Study data were collected as part of a survey of stressors, social supports, and smoking behaviors among Hispanic mothers of young children. The University of New Mexico Health Sciences Center's Human Research Review Committee approved the study protocol. During November 2003 through April 2004, study staff recruited survey participants from waiting rooms of a pediatric emergency room/urgent care clinic, a family practice and pediatric health care facility, and a Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) clinic, all located in southeast Albuquerque, New Mexico.
Clinic appointment logs and sign-in sheets were used to systematically screen women who presented for clinic services. Eligible study participants included U.S.-born or Mexico-born Hispanic mothers of children aged 2 through 12 years. During screening, women were asked, "Which of the following best describes your ethnicity?". Response options were Mexican, Mexican American, Spanish New Mexican, Other Hispanic, and non-Hispanic. Women who responded 'Other Hispanic' were eligible if they were born in the United States or Mexico. A total of 357 eligible Hispanic mothers were identified. Of these, 81 declined to participate and an additional seven interviews could not be completed for other reasons, resulting in a final sample size of 269 (144 U.S.-born; 125 Mexico-born) mothers. The overall response rate was 75.4% (70% for Spanish speakers and 81% for English speakers).
The survey was independently translated into Spanish and verified by back translation into English by two certified translators. After obtaining written informed consent, trained bilingual interviewers conducted interviews with mothers in Spanish (n = 130) or English (n = 139). The survey included questions on maternal and household smoking behaviors and restrictions, sociodemographics, and child ETS exposure.
To increase comparability across studies, smoking restriction and smoking behavior measures were based on items used by Kegler and Malcoe, and were similar to those used by Norman et al. and the 1999 National Health Interview Survey [12, 13, 18]. Home smoking restrictions were measured with one item, "Would you say family members and visitors can: (a) smoke wherever they want in your home, (b) smoke in certain rooms only, or (c) not smoke anywhere inside your home?." Automobile smoking restrictions were similarly assessed with one item, "Would you say: (a) there are no rules about smoking in your family cars, (b) smoking is sometimes allowed in a family car, or (c) smoking is never allowed in any family car?." Respondents could also indicate that they did not have a family car (n = 7).
Two questions assessed mother's smoking status. Respondents were first asked, "Have you smoked at least 100 cigarettes in your entire life?." Those who responded 'no' were classified as nonsmokers; those who responded 'yes' were asked, "How many days per week do you smoke cigarettes now?." Those who reported smoking 1–7 days per week were classified as smokers. Among smokers, the number of cigarettes smoked per day was assessed with one item, "On days you smoke, how many cigarettes do you usually smoke?." This information was combined with the number of days smoked per week to determine the number of cigarettes smoked per week. Respondents were also asked whether any other adults currently living in the home smoke cigarettes. The proportion of each respondent's friends who smoked was assessed by asking, "How many of your friends are smokers, is it: most, about half, less than half, a few, none?."
Child ETS exposure was assessed for one eligible child per household. Each respondent was asked to list the ages of her biological children who currently lived with her. One child aged 2–12 years was randomly selected (based on the child with the most recent birthday) as the 'target child' for questions regarding ETS exposure. Mothers were then asked how many days, during the past seven days, their target child had been in a room with someone who was smoking, and in a car with someone who was smoking.
Social and demographic information collected included the mother's marital status, age, education level, employment status, ethnicity, country of birth (Mexico or United States), and language preference. Data collected on monthly household income and the number of adults and children supported by that income were used to calculate the percent of the 2004 U.S. federal poverty thresholds , which the U.S. government uses to estimate the number of persons in poverty each year.
The survey was designed in Teleforms v.8 (Verity Inc., Sunnyvale, CA) and optically scanned into an electronic database. Data were validated to minimize errors and transferred into SAS v8.2 (Cary, NC) for analysis. Chi square tests were used to assess univariate associations. Multiple logistic regression models were developed using non-automated stepwise modelling techniques to identify multivariate associations between study factors and the primary outcome variables: home or automobile smoking bans (complete ban versus no/partial ban), and child room or automobile ETS exposure. Variables with p values < 0.05 were considered statistically significant; however, all variables that were at least minimally (p ≤ 0.25) associated with outcome variables in univariate analyses were tested for inclusion in final models.