METHODOLOGY REPORT
OF THE 2022 NATIONAL
YOUTH TOBACCO
SURVEY
Recommended Citation
Office on Smoking and Health. 2022 National Youth Tobacco Survey: Methodology Report.
Atlanta, GA: U.S. Department of Health and Human Services, Centers for Disease Control and
Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on
Smoking and Health, 2022.
TABLE OF CONTENTS
CHAPTER 1—NYTS SAMPLING DESIGN.......................................................................... 1
1.1 O
VERVIEW OF THE NATIONAL YOUTH TOBACCO SURVEY (NYTS).............................. 1
1.2 O
VERVIEW OF THE 2022 NYTS METHODOLOGY......................................................... 1
CHAPTER 2—NYTS SAMPLING METHODS..................................................................... 2
2.1 S
AMPLE DESIGN ....................................................................................................... 2
2.2 S
AMPLING FRAME .................................................................................................... 3
2.3 S
AMPLING UNITS AND MEASURE OF SIZE................................................................... 4
2.4 P
LANNED SAMPLE SIZES FOR THE SAMPLE................................................................. 5
2.5 S
AMPLING UNITS ...................................................................................................... 6
2.6 S
TRATIFICATION....................................................................................................... 6
CHAPTER 3—NYTS DATA COLLECTION AND PROCESSING .................................. 8
3.1 S
URVEY INSTRUMENT ............................................................................................... 8
3.2 E
XTERNAL REVIEW AND APPROVALS ........................................................................ 8
3.3 T
ECHNICAL ASSISTANCE PROVIDER (TAPS) STAFFING ............................................... 8
3.4 R
ECRUITMENT PROCEDURES ..................................................................................... 9
3.5 S
URVEY ADMINISTRATION ........................................................................................ 9
3.6 W
EB-BASED DATA COLLECTION MANAGEMENT APPLICATION (DCMA) ....................10
3.7 D
ATA RECORDING ...................................................................................................11
3.8 P
ARTICIPATION RATES.............................................................................................11
3.9 D
ATA MANAGEMENT ...............................................................................................12
CHAPTER 4—WEIGHTING OF NYTS RESPONSE DATA ........................................... 13
4.1 E
STIMATORS AND VARIANCE ESTIMATION................................................................13
APPENDICES
A. Q
UESTIONNAIRE ...................................................................................................................15
B. R
ACE AND ETHNICITY DEFINITIONS ......................................................................................16
CHAPTER 1—NYTS SAMPLING DESIGN
1.1 OVERVIEW OF THE NATIONAL YOUTH TOBACCO SURVEY (NYTS)
Tobacco product indicators included in the NYTS are: tobacco product use (e.g., electronic
cigarettes, cigarettes, cigars [including cigars, little cigars, and cigarillos], smokeless tobacco
[chewing tobacco, snuff, or dip; snus, dissolvable tobacco products], hookahs, pipe tobacco,
bidis, roll-your-own cigarettes, heated tobacco products, and nicotine pouches); exposure to
secondhand smoke and e-cigarette aerosol; smoking cessation; minors’ access to tobacco
products; knowledge and attitudes about tobacco.
1.2 OVERVIEW OF THE 2022 NYTS METHODOLOGY
The 2022 NYTS was conducted using a stratified, three-stage cluster sample design to produce a
nationally representative sample of middle school and high school students in the United States.
Sampling procedures were probabilistic and conducted without replacement at all stages.
Sampling entailed selection of (1) Primary Sampling Units (PSUs) (defined as a county, or a
group of small counties, or part of a very large county) within each stratum; (2) Secondary
Sampling Units (SSUs) (defined as schools or linked schools) within each selected PSU; and (3)
students within each selected school.
The 2022 NYTS was administrated as a web-based survey. Students participated in the survey
while at school, home, or some other location. Using a school-issued or personal internet-
connected device, students logged into a secure website and watched a brief 2-minute
instructional video before completing the survey.
