mental health services,
substance abuse treatment,
Date of Collection:
Unit of Observation:
The civilian, noninstitutionalized population of the
United States aged 12 and older, including residents of
noninstitutional group quarters, such as college dormitories, group
homes, civilians dwelling on military installations, and persons with
no permanent address, such as homeless people and those living in
shelters and rooming houses.
Data Collection Notes:
Data were collected by Research Triangle
Institute, Research Triangle Park, NC, and prepared for release by
National Opinion Research Center, Chicago, IL.
variables, statistical imputation was performed following logical
imputation to replace missing responses. These variables are
identified in the codebook as "...LOGICALLY IMPUTED" and
"...imputed" for the logical procedure or by the designation
"IMPUTATION-REVISED" in the variable label when the statistical
procedure was also performed. The names of statistically imputed
variables begin with the letters "IR". For each imputation-revised
variable there is a corresponding imputation indicator variable that
indicates whether a case's value on the variable resulted from an
interview response or was imputed by the hot-deck technique, which is
described in the codebook.
The "basic sampling weights" are
equal to the inverse of the probabilities of selection of sample
respondents. To obtain "final NHSDA weights," the basic weights were
adjusted to take into account dwelling unit-level and individual-level
non-response and further adjusted to ensure consistency with
population projections from the United States Bureau of the
To protect the confidentiality of respondents, all
variables that could be used to identify individuals have been
encrypted, collapsed, or deleted. These modifications should not
affect analytic uses of the data.
Family and personal income range
variables in the 1994-B questionnaire were constructed from two
imputation-revised variables that were deleted from the public use
dataset and codebook due to confidentiality issues. Users are advised
that the recoded income range variables cannot be replicated from the
imputation-revised variables (IRFINC1 and IRPINC1) contained in the
public use file.
(a) Only the 1994-A file should be used
to analyze trend data from past NHSDAs. (b) Prevalence
estimates for 1994 should be derived only from the 1994-B survey,
using the imputed data rather than raw data from the
questionnaire or drug answer sheets.
(c) Due to the split sample design, users are advised to use the
imputation-revised variables for tabular summaries and other
descriptive analyses for trends with NHSDA surveys prior to 1994. For
analyses of relationships using multiple variables, use of non-imputed
data items may be best. (d) Users who wish to replicate
results published in the NHSDA Main Findings Report or other SAMHSA
reports should use the 1994-A imputation-revised variables for trends
For some drugs that have multiple names, questions regarding the use of that drug may be asked for each distinct name. For example, even though methamphetamine, methedrine and desoxyn are the same drug, their use was measured in three separate variables.
Multistage area probability sample design involving five
selection stages: (1) primary sampling units areas (e.g., counties),
(2) subareas within primary areas (blocks or block groups), (3)
listing units within subareas (housing units or group quarters), (4)
age-group-smoking classes within sampled listing units, and (5)
eligible individuals within sampled age-group-smoking classes. One
hundred twenty-seven Primary Sampling Units (PSUs), including areas of
high Hispanic concentration, were selected to represent the total
United States population. These PSUs were defined as metropolitan
areas, counties, groups of counties, and independent cities. The main
sample consisted of the 117 PSUs used in the NHSDA 1993 sample. Of
these 117 PSUs, 45 were selected with certainty and 72 were randomly
selected with probability proportional to size. The certainty PSUs
were constructed from Metropolitan Statistical Areas (MSAs) or
adjoining MSAs and stratified based primarily on the concentration of
Hispanics. Once a sufficient number of Hispanic dwellings was
identified, the non-Hispanic Black (Black) and non-Hispanic non-Black
(White/others) interviews were proportionally allocated to the
strata. The rural supplement consisted of an additional ten non-MSA
rural PSUs. Based on correlations between past month cigarette smoking
and drug use, cigarette smokers were oversampled to increase the
precision of drug use estimates. The four age groups were: ages 12 to
17, 18 to 34, 35-49, and 50 and older. The three race/ethnic groups
were: Whites/others, Blacks, and Hispanics. Blacks and Hispanics were
oversampled. The age-group-smoking selection probabilities were based
on the desired sample sizes for each age group by racial/ethnic
domain. The probabilities of selecting the person within the
age-group-smoking domain were based on the number of persons found in
the age-smoking group.
Data were weighted based on the
five stages of sampling that were used. Adjustments were made to compensate for nonresponse and
sampling error. Adjustments also included trimming sample weights to
reduce excessive weight variation and a post-stratification to Census
population estimates. The final weight variable to be used in analysis is ANALWT.
personal interviews and self-enumerated answer sheets
The overall interview response rate was
76.5 percent for the 1994-A survey and 78.2 percent for the 1994-B
survey. A completed interview had to contain, at a minimum, data on
the recency of use of marijuana, cocaine, and alcohol. About
four-fifths of respondents received the 1994-B questionnaire, and
one-fifth received the 1994-A questionnaire, based on random
assignment. The 1994-A interview response rates for the three
racial/ethnic groups were: 75.7 percent for Whites/others, 76.7
percent for Blacks, and 78 percent for Hispanics. The 1994-B
interview response rates for the three racial/ethnic groups were: 76.7
percent for Whites/others, 79 percent for Blacks, and 80.4 percent for
Hispanics. The 1994-A study yielded an 85.3 percent eligibility rate for sample households and a 93.8 percent completion rate for screening eligible
households. The 1994-B study yielded an 85.4 percent eligibility rate for sample households and a 93.8 percent completion rate for screening eligible households.
Extent of Processing: ICPSR data undergo a confidentiality review and are altered when necessary to limit the risk of
disclosure. ICPSR also routinely creates ready-to-go data files along with setups in the major
statistical software formats as well as standard codebooks to accompany the data. In addition to
these procedures, ICPSR performed the following processing steps for this data collection:
Performed consistency checks.
Standardized missing values.
Created online analysis version with question text.
Checked for undocumented or out-of-range codes.
Restrictions: Users are reminded by the United States Department of Health and Human Services that these data are to be used solely for statistical analysis and reporting of aggregated information and not for the investigation of specific individuals or treatment facilities.