Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Factors associated with mothers’ protection of their children from environmental tobacco smoke Temple, Beverley A. 2006

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2006-201091.pdf [ 8.18MB ]
Metadata
JSON: 831-1.0092975.json
JSON-LD: 831-1.0092975-ld.json
RDF/XML (Pretty): 831-1.0092975-rdf.xml
RDF/JSON: 831-1.0092975-rdf.json
Turtle: 831-1.0092975-turtle.txt
N-Triples: 831-1.0092975-rdf-ntriples.txt
Original Record: 831-1.0092975-source.json
Full Text
831-1.0092975-fulltext.txt
Citation
831-1.0092975.ris

Full Text

FACTORS ASSOCIATED WITH MOTHERS' PROTECTION OF THEIR CHILDREN F R O M ENVIRONMENTAL TOBACCO SMOKE by i i B E V E R L E Y A. TEMPLE BScN, Brandon University, 1991 ! M N , University of Manitoba, 1997 l A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF ; DOCTOR OF PHILOSOPHY in i The Faculty o f Graduate Studies (Nursing) UNIVERSITY OF BRITISH C O L U M B I A August, 2006 © Beverley A. Temple, 2006 11 ABSTRACT Children's exposure to tobacco smoke is known to have adverse health effects, yet 200,000 Canadian children are still being exposed to tobacco smoke in their homes every day. A cross-sectional descriptive study was conducted to1 identify factors associated with whether mothers/primary care givers provide a smoke free environment for their children and the stage the mothers/primary care givers were in, in relation to their decision to adopt the precaution of ensuring people do not smoke around their children. Data were collected from 571 surveys sent home with kindergarten children in two prairie city school divisions. Mothers or primary care givers completed the survey. In the bivariate analysis, significant factors associated with providing a smoke free environment included education of the mother, number of friends and family who smoke, living with a partner, being a stay at home mother and being a nonsmoker. The strongest predictor in the multivariate logistic regression was the mother's/primary care giver's self-efficacy related to providing a smoke free environment. When mothers/primary care givers had high self-efficacy scores they were more likely to provide a smoke free environment, regardless of smoking status. One surprising finding was that when other variables were controlled for, having a higher education was less likely to be associated with always providing a smoke free environment.. Being in an advanced stage of the precaution adoption process, "deciding to always provide a smoke free environment," was least likely when mothers/primary care givers had more friends and family who smoke, a lower self-efficacy, and a lower knowledge level of the effects of environmental tobacco smoke. The findings from this study expand our understanding of factors associated with mothers/primary care givers who do not always provide a smoke free environment for their children. Most importantly, modifiable factors are associated with the decision-making process and the action of providing a smoke free environment. i i i T A B L E OF CONTENTS A B S T R A C T i i T A B L E OF CONTENTS i i i LIST OF TABLES '. : vi LIST OF FIGURES viii A C K N O W L E D G E M E N T S ix CHAPTER ONE: INTRODUCTION 1 The Health Effects of ETS on Children 2 The Scope of Children Exposed to ETS • 2 The Policy Context 3 The Family's Role in Children's Exposure to ETS 4 Research Purpose 4 Significance of the Study 5 Definitions of Terms 5 CHAPTER TWO: LITERATURE REVIEW 7 The Impact of ETS on Children 7 The Health Effects of ETS on Children 7 The Prevalence of Children Being Exposed to ETS in their Homes 9 The Community Context of Children Being Exposed to ETS 11 Public Policy 11 Workplace Smoking Restrictions ••••• 12 Landlord Tobacco Use Restrictions 13 Rural versus Urban Dwelling 13 Personal Context Factors and ETS Exposure 14 Educational Level 14 Age 15 Occupation, Income and Socioeconomic Status r. 16 Marital Status 17 Crowding in the Home 17 Ethnicity 18 Smoking Patterns in the Home and ETS Exposure 18 Parents' Attempts to Reduce the Amount of Smoke in the Home 19 Tobacco Dependence Level of Mothers 22 Number of Cigarettes Smoked Per Day 22 Partners, Friends, and Relatives Who Smoke 23 Outside Area Available to Smokers 24 Social Interaction Context 24 Family Functioning : 25 Heightened Health Awareness 26 Beliefs and Knowledge of ETS 33 Self-efficacy of Mothers Related to Providing a Smoke Free Environment 36 Measurement of ETS 39 Summary of the Review of the Literature , 42 iv Theoretical Foundations for This Study 43 Synthesized Model 43 Social Cognitive Theory 46 Precaution Adoption Process Model 46 Research Questions 51 CHAPTER THREE: METHODS 52 Design 52 Setting / 52 Sample 53 Procedures 55 Instrumentation '. 57 Dependent Variables 58 Independent Variables 59 Pilot Study { 66 CHAPTER FOUR: FINDINGS 71 Description of the Sample 71 Sample Comparison to Provincial Data 74 Factor Analyses Results of Study Scales... 76 Key Study Findings • 80 Providing a Smoke Free Environment 80 Precaution Adoption Process Stages : 82 Bivariate Analysis 83 Multivariate Analysis 91 Relationship between PAP and PSFE 106 Summary 107 CHAPTER FIVE: DISCUSSION 110 Theoretical Framework 110 Synthesized Model 110 Future Research 114 Education 118 Smoking in Families 120 Beliefs and Knowledge of ETS 123 Critique of Methods Used: Limitations and Contributions 125 Summary 127 Appendix A: University of British Columbia Ethics Certificate 153 Appendix B: Brandon University Ethics Certificate 154 Appendix C: Teacher Letter 155 Appendix D: Advance Notice 157 Appendix E: Information Sheet with Survey 158 Appendix F: Survey Tool 159 V Appendix G: Reminder 1 175 Appendix H: Reminder 2 176 Appendix I: Reminder 3 177 vi LIST OF TABLES Table 1: Age of Participants and Number of Children in Household 72 Table 2: Demographic Characteristics of Participants 73 Table 3: Factor Analysis of Self-efficacy Scale 77 . Table 4: Factor Analysis of Thoughts about Smoking Scale 78 Table 5: Factor Analysis of General Smoking Knowledge and ETS Knowledge Scales : 79 Table 6: Factor Analysis of Family Functioning 80 Table 7: Participant Scores for Providing a Smoke Free Environment 81 Table,8: Proportion of Participants Reporting Being in the Various Stages of the Precaution Adoption Process 83 Table 9: Personal Context Characteristics of Those Who Always Provide a Smoke Free Environment and Those Who Do Not Always Provide a Smoke Free Environment 86 Table 10: Personal Context Characteristics of Participants Identifying Themselves in the Higher or Lower Stages of the Precaution" Adoption Process... 87 Table 11: Predicting Variables by Groups of Those Who Always Provide a Smoke Free Environment (PSFE) and Those Who Do Not Always Provide a Smoke Free Environment (PSFE) 89 Table 12: Predicting Variables by Groups of Those Who Identify Themselves in the Higher and the Lower Stages of the Precaution Adoption Process 90 Table 13: Logistic Regression Analysis Predicting Always Providing a Smoke Free Environment - Block 1 Personal Context 95 Table 14: Logistic Regression Analysis Predicting Always Providing a Smoke Free Environment - Block 2 Smoking Variables 96 Table 15: Logistic Regression Analysis Predicting Always Providing a Smoke Free Environment - Full Model 98 Table 16: Logistic Regression Analysis Predicting Higher Stage of Precaution Adoption Process - Block 1 Personal Context 101 Table 17: Logistic Regression Analysis Predicting Higher Stage of Precaution Adoption Process - Block 2 Smoking Variables. 102 vii LIST OF TABLES continued Table 18: Logistic Regression Analysis Predicting Higher Stage of Precaution Adoption Process - Full Model 104 Table 19: Relationship Between Always Providing a Smoke Free Environment and the Stages of the Precaution Adoption Process 107 viii LIST OF FIGURES Figure 1: Synthesized model of potential factors associated with the "decision" and "action" to provide a smoke free environment for children 50 Figure 2: Reconceptualized model of potential factors associated with the "decision" and the "action" to provide a smoke free environment for children 129 ( ix A C K N O W L E D G E M E N T S I wish to express my sincere gratitude to many people who supported me while I pursued the dream of completing my doctoral work. Without their encouragement I would not have been able to complete my dissertation. First I am grateful to all of the participants of this study, the children of the participants, the kindergarten teachers, and the school divisions for their positive support of the research, without them this study would not have been possible. Also I would like to acknowledge the Canadian Lung Association for their belief in the study through their generous financial support. Secondly I would like to acknowledge the contributions of my dissertation committee. I am especially grateful to Dr. Joy Johnson, chair of the committee for her continued support throughout the process. I thank the committee members for their careful guidance and positive contributions to my work: Dr. Joan Bottorff, Dr. Karen Chalmers, and Dr. Susan Dahinten. Thirdly I would like to dedicate this work to the memory of my parents Lome and Phyllis Temple who taught me the value of hard work, dedication, and perseverance; without their life lessons and ongoing support of my career over the years, I would not have had the courage to start this journey. Finally I would like to thank all of the family and friends who have contributed to this work in more ways than they will ever know; my son Kevin, my best friend and sister Bonnie and husband Russ, brother Doug and his wife Gina, study group (Donna, Elaine, Kathryn, and Sandy), friends Linda, Wendy, and Diane, my research assistants Bobby and Deanne, and many other friends and colleagues who helped with editing and provided ongoing encouragement. The love and support supplied the strength. 1 CHAPTER ONE: INTRODUCTION Despite attention to the physical harm of environmental tobacco smoke (ETS), many children continue to be exposed to ETS daily. Increasing numbers of public policies have been introduced over the last two decades in an attempt to regulate tobacco smoking, and ultimately reduce the number of people exposed to ETS in both their workplaces and public places. Although the policies and regulations have been linked to decreasing smoking rates, it has been suggested that people who are prohibited from smoking in public places will smoke in their homes (Studlar, 2002). As a result, thousands of children continue to be exposed to ETS in their homes on a daily basis. Government health care programs have increasingly placed a high priority on children's health. For instance, legislation has addressed the use of seat belts and bicycle helmets to improve safety for children. Moreover, the broad determinants of health have received growing attention, including the need for clean air and water. Governments have not paid enough attention, however, to children's ETS exposure, and illnesses related to ETS exposure in children continue to be the most common causes of hospitalization for young children (Canadian Institute of Child Health [CICH], 2000). While government regulators recognize the importance of respecting citizen's right to privacy and avoid intervening in people's homes, society has paid attention to physical, emotional, and sexual abuse of children occuring in private homes (Health Canada, 2003; Health Canada etal., 2001). The Canadian Lung Association has described the exposure of children to ETS as a form of abuse. Regulators remain reluctant to promote restrictions within homes and would have difficulty in enforcing such restrictions. As health care providers, we have been ineffective in assisting families to reduce the ETS exposure of children in their homes. In order to develop effective 2 programming, we must come to a fuller understanding of how to assist mothers to protect their children from ETS. The Health Effects of ETS on Children Smoking within a home creates a significant risk to children (Cook & Strachan, 1999; Ontario Medical Association Committee on Population Health, 2002; World Health Organization [WHO] et al., 1999). Children are more sensitive to ETS because the surface area of their lungs is large in proportion to their body weight (American Academy of Pediatrics Committee on Environmental Health, 1997). Research findings have linked ETS to an increased incidence of sudden infant death syndrome, respiratory disease, middle ear infections, and asthma (American Academy of Pediatrics Committee on Environmental Health; Cook & Strachan; WHO et al.). The Scope of Children Exposed to ETS Adults' smoking prevalence rate during the childbearing and childrearing years remains high. Twenty seven percent of 20-24 year olds and 26% of 25-44 year olds currently smoke an average of 13 cigarettes per day (Canadian Tobacco Use Monitoring Survey [CTUMS], 2005). The percentage of current smokers in Manitoba is 21%. Women in the 20-24 and 25-44 year old age groups have similar rates of 21 % and 20% respectively (CTUMS). The majority of children living with a daily smoker are regularly exposed to ETS. In 1999, approximately one in three Canadian children under the age of 12 years was regularly exposed to ETS. In Manitoba, 87% of children living with a daily smoker (slightly above the national average of 85%) were exposed to ETS (Canadian Institute of Child Health [CICH], 2000). A 2001 Canadian Community Health Survey indicates that there continues to be a high percentage of daily smokers within the 3 childbearing and childrearing age groups in Brandon, with 30% of the 20-34 year old age group reporting that they currently smoke (Statistics Canada, 2001). The Policy Context Tobacco legislation across Canada has recently increased since the National Tobacco Control Act was passed in 1997. Graphic displays and messages on cigarette packages convey a message about the dangers of cigarette smoking to pregnant women. The Canadian Lung Association and the Canadian Cancer Society have active anti-tobacco messages in the form of pamphlets and information sessions. Physicians and nurses provide one-on-one counselling, and there are medications that can be used to assist with cessation efforts (Health Canada, 2004). Many provincial governments have provided their own legislation to further reduce the consumption of tobacco. Amendments to Manitoba's Non-Smokers Health Protection Act included tobacco advertising restrictions and fines (Government of Manitoba, 2004). In 2003, the Manitoba Minister of Health established an A l l Party Task Force to recommend ways and means of protecting Manitobans from ETS. Brandon received a total of $125,000 of the tobacco control strategy fund for the purpose of public education when a complete ban on smoking in all enclosed public places was implemented in 2003 (Canadian Tobacco Control Liaison Committee [CTCLC], 2003). Many school programs in Manitoba developed strategies to promote a generation of non-smokers: smokers' help lines, on-line programs, and youth advisory groups were formed to ensure that the tobacco control initiatives were geared to youth's interests and concerns (CTCLC). To date, there has not been an evaluation of the provincial strategy to reduce smoking in Manitoba (CTCLC) or the strategy's effect on reducing exposure of children to ETS. 4 The Family's Role in Children's Exposure to ETS Parents are responsible for producing the majority of ETS within their homes. A greater proportion of the health risks associated with ETS have been attributed to maternal smoking patterns. Significant numbers of children are still being exposed to ETS on a daily basis. Parents understand the risk associated with exposing their children to ETS, but many will not quit smoking or reduce their child's exposure to ETS (Borland, Mullins, Trotter, & White, 1999; Groner, Ahijevych, Grossman, & Rich, 2000; Irvine et al., 1999; Ratner, Johnson, & Bottorff, 2001). Different family situations or contexts may contribute to ETS exposure in the family home. The effects of smoking during pregnancy and in early infancy contribute to negative health consequences for children (Cook & Strachan, 1999). Socioeconomic status and other personal context factors have been linked to smoking patterns in homes; in particular families at a lower socioeconomic level tend to have more smokers in the home (Cook et al., 1994; Irvine et al.,T997; Jaakkola, Ruotsalainen, & Jaakkola, 1994; Jarvis, Strachan, & Feyerabend, 1992; Lund, Skrondal, Vertio, & Helgason, 1998a; Mannino, Caraballo, Benowitz, & Repace, 2001; Merom & Rissel, 2001; Schuster, Franke, & Pham 2002; Stanton & Silva, 1993). There are important health benefits when families are successful at reducing ETS in the home. There is, for example, a documented decrease in the severity of asthma when parents reduce their children's exposure to cigarette smoke (Murray & Morrison, 1988). Research Purpose ETS is harmful to the health of non-smokers. Many mothers do not provide a smoke free environment for their children despite an awareness of the risks of ETS through extensive public health campaigns. Consequently, information is needed 5 regarding the factors which affect behaviours related to protecting children from ETS because there are alarming numbers of children being exposed to ETS, and many of these children are developing the harmful health effects associated with ETS. There is a significant need to identify the factors associated with mothers' or primary care givers' protection of children from ETS in the child's immediate environment (e.g., the home, the family car, other vehicles, and homes frequently visited) so that new effective interventions can be developed and tested. Significance of the Study The task of providing health education is grounded in an understanding of health behaviour and transforming knowledge into effective strategies for healthy outcomes. An understanding of the situations surrounding children will enable health care providers to develop more successful strategies to assist mothers/primary care givers in reducing childhood ETS exposure. Reducing children's exposure to ETS is an important public health issue affecting our future populations. Improving prevention strategies has been recognized as a cost-effective way to deal with many diseases. These strategies need to be comprehensive to include the family and societal contexts where the behaviour takes place. Comprehensive strategies can be developed by having more contextual nformation about children's ETS exposure. Definitions of Terms In order to minimize conceptual confusion, definitions of three key concepts are provided below. Environmental Tobacco Smoke Environmental Tobacco Smoke (ETS) is released when a cigarette burns unattended or is exhaled by a smoker. ETS is sometimes called second-hand smoke and 6 for non-smokers this is often called passive or involuntary smoking (Cunningham, 1996; WHO etal., 1999). Smoke Free Home A Smoke Free Home (SFH) is a home where there is no smoking allowed. There may be smokers living in the home, but they choose to smoke outside only. Visitors are not allowed to smoke in the home. Smoke Free Environment A Smoke Free Environment (SFE) is a setting in which no smoking is allowed. For the purposes of this study we shall focus on environments where children frequently find themselves: the home, the family vehicle, other vehicles, and homes that children frequently visit. 7 CHAPTER TWO: LITERATURE REVIEW In this chapter I examine the research literature related to ETS exposure of children in their homes. This review provided the foundation for the study and identified gaps in current knowledge. The majority of the literature in this review has been published over the previous two decades in the fields of nursing, medicine, psychology, and health education. The data bases searched to acquire research literature included CINHAL, MedLine, and Psyclnfo. Non-English literature was excluded from the search. The first section of this chapter provides a brief review of the literature on the harmful effects of ETS exposure and the prevalence of children's exposure to ETS. The second section focuses on the community context of children's exposure to ETS. The third to fifth sections consist of personal context factors, smoking patterns, and social interaction factors as they relate to ETS exposure. Finally issues related to the measurement of ETS are discussed. The Impact of ETS on Children The Health Effects of ETS on Children Numerous population-based studies have described the health consequences of children being exposed to ETS. The WHO et al. (1999) and the American Academy of Pediatrics Committee on Environmental Health (1997) produced position statements concluding the following conditions are produced or exacerbated by ETS: sudden infant death syndrome (SIDS); respiratory tract infections such as tonsillitis, bronchitis, and pneumonia; asthma symptoms; and middle ear disease. Preliminary links have also been made between ETS exposure and lower serum vitamin C levels in children (Strauss, 2001). Some of the findings that support these conclusions are summarisedbelow. 8 Sudden Infant Death Syndrome The link between ETS exposure and SIDS is widely accepted. For example, WHO et al. (1999) concluded that a relationship exists between an increased risk for SIDS and maternal smoking. The American Academy of Pediatrics Committee on Environmental Health (1997) also identified a link between ETS and SIDS in a policy statement regarding ETS and children's health. The relationship between SIDS and ETS has been demonstrated in families in which only the fathers smoked, eliminating the claim that SIDS is related mainly to the effects of maternal smoking during pregnancy (Cook & Strachan, 1999). Lower Respiratory Infections Children who are exposed to ETS are said to be at higher risk for lower respiratory infections (Li, Peat, Xuan, Berry, 1999). Cook, Strachan, and Anderson (1998) conducted a systematic review of 3,326 references to tobacco smoke pollution derived from the Embase and Medline databases. The risk of developing lower respiratory tract illness in infancy and early childhood associated with ETS exposure was found to have a pooled odds ratio of 1.48 (95% 0=1.40, 1.57). Evidence of a dose response was apparent even after confounding factors were considered. Additionally, there was an increased risk of early childhood respiratory illness when persons other than the mother smoked in homes where the mothers did not smoke (Cook, Strachan, & Anderson). The American Academy of Pediatrics Committee on Environmental Health (1997) conducted a review of epidemiologic studies and found a strong link between parental smoking and respiratory illness. For instance, infants with two parents who smoked were more than twice as likely to have pneumonia and bronchitis as infants who lived in smoke free homes. 9 Asthma ETS exposure and children's risks for developing asthma or having more severe asthma attacks have been extensively studied. Maternal smoking has been found to have a greater effect than paternal smoking (Agabiti et al., 1999; Britton & Weiss, 1997; Fergusson, Horwood, Shannon, & Taylor, 1981; Wright et al., 1991). Maternal smoking is associated with an increased incidence of wheezing illness up to the age of six but appears to be non-atopic in nature (Lewis & Britton, 1998). In cross-sectional surveys, the odds ratio for either parent smoking and children having wheezing illness was 1.35 (95% CI=1.19, 1.54) (Fergusson et al.). Cook and Strachan (1999) concluded that there is strong evidence that maternal smoking increases children's asthma disease severity, attack frequency, medication use, and life-threatening bronchospasm. Middle Ear Disease Systematic reviews have demonstrated the link between middle ear disease and ETS exposure. If either parent smoked, a pooled odds ratio of 1.41 (95% CI=1.19, 1.65) has been reported (Cook, Strachan, & Anderson, 1998). In conclusion, the literature supports the adverse health effects of ETS for children. These conclusions are founded on population-based studies as well as meta-analyses. The relationship between parental smoking and sudden infant death syndrome, acute lower respiratory illness, an increased prevalence of asthma, and middle ear disease is well established. Reducing exposure of children to ETS would result in important reductions in morbidity and mortality rates for infants and children. The Prevalence of Children Being Exposed to ETS in their Homes The prevalence rates of children's ongoing ETS exposure in their homes remain high in the developed nations. Population-based surveys have been used to document the 10 prevalence of children being exposed to ETS in their homes. Prevalence rates are fairly consistent across countries of the developed world and range from 25 - 55% (Cook et al., 1994; Erikson, Sandvik, & Bruusgaard, 1997; Jaakkola et al., 1994; Jarvis et al., 2000; Lund et al., 1998a; Lund & Helagason, 2005; Matt et al., 2004; Soliman, Pollack, & Warner, 2004; Stanton & Silva, 1993; Stronks, van de Mheen, Looman, & Mackenbach, 1997). Studies of parental smoking patterns have been conducted within a number of regions of the United States. Some of these studies have been population-based general health surveys. Using general health-related questions along with smoking-focused questions on the surveys is thought to reduce the problems of social desirability bias and mothers underreporting their smoking status. Parents are more likely to respond truthfully if they are not threatened by the survey tool and its purpose (Fink, 1995; Matt et al., 1999). One American national survey indicated that 35% of children are being exposed to ETS (Schuster et al., 2002), while Soliman et al. (2004) report a decline in ETS exposure of children in homes from 35.6% in 1992 to 25.1 % in 2000. An Oklahoma study found that 88% of smokers did not having smoking bans in the home (Kegler & Malcoe, 2002). Kegler and Malcoe also reported that 77.2% of smokers smoked in the family car. Mannino et al. (2001) reported that 39% of American children were exposed to ETS in their home. Sixty-two percent of the inner city population of Kansas City did not have smoking bans within the home (Okah, Choi, Okuuyemi, & Ahlawalia, 2002). Canadian studies indicate similar numbers of children are being exposed to ETS. The Canadian Tobacco Use Monitoring Survey (CTUMS) is a yearly survey that targets all persons 15 years of age and older living in Canada, excluding the Territories. One recent CTUMS (2004) survey had a sample size of 20,275 persons and results indicate 11 that 12% of Canadian children under 12 years of age were exposed to ETS in the home almost every day (CTUMS). British Columbia had the lowest percentage (4%) of children exposed to ETS and Quebec the highest at 23%. In Manitoba, 13% of children under the age of 12 are being exposed to ETS in their homes. The Canadian literature indicates that high percentages of children continue to be exposed to ETS, despite the fact that the national prevalence of smokers continues to decline (Ashley et al., 1998). The Community Context of Children Being Exposed to ETS Individual behaviour is only partially responsible for health. The environments surrounding an individual also contribute to health and include the structures, institutions, and families within communities (Glanz, Lewis, & Rimer, 2002). In the following section I consider the community contextual factors that influence smoking behaviour, including the protection of children from ETS. Public Policy Tobacco control polices have been implemented in Canada over many years. The regulation of tobacco has primarily occurred through taxation and legislation. Federal warning labels have been present on cigarette packages since the early 1980s and tax increases on cigarettes have increased the price of a package of cigarettes dramatically since 1985 (Studlar, 2002). The tobacco farming industry has not received government subsidies since 1990, and there has been assistance provided to farmers to develop alternatives to tobacco crops. Canadian society's acceptance of the serious health hazards posed by tobacco has been demonstrated by the number of individuals supporting a reduction of tobacco advertising and reduction of tobacco sponsorship of entertainment. Increasingly anti-tobacco groups have grown in Canada, and these organizations are 12 lobbying for support of public regulation to protect nonsmokers from the effects of ETS (Studlar). Canada has demonstrated some success with its tobacco control strategies. Campaigns to educate the public about the hazards of tobacco use and ETS have been federally directed by Health Canada and coordinated through interest groups and the provincial governments (Studlar, 2002). The effectiveness of tobacco control strategies has been evaluated using a variety of methods. Ranson, Prabhat, Chaloupka, Nguyen (2002) found that cigarette price increases are one of the most cost-effective strategies in reducing tobacco use. Conversely, Manitoba has one of the highest taxation rates on tobacco products in the country (Studlar, 2002). Despite the high cost of tobacco, and despite public education campaigns targeting pregnant women and teens, Manitoba continues to have one of the highest female teenage smoking rates in Canada (CTUMS, 2004). In sum, there is increasing public support for control strategies, but the strategies have not necessarily resulted in greater reductions of the smoking rates within certain target groups (Cunningham, 1996; McMillen, Winickoff, Klein, & Weitzman, 2003). Workplace Smoking Restrictions A second community contextual factor that influences smoking behaviour is workplace smoking restrictions. Workplace smoking restrictions affect the smoking patterns of individuals (Cunningham, 1996; Health Canada, 2000). Merom and Rissel (2001) reported that being employed in a smoke free workplace increased the likelihood of smoke free homes both in the current and former smokers' families. Fichtenberg and Glantz (2002) report that totally smoke free workplaces are associated with reductions in the prevalence of smoking-, with 3.8% fewer cigarettes smoked per day per continuing smoker (95% CI=2.8%, 4.7%). 13 Landlord Tobacco Use Restrictions Place of residence can affect smoking patterns. Accompanying the increasing numbers of public policies restricting ETS, restrictions on tobacco use in rented spaces is also occurring. Several court cases have been won where smoking has been restricted in apartments and private homes to protect others from the ETS (Cunningham, 1996). This has made it more acceptable for landlords to impose restrictions on smoking within their buildings for health and safety reasons (Health Canada, 2000). Rural versus Urban Dwelling Children living in rural areas may be at greater risk for ETS exposure. Rural citizens and municipalities have been slower to adopt more restrictive smoking policies in Manitoba (Government of Manitoba, 2004). Until recently, there were few towns that had by-laws enforcing smoke free public places. Consequently, there is no reliable evidence at this time, that all areas of the province are enforcing the provincial law that prohibits tobacco use in public places. This means that many areas of the province still have restaurants and other public places where children may be exposed to ETS outside of the family home. There may be other factors related to rural living which also contribute to greater ETS exposures of children in the rural areas. Atav and Spencer (2002) found that rural adolescents were at greater risk than suburban and urban adolescents of performing risky health behaviours, such as using tobacco. Their study reinforced the need to consider place of residence when studying health risk behaviours. Rural families may not have as many smoking cessation programs available to them and this may also contribute to a greater prevalence of smoking in rural populations (Tilson, McBride, Albright, & Sargent, 2001). 14 The family home is widely considered to be the main source of ETS exposure for children (Ashley et al., 1998; Eriksen & Bruusgaard, 1995; Irvine et al., 1997; Jarvis et al., 2000; Jarvis, Strachan, & Feyerabend, 1992; Kegler & Malcoe, 2002; Knight, Eliopoulos, Klein, Greenwald, & Koren, 1996; Mannino et al., 2001; Matt et a l , 1999; Merom & Rissel, 2001; Okah et al., 2002). Jarvis, Strachan, and Feyerabend found that even i f children had no reported exposure at home, they had elevated cotinine levels, indicating that ETS exposure took place outside of the home. They concluded that the ETS exposure probably occurred in the hallways of apartment blocks and other public places. This public exposure to ETS is particularly prominent in communities where fewer public restrictions are present; as in the case of many rural communities. Personal Context Factors and ETS Exposure Personal context factors are those factors which influence each person's context or experience of individuals, including what are often considered demographic variables. The demographic variables create a context for each person in shaping his or her beliefs and behaviours so are called personal context factors for this study. Many researchers have examined the relationships between the level of ETS exposure in the home and such personal context factors as the parents' occupational status, educational level, income, age, marital status, ethnicity, and crowding in the home (Cook et al., 1994; Irvine et al., 1997; Jaakkola et al., 1994; Jarvis, Strachan, & Feyerabend, 1992; Lund et al., 1998a; Mannino et al., 2001; Merom & Rissel, 2001; Schuster et al., 2002; Stanton & Silva, 1993). Educational Level The level of parental education has been found to consistently predict the level of exposure to ETS in the home (Eriksen & Bruusgaard, 1995; Eriksen, Sorum, & 15 Buruusgaard, 1996b; Gaffney, 2001; Jaakkola et al., 1994; Kegler & Malcoe, 2002; Lund et al., 1998a; Mannino et al., 2001; Merom & Rissel, 2001; Okah et al., 2002; Scherer et al., 2004; Schuster et al., 2002; Soliman, Pollack, & Warner, 2004; Stanton & Silva, 1993). In part this relationship exists because persons with lower levels of education are more likely to use tobacco and associate with those who use tobacco. Lund et al. examined the changes in parental smoking patterns with the presence of children as a modifying factor. While the study findings indicate that mothers with higher levels of education were less likely to have made changes in their smoking behaviour for the sake of their children, the more highly educated mothers were also less likely to smoke. Education of parents was found to be a strong predictor of children's ETS exposure even after adjusting for the amount of smoking in the home and the type of dwelling space (Scherer et al.). Age Several studies have reported that parents of a younger age are more likely to expose their children to ETS (Eriksen, Sandvik, & Bruusgaard, 1997; Merom & Rissel, 2001; Schuster et al., 2002; Stanton & Silva, 1993). An analysis of the American National Health Interview Survey suggests that the lower the parental age, the higher the level of ETS exposure among children (Schuster et al.). When a parent is less than 25 years of age, this parent represented a significant factor in daily smoking occurring in the home. Schuster et al. identified that age remains a significant factor in the exposure of children to ETS, in both homes where parents smoke and those where parents do not smoke. Even non-smoking younger parents do not provide a smoke free environment. 