Participation in the NYTS was voluntary at both the school and student levels. CDC’s
Institutional Review Board (IRB) requires that parents be given the opportunity to opt their
student out of participating in the survey. Schools used either opt-out or active permission forms
at their discretion.
Survey administration began in January and concluded in May 2022. The final NYTS sample
consisted of 574 schools, of which 341 participated, yielding a school participation rate of
59.4%. A total of 28,291 student questionnaires were completed out of a sample of 37,172
students, yielding a student participation rate of 76.1%. The overall participation rate was 45.2%.
A weighting factor was applied to each student record to adjust for nonresponse and for varying
probabilities of selection. Weights were adjusted to ensure that the weighted proportions of
students in each grade matched national population proportions.
1
CHAPTER 2—NYTS SAMPLING METHODS
2.1 SAMPLE DESIGN
The NYTS methodology was designed to produce national estimates at a 95% confidence level
by school level (middle school and high school), by grade (6, 7, 8, 9, 10, 11, and 12), by sex
(male and female), and by race and ethnicity (Hispanic, non-Hispanic White, non-Hispanic
Black, non-Hispanic Asian, and non-Hispanic American Indian/Alaska Native; Appendix B).
The sampling design prohibits subnational analyses.
The universe for the study consisted of all public and private school students enrolled in
middle schools and high schools in grades 6 through 12 in the 50 U.S. states and the District of
Columbia. Alternative schools, special education schools, Department of Defense-operated
schools, Bureau of Indian Affairs schools, vocational schools that serve only pull-out
populations, and students unable to complete the questionnaire without special assistance were
excluded.
The sample was a stratified, three-stage cluster sample. PSUs were stratified by racial/ethnic
status and urban versus nonurban. PSUs were classified as "urban" if they were in one of the 54
largest U.S. Metropolitan Statistical Areas (MSAs); otherwise, they were classified as
"nonurban." Within each stratum, PSUs were chosen without replacement. Table 2.1 presents
key sampling design features.
Table 2.1 Key Sampling Design Features
Sampling
Stage
Sampling Units Stratification
Measure of Size
(MOS)
Designed Sample Size
1
PSUs: Counties,
portions of a
county, or groups
of counties
Urban vs. Nonurban
(2 strata)
Minority concentration
(8 strata)
Aggregate school
size in target
grades
100 Counties, portions of
a county, or groups of
counties
2
Schools
Small, medium, and large
High school vs middle
school
Aggregate
eligible
enrollment
320 SSU (school)
selections*: 240
large schools, 50
medium schools,
and 30 small schools
3
Classes/students
2 Classes per grade in
half of large schools; 1
class per grade otherwise
20,600 student
participants
*In this exhibit, the schools are SSUs or “virtual schools” created by combining actual, physical schools so that each virtual
school unit has a complete set of grades for the level. The virtual schools are expanded to physical schools. The number of
physical schools in the sample was expected to range from 345 to 375.
The first stage of sampling selected PSUs within each stratum for a total of 100 sample PSUs. At
the second sampling stage, a total of 320 SSUs, or schools, were selected from the sample PSUs as
2
follows: two large schools were selected per sample PSU, one per level (middle or high); an
additional large school for each level was selected in a subsample of 40 PSUs, for a total of 240
large SSUs. An additional 50 medium SSUs and 30 small SSUs were selected from subsample
PSUs, for a total of 320 sample SSUs (320 = 240 + 30 + 50). The PSU subsamples were selected
with simple random sampling, and the schools were drawn with probability proportional to the
total number of eligible students enrolled in a school.
Depending on the average design effects, target subgroup sample sizes are between 1,200 and
1,700. Compared to previous cycles, the NYTS sampling design has had both lower effects on
unequal weighting and smaller clustering effects. These factors lead to lower design effects,
particularly for subgroups. Smaller design effects have led to smaller variances and improved
precision.
An appropriate sample size can generate estimates with the required precision by grade, as well
as by sex and school level. Therefore, the precision requirements generally focused on
racial/ethnic subgroups within school level. The targets of n = 700 students per racial and ethnic
minority group by school level (1,400 total per group) correspond to prevalence estimates within
+/- 5% to achieve 95% confidence intervals (CI) for all key racial and ethnic subgroups.
