Tobacco Use Supplement Sample Notes
The Tobacco Use Supplement (TUS), sponsored by the National Cancer Institute (NCI), collects data about current and former use of tobacco products, workplace smoking policies, and personal attitudes around smoking and tobacco use. The first survey was implemented in September 1992, however a small number of questions about tobacco use appeared in the September 1985 and 1989 supplements. The 2003 Tobacco Use Special Cessation Supplement (TUSCS) cycle included a special set of smoking cessation questions, which included questions about methods and resources used to stop smoking, and the use of light cigarettes.
The supplement universe is non-institutional civilians age 15 or older in samples from 1985 and 1992-2007, non-institutional civilians age 16 or older in 1989, and non-institutional civilians age 18 and older in samples from 2010 forward. In 2018, a change was made to interview selection procedures. Beginning in 2018, supplement-eligible respondents who were not randomly selected for a self-interview were not given a proxy-response interview as they had been in 2014 and 2015. For those selected for a self-interview, a proxy interview was sought only after four failed callback attempts.This change resulted in fewer supplement interviews. Variables whose universes include all of those who completed a supplement interview (TCIG100, TINTLANG, TRESP, and TSMKER) are most affected.
More records have "No response" for tobacco use questions in 1985 and 1989 than in the more recent samples. The 1989 documentation states that "No response" codes were given, "If an item should have been filled but was not and could not be plugged based on entries from other items."
The TUS occasionally utilizes the CPS panel design to create longitudinal cohorts or to implement follow-up surveys. There have been nine full sets of Tobacco Use Supplement cycles, and three sets of abbreviated or follow-up series. Each cycle consists of three survey months. CPS Tobacco Use documentation states that state and substate-level analysis will be most reliable when using all three months of data in a survey cycle. All NCI analysis will be based on the average of all three months of a cycle. The user will have to ensure that they have requested all relevant months in a cycle. Note that when combining data from three months to form one cycle, survey weights should be divided by three to make the full sample nationally representative. The survey cycles are below:
- September, 1992; January, 1993; May, 1993
- September, 1995; January, 1996; May, 1996
- September, 1998; January, 1999; May, 1999
- January, 2000; May, 2000 (abbreviated TUS - longitudinal cohort possible, matching to January, 1999 or May, 1999)
- June, 2001; November, 2001; February, 2002
- February, 2003; June, 2003; November, 2003 (2003 TUSCS; longitudinal cohort possible, matching February, 2002 to February, 2003)
- May, 2006; August, 2006; January, 2007
- May, 2010; August, 2010; January, 2011
- May, 2011 (abbreviated follow-up to May, 2010 - longitudinal cohort possible, matching to May, 2010)
- July, 2014; January, 2015; May, 2015
- July, 2018; January, 2019; May, 2019
TBSUPPWT is the weight variables for the tobacco supplements. In addition, for samples from 1992 onward, there are two supplement-specific weights for the Tobacco Use Supplements. TNRWT is a "non-response" weight used when analyzing all responses of the tobacco supplement items. However, use TSRWT when analyzing only self-responses and excluding responses by proxies.