16 Occupation, Income and Socioeconomic Status Several studies have described the relationship between socioeconomic status and ETS exposure in the home. The classification of occupations and socioeconomic status differs by country. Regardless of the differences, the type of occupation (i.e., non-professional) is closely associated with a higher level of ETS exposure found in children in the home (Browne, Shultis, Thio-Watts, 1989; Cook et al., 1994; Irvine et al., 1997; Jaakkola et al., 1994; Jarvis, Strachan, & Feyerabend, 1992; Lund et al., 1998a; Schuster et al., 2002; Mannino et al., 2001; Merom & Rissel, 2001; Schuster et al., 2002; Stanton & Silva, 1993). Cook et al. found a strong association between levels of cotinine in children and social class. Cotinine levels were found to be eight to nine times greater in children from lower social class families than cotinine levels in children from the upper class families. The researchers indicate that the higher cotinine levels in lower class children was found in both families who reported exposure in the home and those who did not report exposure, and these findings are likely related to the practices of the broader community of lower class families (Cook et al). The cotinine levels related only to social class were proposed to be different in the case that children from lower socioeconomic classes live in apartment buildings with poor ventilation and children's ETS exposure is often occurring in other public places. However, these results should be interpreted cautiously as the number of children from the lower-class was not highly represented in the sample (16%). Lund et al. (1998a) reported that low socioeconomic status (SES) parents exposed their children to more smoke, in comparison to higher SES parents, but low SES parents were just as likely to have tried to change their smoking behaviour because of their children. Lund et al. concluded that rules about smoking seem to be less likely to succeed 17 in lower socioeconomic homes due to the relatively high prevalence of daily smoking, a lower level of education, and increased likelihood of a single parent raising the children. Marital Status Single parent households present a greater risk for smoking and exposure of children to ETS in the home (Eriksen, Sorum, & Bruusgaard, 1996b; Eriksen, Sandvik, & Bruusgaard, 1997; Jaakkola et al., 1994; Jarvis et al., 1992; Kegler & Malcoe, 2002; Lund et al., 1998; Schuster et al., 2002). Children are twice as likely to be exposed to ETS in homes with single parents (90% of single parents are mothers) than with two parent families, regardless of the parental smoking status (Jaakkola et al.). Few studies have gone beyond identifying that single mothers produce the most ETS in homes. Crowding in the Home Crowding in the home measured as the number of people living in the home compared to the number of rooms, is often used as a proxy measure of socioeconomic status. A significant relationship has been identified between crowding in the home and the amount of ETS in the home (Jarvis et al., 1992; Scherer et al., 2004). This relationship was significant in both smoking and non smoking homes. The amount of crowding was proportional to the levels of ETS, as measured in children's saliva cotinine. The level of crowding needs to be considered separately from other measures of socioeconomic status, particularly for some transient populations where their temporary living conditions may not reflect their usual socioeconomic status (CICH, 2000). For instance, families may move into a smaller residence temporarily to facilitate seasonal work or school and return to a primary residence where there may be less crowding. 18 Ethnicity Social norms that are present in different cultures may have an impact on the acceptability of smoking for different members of the family. In Chinese cultures, there are fewer mothers smoking than fathers (Chen, L i , Yu, & Qian, 1988). Emmons et al. (2001) found that i f interventions were specifically developed to meet the requirements of the ethnic community, the culturally sensitive interventions would be effective in reducing ETS in the home for both Spanish and English parents, for example. In Canada, there are higher percentages of aboriginal children living in households with daily ETS exposure in comparison to other ethnic groups where ETS exposure is not as prominent (CICH, 2000; Sexton et al., 2004). Therefore, ethnicity is important to consider when examining not only the smoking patterns in the home, but also when planning future interventions to meet the needs of parents. In summary, a number of personal context factors are closely related to ETS exposure. The ETS exposure of children appears to be highest when a parent is young, single, and has a lower level of education and a lower paying job. The literature to date does not provide information about any interactions of the above factors which may be associated with ETS exposure of children. Smoking Patterns in the Home and ETS Exposure There is a clear increase in the likelihood of ETS exposure for children when there are smokers living in the home. The research findings discussed below indicate which factors affect the amount of smoke in the home each day, including: the dependence level of the smokers, the number of cigarettes smoked per day, the number of people smoking in the home, and the availability of an outside area in which to smoke. 19 Parents' Attempts to Reduce the Amount of Smoke in the Home The ETS produced by mothers has the most deleterious health effects on young children in the home (Ashley et al., 1998; Eriksen & Bruusgaard, 1995; Irvine et al., 1997; Jarvis et al., 2000; Jarvis, Strachan, & Feyerabend, 1992; Kegler & Malcoe, 2002; Knight et al., 1996; Mannino et al., 2001; Merom & Rissel, 2001; Okah et al., 2002; Scherer et al., 2004). Researchers suggest one primary reason is that the majority of mothers spend more time in close proximity to young children than fathers, thereby having a greater effect on children's ETS levels (Ashley et al.; Irvine et al.; Jarvis et al.; Kegler & Malcoe; Knight et al.). When both the parents smoke or when only the mother smokes the physiological measures of children's ETS cotinine levels are elevated in either case (Ashley et al.; Irvine et al.; Jarvis et al.; Kegler & Malcoe; Knight et al.; Scherer et al.). The factors influencing potential changes in smoking behaviour and the possible reduction in ETS for children are highlighted in a number of intervention studies (Emmons et al., 2001; Groner et al, 2000). In one such study, Groner et al. (2000) explored the impact of a brief intervention on maternal smoking behaviour. The study involved a three group repeated measures design with random assignment to one of the three groups. Two of the groups received a smoking cessation intervention either focused on the effects of smoking on their child (called the child health group, n = 153 mothers) or focused on the effects of smoking on the mothers themselves (called the mother health group, n = 164 mothers). The third group was the control group of 162 mothers visiting a health clinic in a low-income neighbourhood who were not given smoking cessation advice. The interventions were delivered during the wait times for their pediatric appointment. Mothers were also given a self-help manual on smoking cessation at that 20 time. Reminders of the educational information were sent to members of the intervention groups at two week and four month post-intervention intervals. Participants were also contacted by phone at one and six months post-intervention. The primary outcome measure was smoking status, and the secondary outcomes were changes in smoking patterns, place of smoking, stage of change, and ETS knowledge. Results indicated that all groups had a significant drop in the number of cigarettes smoked, but none of the groups experienced a significant change in smoking status. The child health group showed significant changes in the location of their smoking compared with the control group at six months. The mothers' knowledge of the effects of ETS in the child health group also improved and was retained at the eight-month follow-up. These researchers suggest that while mothers may not be willing or able to quit smoking for their own health, they are willing to make changes in order to protect their children. In another study, motivational interviewing was used to engage parents in protecting their children from ETS (Emmons et al., 2001). Those in the intervention group (n = 150) were recruited from one of eight health clinics over a two-year period. The control group (n=141) received self-help smoking cessation information. The motivational interviews were conducted in the participants' homes, and were followed-up with four counselling phone calls. Motivational interviewing focused on the areas of concern for the participants and aimed at enhancing personal responsibility for change and self-efficacy. Measures of potential exposure to ETS included measuring the levels of household air nicotine and the participants' carbon monoxide levels. Goal setting was used to help parents identify the steps they could take to either reduce or relocate their smoking in order to protect their children. This action followed feedback about the levels of nicotine and carbon monoxide. The control groups were mailed a copy of a smoking 21 cessation manual. Feedback about household nicotine levels was provided after the final follow-up assessment. The outcome measures were the self-reported smoking patterns and the passive smoke found in the homes, which was measured using vapour-phase nicotine emissions with passive sampling monitors in two rooms of the homes. The researchers indicated that the motivational interviewing did reduce the amount of ETS in the homes of families with very young children, but the smoking cessation rates of the participants did not change. The nicotine levels found in the homes of the control group did not change over the study period and the smoking cessation rate of participants also remained stable. A significant treatment effect was found, as the 6-month ETS exposure levels were significantly lower in the motivational interview group (Emmons et al.). Johansson, Hermansson, and Ludvigsson (2004) studied the smoking parents of 366 children aged 2.5 to 3 years old, using a questionnaire. Urine cotinine analyses were used to corroborate the self-report and to identify the smoking patterns which produced the lowest cotinine levels in children. A control group of 433 age-matched children from nonsmoking homes also had cotinine levels assessed. Not living with a partner and high cigarette consumption were the two most important variables in resultantly high levels of ETS in the home. It was also concluded that smoking outdoors was the most effective way to protect children from ETS exposure in homes where there are smokers. These three studies, two intervention and one descriptive, illustrate that the smoking patterns of those living in the home have an impact on the amount of ETS exposure of children in the home. Some interventions are effective in reducing the amount of ETS in the home without necessarily reducing the amount smoked by the parents. The most effective way to achieve a reduction of ETS in the home is for the 22 parents to smoke outside (Emmons et al., 2001; Groner et al., 2000; Johansson et al., 2004). Tobacco Dependence Level of Mothers Not surprisingly, families who have members who smoke have been shown to be the least likely to provide a smoke free environment in their home (Ashley et al., 1999; Eriksen & Bruusgaard, 1995; Irvine et al., 1997; Jarvis et al., 2000; Kegler & Malcoe, 2002; Knight et al., 1996; Mannino et al., 2001; Merom & Rissel, 2001; Okah et al., 2002). It is more difficult for a parent with a personal dependence on tobacco to change his or her patterns of use. The patterns of smoking of members in the home affect the decisions that mothers/primary care givers make about providing a smoke free environment. Two studies have found that the higher the smokers' level of dependence, the less likely the smoker will be able to change his or her habits and reduce the children's exposure to ETS (Ericksen, Sorum, & Bruusgaard, 1996a, 1996b). Unfortunately, researchers have not consistently examined the level of dependence and the smoking patterns of mothers/primary care givers or participants (Helgason & Lund, 2001; Irvine et al., 1997; Jarvis, Strachan, & Feyerabend, 1992; Jarvis et al., 2000; Lacchetti et al., 2001; Mannino et al., 2001; Matt et al., 1999; Merom & Rissel, 2001; Voorhees, Schrieber, Schumann, Biro, & Crawford, 2002). Tobacco dependence needs to be considered in the context of smokers' decisions about providing a smoke free environment. Number of Cigarettes Smoked Per Day Another important predictor of ETS exposure includes the number of cigarettes smoked per day by the parents and others in the home. The number of cigarettes smoked per day has been used as part of the measure of tobacco dependence (Fagerstrom; 23 Fagerstrom & Schneider). When more cigarettes are smoked per day, the variability of dependence scores decreases. This reduced variability results in more similar scores between the number of cigarettes smoked and the dependence scores, thereby proving the number of cigarettes alone can be a reliable measure of dependence (Fagerstrom, 1978; Fagerstrom & Schneider, 1989). Researchers have found that current smokers, particularly heavy smokers (usually defined as more than 10 cigarettes per day) were less likely to have smoking rules pr restrictions in the home (Cook et al., 1994; Eriksen & Bruusgaard, 1995; Eriksen, Sandvik, & Bruusgaard, 1997; Irvine et al., 1997; Kegler & Malcoe, 2002; Knight et al., 1996; Lund et al., 1998a; Mannino et al., 2001; Okah et al., 2002; Scherer et al., 2004). Quit attempts are often related to a reduction in the numbers of cigarettes smoked per day. Current smokers were more likely to have smoking restrictions in the home if they had had at least one quit attempt in the last year. Moreover, those with at least one recent quit attempt were assessed at a higher stage of change, thus indicating a plan to decrease or stop tobacco use (Kegler & Malcoe, 2002; Okah et al.). Partners, Friends, and Relatives Who Smoke While many studies have examined the smoking habits of mothers and spouses, only a few have considered other residents or visitors who are allowed to smoke in the home (Jarvis, Strachan, & Feyerabend, 1992; Kegler, & Malcoe, 2002; Matt et al., 1999; Merom & Rissel, 2001). There is more smoking and fewer restrictions in the home with an increased number of relatives and family members who smoke (Eriksen, Sandvik, & Bruusgaard, 1997; Kegler & Malcoe; Okah et al., 2002; Scherer et al., 2004; Venters et al., 1987). The relative with the greatest influence on mothers' smoking patterns is the 24 partner (Eriksen & Bruusgaard, 1995; Eriksen, Sandvik, & Bruusgaard, 1997; Okah et al.;Ratner etal., 2001). Outside Area Available to Smokers When smokers are making an attempt to change the amount of smoke inside the home, it is important to have an outside area where smokers can go to smoke (Jarvis, Strachan, & Feyerabend, 1992; Kegler & Malcoe, 2002; Matt et al., 1999; Merom & Rissel, 2001). Intervention studies that have proven successful in reducing ETS for children in the home have often not resulted in changes in the smoking patterns of parents (i.e., amount smoked), but rather in the location where the smoking takes place (Groner et al., 2000). In other words, the reductions in ETS in the home often result from parents smoking outside. The housing options for lower socioeconomic status families are more limited, with more crowded living spaces and with fewer options to be outside (Kegler & Malcoe; Lund et al., 1998). Consequently, the availability of an outside area for smokers could be important when aiming to reduce the ETS in the home. Social Interaction Context Family interactions can have an impact on an individual's behaviour. In particular, a mother's behaviour within a family is influenced by the family's overall functioning. The interaction patterns that occur between parents and children have an impact on the way that families interact and resultantly behave (Wright, Bell, & Rock, 1996). When an illness arises, a family's awareness of health is heightened and there is some indication that their patterns of behaviour may change. Researchers have found that this heightened health awareness also occurs when someone living in the home is pregnant and when there is an infant in the home (Eriksen, Sorum, & Bruusgaard, 1996a, 1996b; Jaakkola et al., 1994). 25 Mothers are often the family member who transfers knowledge, beliefs, and influences others' behaviour in the home (Tilson, McBride, Albright, & Sargent, 2001; Wright et al., 1996). The mother's ability to have these interactions within the family and therefore provide this influence could be affected by her beliefs and knowledge of ETS and her self-efficacy related to providing a smoke free environment. The areas of family functioning, heightened health awareness, beliefs and knowledge of ETS and self-efficacy of mothers related to providing a smoke free environment are discussed in this section. Family Functioning Family dynamics or relationships affect a great deal of the health behaviour and understanding of health within a family (Epstein, Bishop, & Levin, 1978; Epstein, Ryan, Bishop, Miller, & Keitner, 2003). Family dynamics are affected by the composition of the family, and interpersonal relationships within the family. Family dynamics affect decision-making processes, including decisions made about providing a smoke free environment (Epstein, Bishop, & Levin; Whitehead & Doherty, 1989; Wright et al., 1989). Family decisions are most commonly based on the pattern the family has adopted for handling behaviour both within and outside the family (Epstein, Bishop, & Levin). In addition, the family will develop a number of functions to enforce what they consider acceptable behaviour in order to maintain their style of behavioural control (Epstein, Bishop, & Levin). The general functioning of the family needs to be considered in relation to decisions about providing a smoke free environment. While the relationships between the adults within a home may have a significant impact on the decisions of individual family members, mothers affect the health beliefs of family members the most (Epstein, Bishop, & Levin; Voorhees, Schrieber, Schumann, Biro, & Crawford, 2002; 26 Whitehead & Doherty, 1989; Wright, Bell, & Rock, 1989). This decision making includes the choices made about smoking restrictions (Whitehead & Doherty; Wright et al.). Decisional and relationship issues have rarely been considered in relation to providing a smoke free environment. Relationship issues within families may create added stress when one member attempts to influence the smoking behaviour of another member. For instance, a mother who asks her abusive husband to reduce his smoking within the home may be placed at greater risk for abuse. Two studies considered relationship issues and smoking behaviour in the home: one study focused on the dynamics of the marital relationship around smoking behaviour when an older spouse is il l (Wright et al., 1989); while the second study examined the extent to which smoking was a bonding or separating factor in relationships and control (Whitehead & Doherty, 1989). Both of the above qualitative studies used a family systems approach to understand the smoking behaviour within families (Whitehead & Doherty; Wright et al.). The researchers found that smoking is indeed an important part of relationships between family members and one member's behaviour influences the other family members within the family system. These family relationships around smoking have not been considered in previous research examining the circumstances under which parents provide a smoke free environment. Heightened Health Awareness Illness in the child Parental assessments of the risk presented for their child have been studied in relation to obesity, asthma, and glycemic control in diabetes (Clark, 1997; Frindik, Williams, Johnson, & Dykman, 2002; Hadad, Franca, & Uchoa, 2002; Hodges, 2003; 27 Yoos, Kitzman, McMullen, & Sidora, 2003). The parent's perceptions of the risks posed by different illnesses or health conditions have often been found to be inaccurate when compared to the health professionals' assessment of the risks for the children in the same situations. For example, in one study, the children were not identified as overweight by many of their parents, despite the fact that the children fit the description for obesity (Hodges). In addition, asthma severity is often not assessed accurately by parents. Parents make their assessment based on an inaccurate appraisal of the state of their child's health. This inaccurate appraisal is founded on their own understanding of the asthma, not the actual state of the child's health. For example, parents may assume that coughing consistently at night is within their child's norm, when it is actually a sign of worsening asthma severity (Clark; Yoos et al.). Studies regarding the parental treatment of diseases such as asthma have shown that parental understanding or interpretations of their child's symptoms influence the care and treatment that parents provide to their children (Clark). Thus, the knowledge of the risk in general, the parental belief about the risk, and the belief in the severity of the risk, all play a part in risk assessment. Finnish researchers examined the factors that determined children's exposure to ETS in the home where one or both parents smoked (Jaakkola et al., 1994). In this population-based cross-sectional study with a sample size of 1,003 children, the researchers found that children who demonstrated some degree of atopy were less likely to be exposed to ETS while asthmatic children were not likely to be exposed at all (Jaakkola et al.). Parents were asked about the child's occurrence of respiratory symptoms, current asthma, and other infectious diseases including ear infections within the last 12 months. The risk of being exposed to ETS was significantly lower for atopic children, indicating that parental knowledge of the child's illness was associated with a 28 change in the smoking behaviour of the parents. This may be attributed to the extensive public education campaigns about the connection between asthma and smoking, and it may be linked to the number of physician visits that children with chronic illness have in comparison to children in good health. The researchers concluded that the parents made the association between smoking and their child's health through education and experiences with the health care system and as a result understood that ETS contributed to their child's illness (Jaakkola et al.). Clark (1997) found that parents who perceived that their child was more severely i l l with their asthma were more likely to make environmental changes to influence their child's health. The major factor influencing the maintenance of these environmental changes, such as the removal of pets and reducing ETS, was the parents' perception of the severity of the child's illness. The more severe the parents perceived their child's illness, the more likely they were to maintain changes in the environment. Butz and Rosenstein (1992) studied the ETS exposure in four groups of children: children with asthma, cystic fibrosis, rheumatoid arthritis and children in good health in a cross-sectional study of 634 families. More than 80% of the families with children with either asthma or cystic fibrosis reported changing their smoking behaviour after the diagnosis of their child's illness as compared to only 39% of the other two groups. It was important to note however that 45% of children with cystic fibrosis and 28% of children with asthma were still reported to have daily exposure to ETS in households where parents smoke in the home. A child's illness may not always be associated with a change in parental smoking behaviour. Kohler, Sollich, Schuster, and Thai (1999) found that 295 children who were admitted to hospital with either cystic fibrosis or asthma were exposed to ETS in the 29 home. Among the families of children with cystic fibrosis and asthma, quantitative and self-report measures were not consistent. Quantitative measures of the ETS exposure of the children with asthma and cystic fibrosis exceeded the parental reports of ETS exposure. In children with cystic fibrosis and asthma, the number of parents admitting that their children were exposed to ETS was 29% when using parental self-report and 65% by physiological measurement. The difference was as little as 18% between self-reports of the parents and the physiological measures in children without respiratory disease. The conflicting information obtained from parents about ETS exposure of their children may be a function of the desire of the parents to reduce their own responsibility for the hospitalization of the child, or an inaccurate assessment of their child's illness and the role of ETS in that illness. Three intervention studies involved parents of children with asthma. In the first study, set in Scotland, families were considered eligible i f they had a child aged 2-12 with documented asthma who lived with a parent or guardian who smoked. Families were randomly assigned to the intervention or control groups (Irvine et al., 1999). The intervention consisted of a brief visit in the home to discuss the effects of passive smoking on children with asthma and potential benefits to the child with reduced ETS, as well as the financial and health benefits of quitting smoking. Information was provided about smoking cessation and about reducing household ETS by changing locations of smoking and asking visitors not to smoke in the home. Parents in the control group were given a pamphlet on smoking cessation but did not receive additional information about the need to reduce ETS to protect their children. Families were revisited after one year. Data analysis was based on 213 intervention families and 222 control families. The primary outcome measure was salivary cotinine concentrations in children and changes in 30 smoking habits of the parents at one year follow-up. A small decrease in salivary concentrations was found in both groups, but this was nonsignificant. The authors concluded that the intervention was ineffective in reducing asthmatic children's exposure to ETS (Irvine et al.). In a second study, families were recruited when children were i l l , and hospitalized with a respiratory illness, most having asthma (Winickoff, Hillis, Palfrey, Perrin, & Rigotti, 2003). A free smoking cessation program was offered to the 71 participants. The intervention included a short counselling session using motivational interviewing, one week of free nicotine replacement therapy, two follow-up telephone calls, a letter to their physician and information about a help quit line. Written information was provided to parents that addressed specific concerns discussed during the counselling session. The main outcome measures were: changes in smoking habits, such as quit attempts and numbers of cigarettes smoked per day; readiness to change smoking behaviour around the child; and attitudes about the dangers of ETS to the child. At the two month follow-up, significantly fewer parents reported smoking in the home and car, and more parents reported having rules prohibiting smoking in the home than at baseline. These significant improvements remained even when the analysis was completed with the continued smokers. Compared to the previously discussed study, there was no control group in this study for comparison of outcomes. The results of this study suggest that parental beliefs about the negative effect of ETS on their child may have more of an impact on their behaviour if the intervention is provided during a time that is more critical to the child's health. In a third study involving parents of children with asthma, Hovell et al. (1994) found that an intervention based on behavioural modification principles was successful in 31 reducing exposure of children with asthma to ETS in the home. The study included three groups: a usual treatment, a monitoring, as well as an experimental group. The usual treatment grouped received only outcome measures, the monitoring group only monitored the smoking, tobacco smoke exposure and the children's asthma symptoms, and the experimental group included the monitoring as well as a 6-month series of counselling sessions designed to decrease the ETS exposure in the home. Exposure to parents' cigarettes in the home decreased in all groups, with a 79% reduction in the experimental, 42% in the monitoring group and 34% in the usual treatment group. Only the experimental group maintained a decrease in children's exposure to cigarettes in the home at the 12 month follow-up. Research evidence to date indicates that there is a link between parents' perception of their child's illness and their smoking behaviour. But, perception of illness is not consistently a factor that influences provision of a smoke free environment for a child. Parental perception of a child with an illness and the perceived severity of that illness needs to be understood in assessing factors influencing a smoke free home, but again this cannot be considered in isolation from other factors in the home. Other times of heightened health awareness for the family also include when there is an infant in the home and when someone pregnant is living in the home. t The presence of an infant in the home A Norwegian intervention study focused on parents who attended a well-child clinic and the intervention involved a brief educational session about the harmful effects of ETS and how to reduce their children's ETS (Eriksen, Sorum, & Bruusgaard, 1996b). Two hundred and twenty-one parents in the intervention group were given brochures focused on reducing ETS in the home and smoking cessation resources available in the area, after a brief clinic session. The parents were also provided with a self-help manual on smoking cessation. Parents completed two questionnaires: including one prior to the intervention and one mailed questionnaire completed one month following the intervention. The authors indicated that there were no significant changes in the intervention or control groups for reductions in ETS in the home, or in smoking behaviour of the parents. The authors also compared families with an infant in the home with those families with older children and found that families with an infant were more likely to make some positive changes (Eriksen, Sorum, & Bruusgaard). Someone pregnant living in the home Society has made smoking while pregnant an unacceptable behaviour (Cunningham, 1996). Women tend to be more likely to quit smoking or reduce their smoking while pregnant (Haslam & Draper, 2000). Haslam and Draper identified that women who were pregnant were more likely to be contemplating quitting smoking and were more likely to agree with the health risks associated with smoking and their pregnancy. In this study there was no significant effect of age, whether the pregnancy was planned, or whether the woman had a child with asthma or respiratory infections. The researchers propose that mothers were more likely to be aware of risks when more ready to quit smoking themselves. According to one Canadian study, continuing to smoke during pregnancy is associated with the presence of other smokers in the household, having other children in the household, and not having postsecondary education (Paterson, Neimanis, & Bain, 2003). Sockrider, Suchanek Hudmon, Addy, and Dolan Mullen (2003) proposed that establishment of a nonsmoking policy in the home early in pregnancy was important in ensuring infant ETS avoidance over time. Other predictors of an ongoing home smoking policy included: the mother's confidence to ask others not to 33 smoke and the mother's perceived difficulty in preventing exposure. Similarly adults tend to modify their behaviour when someone else'who is pregnant is living in the home. The desire to protect the unborn child is portrayed as an important task for society (Cunningham; Eriksen, Sorum, & Bruusgaard, 1996b; Secord, 2000). Beliefs and Knowledge of ETS Parental beliefs and personal behaviours have an impact on the health of children and are especially relevant to children's exposure to ETS in the home. Parental behaviours are influenced by their beliefs about potential negative health outcomes for themselves and for their children (Crone, Reijneveld, Burgmeijer, & Hirasing, 2001; Glanz et al., 2002; Hovell et a l , 1994; Hovell, Zakarian, Matt, Hofstetter, Bernert, & Pirkle, 2000; Roden, 2003). For example, parental risk perception of dangers for their children has been studied in regards to illness, seat belt use, and other preventive measures, such as helmet use for bicycles (Ehrlich, Longhi, Vaughan, & Rockwell, 2001; Hansen, 