The prevalence estimates presented in Chapter 4 show that for all key racial and ethnic
subgroups, prevalence estimates are within +/- 5% for 95% CI (i.e., standard errors are less than
2.5%). Similarly, standard errors are less than 2.5% for all estimates for Black and Hispanic
students at the middle school and high school level.
2.2 SAMPLING FRAME
The 2022 NYTS sample was based on a sampling frame from multiple data sources to increase
the coverage of schools nationally. The frame combined data files obtained from MDR Inc.
(Market Data Retrieval Inc.) and from the National Center for Education Statistics (NCES). The
MDR frame contained school information that included enrollments, grades, race and ethnicity
distributions within the school, district and county information, and other contact information for
public and nonpublic schools across the nation. The NCES frame sources included the Common
Core of Data for public schools and the Private School Survey for nonpublic schools. Including
schools sourced from the two NCES files resulted in substantial coverage increase among all
public and nonpublic high schools. Most of the added schools were smaller schools. Each school
was represented only once in the final sampling frame.
The first step was to remove schools such as Department of Defense schools, vocational schools,
and adult education schools. This resulted in the exclusion of 3.9% of schools (2.8% of public
schools and 8.0% of private schools) and 1.1% of students. Lastly, schools were removed that
had fewer than 40 students enrolled across eligible grades, resulting in the exclusion of 20.4% of
schools (12.8% public and 42.6% private) which had been eligible after the other exclusions.
This exclusion of schools with fewer than 40 students led to the exclusion of only 1.03% of
students of those in eligible schools. Overall, 97.8% of students in middle and high schools
nationally were included in the frame. The frame contained 28,636 high schools and 42,749
middle schools for a total of 71,385 eligible schools.
3
2.3 SAMPLING UNITS AND MEASURE OF SIZE
2.3.1 Sample
The sample was constructed using a three-stage cluster sample design to produce a nationally
representative sample of students in grades 6–12 who attend public and private schools. The
first-stage sampling frame consisted of PSUs made up of counties, groups of smaller, adjacent
counties, or parts of larger counties. For the second stage of sampling, SSUs were defined as a
physical school that can supply a full complement of students in grades 6 through 8 (middle
school) or 9 through 12 (high school) or a school created by linking component physical schools
together to provide all grades for the level.
Schools were stratified into small, medium, and large based on their ability to support less than
one, one, or two class selections per grade. Small SSUs contained fewer than 28 students at any
grade level, and large SSUs contained at least 56 students at each grade level. The remaining
schools were classified as medium sized.
The sampling stages may be summarized as follows:
• Selection of PSUs—One hundred (from approximately 1,257) PSUs were selected from
16 strata with probability proportional to the total number of eligible students enrolled
in all eligible schools located within a PSU.
• Selection of schools—At the second sampling stage, a total of 240 large schools, or
SSUs, were selected from the sample PSUs. Additionally, as described in Section 2.1,
we selected 50 medium schools and 30 small schools, resulting in a total of 320 sample
SSUs (320 = 240 + 50 + 30).
• Selection of students—Students were selected via whole classes whereby all students
enrolled in any one selected class were chosen for participation. Classes were selected
from course schedules provided by each school so that all eligible students had only a
single chance of selection.
The sampling approach used probability proportional to size (PPS) sampling methods, with the
measure of size (MOS) defined as the count of final-stage sampling units—students in intact
classrooms. Coupled with the selection of a fixed number of units, the design resulted in an equal
probability of selection for all members of the universe (i.e., a self-weighting sample). These
conditions were approximated for the NYTS resulting in a roughly self-weighting sample. The
MOS also was used to compute stratum sizes and PSU sizes. By assigning an aggregate measure
of size to the PSU, the sample allocated to the PSU was in proportion to the student population.
The third, and final, sampling stage selected classes within each grade of a sample SSU. We
selected two classes per grade in large schools and one class per grade in the remaining schools.
