2.1. Sample Design
Data for the study were from a population based mail survey of adults aged 18 to 35 residing in three metropolitan areas: Charlotte, NC, Denver, CO, and Topeka, KS. These areas were selected because they were all test markets for dissolvable tobacco products [
14] which were the main focus of the survey.
In order to balance the need for a representative sample and to keep costs down, we used a dual-frame sample, part random address-based sample (ABS) from the U.S. Postal Service Delivery Sequence File and part list frame. An ABS-only random sample would have a less complex sample design for obtaining generalizable estimates of the population, but because we are focusing on young adults, relying on a random ABS only would be inefficient and very expensive. Dual frame samples allow increased efficiency and decreased costs by supplementing the random sample with a list sample that has already been identified to include households likely to have a resident in the targeted age group. The list of addresses was purchased from Marketing Systems Group. It is based on a variety of publicly available and commercial sources, such as driver’s license registries, credit agencies, and marketing databases. Conducting a survey with a dual frame sample involves: (a) removing all addresses from the ABS sample that appear on the list sample; and then (b) drawing unduplicated samples from the three areas, Charlotte, Denver, and Topeka (approximately one-third the sample in each area) from the two sources—the ABS and the purchased list. The survey is conducted with respondents from each of the sources, and with the known probabilities of selection of each household in each frame, weights are developed so that the combined data can be used to make population-based estimates. Using this dual frame design, all residences within the designated geographic areas have a known probability of selection; no one residence excluded—even those with no landlines. By combining the two frames, we do not rely exclusively or too heavily on the cost-saving list sample and its unknown biases. Post-stratification weight adjustments (e.g., by gender, age, race and education) were developed to yield sample estimates that accurately reflect the population. The total sample included 31,999 households (4001 surveys were not delivered due to bad addresses.)
2.3. Measures
Tobacco Use Status: Respondents indicated any lifetime use of a list of 6 types of tobacco, as well as which was the first product tried (including e-cigarettes). Those who indicated any tobacco use were asked if they had smoked 100 cigarettes, and whether they currently smoked every day, some days, or not at all. Those responding affirmatively to the 100 cigarettes question, and checking either “every day” or “some days,” were defined as current smokers. Those who responded affirmatively to the 100 cigarette question, and indicated they now smoked “not at all,” were defined as former smokers. Those who denied having smoked 100 cigarettes in their lifetime, or who said that they had “never used any tobacco products,” were defined as never smokers. Three levels of smoking intensity were coded based on the average number of cigarettes smoked per day: High (16+); Medium (10 to 15); Low (<10).
Expectations of Quitting Smoking: Current smokers were asked how likely they were to still be smoking in 12 months. Those responding “very” or “somewhat” likely were defined as unlikely to quit; those responding “Not very” or “not at all” were classified as likely to quit.
Electronic Cigarette Use: Those who indicated that they had used an e-cigarette “even one time” were defined as having tried an e-cigarette. Those who had tried were asked on how many of the previous 30 days they had used an e-cigarette. If they indicated use on 1 or more of the past 30 days, they were defined as current e-cigarette users. Three levels of frequency of e-cigarette use, based on number of days used in the past month, were computed: None (never tried or tried, but not used in the past 30 days), Medium (1 to 19 days), High (20 to 30 days).
Reasons for Trying E-cigarettes. Those who had tried e-cigarettes were presented with 7 possible reasons for trying and asked to check all that applied. The reasons included curiosity, use by friends, health risks relative to cigarettes, absence of smell, for use where smoking is banned, and to quit or cut down on smoking.
Reasons for Stopping E-cigarette Use. Those who indicated that they had stopped using e-cigarettes were presented with 8 possible reasons for stopping, and asked to check the most important reason. The reasons included health concerns, negative reactions to taste and feeling sick, inferiority to other forms of tobacco, expense, lack of availability and social disapproval.
Sensation Seeking. Sensation seeking was measured using the Brief Sensation Seeking Scale [
15]. Using a five-point Likert scale (1 = strongly disagree to 5 = strongly agree), the eight-item scale measures agreement with statements such as: “I would like to explore new places” and “I prefer friends who are exciting and unpredictable”. Total sensation seeking scores were calculated as the average of all items for individuals who answered a minimum of five questions on the scale. The Cronbach’s alpha for this sample was 0.78.
Demographic Characteristics. The following demographics were measured: gender, age group (18–24), (25–30), (31–35), and race/ethnicity (white non-Hispanic vs. minority).