2023 Adult Tobacco Survey Methods
This report summarizes results from the 2023 Adult Tobacco Survey and trend analyses using previous versions of the Adult Tobacco Survey.
In this section, WYSAC provides a general summary of the methods used to collect and analyze the data for the 2023 Adult Tobacco Survey. For statistical analysis, WYSAC performed logistic regression and multinomial logistic regression to identify associations with time (difference since a reference year or other changes over time longer than two years). Relationships and linear trends noted as significant in the body of the report are statistically significant, p < .05.
Appendix A provides tables reporting Wyoming’s state-level unweighted counts, weighted percentages, and 95% confidence intervals (CIs) for weighted percentages for every survey item and WYSAC-calculated variable.
Appendix B provides the technical details of the methods used to collect the data for the 2023 Adult Tobacco Survey as reported by WYSAC’s Survey Research Center.
Appendix C provides details of statistical analyses summarized in the body of this report.
WYSAC purchased a random sample of telephone numbers in Wyoming and (new to 2023) requested that email and postal mail addresses be connected to the phone numbers. First, on May 2, 2023, WYSAC sent an email with an invitation to complete the survey online when emails were available for the numbers selected. On May 9, 2023, WYSAC sent an email reminder to those who had not responded to the initial email invite.
Second, on May 8, 2023, WYSAC mailed an invitation to complete the survey online to people who had postal addresses in the list and either did not have an email address or had undeliverable email addresses from the initial email invites. On May 23, 2023, WYSAC sent a second invitation letter to those who had not responded.
Finally, WYSAC called people who had not completed a survey online regardless of whether they had email or postal addresses in the sample list. Trained WYSAC telephone interviewers conducted the telephone interviews. Calling began on May 29, 2023, and ended on August 24, 2023. WYSAC completed a total of 2,733 surveys. Adult Tobacco Survey participants completed 1,144 surveys online. WYSAC callers completed 1,589 surveys (75% on cell phones; 25% on landlines).
This three-tier approach maintained the basic sampling process as in previous versions of the survey and sped up data collection by providing adults with options on how to respond.
In some figures and tables, percentages may not total 100% because respondents could choose more than one response. Occasionally, rounding of the actual percentages may result in reporting percentages that do not total 100.
Key Limitations
Most Adult Tobacco Survey items have been tested and validated by the CDC and reused over time. However, the Adult Tobacco Survey relies on self‐reported data, respondents’ memory of events, and their interpretation of the survey items. Therefore, the results presented here might include recall errors or respondent bias (such as not reporting embarrassing or unpopular behaviors).
Unlike previous iterations of the Wyoming Adult Tobacco Survey, which were conducted exclusively via phone, the 2023 Adult Tobacco Survey introduced two response modes: online and phone. In 2023, 42% of respondents completed the survey online, while 58% responded via phone. WYSAC examined the demographic characteristics of respondents (unweighted frequencies) by mode:
- The online survey had a lower proportion of young adults (ages 18-29) than the phone survey.
- The online survey had a higher proportion of women than the phone survey.
- The online survey had a lower proportion of American Indian respondents than the phone survey.
- The online survey had a lower proportion of respondents with low income (annual household income less than $30,000) than the phone survey.
- The online survey had a lower proportion of respondents with behavioral health conditions than the phone survey.
The Adult Tobacco Survey has a complex skip pattern which means not all respondents are asked each question. For example, people who have never smoked are not asked about quitting smoking. In some cases, fewer than 50 people were asked a question. Estimates generated from such small groups are extremely imprecise, so WYSAC does not report them in the body of the report. This follows the example set by the CDC in reporting Behavioral Risk Factor Surveillance System (BRFSS) statistics (https://www.cdc.gov/brfss/brfssprevalence/).