The threshold for double class sampling was based on a simulation study to ensure that the
required numbers of students in specified racial and ethnic minority groups were achieved per
school level. All students in a selected class were eligible to participate in the survey.
4
2.4 PLANNED SAMPLE SIZES FOR THE SAMPLE
In calculating the sample sizes for the 2022 NYTS, we made our approach more robust by
assuming a conservative final rate (student x school) of 60%. The student participation rate was
adjusted to account both for a growing number of ineligible students and parental refusal and for
the new data collection methods (i.e., 100% virtually supported fielding methodology without in-
person survey administrators).
Table 2.4 Planned Sample Sizes for the 2022 NYTS, Sample
PSU Size
# of
SSUs
Number of
Schools
Sampled
Number
of Classes
per
School
Number
of
Students
per Class
Number of
Sampled
Students
Prior to
Attrition
No. of
Students
After 75%
School RR
No. of
Students
After 60%
Final
School and
Student
RR
100
Large
High
School
120
Double
classes: 60
8 24 11520 8640 6912
Single
classes: 60
4 24 5760 4320 3456
Large
Middle
School
120
Double
classes: 60
6 22 7920 5940 4752
Single
classes: 60
3 22 3960 2970 2376
Large
Total
240 29160 21870 17496
25
(subsample)
Medium
High
School
25 4 20 2000 1500 1200
Medium
Middle
School
25 3 20 1500 1125 900
Medium
Total
50 3500 2625 2100
15
(subsample)
Small
High
School
15 4 16 960 720 576
Small
Middle
School
15 3 16 720 540 432
Small
Total
30 1680 1260 1008
Overall
Total
320
34340
25755
20604
The estimated sample yield from these large schools was 29,160 students before school and
student nonresponse, leading to an expected total of 17,496 participating students in large
schools after accounting for nonresponse. The expected yield was 3,500 from medium schools
and 1,680 students from small schools. In total, the expected number of participating students
was 20,604.
5
Within each school, one class was selected from each grade to participate in the survey except in
large schools with high racial and ethnic minority populations, where two classes per grade were
selected. Note that the set of schools with high racial and ethnic minority populations defined for
double class sampling is a subset of the large schools that can support such double class
sampling. For the 2022 NYTS, we implemented double class selection for half of large schools
in the primary sample (randomly selected) to ensure sufficient student yields.
2.5 SAMPLING UNITS
2.5.1 Sampling Units (PSUs)
In defining PSUs, several issues were considered:
• Each PSU should be large enough to contain the requisite numbers of schools and students
by grade, and small enough so as not to be selected with near certainty.
• Each PSU should be compact geographically to control the number of school districts
contacted and recruited.
• Recent data should be available to characterize each PSU.
• PSUs are defined as containing at least four middle and five high schools.
Generally, counties were equivalent to PSUs, with two exceptions:
• Low population counties were grouped to provide sufficient numbers of schools and
students.
• High population counties were divided into multiple PSUs so that the resulting PSUs would
not be selected with certainty.
The PSU frame was screened for PSUs that no longer met the above criteria. The frame was
adjusted by recombining small counties/PSUs as necessary to ensure sufficient size while
maintaining compactness. Near-certainty PSUs were split using an automated procedure built
into the sampling program.
2.5.2 Forming Secondary Sampling Units (SSUs)
Single schools represented their own SSU if they had students in each of grades 6 through 8 or in
grades 9 through 12. Schools that did not have all eligible grades for the level were grouped
together to form an SSU. Linked schools were treated as single schools during sampling.
2.6 STRATIFICATION
The PSUs were organized into 16 strata, based on urban/nonurban location and proportion racial
and ethnic minority enrollment.
• If the percentage of Hispanic students in the PSU exceeded the percentage of non-
Hispanic Black students, the PSU was classified as Hispanic. Otherwise, it was classified
as Black.
• If the PSU was within one of the 54 largest MSAs in the United States, it was classified
as urban, otherwise it was classified as nonurban.
• Hispanic urban and Hispanic nonurban PSUs were classified into four density groupings
depending upon the percentages of Hispanic students in the PSU.
