ELSEVIER Reliability of Proxy Reports of Parental Smoking by Elementary Schoolchildren TRACIE BARNETT, BSc, JENNIFER O’LOUGHLIN, PwD, GILLES PARADIS, MD, MSc, PRCPC, AND LISE RENAUD, PHD PURPOSE: To investigate the reliability of reports of parental smoking by elementary schoolchildren aged 9-13 years, and to identify the correlates of disagreement between student proxy and parent self-reports. METHODS: As part of the evaluation of a school-based heart health promotion program, data on parental smoking status were collected from 1240 student-mother pairs and 898 student-father pairs. RESULTS: Agreement for parental smoking status was 93.1% among student-mother pairs and 86.4% among student-father pairs. Among srudent-mother pairs, reports by students aged 9 years were more likely to disagree with mothers’ self-reports than those of older children (odds ratio (OR) = 3.1). Among student-father pairs, the only significant correlate of disagreement was living in a single-parent family headed by the mother (OR = 2.6). CONCLUSIONS: Children lo-years or older can provide reliable reports of the smoking status of cohabiting parents. Ann Epidemio! 1997;7:396-399. 0 1997 Elsevier Science Inc. KEY WORDS: Reliability, Children, Adolescent, Proxy Respondents, Epidemiologic Methods, Smoking. INTRODUCTION Proxy reports of smoking habits by cohabiting adult next- of-kin have shown excellent agreement with index subjects for broad categories of smoking including “never” or “ever” smoker ( 1, 2) and “current smoker, yes or no” (3,4). They are generally less reliable for self-reported recent quitters, for occasional smokers (5), and when more detailed smoking information is required (2,3, 6, 7). Although the reliability of proxy reports of parental smoking status by adult children has been assessed (6), no reports have yet examined the reliability of reports by young children. If such reports are useful, school-based surveys of young children could be used to obtain estimates of smoking prevalence among parents. Because they target a “captive audience”, school-based sur- veys could yield a higher response than household surveys, particularly in populations with large numbers of recent immigrants, since children of immigmnts are more likely to From the Deoartment of Public Health. Montreal General Hosoital. 4835 Christoph&Zolomb Avenue, Mann&, Quebec, Canada (T.B.,*J.O.; G.P., L.R.); tbe Department of Epidemiology and Biostatisrics, McGill University, 1020 Pine Avenue West, Montreal, Quebec, Canada (J.O., G.P.,), and the Department of Social and Preventive Medicine, Universite de Montreal, C.P. 6128, Succ. A, Montreal, Quebec, Canada (L.R.). Address reprint requests to: Jennifer G’Loughlin, Ph.D., Department of Public Health, Montreal Genera1 Hospital, 4835 Christophe-Cotomb Avenue, Montreal, Quebec, H2J 3G8 Canada. Received Gctober 30, 1996; revised March 31, 1997; accepted April 9, 1997. 0 1997 Elsevier Science Inc. All rights resewed. 6.55 Avenue of the Americas, New York, NY 10010 learn the language of the host country quickly than their parents. Also, because they are logistically less complicated than household surveys, the cost is potentially lower. The purpose of this study was to examine the reliability of proxy reports by elementary sch~lchildren of the current smoking status of their parents. METHODS Data were collected in the baseline survey for the evaluation of the Saint-Louis du Part Heart Health program, a five- year multifactorial heart health promotion program to pre- vent the onset of smoking, and to promote healthy dietary habits and regular physical activity among school-aged chil- dren in a multiethnic, low income inner-city neighborhood in Montreal, Canada (8). Data on smoking, diet and physical activity, as well as on potential correlates of these behaviors, were collected in 24 inner-city elementary schools, in in- class questionnaires completed at school by all students aged 9-13 years in grades 4, 5, and 6. Data on parental lifestyle habits were collected from parents in self-administered ques- tionnaires brought home from school by their children. Data on student gender, age, country of birth (Canada, other), family composition (two-parent, single mother, sin- gle father, or other), household size, and parents’ employ- ment status were collected in the student questionnaires. Proxy reports of parental smoking status were obtained by: “Does your mother smoke cigarettes?” and “Does your father ~047*27971971~~7.00 PII SlO47-2797(97)~045-8