1994; Simpson, Moll , Kassam-Adams, Winston, & Miller, 2002). Education or knowledge about a risk is related to the beliefs that people hold about the risk (Weinstein, 1988). To believe that a threat actually poses a risk, parents must first be aware of the risk through education, and then parents may or may not believe in the potential of that threat as a risk for their children. A parent must first be aware of the disease and then believe that the disease could actually pose a treat for his or her child to influence a parent's behaviour (Clark, 1997; Weinstein, 1988). Parental education and beliefs work together to develop the risk awareness. Parental beliefs about the effect of ETS on the health of their children have been found to influence their actions (Ashley et al., 1998; Crone et al., 2001; Helgason & Lund, 2001; Hovell et al., 1994; Kegler & Malcoe, 2002; Lund et al., 1998b; McMillen et 34 al., 2003). Helgason and Lund assessed the attitudes of parents towards ETS and health risk awareness regarding the potential hazards of ETS for young children. Knowledge of risk awareness was assessed using four questions. Parents were asked i f children who are exposed to ETS are more likely to: start smoking themselves, have inner ear infections, develop respiratory diseases, and be more likely to have asthma attacks. The findings of this study indicate that a dose-response relationship was present between parental level of health risk awareness regarding the potential hazards of ETS for children and ETS exposure in children. The reported level of risk awareness indicated that many parents were still unaware of the potential hazards of ETS for their children. Despite insufficient risk awareness, almost 90% of parents agreed with the statement that children should have the right to live in a smoke free environment. This implies that being aware of the effects of ETS and a belief in the right of a child to a smoke free environment do not always translate into behaviour to provide a smoke free environment. The effects of parental beliefs on the ETS exposure for their children have been explored in different ways. Researchers have assessed the presence of rules about smoking in the home where there are children present, beliefs and understanding of the hazards presented for children by exposure to ETS; and parental attitudes to restrictions being placed on adult smoking to protect children. Some researchers (Lund et al., 1988; Matt et al., 2004) found that parents who smoke make some efforts to reduce the amount they smoke around their children, demonstrating some awareness of the potential effects of ETS on their young children. The Lund et al. study demonstrated that homes were more likely to be smoke free where parents had an understanding of the dangers to children from smoke exposure. Hovell et al. (2000) found that seven sessions of counselling over a 12-month period were effective in not increasing the amount of ETS 35 children were exposed to; the control group had an increase in ETS levels at the end of the study period. The counselling sessions, based on behavioural modification principles, were focused on reducing the child's ETS exposure. The counselling sessions were founded on the understanding of the complexity of changing smoking behaviours and gradually working with mothers to set goals, evaluate them and incorporate some rewards for changed behaviour. Using cross-sectional data from the annual Social Climate Survey of Tobacco Control in the US, McMillen et al. (2003) identified that small important improvements in adult attitudes and practices related to children's ETS exposure occurred between 2000 and 2001. However, there are still significant numbers of adults reporting ignorance of the harmful effects of child ETS exposure (Lund & Helgason, 2005). Smoking bans were reported by participants in 69% to 74% of households and 84% to 88% reported smoking bans in the presence of children (McMillen et al). The desire of mothers to reduce their guilt for exposing their children to ETS was described in a grounded theory study of nine mothers with preschool children and a , history of smoking or having a significant other who smoked (Secord, 2000). Secord found that mothers drew on their knowledge about ETS, their wish for a healthy child, and societal and personal expectations of "good mothering" to establish rules to protect their children from ETS. In response to the competing needs of others, mothers made adjustments to their initial rules over time. Repeated exceptions being made to the initial rules resulted in less restrictive guidelines regarding ETS in their homes. Women used several strategies to deal with the contradictions that making these changes created in their lives. Mothers would try to seek agreement from others, minimize the effects of ETS, hide their smoking, provide explanations about their tobacco dependence, and live 36 in the hope that they would one day quit smoking. From the women's perspective their actions relieved their guilt about exposing their children to ETS, which allowed them to still view themselves as good mothers and to some extent avoid the judgement of others. This study illustrates that these mothers' beliefs and understandings about the effects of ETS influence their behaviour, but the behaviour may not always reflect what the mothers know to be best for their children. Self-efficacy of Mothers Related to Providing a Smoke Free Environment Self-efficacy is an important concept within social cognitive theory (Bandura, 1997). "Perceived self-efficacy refers to beliefs in one's capability to organize and execute the courses of action required to produce given attainments" (Bandura, p 3). Studies have explored parenting self-efficacy for a variety of behaviours. Self-efficacy has been shown to increase as parents become more experienced with parenting, and self-efficacy is predictive of parenting satisfaction and adaptation to the new parenting role with infants (Elek, Hudson, & Bouffard, 2003; Lambden, 2001; Reece, 1992; Reece & Harkless, 1998). Caring for children with a chronic illness (e.g., asthma) has also been considered in relation to parental self-efficacy (Hanson, 1994; Heerman, 1988; Helseth, 1995; Mesters, Meertens, Crebolder, & Parcel, 1993). Targeting parental self-efficacy for specific interventions related to management of their child's asthma resulted in improved parental management skills. Specific patterns of behaviour of family members are also related to smoking in the home. Love, Davoli, and Thurman (1996) conducted a study to examine the degree of consensus among 311 health behaviour change professionals regarding the personal and environmental factors they believe most influence health behaviour decisions related to smoking cessation, exercise, and weight loss. The health professionals provided a 37 judgement rating on the probability that persons shown in vignettes would initiate the behaviour in question. The authors claim that behavioural intentions, self-efficacy, and social support were considered the most powerful determinants of plans to initiate weight loss, regular exercise, and smoking cessation. The success that mothers/primary care givers may have in asking others not to smoke may be influenced by the confidence that they feel in certain situations, and with certain people (Bandura, 1997; Crone et al., 2001). Bandura indicates that self-efficacy comes from several important sources. There is more success in new behaviours when there is mastery experience, such as successful previous experiences. Mothers/primary care givers must feel that they have what it takes to succeed and then be able to persevere in light of adversity. The amount of effort expended to be successful and previous success with behaviour are going to influence their feelings of confidence. If a mother/primary care giver has tried to ask family members previously to not smoke in their home and this resulted in friction and negative emotional responses, they are less likely to feel confident in their ability to establish a smoke free home. Repetition of the behaviour would influence the mother's/primary care giver's confidence for such complex behaviours, such as smoking patterns within families (Bandura; Borelli et al., 2002; Roden, 2003). Crone et al. (2001) found in a cross-sectional study of parents of babies between the ages of 1 and 14 months, from the Netherlands, that the mother's self-efficacy in asking others to not smoke was associated with the ETS exposure of the children. They also found that the self-efficacy of mothers differed between smoking and non-smoking mothers. In a national population-based study, Crone, Reijneveld, Willemsen, and Hirasing (2003) were successful in reducing environmental smoke for infants over a four-year 38 period, with interventions focusing on the opinions, knowledge, motivation, and skills of parents to reduce ETS for infants. The intervention was progressive using three phases: phase one involved training the health care professionals; phase two was directed at parents attending well-baby clinics, with written materials and education provided by the nurses and doctors; and the third phase was directed towards the family and friends with T V advertisements. The second phase of the intervention strategy was based on the premise that parents would have greater self-efficacy to reduce the ETS around young children when their knowledge and motivation was improved. Parents in the precontemplation phase of the Stages of Change were targeted because it was felt that the parents may have taken some action to change and would benefit from specific educational strategies. The outcome measures were the national prevalence rates of ETS exposure among young children, so attributing the change to the intervention is impossible, although no other national campaign about the same topic was implemented during the study time. This study did not measure if parents demonstrated an increase in self-efficacy following the intervention. In summary, family functioning influences the decision-making that occurs in the family and the ability to follow through with plans to provide a smoke free environment. Another one of the factors proposed to be important in influencing a smoke free home is heightened awareness of health. This awareness can occur when someone who is living in the home is pregnant; an infant is in the home; or when a child is i l l . There is often increased contact with the health care system during these times. As families interact with health care providers, families may become more aware of their health and this may have an impact on the smoking patterns and ETS exposure in the home i f health care providers are providing education about the dangers of ETS. Mothers'/primary care givers' beliefs 39 and knowledge of the negative consequences of ETS for their children may affect their desire or motivation to provide a smoke free environment. The amount of ETS knowledge may also have an impact on the development of self-efficacy of mothers/primary care givers related to providing a smoke free environment. Self-efficacy is a concept which is specific for the behaviour being performed; therefore further examination of mothers'/primary care givers' self-efficacy to ask others not to smoke and its relationship with the amount that children are exposed to ETS is important. Confidence in asking others not to smoke in the home and limiting their own smoking in certain social situations is likely to have an effect on the ETS in the home. Measurement of ETS Methodological issues surrounding the measurement of ETS in homes where children are living is complex. A number of these issues need to be addressed when considering a study of ETS in the home. One of these issues is physiological measures of ETS and exposure of children to ETS. Physiological measures involve a variety of methods which have had differing success and most often require taking physical samples from children. A second issue is that some researchers argue that social desirability of responses may interfere with mothers providing an accurate account of their smoking patterns and of the amount of ETS in the home. These issues and their implications for this study are discussed in this section. Physiological measures of ETS exposure of children have often included the measurement of urine, saliva, hair or plasma cotinine levels (Boyd, Windsor, Perkins, & Lowe, 1998; Cattais et al., 2003; Conway, Woodruff, Edwards, Hovell, & Klein, 2004; Johansson, Hailing, Hermansson, Ludvigsson, 2005; Parker, Lasater, Windsor, Wilkins, Upequi, & Heimdal, 2002; Scherer, Meger-Kossien, Riedel, Renner, & Meger, 1999; 40 Vartianen, Seppala, Lillsunde, & Puska, 2002; Willers, Hein, & Jansson, 2004). The use of these types of physiological measures means that parental consent would be required and this may eliminate data on high risk children whose parents would not consent to sample taking. Some researchers have identified increased refusal rates by parents when requests for physiological measures of children are requested (Eriksen, Sorum, & Bruusgaard, 1996a, 1996b). The result could be that a study may be missing important data about a large group of children being exposed to ETS. Researchers have examined the consistency of self-report and the results from physiological measures. The use of cotinine concentrations in urine, saliva and plasma have generally been found to coordinate well with self-report as a measure of smoking status. Measures of smoking status have included a variety of scales and methods of self-report including; the number of people who smoke in the home (Cook et al., 1994; Irvine et al., 1999), the number of cigarettes per day smoked in the home (Boyd et al., 1998; Jarvis, Tunstall-Pedoe, Feyerabend, Vesey, & Siloojee, 1987; Parker et al., 2002; Vartianen et al., 2002; Wills & Cleary, 1997) as well as subjective scales such as the perception of smokiness in the home (Jurado, Munoz, De Dios Luna, & Fernandez-Crehuet, 2004). The use of physiological measures to corroborate reported smoking status has generally produced similar results (Boyd et al., 1998; Cook et al., 1994; Christensen et al., 2004; Glasgow et al., 1998; Irvine et al., 1999; Jarvis et al., 1987; Jurado et al., 2004; Parker et al., 2002; Vartianen et al., 2002; Wills & Cleary, 1997; Woodruffetal.,2003). Social desirability of responses has been a general term used to describe the tendency among respondents to distort answers in ways that make themselves look better or to avoid making them look bad. People may be more reluctant to report behaviour that 41 they feel is less socially acceptable (Bickman & Rog, 1998). Studies of smoking status of pregnant women and also of parents with ill children are examples of the times that this may be a relevant issue. During pregnancy, mothers may not be honest about their smoking to avoid appearing to be a poor mother (Boyd et al., 1998; Secord, 2000). Parents are also vulnerable when they have an ill child, so may be less honest in their responses (Callais et al., 2003; Jaakkola et al., 1994). To address this issue, when sensitive material is being asked of respondents, special attention needs to be paid to respondent's confidence to respond honestly. The use of anonymous surveys are an important way to make respondents feel that they are less vulnerable than when a researcher is speaking directly to them, with either telephone or face-to-face surveys (Bickman & Rog, 1998; Dillman, 2000; Czaja & Blair, 2005). Ensuring the respondents' confidentiality is also important to the likelihood of them responding more honestly (Bickman & Rog). In relation to survey construction, it is important to reduce the extent to which respondents will perceive that particular answers will be interpreted in a negative way. The introduction to a survey needs to reflect the importance of accurate answers to the survey tool and to the importance of the study (Bickman & Rog; Fink, 1995). Gaffhey, Molloy, and Maradiegue (2003) conducted a systematic review of questionnaires used for the measurement of infant ETS exposure using 60 studies published between 1996 and 2002. A l l studies used a brief investigator-developed questionnaire using parental self-reports. The authors identify that some areas not measured include protective factors related to reducing infant's exposures to ETS, and none of the studies recorded exposure to ETS in the family car. Many of the studies used only a few questions to determine exposure and therefore demonstrated little internal 42 consistency. The authors also note that many of the studies in the review included self-report and a biological measure of cotinine and that a strong positive association between these measures of ETS exposure was reported by most researchers. In summary, physiological measures of children's exposure to ETS requires consent from parents and is expensive to administer. Self-report methods of smoking status have been found to be consistent with physiological measures. Therefore self-report measures can be used to describe the actions that mothers are taking to provide a smoke free environment, especially when confidential survey tools with effective questions are used. Summary of the Review of the Literature This chapter has provided a review and analysis of the literature with relevance to children being exposed to ETS. The areas analyzed included the community context of ETS exposure in children, personal context factors, smoking patterns in the home, and social cognitive factors. The literature demonstrating the prevalence of children being exposed to ETS and damaging health effects for children is extensive. The health effects are most profound for young children who spend a great deal of time in the home. The personal context of mothers/primary care givers who expose their children to ETS has been well documented with many population-based studies. The risk of children being exposed to ETS is greatest when mothers are young, single, have a lower level of education, and have lower paying jobs with few smoking restrictions in the workplace. This description of the high risk group, summarizes many of the study findings, but does not go beyond describing the high risk groups where special attention may be required in future intervention studies. The literature identifies many potential factors associated with the exposure of children to ETS, but no studies have used any multivariate analyses to 43 develop a clearer understanding of the modifiable factors which are having the greatest influence. Intervention studies have not proven helpful in preventing ETS exposure in children. The success of interventions to reduce or prevent ETS exposure of children would be more effective with greater understanding of the modifiable, factors, rather than simply a description of the at-risk groups. The exposure of children is not a simple behaviour carried out by one person or several persons within a home. Smoking tobacco is complicated by beliefs and knowledge of the outcomes of the smoking behaviour for others, and by family dynamics that have an impact on mothers'/primary care givers' ongoing decision-making and behaviour about smoking. Preventing ETS exposure is a complex problem and needs to be considered within a theoretical framework that can guide a clearer description of the potential modifiable factors that are associated with mothers'/primary care givers' protection of their children from ETS. Theoretical Foundations for This Study Theories consist of sets of statements that tentatively describe, explain or predict interrelationships between concepts (Knapp, 1998). In order to better explain and understand human needs and experiences, theory based research is required. The theoretical perspective underlying this study draws on a synthesis of the concepts and relationships from the literature review, concepts from the Social Cognitive Theory (Glanz et al., 2002), and the Precaution Adoption Process Model (Weinstein, 1988). Synthesized Model Research results reveal a number of factors associated with mothers'/primary care givers' protection of their children from ETS. Concepts that have been proposed to be associated with mothers/primary care givers protecting their children from ETS have 44 been conceptually grouped and represented in a tentative theoretical model (see Figure 1). Smoking behaviour in a home occurs within a context. The community context is the first key concept and encompasses: where the family lives, whether rural or urban; the public smoking policies being enforced; the workplace smoking policies as well as any landlord smoking restrictions. Many previous studies have identified demographic factors that are associated with an increased risk of exposure of children to ETS. The model includes the most commonly identified demographic factors and this grouping is labelled the personal context. The factors included in personal context are education, age, marital status, ethnicity, occupation, income and crowding in the home. The next most commonly identified area of influence is the smoking patterns found in families. These factors include: the number of smokers who live in the home, the number of family and friends who smoke, spouse smoking, number of cigarettes smoked by the mother/day, tobacco dependency level of the mother, and an outside area in which to smoke. The third area is the social interaction contextual factors. These factors are related to individual and family factors related to behaviour around ETS exposure within a family. The first is the heightened health awareness which is related to mothers'/primary care givers' risk awareness. Previous research has suggested that awareness is heightened when an infant is in the home, when someone pregnant is living in the home, and when there is a child who is i l l . The effect of an ill child in the home becomes more important i f parents perceive that the child may be seriously i l l . The second factor is family functioning. As explained in the literature review, more positive family functioning is more likely related to families making decisions to protect their 45 children from ETS. The other two factors which will affect the interactions within a family about tobacco smoking are related to the mothers'/primary care givers' beliefs and knowledge related to ETS and also the mothers'/primary care givers' self-efficacy related to providing a smoke free environment. The literature indicates that socioeconomic factors are related to smoking, with individuals from a lower socioeconomic home being more likely to smoke. Consequently it is postulated that children who live in low-income families are at greatest risk for smoke exposure and disease morbidity (Emmons et al., 2001). But these effects may be mediated by an increased awareness of the effects of smoking and the way that a family interacts about smoking and its health effects within the family. This model will be used to guide the analysis of possible effects on providing a smoke free environment for young children. The outcomes to be predicted are twofold: the action of providing a smoke free environment and the decision-making process that mothers/primary care givers use related to providing a smoke free environment. The actual behaviours around providing a smoke free environment are often different from the decision-making process but both occur simultaneously. Stage theories have often been used to assist in the explanatory power of decision-making related to health behaviour change. Stage theories recognize that decision-making often does not occur in a linear fashion. Both the behaviour and the decision-making process are influenced by the context which includes the three main areas summarized above. i Within the model two important theories are used to supplement the model. The self-efficacy concept from the social cognitive theory will be highlighted as an important concept in the successful performance of behaviour. The precaution adoption process 46 model will be used as a possible model which could expand the understanding of decision-making in families around providing a smoke free environment Social Cognitive Theory In the area of smoking behaviour change, the models that have been used most extensively include perspectives of personal behaviour change from models such as the Health Belief Model (Groner et al., 2000; Glanz et al., 2002) and the Transtheoretical Model and Stages of Change (Prochaska & Prochaska, 1999), as well as Social Cognitive Theory (Bandura; Gehrman & Hovell, 2003). Some of these models of individual behaviour change have several concepts in common, including the concept of self-efficacy. Self-efficacy is often viewed as the concept which provides a framework for many aspects of the social cognitive theory. A review of empirical evidence related to reducing the home ETS exposure of youth between 1987 and 2002 concluded that interventions can be effective in reducing youths' exposure to ETS (Gehrman & Hovell, 2003) and models for future protection from ETS. This study provides support to the need to continue to examine concepts from the Social Cognitive Theory in relation to different types of health behaviour change. Precaution Adoption Process Model One stage model often used in relation to smoking cessation literature is the Transtheoretical Model, Behaviour Change. This stage theory provides an integrative model for understanding reasons for not changing as well as identifying readiness to change. The stages of change are thought to guide more specific interventions and therefore to be more successful (Prochaska & Prochaska, 1999). This change theory, however, has been applied mainly to individual smoking readiness to change, not to change related to providing a smoke free environment (Patterson, Neimanis, & Bain, 47 2003). The Transtheoretical Model provides one model to assess the expected change behaviours and barriers faced by individuals in moving from one stage of readiness to another. The Precaution Adoption Process model may provide this kind of insight into a family health behaviour change. Stage theories emphasize that people at different points in the process of change behave in qualitatively different ways. Weinstein and Sandman (1992) identified four essential elements of stage theories including: 1) a category system to define the stages; 2) an ordering of the stages; 3) common barriers to change that people face in the same stage; and 4) different barriers to change that people face in different stages. People pass through each stage at very different speeds, thus making the stages undetectable. Progression is not considered linear. Theoretical perspectives from the Precaution Adoption Process Model (PAP) (Weinstein, 1988; Weinstein, Klotz, & Sandman, 1989; Weinstein, Lyon, Sandman, & Cuite, 1998) may be useful in explaining the complex behaviour of mothers/primary care givers providing a smoke free environment. The PAP theory provides a way in which to consider the change of personal behaviour for the health benefit of others. The PAP emphasizes the development of beliefs and intentions that eventually lead to action (Weinstein, 1988; Weinstein, Klotz, & Sandman, 1989; Weinstein, Lyon, Sandman, & Cuite, 1998; Weinstein & Sandman, 1992). The PAP includes seven stages: Stage 1 - A person is unaware of the health issue. Stage 2 - Once they learn something about the issue, they enter a stage of awareness. This does not mean that the issue engages a person. Stage 3 - A decision-making process about the issue is commenced. The decision-making process leads to one of two outcomes (Stage 4 or 5). 48 Stage 4 - No action is taken. OR Stage 5 - The decision to adopt the precaution is made. Stage 6 - Action is taken to adopt the precaution. Stage 7 - The maintenance stage, in which the action or behaviour is maintained over time. This model involving a seven-stage process was first described in 1992 by Weinstein and Sandman. The model has been used in research to describe behaviours related to osteoporosis prevention, mammography, hepatitis B vaccination, and home testing to detect radioactive radon gas (Weinstein, Klotz, & Sandman, 1989; Weinstein, Lyon, Sandman, & Cuite, 1998; Weinstein & Sandman, 1992). Researchers have described some of the factors that may influence movement between stages. These include: media messages about the hazard, perception of the risk, and the influence of significant others, to name a few. Mothers in different stages of decision-making about the risks associated with tobacco smoke will provide different levels of smoke free environments. More attention needs to be given to mediators or factors associated with specific behaviours (Glanz et al., 2002). The use of the PAP provides the opportunity to describe the factors associated with the decision-making process associated with mothers'/primary care givers' behaviours in providing a smoke free environment for their children. The primary purpose of this study was to examine factors that influence whether mothers/primary care givers provide a smoke free environment for their children. Providing a smoke free environment is a complex process and cannot be thought of as a single health behaviour affected by a single influence or factor. The preceding literature review has revealed a number of factors that are associated with the decision to provide a 49 smoke free environment and also the required behaviour(s) to provide a smoke free environment. In summary, the behaviour measured in this study is complex, with many factors that may influence the outcome. Figure 1 presents the proposed model of the factors synthesized from the literature to account for the decision that mothers/primary care givers make about providing a smoke free environment and the action of providing the smoke free environment. 50 Figure 1. Synthesized model of potential factors associated with the 'decision' and the 'action' to provide a smoke free environment for children. ' Community Context: Public Policy; Rural versus Urban; Workplace and Landlord Smoking Restrictions Personal Context Smoking Patterns Social Interaction Context • Educational • Smokers in the • Heightened health level home awareness • Number of • Marital status cigarettes mother • Family functioning smokes/day • Age • Number of friends • Self-efficacy of and family who mother related to • Ethnicity smoke PSFE • Tobacco • Occupation dependency level of • Beliefs and mother knowledge of ETS • Family income • Outside area to smoke • Crowding in the • Spouse smoking home Stage of Precaution Providing a Smoke Free Adoption Process (PAP) Environment (PSFE) Stage of decision-making process Behaviours related to PSFE related to PSFE 51 Research Questions The purpose of this study was to examine factors associated with mothers'/primary care givers' Provision of a Smoke Free Environment in their child's immediate environment (e.g., their home, the family car, others' cars and homes frequently visited). Participants included mothers/primary care givers of children attending kindergarten in two cities in Manitoba. The research questions addressed in this study are as follows: 1. What is the frequency with which participants report Providing a Smoke Free Environment (PSFE) for their children? 2. In relation to providing a smoke free environment, what proportion of participants report being in the various stages of the Precaution Adoption Process (PAP)? 3. What is the relationship between personal context factors and a) the stage of PAP and b) PSFE? 4. What is the relationship between smoking patterns and a) the stage of PAP and b) PSFE? 5. What is the relationship between the social interaction context and a) the stage of PAP and b) PSFE? 6. Which of the following factors - personal context, smoking patterns, and social interaction context - are most predictive of a) being in the highest stage of PAP and b) always PSFE? 7. What is the relationship between the stage of the PAP and PSFE in the identified population? 