6
• Non-Hispanic Black urban and non-Hispanic Black nonurban PSUs were also classified
into four groupings depending upon the percentages of Black students in the PSU.
The density grouping bounds were computed using an optimization algorithm
1
that was
refreshed each cycle to reflect changes in the racial/ethnic distribution of the student population.
The boundaries or cutoffs changed as the frequency distribution (“f”) for the racial and ethnic
groupings changed from one survey cycle to the next. Table 2.6 presents the stratum boundaries
used in the 2022 NYTS.
Table 2.6 Stratum Boundaries: Minority Percentage Cutoffs
Minority
Concentration
Density
Group
Bounds
Urban
Nonurban
1
0%–26%
0%–20%
Black
2
>26%–40%
>20%–34%
3
>40%–54%
>34%–54%
4
>54%–100%
>54%–100%
1
0%–26%
0%–24%
Hispanic
2
>26%–42%
>24%–48%
3
>42%–58%
>48%–68%
4
>58%–100%
>68%–100%
As described earlier, SSUs were stratified into three sizes for small, medium, and large schools
for the primary sample only. For the supplement samples, the frame was restricted to large
schools which support double class sampling at every grade. The two supplement samples were
not stratified explicitly but only implicitly by region and by state. Specifically, the frame was
sorted by region and by state for PPS selection of PSUs. Implicit stratification helps improve the
geographic representation of the supplement samples.
1
The cumulative square root of “f” method developed by Dalenius and Hodges.
7
CHAPTER 3—NYTS DATA COLLECTION AND PROCESSING
3.1 SURVEY INSTRUMENT
The NYTS collects data on key tobacco product prevention and control outcome indicators. The
2022 survey instrument included 166 questions. The web survey was created using ColdFusion
and all data were stored in a MS SQL Server. To take the web survey, students navigated to a
dedicated URL, nyts.cdc.gov, and entered a randomly generated, five-digit access code.
The survey followed a skip-pattern logic based on the student’s responses to questions about ever
and current tobacco product use behaviors. To improve students’ sense of privacy, only one
question was displayed on each screen so that responses to prior questions were not susceptible
to observation. Students were given approximately 35–45 minutes to complete the survey.
Students who could not take the survey on the planned date for administration were asked to take
the survey at the next possible opportunity.
The first five questions on the survey collected student demographic information, and the rest
measured a comprehensive set of tobacco-related topics. Specific areas covered by the survey
included: prevalence of tobacco product use; knowledge of and attitudes toward tobacco product
use; exposure to tobacco media and advertising; minors’ access to tobacco products; nicotine
dependence; cessation attempts; exposure to secondhand smoke; harm perceptions; and exposure
to tobacco product warnings. At the beginning of each tobacco product section, a description of
the product (with example brands) and generic images of specific tobacco products were
provided to assist with product recognition and increase the accuracy of student data. Students
could refer to this description and the images as they answered related questions. The NYTS also
included sociodemographic questions about family affluence, depression and anxiety, and sexual
orientation and gender identity (SOGI).
3.2 EXTERNAL REVIEW AND APPROVALS
Three bodies reviewed and approved the instrumentation, processes, privacy and security
elements, and sampling design of the 2022 NYTS: the Office of Management and Budget (OMB),
ICF’s Institutional Review Board (IRB), and CDC’s Institutional Review Board (IRB).
With the transition to an electronic data collection format for the 2019 NYTS, the Security
Assessment and Authorization (SA&A) approval and Enterprise Performance Life Cycle (EPLC)
review was valid for the 2022 NYTS cycle. The SA&A is a formal methodology for testing and
evaluating the security controls of the system to ensure that it is configured properly to meet the
security mandated by the Federal Information Security Management Act (FISMA). EPLC is a
framework to enhance the Department of Health and Human Services (HHS) IT governance
through rigorous application of sound investment and project management principals, in
conjunction with industry’s best practices.
