and so on. The purpose of these maps is to give a broad picture of how outcomes are distributed nationally; there are no direct comparisons made between individual states. For instance, a map might show that the highest quintile states are clustered in one particular region of the country.

The published NSDUH state estimates are located here:

• State Data Releases: https://www.samhsa.gov/data/nsduh/state-reports
• Interactive NSDUH State Estimates: https://datatools.samhsa.gov/saes/state
• Barometer Reports: https://www.samhsa.gov/data/nsduh/state-reports-barometers

   

The restricted data in DAS can produce state estimates with 2 or more years of pooled NSDUH data. The SAMHSA website also publishes a limited number of state estimates.

In 2022, major revisions were made to the drug treatment and mental health service use sections of the survey. These changes better reflect the current state of treatment services and their delivery. The new measures are not comparable to the measures used in 2021. So, these estimates were based on 2022 data alone.