52 CHAPTER THREE: METHODS Design The study used a descriptive, correlational survey design with a cross-sectional time frame to examine factors associated with mothers'/primary care givers' provision of a smoke free environment. Data were obtained through self-report questionnaires completed by mothers of kindergarten children in two school divisions in the province of Manitoba. This chapter describes the setting, recruitment process and procedures, ethical issues, and instrumentation. J Setting Brandon and Portage La Prairie, two cities in western Manitoba, were chosen for this study. These cities were selected to provide a sample population of mothers/primary care givers from two close geographical settings. The populations of Brandon and Portage La Prairie are 42,242 and 12,976 respectively. Both cities have an agricultural heritage and economy base. The populations of Brandon and Portage La Prairie are made up of a diverse group of people from a wide range of socio-economic groups. The education of the Brandon population ranges from 25% with a university education, 24% with a grade 12 as their highest educational level, and 11% having less than grade 12. In comparison, the Portage population is less educated ranging from, 15% with a university education, 17% with grade 12 education and 39% with less than grade 12. In both cities, respiratory disease is above the provincial average; 15% of the Brandon population were admitted for respiratory disease in the last year while provincially 12% of the population were admitted for respiratory disease. Smoking percentages of Brandon residents for the 20-34 and 35-44 year old age groups are 30% and 21% respectively. The Portage residents are very similar with 30% of the 20-34 year old age group reporting smoking 53 and slightly more of the 35-44 year old age group reporting smoking at 27%. A smoking by-law prohibiting any smoking in Brandon public places was passed in May 2002 and enacted September 2002 (Statistics Canada, 2003). Portage La Prairie had one smoking by-law in place regulating smoking in municipal facilities (Portage La Prairie, By law 7321, April 1991). On October 1, 2004, amendments to the Manitoba Non-Smoking Health Protection Act came into effect prohibiting smoking in enclosed places accessible to the public and indoor workplaces (Government of Manitoba, 2004). Sample The population for this study consisted of mothers/primary care givers of kindergarten children within the Brandon School Division and Portage La Prairie School Division. The responding adults had to read English in order to answer the survey questionnaire. A whole population approach was. used, and all mothers/primary care givers of kindergarten students in the Brandon and Portage La Prairie School Divisions were surveyed. The majority of the schools were within the city limits of the two communities. Two schools from the Brandon School Division, one in a small adjacent town and one at a Canadian Forces Military Base, were included in the surveyed population. There were two schools located in small towns within the Portage La Prairie School Division and they were included in the study population. The population size was approximately 750 children, with 500 in the Brandon School Division and 250 in the Portage la Prairie School Division. This study aimed for a response rate of 50-80 percent to achieve a sample size of 375-600. This sample size was estimated to provide sufficient power to perform the multiple regression analyses (Munro & Page, 1993). The sample which consisted mainly of mothers was accessed through the school divisions as this is one of the only places where access to all mothers of young children is 54 possible. Young children have been shown to be most vulnerable to environmental tobacco smoke. It is important to have the ability to contact the mothers/primary care givers of young children to understand factors influencing the home environment of these children (Groner et al., 2000). Literature suggests that mothers/primary care givers are the people who spend the most time with, and who make the most decisions regarding care of, young children (Eriksen et al, 1997; Merom & Rissel, 2001; Stanton & Silva, 1993). Ethical Considerations The study was approved by the Behavioural Research Ethics Board at the University of British Columbia (Appendix A), and the Brandon University Research Ethics Committee (Appendix B), and was supported by the school boards of both the Brandon and Portage la Prairie School Divisions. The mothers'/primary care givers' consent was assumed if they returned a completed questionnaire. Participation was voluntary. Mothers/primary care givers could return a blank survey form in the envelope and receive no further reminders; this helped to ensure that teachers would not know who had completed or who had not completed the surveys. Data were collected anonymously and no identifying information was solicited, other than basic demographic information. Mothers/primary care givers were reminded that the schools of their child would not have access to the completed surveys and that only grouped data would be reported. The surveys were identified with a case number; lists identifying the student and the case number were maintained for the data collection period to allow appropriate reminders to be sent. The identifying information was destroyed at the end of data collection. Every child's name was entered into a draw for a bicycle at each participating school division. The value of the bicycle and helmet was approximately $300. The other incentive for 55 each child was a pencil, valued at approximately 30 cents. Each kindergarten teacher was provided with information about the school attended by the child who won the bicycle, so that all participants were aware that the bicycle was awarded. There was no known risk to participants in this study. The time required to complete the survey was approximately 30 minutes. There was no direct benefit to the participants for completing the survey questionnaire. A brief report of the survey results will be supplied to the grade one class this year and several presentations will be offered to the teachers in the two school divisions regarding the findings of the study. Procedures Recruitment for the study was conducted in the Brandon and Portage La Prairie School Division kindergarten classes between May 5 and June 2, 2005. Survey forms were sent home along with the children's regular school notices. Approval had first been obtained from the Brandon School Board by sending a letter of request to conduct the study, outlining procedures and supplying the board with a copy of the survey form. A letter of approval was then received from the Brandon School Board. The Portage School Division had a request filtered through one of the Assistant Superintendents who received all of the information and presented this at a Portage la Prairie School Board meeting. Following the Board approval, I supplied the same information to all principals from the division and attended one of the principal's meetings to answer questions about the study. A letter of approval for access was then received from the Assistant Superintendent. Following this access approval process, communication with the school divisions occurred in a couple of different ways. The kindergarten teachers received a letter explaining the survey in the week prior to the data collection period (Appendix C). A notice was sent to mothers/primary care givers the week prior to distribution of the 56 survey (Appendix D). The survey package consisted of a short information notice (Appendix E ) , the survey tool (Appendix F), and a return envelope. The first reminder (Appendix G) was sent home to all participants one week following distribution of the survey package. The second and third reminders (Appendices H and I) were sent home to those participants who had not yet returned the survey form. The third reminder consisted of a new package containing a new survey tool and a new return envelope. The surveys were returned in the sealed return envelope and the teachers did not have access to the survey responses. Nonresponse bias may have occurred for several reasons. During follow-up to the study, one reason for potential nonresponse was revealed by a Brandon school principle. She reported that one teacher had not sent home all of the surveys because she believed that she received very little information back from those families anyway. Bickman and Rog (1998) indicate that with a response rate of 75% or greater, the nonresponse group would need to be very different to affect the total estimates of the sample, therefore for this study with at response rate of 76%, the general effect would be considered small. If individual questions are missed, this is another way that nonresponse bias is introduced into the data. Many people may not have answered a specific question such as the income and introduce bias in this way. There was no indication of a 'pattern' of nonresponse to specific questions in this study. A l l children were given pencils engraved with the saying "Thank you for keeping my air clean" as a thank you for carrying the surveys home to their mothers/primary care givers. A l l children were notified that they were eligible to have their name put into the draw when they returned their surveys. A picture of a bicycle was hung in each classroom as a reminder to return the surveys. The plan was to enter all children's names into the 57 draw but children were not told this at the beginning of the study, providing an 'implied' incentive. Al l children were notified their names were in the draw whether they had returned the survey or not. A l l mothers/primary care givers were encouraged to return their surveys, regardless of whether they had been completed. The same research assistant or the researcher attended the schools so that the staff became familiar with the people involved in the study. Instrumentation The survey tool consisted of 52 questions, divided into the categories of: household information, health status, personal smoking status, smoke free environments, and demographics. The section on household information began with information about the number of people living in the home, the presence of anyone who is pregnant, and the size of the living accommodations. The health status section contained seven questions about the family. The first was about general family functioning, and illnesses diagnosed in children, with follow-up questions about the parental understanding of the risks that the illnesses pose for their children. The personal smoking status was divided into sections to evaluate the respondent's personal smoking status for a total of 15 questions. A l l participants were asked to respond to a 12-item scale describing their thoughts about smoking. The smoke free environments section included 12 questions on topics such as workplace smoking, spouse or partner smoking, other relatives and friends who smoke, smoking within the home, and smoking restrictions that may have been in place in the home. The survey also asked mothers/primary care givers about potential plans to make changes in smoking restrictions, how often they are able to maintain a smoke free environment, and a scale for mothers/primary care givers to describe their approach to providing a smoke free environment. The last section was a 10-item scale asking about 58 the mother's/primary care giver's confidence in providing a smoke free environment in certain social situations. The final section of the survey tool included demographics and consisted of nine questions about the respondent's age, gender, education, ethnicity, income, occupation, marital status, relationship to the kindergarten child, and whether they live within the city limits. The entire questionnaire consisted of 38-46 questions (depending on skip patterns). The survey incorporated a number of items from existing scales and, where necessary, newly developed specific items. Measurement of the study variables, including consideration of item origin and reliability and validity of scales used, are discussed below. Dependent Variables The two dependent variables for this study are Providing a Smoke Free Environment and the stage of the Precaution Adoption Process (PAP). Providing a Smoke Free Environment The first dependent variable, Providing a Smoke Free Environment (PSFE), was measured using a summative index which was developed for this survey to indicate how often the four places frequently populated by children were smoke free: the home, the family vehicle(s), other people's vehicles, and homes frequently visited. The four response choices for each item were "never," "some of the time," "most of the time," and "always." The items were totalled to provide a score ranging from 0-12, with 0 indicating "never" PSFE and twelve "always" PSFE. Stage of Precaution Adoption Process To measure the stage of the Precaution Adoption Process (PAP) we used a self-report item in which respondents were asked to identify which of five statements best represented their current decision making about PSFE. The response choices were: 1) "I 59 haven't given it much thought," 2) "I have decided I don't need to do anything," 3) "I am planning to reduce my child's.exposure to environmental tobacco smoke within the next 6 months," 4) "I have started to reduce my child's exposure to environmental tobacco smoke," and lastly, 5) "I have decided to always provide a smoke free environment for my child." These statements relate to key stages of the PAP and were developed specifically to evaluate the decision-making process that mothers/primary care givers follow in their progression toward always providing a SFE. This item allows mothers/primary care givers to be placed into categories where they believe they fit in the decision-making process. The responses have been developed in a format similar to the identified stage descriptions of other behaviours measured using the Precaution Adoption Process Model (Weinstein, 1988; Weinstein, Klotz, & Sandman 1989; Weinstein, Lyon, Sandman, & Cuite, 1998; Weinstein & Sandman, 1992). Independent Variables The independent variables arise from the research questions. Community context, personal context, smoking patterns, social cognitive factors were the categories of independent variables included in this study. The following section provides a description of the measurement of each of the independent variables. Community Context The policies of the community surrounding a family provide a greater public context for the practices that occur within a family. Workplace smoking restriction. Workplace smoking restriction was measured using Question 32, which asked "If you are working outside the home, are there cigarette smoking restrictions at your workplace?" and gave three response options, "yes," "no," or "not working." 60 City dwelling. The lifestyle of families and how much time they spend outside is affected by where they live - within or outside of city limits. The city dwelling variable produced a dichotomous variable using Question 52, which asked respondents if they lived within the city limits. Landlord restriction. A family who lived within a rented space may not choose to have a smoke free space, unless they are required to do so by a landlord. This variable was measured using Question 7 which asked "If you live in a rented space, does the landlord restrict smoking inside?" Response categories included "yes," "no," and "not applicable." The "yes" response was the most theoretically important response so the "not applicable" and "no" responses were combined. Personal Context Factors. Nine questions were used to measure the personal context factors. Categorical data were obtained from the questions addressing educational level, marital status, ethnicity, occupation, and family income. Crowding in the home and age ' were also assessed. Most questions were derived from the CTUMS (2004) and the National Population Health Survey (Statistics Canada, 1998). Educational level. The mother's/primary care giver's educational level was assessed using Question 46, asking about the highest level of education he or she had attained. There were 10 response categories from "no schooling" to "completed university." Marital status. Question 50 asked about current marital status and had six categories for response: "now married and living with wife/husband", "common-law relationship/live with partner," "separated," "divorced," "widowed," and "never married (single)." Age. An open-ended Question 44, "What is your age?" indicating the response should be a value in years. 61 Ethnicity. Ethnicity was ascertained asking "How would you describe yourself?" (Question 47). This measure of ethnicity was meant to be a general measure of self-proclaimed ethnic origin and not meant to measure cultural behaviour. There were 6 named categories but an "other" category was provided if there was no suitable descriptor for the mother/primary care giver. Occupation. Question 49 asked the participants to indicate their main activity over the last 12 months, allowing them to indicate if they were "looking for work,", "a student,", "retired," "working at a job or business," "raising a family or running the household" or "other." The second part of the question noted that if participants were working they should indicate their occupation. Family income. Participants were asked for an estimate of total household income before taxes and deductions over the last 12 months. The eight response categories of income ranged from "less than $15,000" to "more than $120,000" including a "don't know" category. Crowding. Crowding in the home was defined as the number of people in the home divided by the number of rooms in the home (not including the bathroom(s) and kitchen). Questions 2 and 3 asked "How many people regularly live in the home?" and "How many rooms are in your home, not including the kitchen and bathrooms?" The responses to these two questions were used to calculate a crowding score: the higher the number, the greater the degree of crowding. The concern of crowding in the home arises from studies previously done regarding smoking in the home (Irvine et al., 1997; Jarvis et al., 1992; Mannino etal., 2001). 62 Smoking Patterns Variables related to the smoking patterns category included the presence of smokers in the home, the number of cigarettes smoked by the mother/primary care giver per day, the partner smoking, the number of friends and family who smoke, the dependence level of the mother/primary care giver, and whether an outside area was available to smoke. These variables were assessed using several items from the National Survey - Canadian Tobacco Use Monitoring Survey (CTUMS, 2003), the Fagerstrom Test for Nicotine Dependence (Fagerstrom, 1978, 1982; Fagerstrom et al., 1996), and some specific items developed for this survey. Presence of smokers in the home. This variable was measured using Question 35 which asked "Excluding yourself, how many people who smoke cigarettes live in your home?" There were 5 response categories: "none," "1 person," "2 people," "3-5 people," and "more than 5 people." Number of cigarettes mother/primary care giver smokes per day. A question with four categories of numbers of cigarettes smoked per day was included to be a part of the Fagerstrom dependence score, providing categorical data. Question 24 provided interval data to supplement the Fagerstrom score, part b asked for the respondents' best guess of the number of cigarettes they smoked per day. Partner smoking. Question 33 asked if a spouse or partner smoked cigarettes and the response categories were "Yes" or "No." Number offriends and family who smoke. The number of friends and family who smoked was assessed with Question 34 which used 5 categorical responses: "none," "1 person," "2 people," "3-5 people," or "more than 5 people." 63 Tobacco dependence level of mother/primary care give. The Fagerstrom Test for Nicotine Dependence is a 6-item scale used to measure the level of nicotine dependence (Fagerstrom, 1978, 1982; Fagerstrom, et al., 1996). The total of Questions 21 through 26 provides an overall level of nicotine dependence, with a score greater than 4 being indicative of dependence (Fagerstrom et al., 1996). Evidence supports the reliability and validity of this instrument (Fagerstrom, 1978, 1982; Fagerstrom et al., 1996). Many smoking studies have used the^scale (Fagerstrom, 1978; 1982; Fagerstrom & Schneider, 1989) and criterion related validity has been established through the use of physical testing of cotinine levels correlated with the scores on the scale (Fagerstrom & Schneider). Outside area available to smoke. The availability of an outside area was addressed in Question 6 asking "Is there an outside area where people could smoke if they chose?" Response categories were "yes" or "no." A'similar question was used in other surveys about smoking in the home (Irvine et al., 1997; Jarvis et al., 1992 & Mannino et al., 2001) but this question was edited specifically for this survey. Social Interaction Context Heightened health awareness and family functioning are two of the variables that constitute the social interaction context factors. These two variables receive a score based on several questions from the survey tool. The other two variables are self-efficacy of the mother/primary care giver related to providing a smoke free environment and beliefs and knowledge of ETS. Heightened health awareness. Families are more likely to have heightened awareness of harmful health practices when an infant is present in the home, when someone living in the home is pregnant, or when a child has been diagnosed with an illness (asthma, 64 allergic diseases, ear infections, respiratory illnesses), and especially when that diagnosed illness is believed to be severe. Question 1 asked for the number of children in the home, and the presence of an infant was based on the ages of the children. Question 4 asked if anyone who currently lived in the home was pregnant and had a "yes" or "no" response. Mothers/primary care givers were asked in Question 9 i f their child had been diagnosed by a doctor with any of the diseases: asthma, allergic diseases, ear infections or respiratory illness such as tonsillitis, bronchitis or pneumonia, using a "yes" or "no" response. The mothers/primary care givers were also asked i f a child had been diagnosed, if they have been ill in the last 12 months, using a "yes" or "no" response. These questions have been adapted from previous respiratory studies (CTUMS, 2003; Eriksen et al., 1996; Jaakkola et al., 1994). An additive effect was proposed with a value of 0 - 4 with a score of 0 when none of the situations were present in the home, and 4 when all of the situations were present in the home. The situations were: 1) infant in the home, 2) someone living in the home who was pregnant, 3) a child had been diagnosed with one of the identified illnesses, 4) the child had been il l with one of the illnesses in the last six months. The cumulative effect provided the heightened health awareness score. Family functioning. Family functioning was measured using the general functioning scale of the McMaster Family Assessment Device (Epstein, Baldwin, & Bishop, 1983; Miller, Ryan, Keitner, Bishop, & Epstein, 2000; Miller, Ryan, Keitner, Bishop, Epstein, 2000b). The scale asked the family member to rate agreement or disagreement with how well an item described his or her family by selecting from four alternative responses: "strongly agree," "agree," "disagree," and "strongly disagree." There were 12 items included in the scale. Three examples of statements were: "in times of crisis we can turn to each other for support," "we cannot talk to each other about the sadness we feel," "making decisions is 65 a problem for our family." Scores for the scale range from 1- 4 with 1 reflecting the most healthy functioning and 4 the most unhealthy functioning (Epstein et al.). The McMaster Family Assessment Device (FAD) has been based on extensive qualitative work and the resultant conceptual framework, the McMaster Model of Family Functioning (Epstein et al., 1978; Miller, Kabacoff, Epsein, & Bishop, 1994). Reliability and validity of the scales within the FAD have been established (Epstein et al.; Miller, Epstein, Bishop, & Keitner, 1985; Nelson, 2003). Concurrent validity of the FAD was established by comparing the FAD to two established scales: the FACES II scale, which was developed to measure the dimension of adaptability and cohesion of the family; and the Family Unit Inventory dimensions of family integration and adaptive coping (Miller et al.; Nelson). The general functioning scale of the FAD was hypothesized to have a substantial relationship with both of these other measures (r > .50). The results of the study identified correlations very close to the predicted ranges and provided evidence of concurrent validity of the FAD. The general functioning scale of the FAD correlates with two scales of each model, ranging in values from .48 - .75 with p < .01 (Miller et al.). Smoking restriction self-efficacy. This variable was to measure mothers'/primary care givers' self-efficacy or confidence in their ability to enforce their rules about smoking restrictions in the home around some potentially difficult social situations. Question 43 asked participants how confident they were that they could keep their home smoke free in these 10 specific situations. Examples of the situations were: "when my spouse or close relative wants to smoke in our home," "when I am frustrated or anxious," "when I am feeding young children," "when the children are in bed." The situations are rated using a 5-point Likert Scale with 1 as "not at all confident" and 5 being "very confident". The total of these items provided an overall measure of self-efficacy. The score range was 66 10-50 with 50 being the most confident. The scale was based on Bandura's (1997) theoretical principles, which are specific for the behaviours being studied. Bandura has established a format for items to measure self-efficacy, but indicates that items should be based on supporting literature. The factor analysis of this scale is described in the findings chapter. Similar scales have been used with demonstrated validity and reliability (Velicer, Diclemente, Rossi, & Prochaska, 1990). Beliefs and knowledge of ETS. The mother's/ primary care giver's beliefs and knowledge about the harmful effects of ETS was measured with question 31. The question addressing this variable was a 10-item scale using statements about smoking and its health effects, measured on a 4-point Likert Scale, using "strongly agree" to "strongly disagree" (Jarvis et al., 2000). Reversed scoring was required for 5 of the items and resulted in a total score ranging from 12-48. The larger the score the more aware the mother/primary care giver was of the harmful effects of ETS. This scale has been revised from the CTUMS (2003), using terms consistent with the rest of this study. The factor analysis of this scale is discussed in the findings chapter. . Pilot Study Focus groups were conducted to evaluate the survey tool prior to the pilot study. The focus groups took place in the Wawanesa School and Hartney School of the Southwest Horizon School Division. Wawanesa is a community of 500 people located 50 kilometers southeast of Brandon. Hartney is a town of 460 people located 74 kilometers southwest of Brandon. Both communities have an aging population with about 15% of the population in the 75+ age group. The average income for both towns is low, with an average household income in Wawanesa being $35,605 and average household income in Hartney being $31,285 per year. The Wawanesa School had a kindergarten class of 15 67 students, and Hartney School had a combined kindergarten arid Grade one class of 16 students at the time of the study. The recruitment notices for the focus group were sent home with the children and written responses from the mothers were received. Arrangements for the focus groups were made over the telephone using the number that the mother supplied on her response. The schools provided a quiet space in a small classroom to conduct the focus groups. Mothers signed the consent form and then completed the survey form. Mothers were given the survey assessment form to allow them time to think about the questions asked prior to discussion. A discussion was held about issues that the mothers had encountered while completing the survey tool and any comments that they had about the content of the questions. The surveys were collected and some mothers had comments written on the survey tool. The mothers took an average of 15 - 30 minutes to complete their survey. Question 8 and Question 11 were identified as being more difficult to answer. Negatively worded questions were recognized as taking more time to complete and were reworded as a result of this feedback. It was also noted that the response categories on two questions using similar response categories were reversed (strongly agree to strongly disagree) and it would make it easier to respond if they were the same. The mothers did not find the survey worded in a negative way or in a way that would make other mothers uncomfortable in answering the questions. The mothers acknowledged that there were questions that were sensitive, but they did not have a problem in answering them. Former smoking mothers noted that it was easier to answer the questions in private. Some did not wish to answer the income question. One mother stated that she never answers income questions on surveys. General comments made by 68 mothers about the survey tool were that the questions were clear.; understandable and easy to complete. The pilot study was conducted at the Souris School, also in the Southwest Horizon School Division. Souris is a town of 1613 people located 46 kilometers southwest of Brandon. The average age of the population was 46.3 years and the average household income was $37,075 per year. There were 26 children in the kindergarten class at the time of the study. The kindergarten students attended school every other day, so the surveys were sent home on a Wednesday. The first reminder was sent home on the following Tuesday. The second reminder was sent home for the mothers who had not returned their surveys. The second reminder package contained: the second reminder notice (slightly revised because requesting them to reply by mail), a new survey tool, and a self-addressed stamped envelope; it was sent home two weeks following the initial distribution of the surveys. The second reminder was handled in this way for a couple of reasons: first because the kindergarten had been cancelled one day because of poor weather, and secondly there would be no further chance for returns to the school as the March break week was the last week of March. This Tuesday class was the last chance for interaction with the students prior to the break week because of teacher in-service. The response rate was 18 completed surveys out of 26 for a response rate of 69%. The first collection of surveys included 20 returns of the surveys within the first week, but three of these returned surveys were blank, resulting in an actual response of 17 of 26. Six second reminders were sent out and one of the 6 was returned in the mail, resulting in the 18 responses. Of the 18 surveys returned there were 2 mothers who smoked. The pilot study indicated that an adequate response rate was possible and probably would be greater with three reminders being sent home (Chiu & Brennan, 69 1998). Some changes were made to the survey tool. The changes included minor rewording, and adding "not applicable" categories, which made it easier for mothers to respond (Questions 37, 38, & 39). The data received from the pilot study did not show substantial variability within responses for some questions. For example, the confidence score had very few variations from the full score of 50 i.e. "very confident." There was, however, with just the two smoking mothers some variability in their responses, which was encouraging. The nonsmoking and former smoking mothers would show more confidence in their ability to provide a smoke free environment. Within the pilot study there were a few spouses (n = 3) who smoked while mothers were non-smokers. One mother indicated that her child was exposed to smoke every week but her confidence score indicated that she was very confident in providing a smoke free environment, because it was with her separated partner where her child was being exposed. The pilot study provided information about changes that needed to be made to the survey tool, as well as some rewording of the reminders, and confirmed the need for the three planned reminders. The response rate was deemed acceptable at 69% even without the incentives and the three reminders. Data from the pilot study were not included in the main study analysis. Data Analysis Completed questionnaires were reviewed for completeness; the data were entered into an electronic file, and randomly verified for accuracy. Questions were examined to ensure accuracy: 1) Were all values within the expected range?, 2) Were means and standard deviations possible for continuous variables?, and 3) Were there any numbers out of the acceptable range for discrete variables? (Tabachnick & Fidell, 1996) The 70 missing data were examined to decide i f the data were missing randomly or if there were any patterns. A total of 2 surveys were deleted because of a great deal of missing data. Some missing data were estimated using mean values or knowledge from the answers to other questions on the survey. Details of the missing data are discussed in the findings chapter (de Vaus, 2002; Tabachnick & Fidell). Each variable was examined using univariate statistics to ensure accuracy of the data, quality of the data, and to assess distributional properties of the variables (Munro & Page, 1993). Nonnormal variables were dichotomized for further analysis (de Vaus, 2002; Munro & Page; Tabachnick & Fidell, 1996). Each of the scales had internal consistency calculated before further analysis was completed. The data were analyzed using the Statistical Program for the Social Sciences (SPSS) program. The analysis of the study data involved descriptive, bivariate statistics (i.e., frequencies, crosstabulations) and multivariate statistics (i.e., logistic regression). The analysis' of each of the research questions is delineated in the findings chapter. 