3.3 TECHNICAL ASSISTANCE PROVIDER (TAPS) STAFFING
The role of the Technical Assistance Provider (TAP) was developed for the 2021 NYTS in
response to anticipated complications due to COVID-19 that prohibited data collectors from
conducting in-person survey administration. This role continued for the 2022 NYTS cycle, as
8
well. TAPs provided 100% virtual support to schools and teachers before, during, and after
survey administration to (1) ensure teachers had received all the necessary materials to
administer the survey, (2) answer any questions schools contacts and/or teachers may have prior
to, during, or after survey administration, (3) ascertain that parental consent was properly
obtained prior to the scheduled survey administration date, and (4) provide remote IT support, if
needed. To ensure schools in various time zones would be adequately supported during school
hours, TAPs were hired geographically across the country so that every time zone with sampled
schools had at least one TAP in that part of the country. TAPs were recruited from a pool of
previously trained data collectors. An in-person, 2-day training for TAPs was conducted
December 6–7, 2021.
Key components of the training included the following:
• Pre- and postsurvey communications with the schools and teachers
• Orientation to student and teacher portals
• IT troubleshooting
• Communication with headquarters staff
3.4 RECRUITMENT PROCEDURES
Recruitment began in September 2021 with calls to state departments of education and health,
informing them of the survey effort and sampled schools in their state. After notification at the
state level, district- and school-level recruitment began. For public or diocesan schools, verbal or
written agreement was first obtained by their district or diocese, respectively, before contact was
made with the school. However, private schools were approached directly. A date for survey
implementation was selected that was convenient to the school. Recruiters and TAPs used a
secure web-based calendar to facilitate communication and adjust survey dates upon request by
the school.
3.5 SURVEY ADMINISTRATION
Survey administration began in January and continued through May 2022. While the details of
each data collection varied, there were six core steps followed for every school:
1) Conduct precontact call with the principal or lead contact to confirm survey arrangements
and to answer any questions.
2) Send tailored communications and survey materials to selected teachers.
3) Reach out directly to teachers to confirm receipt of materials, verify intentions to
administer the survey on the scheduled date, confirm parental consent procedures were
followed, and provide additional survey instructions.
4) Virtually monitor survey activities and respond to requests for technical support, as
needed.
5) Follow up with teachers regarding student response rates and class enrollment.
6) Report final progress to school contact and thank them for their school’s participation.
9
Procedures were designed to protect students’ privacy by assuring that student participation was
anonymous and voluntary. Using a school-issued or personal internet-connected device, students
logged into a secure website, watched a brief 2-minute instructional video, and responded to a
question regarding their location (e.g., classroom, home, other location) before completing the
survey. All surveys were submitted directly to a secure SQL server.
3.5.1 Field Procedures
After schools had been recruited, classes selected, and a date for survey administration
scheduled, each school received a mailing with presurvey materials containing instructions for
the school contact and packets for the teacher of each selected class. Teacher packets contained
the parental permission forms to be distributed to all students in the selected classes prior to data
collection. The timing of these presurvey materials was determined in part by the type of
permission form being used by the school; this decision was made by the school district or
individual school. Opt-out parental permission forms (i.e., forms returned only if the parents do
not want their child to participate) were sent approximately 2 weeks prior to the scheduled date
of data collection in the majority of schools. Active parental permission forms (i.e., forms that
must be returned with the parent’s signature for the child to participate) were sent out 4 weeks
prior to the scheduled date of data collection for schools that require active consent. TAPs
conducted follow-up calls and sent emails to the selected schools to answer any questions and to
make sure materials were received and distributed to selected classes and students.
3.5.2 Classroom Selection
Students were selected for participation by default via the selection of whole classes (i.e., all
students enrolled in a selected class were eligible to take the survey). The frames from which
classes were chosen were constructed so that eligible students had one, and only one, chance of
being selected. However, at times the specific method of selecting classes varied from school to
school, according to how a school’s class schedule was structured. Typically, classes were
selected from a list of required core courses such as English, social studies, math, or science.