71 CHAPTER FOUR: FINDINGS In this chapter the findings are described. A description of the validation of the study measures and sample characteristics is provided. Secondly an explanation of the bivariate relationships and multivariate analysis is delineated. The findings for each of the seven research questions are addressed. Description of the Sample In total 727 surveys were distributed through 30 kindergarten classes from two school divisions. Of the 574 surveys that were returned 21 of these were blank and two were eliminated because of missing data. The final return of 551 completed surveys yielded a response rate of 76%. Responses within the first week after the initial survey distribution were 374 completed surveys or 51%, within one week of the first reminder 94 more surveys were returned, within one week after the second reminder 100 more surveys were returned and within one week after the third reminders only 6 more surveys were returned. The response rates for individual classrooms ranged from 32% to 90%, with an average in the Portage School Division of 66% and the average for the Brandon School Division of 76%. The study sample consisted of 532 women and 18 men. The average age of the sample was 33.1 years with a range of 20-59 years (see Table 1). Individuals who identified themselves as primary care givers other than mothers (n= 32) were mainly grandparents. These primary care givers ranged in age from 28-59 years. The number of children living in households ranged from one to six (M=2.4). Two hundred and ninety-seven (53.9%o) of the families had two children. Eighty percent of the sample reported 72 living with a partner; not living with a partner (separated, divorced, and never married) constituted 20.2% of the sample (see Table 2). Only 10.5%) of participants reported having less than a high school education. Those who had completed high school comprised 17.1% of the sample and the most frequent category was completion of community college (26.3%). Another 16.7% of the sample had a completed university degree. The majority of the sample (55.5%) identified their main activity of the last twelve months as working at a job outside the home. There were also many mothers/primary care givers (32.7%) who identified their main activity as raising a family. Table 1 Age of Participants and Number of Children in Household Range Mean (M) Std. Deviation (s.d.) Age 20-59 33.09 5.79 Number of children 1-6 2.36 .92 Of the 88%) of respondents who chose to respond to the income question, 23.6%> of the respondents had a household income of less than $30,000 per year (see Table 2). The middle ranges of income including three categories between $30,000 and $79,999 each accounted for an average of 20.2% of the sample, with a total for the three categories of 60.6%. Over 15% of participants indicated an income of over $80,000. Most of the participants identified themselves as White/Caucasian (84.2%) and 13.8% of the sample was Aboriginal/First Nations. There were 153 current smokers (27.8%). Never smokers made up 38.1% of the sample (n=210) and former smokers 32.6% of the sample (n=T88). Those who identified that they had a spouse who smoked accounted for 73 26.3% of the sample (n=153), 82 of the sample (53.6%) were smokers who identified that their spouse also smoked. Table 2 Personal Context Characteristics of Participants (N=551) N ' Percent Marital Status Married and living with 356 64.6 partner Common law 84 15.2 Separated 26 4.7 Divorced 35 6.4 Never married 50 9.1 Education Less than high school 58 10.5 Completed high school 94 17.1 Some community college 79 14.3 Completed community 145 26.3 college Some university 83 15.1 Completed university 92 16.7 Main Activity of last 12 months Looking for work 11 2.0 Working at a job or 306 55.5 business outside of the home Student 31 5.8 Raising a family 180 , 32.7 Other 22 4.0 Continued next page , 74 Table 2 continued N Percent Household Income/ year Less than $15,000 53 10.9 $15,000-$29,999 61 12.7 $30,000 - $44,999 89 18.3 $45,000 - $59,999 112 23.0 $60,000 - $79,999 93 19.3 $80,000 - $99,999 55 11.3 More than $100,000 22 4.5 How participants describe themselves White/Caucasian 464 84.2 Aboriginal/First Nations 76 13.8 A l l other 11 2.0 Smoking status Never smoker 210 38.1 Former smoker 188 32.6 Current smoker 153 27.8 Spouse/partner smokes 145 26.3 Current smokers whose 82 53.6 spouse smokes Sample Comparison to Provincial Data The study population characteristics were compared to data available regarding the population of Manitoba. Participants had an average of 2.4 children per household whereas the average number of children in Manitoba homes is 1.2 (Statistics Canada, 2001). Slightly more study participants reported the lowest income bracket compared to provincial percentages (10.9% versus 7.8%), with other income categories largely equivalent. The educational preparation of participants was slightly higher than that of Manitobans: high school or less 23.0% versus 10.5%, completed high school 20.0% versus 17.1%, some community college 16.8% versus 14.3%, completed community college 15.2%o versus 26.3%, some university 12.0% versus 15.1%, and completed university 13.0% versus 16.7% (Statistics Canada, 2001). The most recent statistics for all Manitobans indicate 66.2% were employed for July 2005 (Statistics Canada, 2005) 75 which is slightly higher than the 55.5% of the sample employed. This difference is likely the result of the study participants being female and more being stay-at-home mothers. The number of working Manitoba females from the 2001 Census data was 46.7%), which is slightly lower than the rate from this study but it must be considered that the Census data is four years old. Also the rate of stay at home mothers may be higher in rural Manitoba. Marital status reported from the 2001 Census data indicated 46.5% of Manitobans were married, which included legally married, separated, and common-law relationships, which is lower than the study population of 84.5% of participants within these three categories. The difference in the study population was that they were mainly mothers of kindergarten children, so did not reflect all families within the population. When considering families with children at home, 56% of Manitobans lived in nuclear families with children at home. Divorce statistics were slightly higher for the study with 6.4%) of the study population divorced and 4% for all Manitobans divorced. The CTUMS (2005) data indicated that the current smoking rate for all Manitobans in the 20-24 year old age group is 32.0% and for the 25-44 year old age group is 25.0%. The study population was within this range with 27.8% current smokers. When only Canadian females were reported, in the 15-24 year old age group the current smoking rate was 19.0 - 21.0% and for the 25+ age group the rate is 14.0 - 20.0% (CTUMS, 2000). The CTUMS notes that there was large variability in the sampling for the female-only statistics and data should be interpreted cautiously. Statistics regarding smoking rates vary according to the criteria used to define current smokers and may account for some for some of the differences identified (Cunningham, 1996). 76 Data Screening Factor Analyses Results of Study Scales Exploratory factor analyses were performed on all scales including the Smoking Restriction Self-Efficacy Scale, Thoughts about Smoking Scale, and Family Functioning Scale. The principal components extraction method was used for all factor analyses to verify item correlation as coherent subsets (Tabachnick & Fidel, 1996). The number of factors was decided based on eigenvalues greater than 1. If the initial unrotated factor analyses produced more than one factor, the solution was rotated using varimax rotation which provides a clearer pattern of factor loadings (deVaus, 2002). Smoking Restriction Self-efficacy Scale Results of the factor analysis for thelO-item self-efficacy scale are found in Table 3. The total was used as a single score. The inter-item correlations were all over 0.60 with the exception of three values; two of these low correlations were related to the item, "When I am feeding young children." The component matrix indicated factor loadings for all items over 0.77. Foster (2002) identifies any loading over 0.50 as a high loading for the factor. A Cronbach's alpha of 0.96 was obtained for the scale and based on these findings the scale was used without changes in subsequent data analysis. The total variance explained by the components in this factor was 82.8%. 77 Table 3 Factor Loadings for Exploratory Factor Analysis of the Smoking Restriction Self-efficacy Scale Items Component Matrix3 Cronbach's Alpha i f item Factor Loadings deleted When my spouse or, a close 0.816 0.952 relative wants to smoke in our home When I have friends in for 0.887 0.948 coffee and we are relaxing When I am frustrated and 0.864 0.949 anxious When the family is 0.880 0.950 watching television When we have friends in 0.825 0.952 for a party When the family has just 0.914 0.947 finished eating When there are arguments 0.910 . 0.947 and conflicts in my family When I am feeding young 0.770 0.955 children When no one else is home 0.772 0.954 When the children are in bed 0.901 0.948 a Extraction method: Principal Component Analysis. 1 component extracted. N = 551 Thoughts about Smoking Scale The 12-item Thoughts about Smoking Scale, which taps into knowledge of smoking and its effects, was reduced to nine items after initial unrotated factor analysis. Three items were dropped because of a large amount of missing data and because they were not as theoretically important as the other items in the scales. A factor analysis with varimax rotation of the nine items revealed two components (see Table 4). One additional item, "smoking is addictive," was dropped because it did not load on either factor (factor loadings < 0.60) (Foster, 2002). A Cronbach's alpha of the scale was 0.88. The total variance explained by the factor was 53.4%. The resultant two new scales had four items each and were renamed: General Smoking Knowledge and Environmental Tobacco 78 Smoke Knowledge (see Table 5). Cronbach's alphas for the two scales were: 0.86 for General Smoking Knowledge and 0.96 for Environmental Tobacco Smoke Knowledge. The total variance explained by the General Smoking Knowledge Scale was 66.2% and the Environmental Tobacco Smoke Knowledge was 69.9%. The two revised scales were used for further data analysis. Table 4 Factor Analysis of Thoughts about Smoking Scale Items Rotated Component Matrix Smoking is addictive -0.136 0.584 Smoking is dangerous to the smoker's health 0.325 0.676 Smoking is physically dangerous 0.429 0.649 Smoking is likely to cause heart disease 0.398 0.728 Smoking is likely to cause cancer 0.442 0.736 Smoke is dangerous to those who inhale ETS b 0.640 0,471 ETS is linked to asthma in children 0.865 0.267 ETS is linked to ear infections in children 0.813 0.066 ETS is linked to respiratory infections in children 0.881 0.241 a Extraction method: Principal Component Analysis. 2 components extracted. TV = 551 b Environmental Tobacco Smoke 0 Rotation method: Varimax ) ( 79 Table 5 Factor Analysis of General Smoking Knowledge and ETS Knowledge Scales Items Component Matrix for two scalesa' General Smoking ETS Knowledge11 Knowledge0 Smoking is dangerous to the 0.747 smoker's health Smoking is physically dangerous 0.828 Smoking is likely to cause heart 0.866 disease Smoking is likely to cause cancer 0.891 Smoke is dangerous to those who 0.786 inhale ETS ETS is linked to asthma in 0.907 children ETS is linked to ear infections in 0.767 children ETS is linked to respiratory 0.918 infections in children Extraction method: Principal Component Analysis. 2 components extracted. TV = 551 Rotation method: Varimax Scale Cronbach's alpha 0.858 7V= 551 d Scale Cronbach's alpha 0.955 7V= 551 Family Functioning Scale The Family Functioning Scale included in this survey is a well established scale previously used in many studies (Epstein et al., 1983; Epstein et al., 1978). One component was extracted with the exploratory unrotated factor analysis (see Table 6). Even though the factor loadings were less than 0.60 for three of the items, they were left in the scale because they were conceptually coherent, loaded above 0.50 and because previous research suggests the scale is reliable and valid. Loadings above 0.30 can be considered related to the factor but over 0.50 is considered a high loading (Foster, 2002; Tabachnick & Fidell, 1996). The Cronbach's alpha for the scale was 0.88. The variance explained by the components was 44%. 80 Table 6 Factor Analysis Of Family Functioning Scale Items Component Matrix3 Cronbach's alpha i f Item Factor Loadings Deleted13 Planning family activities is 0.576 0.869 difficult because we misunderstand each other In times of crisis we can 0.595 0.869 turn to each other for support We cannot talk to each 0.552 0.874 other about the sadness we feel Individuals are accepted for 0.603 0.868 what they are We avoid discussing our 0.662 0.864 fears and concerns We can express feelings to 0.697 0.863 each other There are lots of bad 0.676 0.864 feelings to each other We feel accepted for what 0.702 0.863 we are Making decisions is a 0.680 0.864 problem for our family We are able to make 0.665 0.865 decisions about how to solve problems We don't get along well 0.703 0.863 together We confide in each other 0.725 0.861 Extraction method: Principal Component Analysis. 1 component extracted. N = 551 b Cronbach's alpha 0.875 Key Study Findings Providing a Smoke Free Environment The frequency of participants reporting that they Provide a Smoke Free Environment (PSFE) for their children was recorded in relation to four areas: their entire home, their vehicle, other people's vehicles that children travel in, and homes their child visits at least once per week. The scores ranged from 0, "never providing a smoke free 81 environment," to 12, "always providing a smoke free environment" (see Table 7). Those who identified that they always PSFE for their children in all environments was 34.5% of the sample (n=190). Only 1.4% (n=8) of the sample indicated that they never PSFE; 29.1% scored 8 or less on the scale. Data for this variable were skewed, so the variable was collapsed into a binary variable for subsequent data analysis. Some researchers would indicate that normality of a variable is not always required with a large data set (de Vaus, 2002; Myers, Gamst, & Guarino, 2006). A conservative approach was used and a binary variable was created rather than ignore the skewness or transform the variable. A category was created using the participants with a score of 12 for those who "Always Provide a Smoke Free Environment" and the other category was called "Those Who Do Not Always Provide a Smoke Free Environment." Theoretically it was postulated that the characteristics of participants who are always providing a smoke free environment would be different from those who are not able to accomplish a smoke free environment all of the time. Table 7 Participant Scores for Providing a Smoke Free Environment (N=551) Scores n Percent (%) 0 8 1.5 1 2 0.4 2 5 0.9 3 6 1.0 4 " 16 2.9 5 13 2.4 6 36 6.5 7 21 3.8 8 53 9.6 9 57 10.3 10 81 14.7 11 63 11.4 12 190 34.5 82 Precaution Adoption Process Stages The second research question - "In relation to providing a smoke free environment what proportion of participants report being in the various stages of the Precaution Adoption Process (PAP)?" - is addressed in this section. The largest proportion of participants (n=336, 61%) were in the highest stage of the Precaution Adoption Process (PAP), "I have decided to always provide a smoke free home" (see Table 8). The next most commonly endorsed category was "I have started to reduce my child's exposure to environmental tobacco smoke" (ETS) (n=106, 19.8%). Approximately 10% of the sample identified that they had not given the issue much thought. As is evident in Table 8, the data were skewed. Therefore for subsequent data analysis the variable was collapsed into a binomial variable with the fifth category of "I have decided to always provide a smoke free environment" as the highest stage of the PAP, with the other categories collapsed into a category called the lower stages of the PAP. It made theoretical sense to be able to detect differences in those participants who identify themselves in the highest stage of the PAP and those who do not. Presumably there would be differences in participants who had already made the decision to always PSFE and those who had not yet decided to make this commitment. 83 Table 8 Proportion of Participants Reporting Being in the Various Stages of the Precaution Adoption Process Categories n Percent (%) I haven't given it much 54 9.8 thought I have decided I don't need 26 4.7 to do anything I plan to reduce my child's 29 5.3 exposure to ETS within 6 months I have started to reduce my 106 19.2 child's exposure to ETS I have decided to always 336 61.0 provide a smoke-free environment _____ Bivariate Analysis The third, fourth, and fifth research questions ask about the relationship between three areas including: personal context, smoking patterns, and social interaction context and a) the stage of PAP and b) PSFE. These relationships were analyzed using bivariate data analysis techniques. Bivariate relationships were first examined between the two dependent variables PSFE and the Stage of the PAP and each of the independent variables. Tables 9 and 10 contain findings of analyses in which the personal context characteristics of the two groups are considered in relation to each dependent variable. Tables 11 and 12 provide results of the bivariate analysis examining the effects of the other predictor variables related to the two dependent variables of interest. Depending upon the level of measurement of the variables (either continuous or discrete) a t Test or Chi-square test was performed to examine the bivariate relationships. Missing data on all independent variables were handled in the most common and conservative way, imputing the mean or median score where possible (Knapp, 1998; 84 Myers, Gamst, & Guarino, 2006; Tabachnick & Fidel, 1996). Myers, Gamst, and Guarino (2006) indicate that many large surveys report a loss of data for such common variables as income. By imputing the mean of the available sample it is acknowledged that individual scores would obviously fall lower or higher than the mean, but it is estimated that the average of these missing values would be equal to the average of the valid means. The mean was imputed for 12% of the income variable. This series-mean approach does have the effect of reducing the variability. This limitation did not have a significant effect however as the categories were later collapsed for further data analysis (de Vaus, 2002; Tabachnick & Fidel, 1996). For missing values within scales, when there were less than 50% of values missing, values were imputed on a case by case basis, either using data from other questions in the survey or by imputing the mean value from the other scores. Imputation is based on a narrower number of cases when the calculated mean of the subgroup or an individual score based on other survey data is used (de Vaus, 2002; Myers, Gamst, & Guarino, 2006). For the dependent variable, PSFE for example, i f a respondent missed providing one of the scores or provided a "not applicable" answer, a score was imputed based on the modal response from the respondent. If the family did not own a car but provided a smoke free environment in the other places all of the time, it would make theoretical sense that they would also provide a smoke free environment in the car if they had owned one. Values were only imputed for these kinds of circumstances to provide an overall score for that scale which reflected the practices of that particular respondent. It was important to impute values for the circumstances where participants did not own a car as this would not indicate a random loss of the data. Individuals who do not own a car would not be a random or common group in the demographics of the study region. If all of these cases were eliminated a crucial portion of the sample would be 85 eliminated, creating a greater nonresponse error (Cohen & Cohen, 1987; Myers, Gamst, & Guarino, 2006; Patrician, 2002). Bivariate analysis of personal context variables and PSFE and PAP In relation to PSFE, those who always PSFE were found to differ from those who do not always PSFE on all personal context variables except ethnicity (see Table 9). Those who do PSFE tend to be more educated, have higher household incomes, are older, live with a partner, and are less likely to be smokers. The personal context characteristics of those in the higher stage of the PAP were compared with those in the lower stages PAP and showed significant differences for all personal context variables except occupation (see Table 10). Participants in the higher stage of PAP tended to be better educated, married or living with a partner, have a higher income, were older, and were nonsmokers. 86 Table 9 -Personal Context Characteristics of Those Who Always Provide a Smoke Free Environment (PSFE) and Those Who Do Not Always Provide a Smoke Free Environment Variables Always Does not Statistics PSFE always PSFE t test X2{&f) n = 190 n = 361 Highest Education 17.52 (3) Achieved (%) < High school 17(8.9) 41 (11.4) Completed high school/ 45 (23.7) 128 (35.5) some community college Completed community 81 (42.7) 147 (40.6) college/some university Completed university 47 (24.7) 45(12.5) Ethnicity (%) 2.96(1) White/Caucasian 167 (87.9) 297 (82.3) Other 23 (12.1) 64 (17.7) Family Yearly Income (%) 37.22" (4) < $30,000 24(12.6) 90 (24.9) $30,000 - $44,999 22 (11.6) 67(18.6) $45,000 - $59,999 65 (34.3) 113(31.3) $60,000 - $79,999 32(16.8) 61 (16.9) > $80,000 47 (24.7) 30 (8.3) Marital Status (%) 13.23**(1) Living with partner - 168 (88.4) > 272(75.3) ' married or common-law Separated, divorced, 22(11.6) 89 (24.7) widowed, single never married 6.15*(2) Occupation (%) Working outside the home 109 (57.4) 219 (60.7) Looking for work, student 9 (4.7) 34 (9.4) or retired At home raising a family 72 (37.9) 108 (29.9) Age M{SD) 34.05 (5.1) 32.6(6.1) 2.84* Smoking Status 48.39** (1) Currently smokes cigarettes 18(9.5) 135(37.4) Does not smoke cigarettes 172 (90.5) 226 (62.6) p<.001; p<.05 87 Table 10 Personal Context Characteristics of Participants Identifying Themselves in the Higher or Lower Stages of the Precaution Adoption Process Variables Higher stage Lower stages Statistics of PAP of PAP n = 336 n = 215 rTest X2(df) Highest Education Achieved (%) 40.82* (3) < High school 26(7.7) 32 (14.9) Completed high school/ 83 (24.7) 90 (41.9) some community college Completed community 150(44.6) 78 (36.2) college/some university Completed university 77 (23.9) 15 (7.0) Ethnicity (%) 11.33*(1) White/Caucasian 297 (88.4) 167 (77.7) Other 39(11.6) 48 (22.3) Family Yearly Income (%) 38.38* (4) < $30,000 51 (15.2) 63 (29.3) $30,000 - $44,999 45 (13.4) 44 (20.5) $45,000 - $59,999 107(31.8) 71 (33.0) $60,000 - $79,999 71 (21.1) 22 (10.2) > $80,000 62 (18.5) 15 (7.0) Marital Status (%) 14.83* (1) Living with partner- 286(85.1) 154(71.6) married or common-law Separated, divorced, 50 (14.9) 61 (28.4) widowed, single never married Occupation (%) 5.26 (2) Working outside the home 190 (56.5) 138 (64.2) Looking for work, student or 24 (9.0) 19 (8.8) retired At home raising a family 123 (36.8) 56(26.5) Age M(SD) 33.95(5.6) 31.75 (5.9) 4.43* Smoking Status 100.08* (1) Currently smokes cigarettes 42 (12.5) 111 (51.6) Does not smoke cigarettes 294 (87.4) 104 (48.4) *p < .001 88 Bivariate relationships between predicting and dependent variables Analysis was undertaken to examine whether other hypothesized variables, from the theoretical model, were bivariately associated with the two dependent variables. The variables found to be positively associated with always PSFE (see Table 11) included self-efficacy score, general smoking knowledge, family functioning score, number of smokers in the home, number of family and friends who smoke, and availability of an outside area to smoke. Participants who always PSFE had lower scores for crowding in the home, Fagerstrom scores for smoking mothers, and number of cigarettes smoked per day. Two variables were found to be non-significant and included whether the family lived within the city limits and the heightened health awareness score. Self-efficacy was identified as a statistically significant predictor in relation to stage of PAP (see Table 12). Other variables found to be associated with the higher stage of PAP included higher scores for ETS knowledge, general smoking knowledge, and family functioning. For those in the higher stage of PAP compared to those in the lower stages of the PAP, the Fagerstrom scores tended to be lower, number of cigarettes smoked each day tended to be fewer, there was less crowding in the home, and there were fewer smokers in the home as well as fewer friends and family who smoke. Living within the city limits and having an outside area available to smoke were not associated with any particular stage of PAP. Table 11 89 Predicting Variables by Groups of Those Who Always Provide a Smoke Free Environment (PSFE) and Those Who Do Not Always Provide a Smoke Free Variables Always PSFE Does not always Statistics PSFE n = 190 n = 361 f Test A*(df) Lives with in the city limits (%) 0.01 (1) Yes 163 (85.8) 311 (86.1) No 27(14.2) 50(13.9) Crowding in the home M(SD) 0.76 (0.37) 0.86 (0.38) -2.70* Heightened Health Awareness M (SD) 1.78 (1.79) 1.81 (1.76) -0.14 Family Functioning M (SD) 3.60 (0.39) 3.43 (0.41) 4.53** Self efficacy M (SD) 49.36(3.23) 44.76 (8.3) 7.34*'* ETS knowledge M (SD) 14.58 (1.75) 13.68 (2.23) 4.85** General smoking knowledge M (SD) 15.24(1.58) 14.58 (1.94) 4.04** Number of smokers in the home (%) 40.98**(2) none 168 (88.4) 226 (62.6) 1 person 16(8.4) 107 (29.6) 2 people or more 6(3.2) 28(7.8) Number of friends and , family who smoke (%) 108.14**(4) None 71 (37.4) 27 (7.5) 1 person 36(18.9) 31 (8.6) 2 people 23 (12.1) 68(18.8) 3-5 people 36(19.0) 96 (26.6) More than 5 people 24(12.6) 139 (38.5) Outside area available to smoke 10.31*(1) Yes 172 (90.5) 350 (97) No 18(9.5) 11 (3.0) Fagerstrom score for smoking mothers M 0.19(0.75) 1.1 (1.64) -6.39** (SD) Number of cigarettes smoked/day M (SD) 0.62 (2.20) 3.67 (6.12) -6.66** PS001; p<.05 90 Table 12 Predicting Variables by Groups of Those Who Identify Themselves in the Higher and the Lower Stages of the Precaution Adoption Process Variables Higher stage of Lower stages Statistics PAP of PAP n = 336 n = 215 f Test X2(Af) Live within the city limits (%) 0.55(1) Yes 292 (86.90) 182 (84.70) No 44(13.10) 33 (15.30) Crowding in the home M(SD) 0.77 (0.35) 0.90 (0.41) -3.61** Heightened Health 2.12* Awareness M (SD) 1.92 (1.80) 1.60(1.70) Family Functioning M(SD) 3.54 (0.41) 3.40 (0.40) 4.12** Self efficacy M (SD) 49.12(3.18) 42.02 (9.52) 12.62** ETS knowledge M (SD) 14.57(1.64) 13.08 (2.45) 8.55** General smoking knowledge M (SD) 15.24(1.44) 14.12(2.18) 7.25** Number of smokers in the home (%) 51.75**(2) None 276 (82.10) 118(54.90) 1 person 52(15.50) 71 (33.00) More than 2 people 8 (2.40) 26(12.10) Number of friends and family who smoke (%) 71.86**(4) None 88 (26.20) 10(4.70) 1 person 50(14.90) 17(7.90) 2 people 59(17.60) 32 (14.90) 3-5 people 74 (22.00) 58 (26.90) More than 5 people 65 (19.30) 98 (45.60) Outside area available to smoke 2.84(1) Yes 314(93.50) 208 (96.70) No 22 (6.50) 7 (3.30) Fagerstrom score for smoking mothers M 0.23 (0.81) 1.65 (2.16) -10.88** (SD) Number of cigarettes smoked/day M (SD) 0.75 (2.29) 5.44 (7.10) -11.49** p<.001; p<.05 91 Multivariate Analysis The sixth research question addresses which of the variables from the three areas -personal context, smoking patterns, and social interaction context - are most predictive of a) stage of PAP and b) PSFE. Logistic regression analyses were used to determine the significance of various independent variables associated with the two dependent variables, PSFE and stage of PAP. The dependent variables were dichotomized into variables with two categories labeled in the model as follows: PSFE 1 = always provides a smoke-free environment 0 = does not always provide a smoke-free environment PAP 1 = higher stage of Precaution Adoption Process 0 = lower stages of Precaution Adoption Process The independent variables included in the regression model analysis included all significant variables from the bivariate analysis (see Tables 9-12) . To use categorical variables in a regression analysis, the variables were converted to dummy variables. The interpretation of a regression analysis requires that arbitrary coding of categorical variables be provided rather than using a numbering system of the categories which would represent ordinal data. It was important to include the categorical variables as possible predictors though because of their theoretical influence in previous research. Because the dummy variables need to be independent of each other, the number of the levels of the variables was one less than the number of categories in each variable. The baseline group in all of the nominal variables was selected as the reference or baseline category, coded with the "0" value of each variable. The coding of the variables used in the regression analysis is explained below. The coding of the education variable results in the following vectors, using 4 dummy variables: 0) less than high school; completed high school and some community college; completed community college and some university; 92 and completed university. The vectors of income category were (4 dummy variables): 0) < $30,000; $30,000 - $44,999; $45,000 - $59,999; $60,000 - $79,999; > $80,000. The occupation variable vectors (2 dummy variables): 0) working at a job; looking for work, student, or retired; and raising a family. The number of family members who smoke was coded into the following vectors (4 dummy variables): 0) none; 1 person; 2 people; 3-5 people; and more than 5 people. The number of smokers living in the home not counting the participant was coded (2 dummy variables): 0) none; 1 person; and 2 or more people. The variables were entered into the logistic regression model in 3 blocks, according to the theoretical model. The first block included all personal context characteristics: living within the city limits, education, marital status, age, ethnicity, occupation, income, and crowding in the home. Block 2 included the smoking variables: a yard available for smokers to smoke outside i f they wished, number of cigarettes smoked per day, and Fagerstrom score for smoking mothers, spouse smoking, number of friends and family they know who smoke, the number of people who live in the home who smoke (besides the participant). Block 3 included the social interaction context factors which include scores for family functioning, self-efficacy, general smoking knowledge, and ETS knowledge. The logistic regression model examining the significance of the independent variables associated with the dependent variable PSFE is shown in Tables 13-15. Each table identifies the statistics produced with entering Block 1 (Table 13), Block 2 (Table 14), and the full model (Table 15). Variables were considered to provide a significant effect at alpha levels of < 0.05. The Chi-square test for Block 1 was significant^ (14, N = 551) = 52.8, p<.001. The variability explained by Block 1 using the Cox & Snell R square and Nagelkerke R 93 square showed a range of 9.1% to 12.6%. The odds of PSFE increased with an income of > $80,000 compared to an income of < $30,000, and being a stay at home mother as compared to working. The second Block entered into the logistic regression included the smoking variables: yard available to smoke outside, number of cigarettes smoked per day, Fagerstrom score, spouse smoking, number of family and friends who smoke, and the number of people who live in the home who smoke. Table 14 reports the results for Blocks 1 and 2, showing that there were changes in significance in variables from the first block after the second block of variables were introduced into the model. The Block 2 significance was X2 (24, N=551) = 159.8, p<.001. The model with two blocks entered explained an increased amount of variance from 25.2% to 34.8% (according to the Cox & Snell R square and Nagelkerke R square). The results involving the education variable indicate that participants with completed high school were less likely to PSFE as compared to having less than high school education. The significant smoking variable was the number of family and friends who smoke, with PSFE less likely with any numbers of family and friends who smoke as compared to no family smoking. The effects of income and occupation, present with only the first block of demographic variables entered into the model, were lost with the smoking variables entered. The full model of the logistic regression for the dependent variable, PSFE, is shown in Table 15. The Chi-square score for the model was (28, N=551) = 191.0, p<.001. The model explained 29.3% to 40.5 % of the variance (according to the Cox & Snell R square and Nagelkerke R square). An improved fit of the model was indicated by the -2 Log likelihood model ratio of fit value decreasing from the first block being entered to the final model (657 to 518). Using the Hosmer and Lemeshow test of 94 predicted versus observed classifications, the percentage correct increased from the first block being 68.8% correct to 76.4% correct in the final model. The significant variables included education as well as the number of family and friends who smoke but also included the self-efficacy score, once the final block of variables (family functioning, self-efficacy, ETS knowledge, and general smoking knowledge) was entered. With any education greater than high school, there were similar odds (.345 to .386) of being less likely to always PSFE. With an increasing number of friends and family who smoke, respondents were less likely to report that a smoke free environment was provided. The greater the self-efficacy of the mother/primary caregiver the more likely that respondents reported that a smoke free environment was provided (odds ratio 1.18, 95% CI 1.08, 1.28). Table 13 Logistic Regression Analysis Predicting Always Providing a Smoke Free Environment Block 1 Personal Context Predictor B SE Wald df Sig. Exp 95% conf Variable stat ( B ) interval City living -.261 .277 .884 1 .347 .770 .447 1.327 Education -< high school 4.974 3 .174 -Completed High -.537 .369 2.115 1 .146 .584 .283 1.205 School -Completed -.173 .3621 .229 1 .632 .841 .413 1.710 -Community College -Completed .069 .414 .027 1 .868 1.071 .476 2.412 University Lives with -.563 .297 3.247 1 .072 .585 .327 1.048 partner Age .014 .018 .660 1 .417 1.014 .980 1.050 Ethnicity -.033 .314 .011 1 .917 .968 .524 1.790 Crowding -.196 .286 .468 1 .494 .822 .469 1.440 Income < $30,000 12.812 4 .012 $30,000 - $44,999 .025 .363 .005 1 .945 1.025 .503 2.091 $45,000-$59,999 .441 .327 1.817 1 .178 1.554 .819 2.950 $60,000 - $79,999 .245 .375 .424 1 .515 1.277 .612 2.666 > $80,000 1.194 .407 8.622 1 .003 3.300 1.487 7.323 Occupation - Working 5.204 2 .074 -Looking for work, -.158 .423 .139 1 .709 .854 .372 1.959 student, retired -Raising a family .439 .205 4.572 1 .032 1.550 1.307 2.318 Table 14 Logistic Regression Analysis Predicting Always Providing a Smoke Free Environment Block 2 Smoking Variables P r e d i c t o r B SE W a l d d f S ig . E x p 95% c o n f V a r i a b l e t stat (B) i n t e r v a l City living .044 .313 .020 1 .888 1.045 .565 1.932 Education •-< high school 4.979 3 .173 -Completed High School -.927 .417 4.938 1 .026 .396 .175 .896 -Completed Community College -Completed University -.761 -.817 .416 .476 3.346 2.951 1 .067 .086 .467 .442 .207 .174 1.056 1.122 Lives with -.528 .341 2.401 1 .121 .589 .302 1.150 partner Age -.020 .021 .936 .333 .980 .941 1.021 Ethnicity • .274 .367 .557 1 .455 1.315 .641 2.698 Crowding -.285 .344 .687 1 .407 .752 .383 1.476 Income < $30,000 6.480 .166 $30,000 - $44,999 -.283 .410 .476 1 .490 .754 .338 1.683 $45,000- $59,999 -.007 .383 .000 1 .986 .993 .469 2.103 $60,000 - $79,999 -.197 .434 .205 1 .651 .822 .351 1.925 > $80,000 .649 .473 1.877 1 .171 1.913 .756 4.838 Occupation - Working 2.685 .261 -Looking for work, student, retired -.802 .493 2.648 1 .104 .448 .171 1.178 -Raising a family -.039 .240 .026 1 .872 .962 .601 1.541 Has a Yard .587 .462 1.613 1 .204 1.798 .727 4.445 Fag. score -.140 .181 .600 1 .439 .869 .609 1.240 #cigs -.087 .063 1.924 1 .165 .917 .811 1.037 97 Table 14 continued P r e d i c t o r B SE W a l d d f Sig. E x p 95% c o n f V a r i a b l e stat (B ) i n t e r v a l Spouse smoking .077 .520 .022 1 .882 1.080 .390 2.993 Family smoking none 40.137 4 .000 1 person -.772 .356 4.695 1 .030 .462 .230 .929 2 people -1.91 .359 28.302 1 .000 .148 .073 .299 3-5 people -1.51 .344 19.263 1 .000 .221 .112 .433 > 5 people -2.05 .383 28.500 1 .000 .129 .061 .274 #Smoking in home none 3.023 2 .221 1 person -.696 .521 1.781 1 .182 .499 .179 1.385 2 or more people .138 .669 .042 1 .837 1.148 .309 4.278 98 Table 15 Logistic Regression Analysis Predicting Always Providing a Smoke Free Environment -Full model P r e d i c t o r V a r i a b l e B SE W a l d stat d f S ig . E x p ( B ) 95% c o n f i n t e r v a l Gity living .011 .323 .001 1 .973 1.011 .537 1.903 Education 5.583 3 .134 -< high school -Completed High School -1.02 .451 5.157 1 .023 .359 .148 .869 -Completed Community -.952 .451 4.558 1 .035 .386 .160 .934 College -Completed University -1.06 .509 4.359 1 .037 .345 .127 .927 Lives with -.524 .353 2.200 1 .138 .592 .296 1.183 partner Age -.016 .021 .550 1 .458 .984 .944 1.026 Ethnicity .435 .393 1.225 1 .268 1.544 .715 3.333 Crowding -.419 .359 1.367 1 .242 .657 .326 1.328 Income < $30,000 6.009 .198 $30,000 - $44,999 -.367 .429 .730 1 .393 .693 .299 1.607 $45,000-$59,999 -.084 .410 .042 1 .838 .919 .412 2.053 $60,000 - $79,999 -.363 .458 .627 1 .428 .696 .284 1.707 > $80,000 .490 .494 .982 1 .322 1.632 .619 4.300 Occupation -Working 3.594 .166 -Looking for work, -.932 .503 3.434 1 .064 .394 .147 1.055 student, retired -Raising a family -.201 .252 .637 1 .425 .818 .500 1.339 Has a yard .621 .476 1.702 1 .192 1.861 .732 4.730 Fag. score -.045 .207 .047 1 .829 .956 .638 1.434 #cigs -.085 .068 1.554 1 .213 .919 .804 1.050 Spouse smoking -.121 .589 .042 .837 .886 .279 2.809 Family smoking none 35.175 .000 1 person -.768 .364 4.458 1 .035 .464 .227 .946 2 people -1.81 .369 24.112 1 .000 .230 .079 .336 3-5 people -1.47 .354 17.176 1 .000 .230 .115 .461 > 5 people -2.02 .395 26.062 1 .000 .133 .061 .289 99 Table 15 continued P r e d i c t o r V a r i a b l e B SE W a l d stat d f Sig. E x p (B) 95% c o n f i n t e r v a l #Smoking in home none 3.985 2 .136 1 person -.491 .590 .692 1 .406 .612 .193 1.946 2 or more people .780 .733 1.133 1 .287 2.181 .519 9.172 Family Function .405 .292 1.932 1 .164 1.500 .847 2.656 Self Efficacy .163 .043 14.606 1 .000 1.177 1.083 1.280 ETS Knowledge -.014 .073 .038 1 .846 .986 .854 1.138 General Smoking .090 .082 1.196 1 .274 1.094 .931 1.286 Knowledge Constant -7.54 2.57 8.647 1 .003 .001 A logistic regression model was also developed using independent variables to predict the second dependent variable, stage of PAP. Coding was the same as previously described for use in the model for the first dependent variable. The blocks were entered using the same variables in each block. The statistics for the regression models for each block are shown in Tables 16-18. Following entering the first block, personal context, participants who completed university were two and one-half times as likely to have placed themselves in the higher PAP stage than those with less than high school, participants in the third category of income ($60,000 - $79,999) were also two and half times as likely to have placed themselves in the higher PAP stage than those people indicating an income of < $30,000. Participants in the categories of looking for work, being a student or retired as well as being a stay at home mother were twice as likely to place themselves in the higher PAP stage than participants working at a job outside of the home. The first block provided a significant X2 (14, N=551) = 89.1, p<.001, contribution to the regression model, explaining 15.1% to 39.0% of the model. 100 The model contributed 28.7% to 39.0% of the variance X2 (25, N=551) =184.6, p < .001 once Block 2, smoking variables, was entered. Education and income, found to be previously significant, did not continue to contribute to the model (see Table 17). Again both occupation categories of looking for work, a student or retired and stay at home mothers were found to be more likely to be in the higher stage of PAP than those working at a job outside of the home. Contributing most to the explanation of the model from the smoking variables was the number of cigarettes smoked per day, as well as the number of family and friends who smoke. The final logistic regression model for the dependent variable stage of PAP explained 37.2% to 50.5% of the variance; block X2 (4, N=551) = 77.5, p < .001; model X2(29, N=551) = 253.5, p< .001. The -2 Log likelihood ratio of fit of the model for the data decreased from the first block to the full model (638 to 474), indicating an increasingly better fit of the model. Using the Hosmer and Lemeshow test of observed versus expected, the percentage of correct estimates improved from 61.3% with the first block to 80.7% with the final model. The variables which explained the greatest variance in the model, were identified as the amount of crowding in the home, the number of cigarettes smoked per day, the number of family and friends who smoke, the self-efficacy score and the ETS knowledge score. In the final model the effects of occupation, as well as the number of people who smoke in the home, no longer had a significant effect. The mothers/primary care givers with the highest self-efficacy score were the most likely to place themselves in the higher stage of PAP, and the more cigarettes smoked per day as well as the more family and friends who smoke, the less likely the mother/primary care giver was to place him or herself in the higher stage of the PAP. The personal context variables except crowding lost all significance in the model and only two smoking 101 variables remained significant. The self-efficacy variable provided the most explanation in the model. Table 16 Logistic Regression Analysis Predicting Higher Stage of Precaution Adoption Process -Block 1 Personal Context Predictor B S E W a l d d f Sig. E x p 9 5 % conf Var iable stat (B) interval City living -.261 .277 .884 1 .347 .770 .447 1.327 Education -< high school 18.192 3 .000 -Completed High -.249 .345 .520 1 .471 .780 .397 1.533 School -Completed .325 .344 .889 •1 .346 1.384 .704 2.718 Community College -Completed .892 .375 5.644 1 .018 2.439 1.434 8.078 University Lives with -.351 .263 1.780 1 .152 1.026 .421 1.179 partner Age .025 .018 2.053 1 .055 1.035 .991 1.062 Ethnicity -.344 .276 1.552 1 .213 .709 .412 1.218 Crowding -.196 .286 .468 1 .494 .822 .469 1.440 Income < $30,000 9.438 4 .051 $30,000 - $44,999 .085 .320 .071 1 .790 1.089 .581 2.040 $45,000 -$59,999 .137 .301 .207 1 .649 1.146 .636 2.067 $60,000 - $79,999 .892 .375 5.644 1 .018 2.439 1.169 5.090 > $80,000 .769 .403 3.632 1 .057 2.157 .978 4.754 Occupation - Working 15.133 2 .001 -Looking for work, .712 .349 4.164 1 .041 2.038 1.028 4.037 student, retired -Raising a family .824 .223 13.6962 1 .000 2.279 1.473 3.526 102 Table 17 Logistic Regression Analysis Predicting Higher St age of Precaution Adoption Process Block 2 Smoking Variables P r e d i c t o r B S E W a l d d f S i g . E x p 95% c o n f V a r i a b l e stat (B) i n t e r v a l City living -.109 .322 .114 1 .735 .897 .477 1.686 Education < high school 7.254 3 .064 -Completed High -.494 .393 1.577 1 .209 .610 .283 1.319 School -Completed -.158 .393 .162 1 .687 .854 .395 1.844 Community College -Completed University .523 .496 1.111 1 .292 1.687 .638 4.459 Lives with -.426 .301 2.002 1 .157 .653 .362 1.178 partner Age .004 .020 .036 1 .850 1.004 .965 1.044 Ethnicity -.309 .315 .961 1 .327 .734 .396 1.361 Crowding -.182 .114 2.577 1 .108 .833 .667 1.041 Income < $30,000 5.659 4 .226 $30,000-$44,999 -.344 .354 .942 1 .332 .709 .354 1.420 $45,000-$59,999 -.309 .344 .807 1 .369 .734 .374 1.441 $60,000- $79,999 .400 .417 .918 1 .338 1.492 .658 3.380 > $80,000 .091 .454 .040 1 .841 1.096 .450 3.380 Occupation - Working 8.973 2 .011 -Looking for work, .835 .410 4.154 1 .042 2.305 1.033 5.144 student, retired -Raising a family .662 .257 6.616 1 .010 1.938 1.171 3.208 Has a yard .357 .565 .400 1 .527 1.429 .472 4.325 Fag. score .005 .136 .001 1 .973 1.005 .769 1.313 #cigs -.213 .059 13.135 1 .000 ,808 .720 .907 103 Table 17 continued Predictor B SE Wald df Sig. Exp 9 5 % conf Variable stat (B) interval Spouse smoking -.513 .475 1.166 1 .280 .599 .236 1.518 Family smoking none 11.809 4 .019 1 person -1.080 .484 4.984 1 .026 .339 .131 .876 2 people -1.362 .443 9.441 1 .002 .256 .107 .611 3-5 people -1.242 .448 7.699 1 .006 .289 .120 .694 > 5 people -1.502 .453 10.969 1 .001 .223 .092 .542 #Smokmg in home none 4.222 2 .121 1 person -.838 .455 3.394 1 .065 .433 .177 1.055 2 or more people -1.113 .612 3.309 1 .069 .328 .099 1.090 104 Table 18 Logistic Regression Analysis Predicting Higher Stage of Precaution Adoption Process -Full Model P r e d i c t o r B S E W a l d d f S i g . E x p 95% c o n f V a r i a b l e s t a t ( B ) i n t e r v a l City living -.100 .355 .079 1 .779 .905 .451 1.815 Education 5.086 3 .166 -< high school -Completed High School -.532 .435 1.491 1 .222 .588 .250 1.379 -Completed -.191 .430 .198 1 .657 .826 .356 1.918 Community College -Completed University .356 .533 .447 1 .504 1.428 .503 4.055 Lives with -.389 .326 1.423 1 .233 .677 .357 1.285 partner Age .001 .022 .009 1 .923 .966 .959 1.044 Ethnicity -.034 .353 .829 1 .363 .708 .337 1.488 Crowding -.248 .127 3.835 1 .050 .780 .608 1.000 Income < $30,000 3.799 4 .434 $30,000-$44,999 -.318 .380 .703 1 .402 .727 .346 1.531 $45,000-$59,999 -.322 .372 .746 1 .388 .725 .350 1.504 $60,000 - $79,999 .296 .452 .427 1 .513 1.344 .554 3.260 > $80,000 .076 .496 .023 1 .878 1.079 .278 2.160 Occupation - Working 5.103 2 .078 -Looking for work, student, retired .710 .464 2.344 1 .126 2.034 .820 5.049 -Raising a family .550 .283 3.777 1 .052 1.734 .995 3.021 Has a yard .113 .640 .031 1 .860 1.119 .320 3.922 Fag. score -.022 .153 .020 1 .887 .978 .724 1.321 #cigs -.176 .061 8.306 1 .004 .838 .774 .945 105 Table 18 continued Predictor B SE Wald df Sig. Exp 95% conf Variable stat (B) interval Spouse smoking -.633 .510 1.537 1 .215 .531 .195 1.444 Family smoking none 11.069 4 .026 1 person -1.368 .533 6.584 1 .010 .358 .195 1.444 2 people -1.441 .498 8.377 1 .004 .237 .089 .628 3-5 people -1.262 .504 6.266 1 .012 .283 .105 .760 > 5 people -1.620 .507 10.202 1 .001 .198 .073 .535 #Smoking in home none 1.921 2 .383 1 person -.631 .490 1.663 1 .197 .532 .204 1.388 2 or more people -.766 .675 1.288 1 .256 .465 .124 1.746 Family Function .214 .406 .279 1 .597 1.239 .560 2.745 Self Efficacy .106 .020 28.159 1 .000 1.112 1.069 1.157 ETS Knowledge .118 .049 5.786 1 .016 1.125 1.022 1.238 General Smoking .077 .072 1.167 1 .280 1.080 .939 1.243 Knowledge Constant -4.238 1.96 4.677 1 .031 .014 Interaction effects Two criteria were used to examine the interaction effects, the variables were . significant and the hypothesized interaction was theoretically feasible. Several interaction effects were tested for both dependent variables based on these two criteria. The interactions for each dependent variable will be discussed separately. Two interaction effects were examined for the first dependent variable PSFE: interaction between self-efficacy and number of family smoking; and level of education and self-efficacy. The only significant interaction effect found was between self-efficacy and number of family members smoking. The other variables that remained significant in the model with the introduction of this interaction were self-efficacy, the number of 106 family smoking, and education. The amount of overall variance explained by the model improved with the introduction of the significant interaction from a range of 29.1% to 40.2% to a range of 29.9% to 41.2%; interaction block X2 (1, N=551) =4.3, p = .04; model X2 (28, N=551) =195.4, p < .001. This interaction effect indicates that having fewer friends and family who smoke improves mothers'/primary care givers' self-efficacy related to providing a smoke free environment. Three interactions among predictors were examined for their effect on the second dependent variable, stage of PAP. The interaction effects included: number of cigarettes smoked and number of family smoking, self-efficacy and number of family smoking, self-efficacy and ETS knowledge. None of the interactions examined for this dependent variable was significant. Relationship between PAP and PSFE To address the final research question - the relationship between the two dependent variables - PFSE and the stage of PAP was examined (see Table 19). The number of participants who always PSFE or those who do not always PSFE were compared within each of the stages of the PAP. Chi-square is not considered reliable i f more than one cell has no data; the data analysis was completed, however, because only one cell had missing data (Hosmer & Lemeshow, 2000; Tabachnick & Fidel, 1996). A significant difference was found between the groups of participants who always PSFE in relation to their stage of PAP. Ninety percent (n=171) of the participants who always PSFE also identified themselves in the higher stage of the PAP. Among those who always PSFE, only 14 participants identified themselves in the lower two stages of the PAP. A very different picture was seen for those who do not always PSFE, where there is a range of responses in all categories of the stages of PAP. One hundred and sixty-five 107 participants (45.7%) identify themselves in the higher stage of the PAP but do not always PSFE. The lowest percentage of participants (5.8%) identified themselves in the second stage of the PAP as deciding not to do anything about PSFE. Table 19 Relationship Between Always Providing a Smoke Free Environment and the Stages of the Precaution Adoption Process Stage of Precaution Adoption Always Does not Statistic Process provides a smoke free environment always provide a smoke free environment ^(df) n= 190 n = 361 107.15 (4) 1.1 haven't given it much thought (%) 2. I have decided I don't need to 9 (4.7) 45 (12.5) do anything (%) 3. I plan to reduce my child's 5 (2.6) 21 (5,8) exposure to ETS within 6 months (%) 0 (0.0) 29 (8.0) 4. I have started to reduce my child's exposure to ETS (%) 5.1 have decided to always 5 (2.6) 101 (28.0) provide a smoke-free 171 (90.0) 165 (45.7) environment (%) Summary The purpose of this chapter was to provide a detailed report of the quantitative findings of this study. Assessments of the psychometric properties of the scales used in the study were provided, followed by a description of the findings for each of the research questions. The most significant findings are highlighted below. In the bivariate data analysis there was a significant difference between groups of participants who always PSFE and those who do not always PSFE and also between those participants who identified themselves in the higher versus lower stages of the PAP. The differences were identified in the personal context characteristics as well as on 108 predictor variables in relation to each of the dependent variables. Those participants, who always PSFE, as well as those who identify themselves in the highest level of the PAP tend to have a higher education and income, live with a partner, were older and did not smoke. The predictor variables that were found to be positively associated with always PSFE included higher scores for self-efficacy and family functioning. Participants were more likely to always PSFE when no family smoked. Higher Fagerstrom scores and a greater number of cigarettes smoked were negatively related to always PSFE. The higher stage of PAP was positively associated with higher scores.for self-efficacy, ETS knowledge, and general smoking knowledge, as well as family functioning scores. Negatively associated with identifying themselves in the highest level of the PAP were number of cigarettes smoked per day, crowding in the home, more smokers in the home, and more family and friends who smoked. Logistic regression analysis for the dependent variable, always PSFE, revealed that the significant predictive factors were education, number of family and friends who smoke, and the mother's self-efficacy score. Two of these variables also remained significant for the second dependent variable, stage of PAP, the number of friends and family who smoke, and the self-efficacy score, but the third significant variable was the number of cigarettes smoked per day. When interaction effects were introduced into the logistic models, only one was found to be significant. Always PSFE was associated with higher self-efficacy scores, especially when there were fewer family and friends who smoked. There were also significant findings related to the relationship between the two dependent variables. As expected, a major proportion of participants who identified that they always PSFE also identified themselves in the highest level of the PAP. There was greater variety in what stage participants placed themselves in the PAP for those participants who did not always PSFE. These findings are discussed in the following chapter. 110 CHAPTER FIVE: DISCUSSION This study examined factors that are associated with mothers'/primary care givers' Provision of a Smoke Free Environment (PSFE) for their kindergarten child. In this chapter I discuss key findings from the study. Recommendations for research, practice, and education, as well as policy are highlighted. Study limitations are also discussed. This study provides an analysis that extends beyond a basic demographic description of families and homes where children are more likely to be exposed to environmental tobacco smoke (ETS) in that it includes an examination of the effect of a number of modifiable variables which expand our understanding of ETS exposure of children. Previous studies have determined that marital status, occupation, age, and education (Eriksen et al., 1996a) are significantly associated with the level of ETS in the home. ETS exposure is said to be greatest when mothers are young, single, and have a lower level of education, as well as a lower paying job. Using a theoretical framework, this study explored the presence and the strength of associations between the personal context and other modifiable factors in relation to PSFE. The use of this framework assisted in explaining the decision-making process of mothers/primary care givers related to protecting their children from ETS, and the actions associated with establishing a smoke free environment (SFE). Theoretical Framework Synthesized Model Previous literature has provided some insight into the potential factors which may be associated with mothers'/primary care givers' protection of their children from ETS. The findings from this study suggest that the original model used to guide the study I l l requires some refinement. The participants in this study did not live or act in isolation of the larger environment around them including the community context, a personal context and their family and friends. A l l of these contextual factors were included in the original model but the findings from this study suggest that these contextual factors may need to be reconsidered in their positioning within the model (see Figure 2). Previous studies have identified that individuals are more likely to smoke when they are in situations associated with material disadvantage (Greaves et al., 2003) and this is reflected in the model, where community and personal context directly affect smoking patterns. The smoking patterns in turn are hypothesized to affect the provision of a SFE. The most significant smoking variable identified in this study was the effect of the number of family and friends who smoke and will be discussed in more detail later. The social interaction context block of factors highlights the issues that occur within families which may have an impact on the behaviour and decision-making of mothers/primary care givers around PSFE. The variable within the social interaction context that was found to be significant in the likelihood of a SFE was the self-efficacy of mothers/primary care givers related to PSFE. The model continues to reflect the importance of the interaction between the decision-making process and the action of PSFE. Precaution Adoption Process The Precaution Adoption Process (PAP) has previously been used to assist in understanding the decision-making process of parents in relation to health protective behaviours such as accessing radon testing in homes (Weinstein, Klotz et al., 1989; Weinstein, Lyon et al., 1998; Weinstein & Sandman, 1992). Accessing radon testing is a short term behavior and does not require ongoing decision-making around a complex 112 behaviour such as exposure of children to ETS. This is the first study to use this decision-making model related to PSFE. The model suggests that the development of beliefs and intentions leads to action or inaction (Weinstein, 1988; Weinstein, Klotz et al., 1989; Weinstein, Lyon et al., 1998; Weinstein & Sandman, 1992). Individuals found to be in different stages of the process are thought to behave in qualitatively different ways. Weinstein and Sandman identified the four essential elements of stage theories such as PAP including the need for: 1) a category system to define the stages; 2) an ordering of the stages; 3) common barriers to change that people face in the same stage; and 4) different barriers to change that people face in different stages. In this study mothers/primary care givers were asked to place themselves in a stage of the decision-making process which they thought best described their approach to PSFE. Response options ranged from "I have decided to always provide a smoke free environment" to "I haven't given it much thought." In the bivariate analysis, 90% of the participants who always PSFE also indicated that they were in the highest stage of PAP, which is what would be expected, that the action matches the decision-making process. Among those who did not always PSFE, participants identified themselves in the lower stages of the PAP; there was some distribution over all stages of decision-making. Researchers have found that people are in different stages of readiness or decision-making prior to consistently acting to change health behaviour (Glanz et al., 2002; Love et al., 1996). The mothers/primary care givers who identified themselves in the highest stage of the PAP, but still did not always PSFE could be hypothesized to be more likely to make changes in the near future, than those who identify themselves in a lower stage of the decision-making process. 1 1 3 There were five respondents in this study who in relation to PSFE indicated "I have decided I don't need to do anything," yet these individuals also indicated that they always PSFE. These individuals were perhaps not faced with having to make a decision about exposing their children to ETS; perhaps because their families do not interact with smokers, and nonsmoking public policies keep them from having to make choices when they visit public places. The decision to protect their children may rarely have to be -considered by these mothers/primary care givers. In this study, because the participants were not well distributed over all stages of the PAP, the analysis did not provide insight into the characteristics of individuals in the different stages of decision-making in the multivariate analysis. The sampling frame was designed to intentionally include smokers and nonsmokers. The anticipated benefit was to be able to differentiate between the decision-making of mothers/primary care givers who smoke and those who do not. But with so many smokers and nonsmokers identifying that they had made the decision to always PSFE, the differences could not be detected. Mothers/primary care givers, whether they are smokers or nonsmokers, indicate similar awareness of the need to PSFE even if they have not accomplished the action of always PSFE. The multivariate analysis in this study provided information about those who were in the highest stage of the decision-making process and those grouped into the other lower stages of the decision-making process. This collapsing of stages meant that a great deal of information was lost about the characteristics of individuals in the other stages of the decision-making process. 114 Future Research The aim of using theoretical frameworks is to improve our ability to predict action and develop more effective interventions to assist individuals to make changes in behaviour. This cross-sectional study provided evidence of key factors associated with mothers'/primary care givers' placement within a certain PAP stage, but did not allow testing of how these factors might change over time. Future research using longitudinal studies would strengthen the support for using a stage model to predict which homes will become smoke free. Longitudinal studies could assist in assessing the changes in stage of mothers/primary care givers in the decision-making process and improve predictability of those who are more likely to go on to ideally PSFE. To evaluate a theoretical framework based on a process, one must be able to evaluate participants' actions over time (Gehrman & Hovell, 2003). The factors that mediate the movement between stages could also be tested. If the stages of the decision-making process were changed to a continuum instead of restrictive stages of decision-making, more subtle changes in decision-making may be detectable. With finer distinctions, repeated measures in a longitudinal study might detect movement along the continuum which may not be detectable when participants identify themselves in one stage or another. Some authors have hypothesized that certain factors influence the movement between stages in process models. These include: media messages about the hazard, perception of risk, and the influence of significant others (Crone et al., 2003; Weinstein, Klotz et al., 1989; Weinstein, Lyon et al., 1998; Weinstein & Sandman, 1992). The findings from this study provide insight into factors associated with participants who identify themselves within the highest PAP stage at one point in time, but do not shed light on changes that may occur over time. The PAP requires more study in order to 115 evaluate its usefulness in predicting mothers'/primary care givers' decision-making related to the behaviour of PSFE for their children. The reconceptualized model (figure 2) provides some insight into potential methods for future research. Two methods could include hierarchical linear modeling or path analysis. Hierarchical linear modeling would be one way to assess the effect of some higher level as well as lower level variables on the dependent variables. Using hierarchical linear modeling allows examination of nested effects, allowing the researcher to examine children being exposed to ETS within certain homes or other nested levels of variables. Hierarchical linear modeling method would also allow investigation of relationships within a particular level of variables as well as between levels. The other possible method for future research is using path analysis. The importance of direct and indirect effects of variables could be detected using similar cross-sectional data in a path analysis, with the potential paths indicated by the arrows in the figure. A more detailed theoretical understanding could be possible if future research considers some of these methods. Self-Efficacy The self-efficacy scores of the mothers/primary care givers were the most influential predictor of PSFE and the stage of the PAP. Self-efficacy is the belief that one can perform a skill under a specific set of conditions (Bandura, 1997). Bandura indicates that self-efficacy is one of the most important factors in the performance of difficult tasks. Positive efficacy beliefs are related to the person's sense of mastery from previous experience. When mothers/primary care givers have family and friends who smoke, it is helpful i f they have had some previous positive experiences with asking others not to smoke in their home, to build their capacity to feel confident in future situations. 116 In the multivariate analysis of this study the only interaction effect found to be significant was the self-efficacy score and number of family and friends who smoke related to always PSFE. This provides support for the need for mothers to have positive experiences to build their self-efficacy to ask others not to smoke in their homes. When fewer friends and family smoke, mothers/primary care givers were more likely to have a positive self-efficacy and were more likely to always PSFE. ETS knowledge, number of cigarettes smoked and crowding were also significant variables in the multivariate analysis for the highest stage of the decision-making process. This finding would suggest that while mothers/primary care givers are making the decision about PSFE in their home, there are more factors which influence their decision-making. The negative influences of crowding in the home, associating with smokers and the number of cigarettes the mother/primary care giver smokes each day speaks to the social context of smoking. Poland et al. (2006) describe how we can only explain human behaviour by dealing with the smaller issues of everyday life. Smoking behaviour is only one small portion of the many activities within a day for these families and this smoking behaviour occurs within the context of that family and the meanings constructed in the family. Interactions within families includes smoking behaviours, and these behaviours are developed through certain collective patterns of behaviour and power relationships in families (Poland et al., 2006) A mother/primary care giver who faces more people living in the home and more people who smoke and also smokes more cigarettes in a day may have a difficult time coming to the decision to always protect children from ETS. Once mothers/primary care givers make the decision to always protect children from ETS, self-efficacy and the number of family and friends who smoke remain the most important variables in the action of always PSFE. 117 Sockrider et al. (2003) identified that one of the most important predictors of having a smoking policy in the homes of families where smoking took place, was the self-efficacy of the mother in limiting infant exposure to ETS. Crone et al. (2001) also found that mothers' self-efficacy in asking others not to smoke was associated with the level of ETS exposure of children. Crone et al. identified that the self-efficacy of mothers differed between smoking and nonsmoking mothers. In this study the number of smoking mothers was too low (N = 153) to be able to do a further sub-group multivariate analysis. But there is growing support for the relationship between mothers'/primary care givers' self-efficacy and the level of ETS exposure of children in the home. Future research focused on the self-efficacy of mothers/primary care givers must consider the context of the family and community where self-efficacy develops. More studies need to involve the family and the large community context such as the Crone et al. (2003) study. The success that mothers/primarycare givers have in asking others not to smoke is associated with mastery of the skill. Self-efficacy to ask others not to smoke could be more successful in environments where more members of the families also recognize the importance of smoke free homes, and families as a whole expect to protect their children from ETS. When everyone is held responsible for protecting children from ETS, there will be fewer expectations placed on only mothers/primary care givers. A, few participants in this study had high self-efficacy scores but did not always PSFE. This finding may be related to the self-efficacy measure used in the study. For example, the self-efficacy measure focused on only the mothers'/primary care givers' ability to keep their home smoke free; other areas were not measured. The scale asked about a mother's/primary care giver's confidence in keeping his or her home smoke free under a variety of situations, including personally stressful situations (e.g., when 118 frustrated and anxious; when there are arguments in the home); social situations (e.g., having friends in for coffee or for a party); and normal family times (e.g., when feeding young children, or watching TV). The self-efficacy scale used for this study, did not measure the other areas where children could be exposed to ETS, such as in the family car, other people's vehicles, and other's homes where the family would visit which is where the ETS exposure of the children was occurring. As a result while mothers/primary care givers might consistently provide a smoke free home and have high self-efficacy, their children may be exposed to ETS in one of the other areas of their environments (e.g. the family car, other people's vehicles or other's homes where the family would visit). We need to continue to develop reliable tools to measure self-efficacy related to behaviours to protect children in all areas where ETS exposure is occurring. Bandura (1997) indicates that self-efficacy tools need to be specific for the behaviour being measured, so future tools need to account for the factors which have an influence on the areas where children are being exposed to ETS, beyond the confines of the home. Qualitative work with mothers, whose children are exposed to ETS in the family car, or other homes, may lead to answers about some of the issues that make it difficult for mothers/primary care givers to ask others not to smoke in those places. Education The exposure of young children to ETS has been associated with the educational level of mothers. Similarly in this study, level of education of the mother/primary care giver was associated with ETS exposure. The less education the mother/primary care giver has, the more likely her children will be exposed to ETS in the home. The action of PSFE was significantly associated with the level of education of the mother/primary care giver. 