Among middle school students, and among high school students in a few states, physical
education and/or health also were considered core courses. However, in a small number of
schools, it was difficult to develop an appropriate frame using this approach. Therefore, in these
schools, classes were selected by using a time of day (e.g., second period) when all eligible
students were scheduled to be attending a class as the frame, and randomly selecting from all
classes held at this time. Lastly, in some schools, homerooms or advisory periods were used as
the frame for class selection.
3.6 WEB-BASED DATA COLLECTION MANAGEMENT APPLICATION (DCMA)
For multiple cycles of the NYTS, a web-based data collection management application (DCMA)
has been used to help: centralize the management of the study; facilitate information exchange
with project staff; and allow all members of the project management teams, recruitment teams,
supervisory teams, and remote staff access to information necessary to implement the study. The
system is designed with differing levels of access depending on the user’s role on the study. The
system’s primary functions include generating invitation letters, tracking recruitment progress,
10
11
scheduling data collection, registering student records submitted to the central repository, and
tracking school and student response rates.
3.7 DATA RECORDING
Preliminary student participation rates were calculated based on (1) class enrollment numbers
from teachers of selected classes and (2) the number of surveys received in the central repository.
If teachers reported a different number of expected completes than what was received in the
central repository, a TAP followed up to resolve discrepancies and determine ways to maximize
student participation. As additional surveys were received after the initial survey administration
date, the DCMA automatically updated the number of records received; participation reporting
was revised accordingly.
3.8 PARTICIPATION RATES
Participation rates for the NYTS were calculated at the school and student levels.
3.8.1 School-Level Participation Rates
The sample includes 574 schools that were selected across 243 districts in 41 states and the
District of Columbia. During sample validation, 34 schools were deemed ineligible and were
replaced. In total, 341 schools (59.4%) participated in the study. Of refusals, 145 of them were
due to district-level refusals to allow contact with schools to discuss participation, and 88 were
school-level refusals.
3.8.2 Student-Level Participation Rates
Initial student-level participation rates were calculated from the field as teachers reported
enrollment information and submitted surveys registered in the central repository. In follow-ups
between teachers and TAPs, further refinements were made to (1) revise the number of eligible
students based on available documentation, (2) correct mathematical errors, (3) review counts of
surveys received by the database, and (4) account for make-ups as they were received from
students and classes that did not participate on the initial day of survey administration.
The final student participation rate for the 2022 NYTS sample was 76.1%. Overall, 37,172
eligible students from the 341 participating schools were invited to participate in the survey, and
28,291 did so.
Table 3.1 Overall NYTS 2022 Student Participation Rate
# Eligible
# Completed
Participation %
Final Sample
37,172
28,291
76.1%
The 2022 NYTS final sample attained an actual school participation rate of 59.4% and a student
participation rate of 76.1%. The overall participation rate was 45.2% for the final sample.
3.9 DATA MANAGEMENT
To take advantage of the electronic format of the NYTS, the dataset was designed to be self-
cleaning based on programming logic. However, to ensure accuracy, CDC created a series of
data-cleaning specifications that were applied to eliminate internal inconsistencies. These
cleaning specifications also computed certain analytic variables and recoded race and ethnicity
values to match CDC-required classifications. Data “missingness” was categorized into one of
four types: as a legitimate skip based on programmed logic, as item-level refusal if a question
was presented to a student on screen but not answered, as not answered because the student was
never shown a question on screen (e.g., partial complete), or as recoded to missing due to edit
checks. Missingness is distinguished in the data set as follows:
• .S – Legitimate skip
• .N – Displayed, not answered (item-level refusal)
• .Z – Not displayed (partial complete)
• .E – Missing due to edit check
The survey data file preparation for weighting involved a series of data file linking steps. These
steps ensured that the data files merged the school information compiled during frame
construction, sample selection, replacement of ineligible schools, recruitment, and data collection
using a common school identifier.