119 In many studies the likelihood of PSFE is described in relation to personal context variables. The demographic variables of education, income, age, and occupation identify the high risk groups where children are more likely exposed to ETS. It is interesting that when the smoking variables were entered into the multivariate analysis related to PSFE, education became a significant variable but was negatively associated with PSFE. The greater the education, in comparison to less than high school, the less likely a SFE was always provided. Could it possibly be that when the mother/primary care giver had a higher educational level and knows more people who smoke, that they have made a conscious decision NOT to provide a SFE? Are they possibly the people who still smoke, are resistant to the nonsmoking messages, are not allowed to smoke at work, so are smoking at home? The other issue to consider about this study population is that we measured the educational level of the mother/primary care giver. The education of the mother/primary care giver may not reflect the social context of the family as a whole. A mother/primary care giver may have a higher education which we might associate with a higher socio-economic family where fewer smokers are usually present. Many rural and farming families have different perspectives about smoking. The social context of the family may be more reflective of the group of individuals most in contact with the family who may still smoke more than other families where the mother has a high level of education. It is also interesting to note in this study that when controlling for other personal context factors and also smoking factors, education added to the predictive value in the final logistic regression model. Higher levels of educational attainment influence a mother's/primary care giver's self-efficacy to ask others not to smoke in the presence of children in his or her home. A third potential explanation for the negative association between always PSFE and education is that mothers with higher education may have had 120 a greater understanding of the areas where ETS may be found, such as in upholstery and carpet. These mothers may have believed that it was almost impossible not to expose their children to some ETS at some points in their lives, so avoided the "always" response to the question of how often they could keep places smoke free for their children. When we compare the influence of education on the two dependent variables, we recognize that education is associated with mothers'/primary care givers' action to protect their children from ETS but is not a significant factor for making the decision to PSFE. In other recent studies, educational level of mothers was found to influence their success in reducing their child's exposure to ETS, whether they smoked or not (Crone et al., 2001; Soliman et al., 2004). Health care professional educators and practicing professionals need to be aware of the influence of education on mothers'/primary care givers' confidence levels in the action of PSFE. Interventions to improve self-efficacy may be more likely effective when mothers have higher educational levels, especially i f the mothers/primary care givers have made the decision to always PSFE. Smoking in Families The findings from this study related to smoking within families demonstrated some expected and unexpected findings. As expected 90% of nonsmokers indicate that they always PSFE, but 63% of the participants who do not always PSFE, were also nonsmokers. Clearly there are many more factors affecting the PSFE for children than whether a mother/primary care giver smokes. Indeed, the regression analysis examining predictors of whether one always PSFE, revealed that smoking status was not a significant variable. The only smoking variable that was significant was the number of family and friends that the mother/primary care giver knew who smoked. This finding 121 suggests that the social milieu or social context of cigarette smoking is important (Love et al., 1996; Poland et al., 2006). With an increased number of smokers entering a home and requiring a mother to ask them not to smoke, there is an increased likelihood that there will be ETS in the home. Behaviours of nonsmokers who could not PSFE are also affected by the number of friends and family they know who smoke. Any changes in individual behaviours around smoking in the home are affected by the individual family context and also the larger community context. The family context around smoking behaviours may present unique challenges to family interactions, possibly making some women vulnerable in certain family situations, illustrating the power relationships sometimes present (Bottorff, Kalaw, Johnson, Steward, & Greaves, 2005; Poland, et al., 2006). The finding that the number of family and friends who smoke influences the likelihood of ETS exposure of children in the home is related to the larger social context that occurs around smoking behaviours themselves. The mother/primary care giver is only one person within the interactions that occur around the smoking behaviours in the home, so we must consider the larger social context of children being exposed to ETS in their homes. Poland et al. (2006) suggest that smoking behaviours are grounded in personal experience and should not be considered an individual behaviour but a collective social practice. When smoking is viewed in this way a much greater responsibility is placed on others to also be responsible for protecting children from ETS. In relation to the PAP, in the bivariate analysis smokers were less likely to be in a higher stage of PAP. In the logistic regression model, only two smoking variables remained significant for participants identifying themselves in the lower stages of the PAP, and that was number of cigarettes smoked per day and the number of friends and family who smoked. The decision to always PSFE is influenced by mothers'/primary care 122 givers' number of cigarettes they smoke per day. However, a high percentage of nonsmokers also identify themselves in the lower stages of the PAP. The nonsmoking mothers/primary care givers are obviously not affected by how many cigarettes are smoked per day, so the number of friends and family is the most influential smoking factor for the nonsmoking mothers. Again the social context of smoking behaviour appears to affect the decision to always protect their children from ETS (Crone et al., 2003; Poland et al., 2006). The implications for nursing practice, education, research, and policy around the smoking factors related to children's exposure to ETS are important. Addressing the social milieu or social context of smoking behaviour is accomplished to some extent with public policy, and gradually changing people's attitudes towards the acceptability of smoking behaviours. McMillen et al. (2003) indicate that attitudes among smoking and nonsmoking adults are becoming more positive about public smoking bans and more people agree that children should not be exposed to smoke. The public smoking bans, including those in workplaces, seem to be having an effect on the general smoking rate. This may gradually decrease the number of smokers in society and could potentially have a positive effect on the home environments as well (Ashley et al., 1998). Public policies that increase the awareness of the need to move smoking away from others, such as nonsmoking grounds of workplaces, will assist in making smoking around others less acceptable. However, public policy developed to care for children needs to be considered carefully regarding the implications for those affected. For example, a public policy banning smoking in all rental housing may put children in further danger by not considering that mothers may leave children unattended to go outside to smoke, or that by having smoking cessation imposed, stress may be increased in lower socioeconomic 123 homes. Therefore public policy should not be developed without consideration of all individuals being affected. Health care providers need to be sensitive to avoid stigmatizing smoking parents, which often just serves to decrease their desire to quit or change any habits regarding smoking in their homes (Botelho & Fiscella, 2005). From this study it is clear that it is not just the issue of whether a mother/primary care giver smokes that influences the presence of ETS. Emphasis must be also be placed on the social context of the smoking behaviour in the home and also include other predictive variables identified by this study. Beliefs and Knowledge of ETS Beliefs and knowledge of the effects of ETS have previously been identified as having an influence on protective behaviours (Lund et al., 1998a; McMillen et al., 2003; Tilson et al., 2001). Findings from the bivariate analysis of this study demonstrate that participants who always PSFE had significantly greater ETS knowledge, as well as greater general smoking knowledge than those participants who did not always PSFE. Participants who also identified themselves in the higher stages of the PAP also had greater ETS knowledge as well as a significantly greater general smoking knowledge. ETS knowledge remained significant in the final model for participants who identified themselves in the higher stage of the PAP, but neither smoking knowledge factors, general or ETS, remained a significant factor in the action of PSFE. Bandura (1997) claims that knowledge is required to perform a skill but knowledge does not always contribute to the actual performance of the skill. In other words, ETS smoking knowledge is related to the decision-making process, as part of the cognitive preparation for making a decision about the health consequences produced by smoking. Mothers use their knowledge in the development of a skill when moving 124 through this decision-making process. This is in keeping with the other factors which we have identified as influencing the stage of PAP in the full logistic regression model, some of which were the number of cigarettes smoked, the number of friends and family who smoke and the mother's self-efficacy. Health care providers need to continue to teach mothers/primary care givers about the effects of ETS. This action needs to be consistently completed by more health care professionals who mothers/primary care givers encounter. Klerman (2004) reviewed the literature to provide a summary of the current status of efforts to reduce ETS exposure among infants and young children. Interventions with caregivers were found to be effective at times but were not cost effective in their outcomes. Klerman questions whether the intensity of the interventions was justified for the amount of change that was produced. According to Klerman few consistent messages are being provided by health care providers. The consistency of messages provided by more health care professionals may be more cost effective than expecting one group of health care professionals to be responsible for time intensive interventions. This will only be accomplished by incorporating training techniques in all educational programs about reducing tobacco dependence and the harm of ETS to children (Klerman). Future research needs to continue to incorporate measures of knowledge and attitudes of mothers/primary care givers as well as others in the home regarding protecting their children from ETS. This will assist in predicting those families who will be successful at maintaining a SFE and assist in identifying those who need more intensive interventions in order to be successful. 125 Critique of Methods Used: Limitations and Contributions Limitations associated with this study arise from a variety of sources related to the use of survey research including: issues of sampling, response rate, bias, and measures. The relationship between these potential areas of limitation and the validity of study findings are discussed in this section. Methodological contributions of this study will also be discussed. Whole population sampling was used for this study; therefore generalizability of the findings beyond the study participants is not possible. A topic such as smoking in the home and exposing children to ETS is generally considered a sensitive topic and some would think that this is not anyone else's business. The methodology used resulted in a response from 76% of the population of kindergarten mothers/primary care givers. This relatively high response rate may have been obtained because of several positive techniques that contributed to more mothers/primary care givers willing to complete the survey. These included survey constructed using an explicit nonblaming approach; carefully worded reminders; advance notice; and incentives. This methodology has been recommended for surveys and follows suggestions to improve response rates (Brennan, 1998; Chiu & Brennan, 1998; Dillman, 2000; Mangione, 1995). Social desirability response bias and self-report of behaviours must also be considered with a survey, although self-report surveys have the least probability for producing socially desirable answers (Dillman, 2000). Social desirability bias is always a concern when the behaviour being asked about is not considered socially acceptable. Society is becoming less tolerant of people smoking around children and often views this as providing poor care for children (Government of Manitoba, 2004; Gross, Fogg, & Tucker, 1995; Gupta & Dwyer, 2001; Irvine et al., 1999; Reuter, Dennis, & Wilson, 126 2001; Secord, 2000). We can be fairly confident, however, that respondents were honest in their responses because the smoking rate for participants in this study was higher than the provincial smoking rate (CTUMS, 2004). Previous studies have reported that parental self-report on confidential surveys has correlated well with biological markers of exposure. A recent review of 60 studies, all including some form of self-report, indicated there was consensus of self-report and biological markers (Gaffney, Molloy, & Mandiegue, 2003). Many of the measures utilized in this study have been previously used and found to be valid and reliable: such as the family functioning scale arising from the McMaster Family Assessment Model (Miller, Ryan, Keitner, Bishop, & Epstein, 2000); demographic measures from Canadian Tobacco Use Monitoring Use Survey (2003) and the Fagerstrom Dependence Scale (Fagerstrom & Schneider, 1989). The factor analysis of the scales used for this study showed satisfactory internal consistency. The two dependent variables were measured using scales developed specifically for this study. The scale measuring the action of PSFE was a simple addition of the amount of time that mothers/primary care givers provided a SFE in four areas. Responses using this scale were validated by the response to another survey question which asked about the specific smoking restrictions in the home. The other dependent variable, the stage of PAP, was a scale based on the stages of the theoretical model specifically worded for the behaviour of PSFE. This study provides a foundation for future work with validation of a model to assess mothers' decision-making process around PSFE. The other scale which influenced the results of this study is the self-efficacy scale; the limitations of this scale were discussed with the self-efficacy findings. 127 Summary This study is a first to consider factors within a multivariate analysis associated with mothers'/primary care givers' protection of their children from ETS. This study considered community factors, personal context, smoking, and social interaction context factors as they relate to mothers'/primary care givers' PSFE and where they place themselves in the PAP. The personal context factors which remained significant were different for the two dependent variables, with the amount of crowding significant in the decision-making variable, and education significant for the action of PSFE. ETS knowledge was associated with the decision-making stage of the PAP, as well as the number of cigarettes the mother/primary care giver smoked per day. However, the most significant variable associated with both dependent variables was the mothers'/primary care givers' self-efficacy. The other notable factor was the number of friends and family who smoke which is a significant indicator of the social nature of smoking, not only in society as a whole but the social nature of smoking behaviour in homes. The one surprising finding that is beyond the scope of this study to adequately explain was the negative association between the variable always PSFE and the levels of advanced education. The greater the education compared to having less than high school, the less likely the respondents were to always PSFE. Several potential discussion points were provided but this finding requires further examination prior to arriving at any conclusions. A number of factors associated with the provision of a SFE are documented by this research. The knowledge gained from this study provides support for the development of more effective nursing interventions to reduce children's exposure to 128 ETS. Nurses play an important role in identifying homes where children are being exposed to ETS and finding ways to reduce the ETS in homes. The need to consider the influences of the community, the family, and individuals will improve the chances of reducing ETS in homes. Reducing ETS in homes and other places where children visit would have important positive health outcomes for children. 129 Figure 2. Reconceptualized model of potential factors associated with the 'decision' and the 'action' to provide a smoke free environment for children Community Context: Public policy; Rural versus urban; Workplace and Landlord smoking restrictions Personal Context: Educational level, Marital status, Age, Ethnicity, Occupation, Family income, Crowding in the home Smoking Patterns u Smokers in the home m Number of cigarettes mother smokes/day m Number of friends and family who smoke u Tobacco dependency level of mother u Outside area to smoke a Spouse smoking Providing a smoke free environment PSFE Stage of PAP Stage of decision-making process Social Interaction Context Heightened health awareness Family functioning Self-efficacy of mother related to PSFE Beliefs and knowledge of ETS 130 REFERENCES Agabiti, N . , Mallone, S., Forastiere, F., Corbo, G., Ferro, S., Renzoni, E., Sestini, P., et al. (1999). Impact of parental smoking on asthma and wheezing. Epidemiology, 10(6), 692-698. American Academy of Pediatrics Committee on Environmental Health. (1997). Environmental tobacco smoke: A hazard to children. Pediatrics, 99(A), 639-642. Ashley, M . , Cohen, J., Ferrence, R., Bull, S., Bondy, S., Poland, B., & Pederson, L. (1998). Smoking in the home: Changing attitudes and current practices. American Journal of Public Health, 88(5), 797-800. Atav, S., & Spencer, G. (2002). Health risk behavior among adolescents attending rural, suburban, and urban schools: A comparative study. Family Community Health, 25(2), 53-64. Bandura, A. (1997). Self-efficacy. The exercise of control. New York: W.H. Freeman and Company. Baranowski. T., Perry, C , & Parcel, G. (2002). How individuals, environments, and health behavior interact. In Glanz, K., Rimer, B. , Lewis, F. Health behavior and health education. Theory, research, and practice. (3 r d ed.). San Francisco: Jossey-Bass. Benter, S. (1991). Perceived seriousness of children's surgery and family adaptability and cohesion. Journal of Child and Adolescent Psychiatric and Mental Health Nursin,. 4(4), 137-142. 131 Berman, B., Wong, G., Bastani, R., Huang, T., Jones, C., Goldstein, D., et al. (2003). Household smoking behavior and ETS exposure among children with asthma in . low-income, minority households. Addictive Behaviors, 23, 11-128. Bickman, L. , & Rog, D. (Eds.). (1998). Handbook of applied social research methods. Thousand Oak: Sage Publications. Borland, R., Mullins, R., Trotter, L. , & White, V. (1999). Trends in environmental tobacco smoke restrictions in the home in Victoria, Australia. Tobacco Control, 8, 266-271. Borrelli, B., McQuaid, E, Becker, B., Hammond, K., Papandonatos, G., Fritz, G., et al. (2002). Motivating parents of kids with asthma to quit smoking: the PAQS project. Health Education Research, 17(5), 659-669. Botelho, R., & Fiscella, K. (2005). Protect children from environmental tobacco smoke, but avoid stigmatization of parents. A commentary. Family Systems and Health, 23 (1), 17-20. Bottorff, J., Kalaw, C., Johnson, J., Stewart, M . , & Greaves, L. (2005). Tobacco use in intimate spaces: Issues in the study of couple dynamics. Qualitative Health Research, 15(4), 564-577. Boyd, N . , Windsor, R., Perkins, L. , & Lowe, J. (1998). Quality of measurement of smoking status by self-report and saliva cotinine among pregnant women. Maternal and Child Health Journal, 2(2), 77-83. 132 Brennan, M . (1998). Techniques for improving mail survey response rates. Marketing Bulletin, 3, 24-37. Britton, J., & Weiss, S. (1997). Parental smoking and prevalence of respiratory symptoms and asthma in school age children. Thorax, 52(12), 1081-1094. Browne, A. , Shultis, J., & Thio-Watts, M . (1999). Solution-focused approaches to tobacco reduction with disadvantaged prenatal clients. Journal of Community Health Nursing, 16(3), 165-177. Butz, A. , & Rosenstein, B. (1992). Passive smoking among children with chronic respiratory disease. Journal of Asthma, 29(4), 265-272. Callais, F., Momas, I., Roche, D., Gauvin, S., Reungoat, P., & Zmirou, D. (2003). Questionnaire or objective assessment for studying exposure to tobacco smoke among asthmatic and healthy children: The French VESTA study. Preventive , Medicine, 36(\), 108-113. Canadian Institute of Child Health. (2000). The health of Canada's children. (3 r d ed.). Ottawa: Author. Canadian Tobacco Control Liaison Committee. (2003/ The national strategy: Moving forward. The 2003 progress report on tobacco control. Ottawa: Minister of Health. Canadian Tobacco Use Monitoring Use Survey. (2004). Retrieved February 22, 2006, from http://www.hc-sc.gc.ca/hl-vs/pubs/tobac-tabac/ctums-esutc-2004/tabl()9 c.html 133 Canadian Tobacco Use Monitoring Use Survey. (2005). Retrieved February 21, 2006 from http://www.hc-sc.gc.ca/hl-vs/pubs/tobac-tabca/ctums-esutc-2005/index e.html i Chiu, I., & Brennan, M . (1998). The effectiveness of some techniques for improving mail survey response rates: A meta-analysis. Marketing Bulletin, 1, 13-18. Chen,Y., L i , W., Y u , S., & Qian, W. (1988). Chang-Ning epidemiological study of children's health 1: Passive smoking and children's respiratory disease. InternationalJournal of Epidemiology, 77(2), 348-355. Christensen, A. , Tobiassen, M . , Jensen, T., Wielandt, H., Bakketeig, L., & Host, A. (2004). Methodology repeated validation of parental self-reported smoking during pregnancy and infancy: A prospective cohort study of infants at high risk for allergy development. Paediatric and Perinatal Epidemiology, 18(\), 73-80. Clark, B. (1997). Management of a preschool child's asthma: The parents' perspective. Unpublished master's thesis, University of Manitoba, Winnipeg, Manitoba, Canada. Cohen, J., & Cohen, P. (1987). Applied multiple regression/correlation analysis for the behavioral sciences. (2 n d ed.). London: Lawrence Erlbaum Associates, Publishers. Cook, D., Strachan, D., & Anderson, H. (1998). Systematic quantitative review of the effect of environmental tobacco smoke exposure on respiratory health in children. London, England: Scientific Committee on Tobacco and Health. 134 Cook, D., & Strachan, D. (1999). Summary of effects of parental smoking on the respiratory health of children and implications for research. Thorax, 54(April), 357-366. Cook, D., Whincup, P., Jarvis, M . , Strachan, D., Papacosta, O., & Bryant, A. (1994). Passive exposure to tobacco smoke in children aged 5-7 years: Individual, family and community factors. British Medical Journal, 308, 384-389. Conway, T., Woodruff, S., Edwards, C , Hovell, M . , & Klein, J. (2004). Intervention to reduce environmental tobacco smoke exposure in Latino children: Null effects on hair biomarkers and parent reports. Tobacco Control, 73(1), 90- 92. Crone, M . , Reijneveld, S., Burgmeijer, R., & Hirasing, R. (2001). Factors that influence passive smoking in infancy: A study among mothers of newborn babies in the Netherlands. Preventive Medicine, 32, 209-217. Crone, M . , Reijneveld, S., Willemsen, M . , & Hirasing, R. (2003). Parental education on passive smoking in infancy does work. European Journal of Public Health, 13, 269-274. Cunningham, R. (1996). Smoke and mirrors. The Canadian tobacco war. Ottawa: International Development Research Centre. Czaja, R., & Blair, J. (2005). Designing surveys. A guide to decisions and procedures. (2 n d ed.). Thousand Oaks: Sage Publications. de Vaus, D. (2002). Analyzing social science data. Thousand Oaks: Sage Publications. 135 Dillman, D. (2000). Mail and internet surveys. The tailored design method. (2 n d ed.). Toronto: John Wiley & Sons, Inc. Ehrlich, P.F., Longhi, J., Vaughan, R., & Rockwell, S. (2001). Correlation between parental perception and actual childhood patterns of bicycle helmet use and riding practices: Implications for designing injury prevention strategies. Journal of Pediatric Surgery, 36(5), 763-766. Elek, S., Hudson, D., & Bouffard, C. (2003). Marital and parenting satisfaction and infant care self-efficacy during the transition to parenthood: The effect of infant sex. Issues in Comprehensive Pediatric Nursing, 26(1), 45-58. Emmons, K. , Hammond, K., Fava, J., Velicer, W., Evans, J., & Munroe, A. (2001). A randomized trial to reduce passive.smoke exposure in low-income households with young children. Pediatrics, 108(\), 18-24. Epstein, N . , Baldwin, L. , & Bishop, D. (1983). The McMaster family assessment device. Journal of Marital and Family Therapy, 9(2), 171-180. Epstein, N . , Bishop, D, & Levin, S. (1978). The McMaster model of family functioning. Journal of Marriage and Family Counseling, 4(4), 19-31. Epstein, N . , Ryan, C , Bishop, D., Miller, I., & Keitner, G. (2003). The McMaster model: A view of health family functioning. In Walsh, F. (Ed.). Normal family processes: Growing diversity and complexity (3 r d ed., pp 581-607). New York: Guilford Press. 136 Eriksen, W., & Bruusgaard, D. (1995). Smoking behaviour in young families. Do parents take practical measures to prevent passive smoking by the children? Scandinavian Journal of Primary Health Care, 13, 275-280. Eriksen, W., Sandvik, L. , & Bruusgaard, D. (1997). Social support and the smoking behaviour of parents with preschool children. Scandinavian Journal of Social Medicine, 25(2), 93-99. Eriksen, W., Sorum, K. , & Bruusgaard. D. (1996a). Effects of information on smoking behaviour in families with preschool children. Acta Pediatric, 85, 209-212. Eriksen, W., Sorum, K. , & Bruusgaard, D. (1996b). Is there an increased lability in parents' smoking behaviour after a childbirth? Scandinavian Journal of Primary Health Care, 14, 86-91. Fagerstrom, K. (1978). Measuring degree of physical dependence to tobacco smoking with reference to individualization of treatment. Addictive Behaviors, 3, 235-241. Fagerstrom, K. (1982). Effects of a nicotine-enriched cigarette on nicotine titration, daily cigarette consumption, and levels of carbon monoxide, cotinine, and nicotine. Psychopharmacology, 77, 164-167. Fagerstrom, K., Kunze, M . , Schoberger, R., Breslau, N . , Hughes, J., Hurt, R., et al. (1996). Nicotine dependence versus smoking prevalence comparisons among countries and categories of smokers. Tobacco Control, 5, 52-56. 137 Fagerstrom, K. , & Schneider, R. (1989). Measuring nicotine dependence: A review of the Fagerstrom Tolerance Questionnaire. Journal of Behaviour Medicine, 12(2), 159-182. Fergusson, D., Horwood, F., Shannon, F., & Taylor, B. (1981). Parental smoking and lower respiratory illness in the first three years of life. Journal of Epidemiology and Community Health, 35(3), 180-184. Fichtenberg, C , & Glantz, S. (2002). Effect of smoke-free workplaces on smoking behaviour: Systematic review. British Medical Journal, 325, 188-194. Fink, A. (1995). The survey handbook. Thousand Oaks, California: Sage Publishing, Inc. Foster, J. (2002). Data analysis using SPSS for windows. Thousand Oaks: Sage Publications. Frindik, J., Williams, J., Johnson, R., & Dykman, R. (2002). Parental perceptions^and factors associated with glycemic control in pediatric patients with type 1 diabetes. Endocrine Practitioner, 8(2), 105-8. Gaffney, K. (2001). Infant exposure to environmental tobacco smoke. Journal of Nursing Scholarship, 33(4), 343-347. Gaffney, K. , Molloy, S., & Maradiegue, A. (2003). Questionnaires for the measurement of infant environmental tobacco smoke: A systematic review. Journal of Nursing Measurement, 11(3), 225 - 239. Gendall, P. (1994). Which letter worked best? Marketing Bulletin, 5, 53-63. 138 Gehrman, C , & Hovell, M . (2003). Protecting children from environmental tobacco smoke (ETS) exposure: A critical review. Nicotine and Tobacco Research, 5, 289-301. Glanz, K. , Lewis, F., & Rimer, B. (Eds.). (2002). Health behavior and health education. (3rd ed.). San Francisco: Jossey-Bass. Glasgow, R., Foster, L., Lee, M . , Hammond, S.K., Lichtenstein, E., & Andrews, J. (1998). Developing a brief measure of smoking in the home: Description and preliminary evaluation. Addictive Behaviors, 23(4), 567-571. Government of Manitoba. (2004, October 7). The non-smoking health protection act. Retrieved Oct. 11, 2004 from http://web2.gov.mb.ca/laws/statutes/ccsm/sl25e.php Greaves, L. , Cormier, R., Devries, K., Bottorff, J., Johnson, J., Kirkland, S., & Aboussafy, D. (2003). A best practices review of smoking cessation interventions for pregnant and postpartum girls and women. Vancouver: British Columbia Centre of Excellence for Women's Health. Groner, J., Ahijevych, K., Grossman, L., & Rich, L. (2000). The impact of a brief intervention on maternal smoking behavior. Pediatrics, 105(\), 267-271. Gross, D., Fogg, L., & Tucker, S. (1995). The efficacy of parent training for promoting positive parent-toddler relationships. Research in Nursing and Health, 18(6), 489-499. Gupta, R., & Dwyer, J. (2001). Focus groups with smokers to develop a smoke-free home campaign. American Journal of Health Behaviour, 25(6), 564-571. 139 Hadad, S., Franca, E., & Uchoa, E. (2002). Preventable infant mortality and quality of health care: Maternal perception of the child's illness and treatment. Cad. Saude Publico, Rio de Janerio, 18(6), 1519-1527. Haslam, C , & Draper, E. (2000). Stage of change is associated with assessment of the health risks of maternal smoking among pregnant women. Social Science and Medicine, 51, 1189-1196. Hansen, B. (1994). Acute illnesses in children. A description and analysis of parents' perception of illness threat. Scandinavian Journal of Primary Health Care, 12, 15-19. Health Canada. (2000) Statistical report on the health of Canadians. Ottawa: Author. Health Canada. (2003). The national strategy: Moving forward. The 2003 report on tobacco control. Ottawa: Author. Health Canada. (2004). Workplace tobacco control: Effective policies and programs. Retrieved March 17, 2004, from Health Canada Web site: http://www.hc-sc.gc.ca/hecs-sesc/tobacco/facts/blueribbon/healthier/06.html Health Canada, Canadian Institute of Health Information, Canadian Lung Association, & Statistics Canada. (2001). Respiratory disease in Canada. Editorial Board of Respiratory Disease in Canada. Ottawa: Health Canada. Heermann, J. (1988). Problem solving instruction: Examining its effect on the control of childhood asthma. Unpublished doctoral dissertation, University of Nebraska, Nebraska. 140 Helgason, A. , & Lund, K. (2001). Environmental tobacco smoke exposure of young children - attitudes and health-risk awareness in the Nordic countries. Nicotine and Tobacco Research, 3, 341-345. Helseth, S. (1995). Infant colic and parental self-efficacy - developing a theoretical model. Nursing Science and Research in the Nordic Countries, 15(2), 4-9. Hodges, E. (2003). A primer on early childhood obesity and parental influence. Pediatric Nursing, 29(1), 19-22. Hosmer, D., & Lemeshow, S. (2000). Applied logistic regression. (2 n d ed.). Toronto: John Wiley & Sons, Inc. Hovell, M . , Meltzer, S., Zakarian, J., Wahlgren, D., Emerson, J., Hofstetter, R., et al. (1994). Reduction of environmental tobacco smoke exposure among asthmatic children: A controlled trial. Chest, 106, 440-446. Hovell, M . , Zakarian, J., Matt, G., Hofstetter, R., Bernert, J., & Pirlke, J. (2000). Effect of counseling mothers on their children's exposure to environmental tobacco smoke: Randomized controlled trial. British Medical Journal, 321, 337-342. Irvine, L., Crombie, I., Clark, R., Slane, P., Feyerabend, C , Goodman, K., et al. (1999). Advising parents of asthmatic children on passive smoking: Randomised controlled trial. British Medical Journal, 318(7196), 1456-1459. Irvine, L., Crombie, I., Clark, R., Slane, P., Goodman, K. , Feyerabend, C , et al. (1997). What determines levels of passive smoking in children with asthma? Thorax, 52, 766-769. 141 Jaakkola, N . , Ruotsalainen, R., & Jaakkola, J. (1994). What are the determinants of children's exposure to environmental tobacco smoke at home? Scandinavian Journal of Social Medicine, 22(2), 107-112. Jarvis, M . , Goodman, E., Higgins, V. , Feyerabend, C., Bryant, A. , & Cook, D. (2000). Children's exposure to passive smoking in England since the 1980's: Cotinine evidence from population surveys. British Medical Journal, 527(7257), 343-345. Jarvis, M . , Strachan, D., & Feyerabend, C. (1992). Determinants of passive smoking in children in Edinburgh, Scotland. American Journal of Public Health, 82(9), 1225-1229. Jarvis, M . , Tunstall-Pedoe, H., Feyerabend, C , Vesey, C , & Saloojee, Y . (1987). Comparison of tests used to distinguish smokers from nonsmokers. American Journal of Public Health, 77(11), 1435-1438. Johansson, A. , Hailing, A. , Hermansson, G., & Ludvigsson, J. (2005). Assessment of smoking behaviors in the home and their influence on children's passive smoking: Development of a questionnaire. Annals of Epidemiology, 15, 453-459. Johansson, A. , Hermansson, G., & Ludvigsson, J. (2004). How should parents protect their children from environmental tobacco-smoke exposure in the home? Pediatrics, 113(4), e291-e295. Jurado, D., Munoz, C , De Dios Luan, J., & Fernandez-Crehuet, M . (2004). Environmental tobacco smoke exposure in children: Parental perception of smokiness at home and other factors associated with urinary cotinine in preschool 142 children. Journal of Exposure Analysis and Environmental Epidemiology, 14(A), 330-336. Kegler, ML, & Malcoe, L. (2002). Smoking restrictions in the home and car among rural native American and white families with young children. Preventive Medicine, 35,334-342. Klerman, L. (2004). Protecting children: Reducing their environmental tobacco smoke exposure. Nicotine and Tobacco Research, 6(2), S239-S252. Knapp, T. (1998). Quantitative nursing research. Thousand Oaks: Sage Publications. Knight, J., Eliopoulos, C , Klein, J., Greenwald, M . , &, Koren, G. (1996). Passive smoking in children. Racial difference in systematic exposure to cotinine by hair and urine analysis. Chest, 109(2), 446-450. Kohler, E., Sollich, V. , Schuster, R., & Thai, W. (1999). Passive smoke exposure in infants and children with respiratory tract diseases. Human and Experimental Toxicology, 18(A), 212-217. Lacchetti, C , Cohen, J., Ashley, M.J. , Ferrence, R., Bull, S., de Groh, M . , et al. (2001). Is nicotine dependence related to smokers' support for restrictions on smoking? Nicotine and Tobacco Research, 3, 257-260. Lambden, M . (2001). 