12
CHAPTER 4—WEIGHTING OF NYTS RESPONSE DATA
4.1 ESTIMATORS AND VARIANCE ESTIMATION
Weighted estimates of means, percentages, and totals can be computed using the final weights in
the analysis file. If w
i
is the weight of case i (the inverse of the probability of selection adjusted
for nonresponse and poststratification adjustments) and x
i
is a characteristic of case i (e.g., x
i
= 1
if student i smokes, but is zero otherwise), then the mean of characteristic x is estimated as (Σ
w
i
x
i
)/(Σ w
i
). A weighted population total estimate is computed similarly as (Σ w
i
x
i
). The weighted
population estimates can be computed with the Statistical Analysis System (SAS) as well as with
other statistical software.
These estimates are accompanied by measures of sampling variability, or sampling error, such as
variances and standard errors, that account for the complex sampling design. These measures
support the construction of confidence intervals and other statistical inference such as statistical
testing (e.g., subgroup comparisons or trends over successive NYTS cycles). Sampling variances
can be estimated using the method of general linearized estimators
2
as implemented in SAS
survey procedures. These software packages must be used because they permit estimation of
sampling variances for multistage stratified sampling designs. They also account for unequal
weighting and for sample clustering and stratification.
The final weight files also include PSU and strata variables which support the analysis of
clustered survey data and accurate variance estimation. As in previous cycles, a variable for
“variance strata” was added, which may differ from the design strata, to ensure that all variance
strata had at least two PSUs.
3
2
Skinner CJ, Holt D, Smith TMF. Analysis of Complex Surveys. John Wiley & Sons; 1989: 50.
3
Specifically, two strata were combined into one variance stratum because the original stratum had only one PSU
when analyzed at both the middle and high school level.
13
*Example SAS and SUDAAN code will generate estimates of ever use and current (past 30 day use) of e cigarettes,
cigarettes, cigars, smokeless tobacco products (chewing tobacco, snuff, or dip), and hookah tobacco. This is not an
exhaustive list of all tobacco products assessed in the NYTS
- -
Exhibit 4.1 Example SAS and SUDAAN Code for Generating Weighted Tobacco Product Use
Estimates (Ever Use, Current Use)* and Standard Errors
SAS:
Proc Surveymeans Data=nyts2022 mean;
Var eelcigt ecigt ecigar eslt ehookah celcigt ccigt ccigar cslt chookah;
Class eelcigt ecigt ecigar eslt ehookah celcigt ccigt ccigar cslt chookah;
Stratum v_stratum2;
Cluster psu2;
Weight finwgt;
Domain SCHOOLTYPE SCHOOLTYPE*Sex SCHOOLTYPE*Race_S;
Title “NYTS 2022, Tobacco Product Use Estimates by School Type, by School Type and Sex Cross-Classified,
and by School Type and Race/Ethnicity Cross-Classified”;
run;
SUDAAN:
Proc Descript Data=nyts2022 Filetype= SAS Design=WR;
Var eelcigt ecigt ecigar eslt ehookah celcigt ccigt ccigar cslt chookah;
Catlevel 1 1 1 1 1 1 1 1 1 1;
Nest v_stratum2 PSU2 / Missunit;
14
APPENDIX B. RACE AND ETHNICITY DEFINITIONS
Non-Hispanic American Indian/Alaska Native—A person having origins in any of the original
peoples of North and South America (including Central America) and who maintains cultural
identification through tribal affiliation or community attachment.
Non-Hispanic Asian—A person having origins in any of the original peoples of the East Asia,
Southeast Asia, or the Indian subcontinent.
Non-Hispanic Black—A person having origins in any of the Black racial groups of Africa;
African American.
Non-Hispanic Pacific Islander
4
—A person having origins in any of the original peoples in the
Pacific Islands. This area includes, for example, Guam, Hawaii, Samoa, and other Pacific
Islands.
Hispanic—A person of Mexican, Puerto Rican, Cuban, Central or South American, or other
Spanish culture or origin, regardless of race.
Non-Hispanic White—A person having origins in any of the original peoples of Europe, North
Africa, or the Middle East.
4
Our design and estimation processes separate out the two subgroups, Asian and Pacific Islander, as per Final
Standards, Office of Minority Health (https://minorityhealth.hhs.gov/omh/browse.aspx?lvl=3&lvlid=53).
16