77ze mediational role of working mother perceived self-efficacy. Unpublished doctoral dissertation, University of Texas at Austin, Texas. 143 Lewis, S., & Britton, J. (1998). Consistent effects of high socioeconomic status and low birth order, and the modifying effect of maternal smoking on the risk of allergic disease during childhood. Respiratory Medicine, 92(10), 1237-1242. L i , J., Peat, J., Xuan, W., & Berry, G. (1999). Meta-analysis on the association between environmental tobacco smoke (ETS) exposure and the prevalence of lower respiratory tract infection in early childhood. Pediatric Pulmonology, 27(\), 5-13. Love, M . , Davoli, G., & Thurman, Q. (1996). Normative beliefs of health behavior professionals regarding the psychosocial and environmental factors that influence health behavior change related to smoking cessation, regular exercise, and weight loss. The Science of Health Promotion, 10(5), 371-379. Lund, K., & Helgason, A. (2005). Environmental tobacco smoke in Norwegian homes, 1995 and 2001: Changes in children's exposure and parents' attitudes and health risk awareness. European Journal of Public Health, 15(2), 123-127. Lund, K. , Skrondal, A. , Vertio, FL, & Helgason, A . (1998a). Children's residential exposure to environmental tobacco smoke varies greatly between the Nordic countries. Scandinavian Journal of Social Medicine, 26(2), 115-120. Lund, K., Skrondal, A. , Vertio, H. , & Helgason, A . (1998b). To what extent do parents strive to protect their children from environmental tobacco smoke in the Nordic countries? Tobacco Control, 7, 56-60. Mangione, T. (1995). Mail surveys. Improving the quality. Thousand Oaks: Sage Publications. 144 Mannino, D. , Caraballo, R., Benowitz, N . , & Repace, J. (2001). Predictors of cotinine levels in U.S. children. Chest, 120, 718-724. Matt, G., Wahlgren, D., Hovell, M . , Zakarian, J., Bernert, J., & Meltxer, S. (1999). Measuring environmental tobacco smoke exposure in infants and young children through urine cotinine and memory-based parental reports: Empirical findings and discussion. Tobacco Control, 8, 282-289. Matt, G., Quintana, P., Hovell, M . , Bernert, J., Song, S., Novianti, N . , et al. (2004). Households contaminated by environmental tobacco smoke: Sources of infant exposure. Tobacco Control, 13, 29-31. McLeroy, K., Bibeau, D., Steckler, A. , & Glanz, K. (1988). An ecological perspective on health promotion programs. Health Education Quarterly, 15, 351-377. •f-McMillen, R., Winickoff, J., Klein, J., & Weitzman, M . (2003). US adult attitudes and practices regarding smoking restrictions and child exposure to environmental tobacco smoke: Changes in the social climate from 2000 - 2001. Pediatrics 112 (1), e55-e60. Merom, D., & Rissel, C. (2001). Factors associated with smoke-free homes in NSW: Results from the 1998 NSW health survey. Australian and New Zealand Journal of Public Health, 25(4), 339-346. Mesters, I., Meertens, R., Crebolder, H., & Parcel, G. (1993). Development of a health education program for parents of preschool children with asthma. Health Education Research, 8(\), 53-68. 145 Miller, I., Epstein, N . , Bishop, D. & Keitner, G. (1985). The McMaster Family Assessment Device: Reliability and validity. Journal of Marital and Family Therapy, 11(A), 345-356. Miller, I., Kabacoff, R., Epstein, N . , & Bishop, D. (1994). The development of a clinical rating scale for the McMaster Model of Family Functioning. Family Process, 33(\), 53-69. Miller, I., Ryan, C., Keitner, G., Bishop, D., & Epstein, N . (2000). Factor analysis of the family assessment device by Ridenour, Daley and Reich. A commentary. Family Process, 39(\), 141-144. Miller, I., Ryan, C., Keitner, G., Bishop, D., & Epstein, N . (2000). The McMaster approach to families: Theory, assessment, treatment, and research. Journal of Family Therapy, 22, 168-189. Miller, 1., Ryan, C., Keitner, G., Bishop, D:, & Epstein, N . (2000). Why fix what isn't broken? A rejoinder to Ridenour, Daley, & Reich. Family Process, 39(3), 381 -384. Munro, B., & Page, E. (1993). Statistical methods for health care research. (2 n d ed.). Philadelphia: J.B.Lippincott company. Murray, A . , & Morrison, B. (1988). Passive smoking and the seasonal difference of severity of asthma in children. Chest, 94(4), 701-708. Myers, L. , Gamst, G., & Guarino, A. J. (2006). Applied multivariate research. Design and interpretation. Thousand Oaks: Sage Publications. 146 Nelson, D. (2003). Family functioning measures: Convergent and discriminant validity. The Journal of Theory Construction and Testing, 7(1), 18-26. Okah, F., Choi, W., Okuuyemi, K., & Ahlawalia, J. (2002). Effect of children on home smoking restriction by inner city smokers. Pediatrics, 109(2), 244-249. Ontario Medical Association Committee on Population Health. (2002). OMA position paper on second-hand smoke. Ottawa: Ontario Medical Association. Parker, D., Lasater, T., Windsor, R., Wilkins, J., Upequi, D., & Heimdal, J. (2002). The accuracy of self-reported smoking status by cotinine test strips. Nicotine and Tobacco Research, 4, 305-309. Patrician, P. (2002). Multiple imputation for missing data. Research in Nursing and Health, 25, 76-84. Patterson, J., Neimanis, I., & Bain, E. (2003). Stopping smoking during pregnancy: Are we on the right track? Canadian Journal of Public Health, 94(4), 297-299. Phung, H. , Bauman, A. , Young, L., Tran, M . , & Hillman, K . (2002). Ecological and individual predictors of maternal smoking behaviour. Looking beyond individual socioeconomic predictors at the community setting. Addictive Behaviors, 28, 1333-1342. Poland, B. , Frohlich, K., Haines, R., Mykhalovskiy, E., Rock, M . , & Sparks, R. (2006). The social context of smoking: The next frontier in tobacco control? Tobacco Control, 15, 59-63. 147 Prochaska, J.O., & Prochaska, J.M. (1999). Why don't continents move? Why don't people change? Journal of Psychotherapy Integration, 9(\), 83-102. Ratner, P., Johnson, J., & Bottorff, J. (2001). Mothers' efforts to protect their infants from environmental tobacco smoke. Canadian Journal of Public Health, 92(\), 46-47. Ranson, M.K. , Prabhat, J., Chaloupka, F., & Nguyen, S. (2002). Global and regional estimates of the effectiveness and cost-effectiveness of price increases and other tobacco control policies. Nicotine and Tobacco Research, 4, 311-319. Reece, S. (1992). The parent expectations survey: A measure of perceived self-efficacy. Clinical Nursing Research, 7(4), 336-346. Reece, S., & Harkless, G. (1998). Self-efficacy, stress, and parental adaptation: Applications to the care of childbearing families. Journal of Family Nursing, 4(2), 198-215. Reutter, L., Dennis, D., & Wilson, D. (2001). Young parents' understanding and actions related to the determinants of health. Canadian Journal of Public Health, 92(5), 335-339. Roden, J. (2003). Capturing parents' understanding about the health behaviors, they practice with their preschool-aged children. Issues in Comprehensive Pediatric Nursing, 26, 23-44. Scherer, G., Kramer, U . , Meger-Kossien, I., Riedel, K., Heller, W., Link, E., et al. (2004). Determinants of children's exposure to environmental tobacco smoke (ETS): A 148 study in Southern Germany. Journal of Exposure Analysis and Environmental Epidemiology, 14, 284-292. Scherer, G., Meger-Kossien, I., Riedel, K., Renner, T., & Meger, M . (1999). Assessment of the exposure of children to environmental tobacco smoke (ETS) by different methods. Human Experimental Toxicology, 18, 297-301. Schuster, M . , Franke, T., & Pham, C. (2002). Smoking patterns of household members and visitors in homes with children in the United States. Archives of Pediatric and Adolescent Medicine, 156, 1094-1100. Secord, S.A. (2000). How women protect their children from environmental tobacco smoke: A smoldering issue. Unpublished master's thesis, University of British Columbia, Vancouver, British Columbia, Canada. Sexton, K. , Adgate, J., Church, T., Hecht, S., Ramachandran, G., Greaves, I., et al. (2004). Children's exposure to environmental tobacco smoke: Using diverse exposure metrics to document ethnic/racial differences. Environmental Health Perspectives, 112, 392-397. Simpson, E., Moll , E., Kassam-Adams, N . , Winston, F., & Miller, G. (2002). Barriers to booster seat use and strategies to increase their use. Pediatrics, 110(4), 729-737'. Sockrider, M . , Suchanek Hudmon, K. , Addy, R., & Dolan Mullen, P. (2003). An exploratory study of control of smoking in the home to reduce infant exposure to environmental tobacco smoke. Nicotine and Tobacco Research, 5(6), 901-910. 149 Soliman, S., Pollack, H. , & Werner, K. (2004). Decrease in prevalence of environmental tobacco smoke exposure in the home during the 1990's in families with children. American Journal of Public Health, 94(2), 314-320. Smith, T. (1999). Developing nonresponse standards. International Conference on Survey Nonresponse. Portland, October, 1999. Stanton, W., & Silva, P. (1993). Tracking change in the patterns of parental smoking. Royal Society of Health Journal, (February), 12-16. Statistics Canada (1998). National Population Health Survey, 1996-97. Ottawa: Minister of Industry (Statistics Canada Cat. No. 82-567-XPB). Statistics Canada. (2001) Canadian Community Health Survey, 2000/01. Retrieved January 13, 2004 from http://www.statcan.ca/english/sdds/3226.htm. Statistics Canada (2003). Canadian Community Health Survey, 2002/03. Retrieved July 15, 2004 from http://www.statcan.ca/english/sdds/3226.htm. Statistics Canada. (2005). Canadian Statistics: Labour force, employed and unemployed by province. Retrieved February 28, 2006 from http ://w w w40. statcan. ca/101 /gsto 11\ abor07b .htm Stotts, A. , Diclemente, C , Caronari, J., & Mullen, P. (1996). Pregnancy smoking cessation. A case of mistaken identity. Addictive Behaviors, 21(A), 459-471. Strauss, R. (2001). Environmental tobacco smoke and serum vitamin C levels in children. Pediatrics, 107(3), 540-542. 150 Stronks, K., van de Mheen, K., Looman, C , & Mackenbach, J. (1997). Cultural, material, and psychosocial correlates of the socioeconomic gradient in smoking behavior among adults. Preventive Medicine, 26, 754-766. Studlar, D. (2002). Tobacco control. Comparative politics in United States and Canada. Peterborough, ON: Broadview Press, Ltd. Tabachnick, B., & Fidell, L. (1996). Using multivariate statistics. (3 r d ed.). New York: Harper Collins Publishers Inc. Tilson, E., McBride, C , Albright, J., & Sargent, J. (2001). Attitudes toward smoking and family-based health protection among rural mothers and other primary care givers who smoke. Journal of School Health, 77(10), 489-495. Vartianen, E., Seppala, T., Lillsunde, P., & Puska, P. (2002). Validation of self-reported smoking by serum cotinine measurement in a community-based study. Journal of Epidemiology and Community Health, 56(3), 167-170. Velicer, W., Diclemente, C , Rossi, J., & Prochaska, J. (1990). Relapse situations and self-efficacy: An integrative model. Addictive Behaviors, 15, 271-283. Venters, M . , Solberg, L., Kottlke, T., Brekke, M . , Perchacek, T., & Grimm, R. (1987). Smoking patterns among social contacts of smokers, ex-smokers, and never smokers: The doctors helping smokers study. Preventive Medicine, 16, 625-635. Voorhees, C , Schrieber, G:, Schumann, B., Biro, F., & Crawford, P. (2002). Early predictors of daily smoking in young women: The national heart, lung, and blood institute growth and health study. Preventive Medicine, 34, 616-624. 151 Weinstein, N . (1988). The precaution adoption process. Health Psychology, 7(4), 355-• 386. Weinstein, N . , Klotz, M . , & Sandman, P. (1989). Promoting remedial response to the risk of radon: Are information campaigns enough? Science, Technology, and Human Values, 14(4), 360-379. Weinstein, N . , Lyon, J., Sandman, P., & Cuite, C. (1998). Experimental evidence for stages of health behavior change: The precaution adoption process model applied to home radon testing. Health Psychology, 17(5), 445-453. Weinstein, N . , & Sandman, P. (1992). A model of the precaution adoption process: Evidence from home radon testing. Health Psychology, 11(3), 170-180. Whitehead, D., & Doherty, W. (1989). Systems dynamics in cigarette smoking: An exploratory study. Family Systems Medicine, 7(3), 264-273. Willers, S., Hein, H. , & Jansson, L. (2004). Assessment of environmental tobacco smoke exposure: Urinary cotinine concentrations in children are strongly associated with the house dust concentrations of nicotine at home. Indoor Air, 14, 83-86. Wills, T., & Cleary, S. (1997). The validity of self-reports of smoking: Analysis of race/ethnicity in a school sample of urban adolescents. American Journal of Public Health, 57(1), 56-61. Woodruff, S., Conway, T., Edwards, C , & Hovell, M . (2003). Acceptability and validity of hair collection from Latino children to assess exposure to environmental tobacco smoke. Nicotine and Tobacco Research 5, 375-385. 152 Winickoff, J., Hillis, V . , Palfrey, J., Perrin, J., & Rigotti, N . (2003). A smoking cessation intervention for parents of children who are hospitalized for respiratory illnesses. The Stop Tobacco Outreach Program. Pediatrics, 111 (1), 140-145. World Health Organization, Department of Noncommunicable Disease, & Tobacco Free Initiative. (1999) International consultation on environmental tobacco smoke (ETS) and child health. Geneva, Switzerland: World Health Organization. NCD/tfi/ets/99.2. Wright, L., Bell, J., & Rock, B.L . (1989). Smoking behavior and spouses: A case report. Family Systems Medicine, 7 (2), 158-171. Wright, A. , Holberg, C , Martinez, F., Taussig, L., & Group Health Medical Associates. (1991). Relationship of parental smoking to wheezing and nonwheezing lower respiratory track illnesses in infancy. Journal of Pediatrics, 118, 207-214. Wright, L., Watson, W., & Bell, J. (1996). Beliefs. The heart of healing in families and illness. New York: Basic Books. Yoos, B., Kitzman, H. , McMullen, A . , & Sidora, K. (2003). Symptom perception in childhood asthma: How accurate are children and their parents? Journal of Asthma, 40(\), 27-39. Appendix C: Teacher Letter 155 Date Dear Kindergarten Teachers, There are still 200,000 Canadian children under the age of 12 being exposed to environmental tobacco smoke in their homes every day. Public policies and government programs have placed children's health as a priority, but there has been little success in reducing young children's exposure. Environmental tobacco smoke has severe health consequences for children and is strongly linked to asthma severity, middle ear disease, respiratory infections and sudden infant death syndrome. I am conducting a survey as part of a PhD program from University of British Columbia, and also teach nursing at Brandon University. This research has received ethical approval from both Brandon University and University of British Columbia. M y previous research work has been with parents' management of asthma and I initiated the pediatric asthma clinics at the Brandon Regional Health Centre as a consequence. This survey is about factors that influence mothers' protection of their children from environmental tobacco smoke in the child's immediate environment, including their home, family vehicle, others' vehicles or homes children frequently visit. I am interested in identifying the most influential and interacting factors, in order to develop more specific interventions to assist mothers' in their efforts to provide a smoke-free environment for their children. The Brandon School Division Board has given permission for distribution of the surveys in the Brandon Schools, and all mothers of kindergarten children will be invited to participate. In the next few weeks a research assistant or myself will be visiting your classroom, bringing advance notices of the study to be distributed through the children in your classroom to their mothers. At that time we will collect a class list. The list will be used only during data collection procedures and then destroyed and no information will be retained that identifies the children with the surveys. th The survey package will arrive to be distributed on May 5 and contain an invitation to participate, the survey tool and a return envelope. A research assistant or myself will visit the classroom frequently to gather the surveys that have been returned to you. The mothers are asked to complete the survey and return it to the school with their child. When the child has returned the surveys, they will receive a pencil from the research assistant. The mothers may refuse to participate, by not filling out the survey, but are asked to return it in the sealed envelope anyway. This allows mothers to feel that the teachers are also not aware of who has actually participated even if they return their envelope. 156 Three reminders will be used. One week following the survey distribution, a reminder notice will be sent home in the same way. Two weeks after the survey distribution a second reminder will be sent out to those who have not returned the survey. The third reminder will also be sent home to those who have not returned the survey, but will include a new survey package. The reminder list for the children will be maintained by the research team until the survey collection is complete and then the list will be destroyed. A draw will be held for all children within the School Division for a bicycle and helmet as a thank you for their time. If you receive any questions about the survey from mothers, please indicate to the mothers that they should contact me at (XXX) X X X - X X X X for more information, for another survey or for assistance in completing the survey. If you have any questions about the survey, do not hesitate to call or e-mail me. Sincerely, Bev Temple, R N , M N , PhD (c) Assistant Professor, Department of Nursing, Health Studies Brandon University 157 Appendix D: Advance Notice NOTICE TO MOTHERS OR PRIMARY CARE GIVERS OF KINDERGARTEN CHILDREN IN THE BRANDON SCHOOL DIVISION We need your help by completing a "Child Health Survey." Surveys will be sent home with your child on May 5 t h and we ask that you complete the survey as soon as possible and return in the attached envelope. Your help is important to our work of assisting mothers or primary care givers with keeping their children healthy. Bev Temple, RN, M N Assistant Professor, Brandon University, PhD Student, University of British Columbia 158 Appendix E: Information Sheet with Survey Young children's health is important to us all. We would like the benefit of your knowledge as a mother and a few minutes of your time. Only mothers of kindergarten children or the person who spends the most time with your child can help us. It is important to our work that we gain from the experiences and opinions of all of you. Your responses will be kept confidential. Please read the attached letter of invitation prior to filling out the survey and returning it. Thanh you for assisting us. Appendix F: Survey Tool The Child Health Survey BRANDON UNIVERSITY Founded 1899 This survey uses the term Environmental Tobacco Smoke to mean the same as second hand smoke. H O U S E H O L D I N F O R M A T I O N This section focuses on people within your household and where you live 1. Please list the number of children and ages of those children living in your home. Number Ages in years (If less than one year, list age in months) 2. How many people regularly live in your home? 3. How many rooms are in your home (not including the kitchen and bathroom/s)? 4. Currently, is there anyone living in your home who is pregnant? • Yes • No 5. Do you have a yard, balcony or deck attached to your home where children can play? • Yes • No 6. Is there an outside area where people could smoke if they chose? • Yes • No 7. If you are living in a rented space, does the landlord restrict smoking inside? • Yes • No • Not applicable H E A L T H S T A T U S This section focuses on communication within the family and the families' health 8. Please rate your agreement or disagreement about how well the following statements describe your family: Strongly Strongly Agree Agree Disagree Disagree a) Planning family activities is difficult because we misunderstand each other. 1 2 3 4 b) In times of crisis we can turn to each other for support. 1 2 3 4 c) We cannot talk to each other about the sadness we feel. 1 2 3 4 d) Individuals are accepted for what they are. 1 2 3 4 e) We avoid discussing our fears and concerns. 1 2 3 4 f) We can express feelings to each other. g) There are lots of bad feelings toward each other. h) We feel accepted for what we are. 1 2 3 4 i) Making decisions is a problem for our family. 1 2 . 3 4 j) We are able to make decisions about how to solve problems. 1 2 3 4 k) We don't get along well together. 1 2 3 4 I) We confide in each other. 1 2 3 4 9. Have any of your children been diagnosed by a doctor with any of the following and been ill in the last 12 months with the illness? D i a g n o s e d III w i t h i n 12 m o n t h s . Asthma • Yes • No • Yes • No Allergic diseases • Yes • No • Yes • No Ear Infections • Yes • No • Yes • No Respiratory illnesses such as Tonsillitis, Bronchitis, or Pneumonia • Yes • No • Yes • No 10. Do you believe that any of the above-diagnosed illnesses in your children could be made worse by their exposure to tobacco smoke? • Yes • • No • My children have not been diagnosed 11. Do you believe that any of the above-diagnosed illnesses are a major health risk to your children? • Yes • No • My children have not been diagnosed 12. Do you believe that any of these illnesses could be a threat to your child's life? • Yes • No • My children have not been diagnosed 13. Do you believe that your child is less healthy, about the same or more healthy than other children their age? • Less healthy • The same • More healthy 14. Do you believe that reducing the amount of Environmental Tobacco Smoke your children are exposed to will improve their health? • Yes • No • Not applicable 164 P E R S O N A L S M O K I N G S T A T U S This section focuses on your personal experience with tobacco use 15. Have you ever smoked cigarettes? • Yes • No (If NO, skip to Question 29) 16. Have you ever smoked more than 100 cigarettes in your lifetime? • Yes • No 17. How old were you when you first started to smoke cigarettes? years 18. Do you currently smoke cigarettes? • Yes • No (If NO, skip to Question 27) 19. How many days per week do you smoke? days per week 20. How many cigarettes do you smoke each day? cigarettes per day CURRENT SMOKERS 21. How soon after you wake up do you smoke your first cigarette? • Within 5 minutes Q 6 - 3 0 minutes • 30 - 60 minutes • After 60 minutes 22. Do you find it difficult to refrain from smoking cigarettes in places where it is forbidden? • Yes • No 23. Which cigarette would you hate the most to give up? • First one in the morning • Any other 24. How many cigarettes per day do you smoke? • Less than 10 cigarettes • 11 - 20 cigarettes • 21 - 30 cigarettes • More than 31 cigarettes 165 25. Do you smoke cigarettes more frequently during the first hours after waking than during the rest of the day? • Yes • No 26. Do you smoke cigarettes even if you are so ill that you are in bed most of the day? • Yes • No (Skip to Thoughts about Smoking, question 31) FORMER SMOKERS 27. When did you quit smoking? OR (month/day/year) (years ago) 28. Why did you quit smoking? (Please select all that apply) • Because of an allergy • Bad for my health in general • Bad for my health because of a specific health problem (e.g. respiratory problems, heart disease, diabetes) • Bad for my children's health • Bad for the rest of my families'health • Relative/close friend had smoking related illness • Someone in the home was pregnant • Didn't like the taste/smell • Because it costs too much • Does not suit my lifestyle • Other (Skip to Thoughts about Smoking Question 31) 1 6 6 NON SMOKERS This section focuses on smoking in the your environment 29. What are the reasons why you are a non smoker? (Please select all that apply) • Was never interested in starting • Does not suit my lifestyle • Because of an allergy • Bad for my health in general • Bad for my health because of a specific health prob lem (e.g. respiratory problems, heart disease, diabetes) • Bad for my children's health • Bad for the rest of my families' health • Relative/close fr iend had smoking related illness • Someone in the home was pregnant • Because it costs too much • Other 30. Have you ever had illness symptoms that you believed were related to exposure to Environmental Tobacco Smoke? a) QYes • No b) Please describe the symptoms 167 THOUGHTS A B O U T SMOKING 31. Please circle the number that best describes your level of agreement or disagreement with the following statements: a) Smoking makes people feel lively and awake b) Smoking is addictive c) Smoking helps people relax d) Smoking is dangerous to the smokers' health e) Smoking is physically dangerous f) Smoking is likely to cause heart disease g) Smoking causes cancer h) Smoking is helpful for weight loss i) Smoke is dangerous to those who inhale Environmental Tobacco Smoke j) Environmental Tobacco Smoke is linked to asthma in children k Environmental Tobacco Smoke is linked to ear infections in children I) Environmental Tobacco Smoke is linked to respiratory infections in children Strongly Agree Agree 2 2 2 2 2 2 2 2 Strongly Disagree Disagree 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4 168 S M O K E F R E E E N V I R O N M E N T S This section focuses on smoking in the your environment 32. If you are working outside the home, are there cigarette-smoking restrictions at your workplace? • Yes • N o • Not working outside my home 33. Does your spouse or partner smoke cigarettes? • Yes • No Q Live alone 34. Including both family members and friends who you see weekly, about how many people do you know who smoke cigarettes? • None • 1 person • 2 people • 3 - 5 people • more than 5 people 35. Excluding yourself, how many people who smoke cigarettes live in your home? • None • 1 person • 2 people • 3 - 5 people • more than 5 people 36. On a typical day, how many cigarettes are smoked INSIDE your home? • none • less than 10 cigarettes • 11 - 20 cigarettes • 21 - 30 cigarettes • more than 31 cigarettes 37. Is cigarette smoking restricted in any way inside your home? a) QYes • No b) If yes, how is cigarette smoking restricted inside your home? (Please check all that apply) • Smoking is not allowed in any room of the house • Smoking is allowed in certain rooms only • Smoking is only restricted in the presence of young children • Smoking takes place near an open window • There are no ashtrays ' • There are smokeless ashtrays • Other - please specify 169 38. Under what conditions, is smoking allowed inside your home? • Not applicable - there is no smoking in my home Please check all conditions that apply: • Only when certain people come to my home, please specify who • Only certain times of the day, please specify when • Only in certain rooms of the home, please specify which rooms • Only in winter, when it is too cold outside • Any other times smoking is allowed, please describe 39. If you have no smoking restrictions in your home, have you started to make plans about having some smoking restrictions within your home in the next 6 months? • Yes • No • Not applicable If yes, could you describe how you plan to do this? 1 7 0 40. How often are you able to keep the following places smoke free? (Please circle the number that best describes how often for each situation) Never Some of Most of Always N/A the time the time Your entire home Your vehicle Other peoples' vehicles that children travel in 1 e.g. Don't have a vehicle Homes your child visits at least once/week 41. How many hours per week is your child exposed to Environmental Tobacco Smoke in these places (your home, your vehicle, other vehicles and other homes)? • None • 1 - 3 hours • 4 - 5 hours • More than 5 hours ' 42. Please circle the number that indicates which of the following best represents your approach to providing a Smoke Free Environment for your child? I haven't given it much thought. I have decided I don't need to do anything. I am planning to reduce my child's exposure to environmental tobacco smoke within the next 6 months. I have started to reduce my child's exposure to environmental tobacco smoke. 5) I have decided to always provide a smoke-free environment for my child. D 2) 3) 4) 1 171 43. People often find that there are certain times when it is harder to be firm about their plan to have a smoke-free environment. Please indicate how confident you are in situations to be firm about your plan, with 1 being not at all confident and 5 being very confident. (Please circle the number that best describes your confidence level for each situation) How confident are you that you could keep your home 'smoke free' in the following situations? a) When my spouse or a close relative wants to smoke in our home b) When I have friends in for coffee and we are relaxing c) When I am frustrated and anxious d) When the family is watching television e) When we have friends in for a party f) When the family has just finished eating g) When there are arguments and conflicts in my family h) When I am feeding young children i) When no one else is home j) When the children are in bed Not at all Confident 2 Very Confident 4 5 3 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4 4 D E M O G R A P H I C S This section explores demographic information and provides a profile of those who participate in the survey 44. What is your age? years. 45. What gender are you? • Male • Female • Transgender 46. What is the highest grade or level of education you have ever attained? • No schooling • Some elementary • Completed elementary • Some high school • Completed high school • Some community college, technical college • Completed community college, technical college • Some university • Completed university • Other education or training, . please list . . 47. How would you describe yourself? (Please select all that apply) • White/Caucasian • Japanese • Aboriginal/First Nation • South Asian • Korean • Arab/West Asian • Black • Latin American • Filipino • Other ; 48. What is your best estimate of your total household income for the last 12 months before taxes and deductions? Please include income from all household members and from all sources.Was it... • Less than $15,000 • $15,000 to $29,999 • $30,000 to $44,999 • $45,000 to $59,999 • $60,000 to $79,999 • $80,000 to $99,999 • $100,000 to $119,999 • More than $120,000 • Don't know 49. Which of the following best describes your MAIN activity during the last 12 months? Were you... • Looking for work Q Working at a job or business outside the home • A student Q Raising a family or running a household • Retired • Other If working please describe your occupation 50. What is your current marital status? • Now married and living with wife/husband • Common Law relationship/live with partner • Separated • Divorced • Widowed • Never married (single) • 51. What is your relationship to the kindergarten child? • Mother • Father • Foster Parent • Step Parent • Other related - please specify • Unrelated 52. Do you live within the Brandon city limits? • Yes • No Thank you for c o m p l e t i n g f H © C| €* S11 ci P fi ci I r P . P l e a s e c h e c k t o s e e t h a t y o u h a v e a n s w e r e d a l l q u e s t i o n s t h a t r e l a t e t o y o u . P l e a s e r e t u r n t h e s u r v e y t o t h e s c h o o l i n t h e e n c l o s e d e n v e l o p e . 175 Appendix G: Reminder 1 May 12 t h, 2005 About one week ago we sent you a survey about your child's health. A l l mothers' of kindergarten children in the Brandon School Divisions were invited to participate. If you have already returned the survey, please accept our sincere thanks. If not, please do it today. We need your survey to gain an accurate account of mothers' opinions and experiences. If you smoke or do not smoke it is very important for us to hear your opinions and gain an understanding of what supports you in how you are protecting your child. Your confidentiality will be maintained and there will be no risk to your child's education because their teacher will not know if you completed the survey or not. Your child's name will be entered in the draw for the bicycle when you return your survey. If you did not receive the survey, or it got misplaced, please call me at ( X X X ) X X X -X X X X and leave a message with your child's name and the school where they attend. We will send another one home with your child as soon as possible. Sincerely, Beverley Temple, RN, M N , PhD (c) Brandon University, Department of Nursing PhD Student, University of British Columbia 176 Appendix H: Reminder 2 May 19 t h, 2005 About 2 weeks ago we sent you a survey about your child's health and their environment. We have not received your survey yet. If you have completed the survey we thank you. If you have misplaced the survey and you would like another sent to you, please check the box below and return this notice with your child to the school and another survey will be sent out as soon as possible or simply call ( X X X ) X X X - X X X X and another survey will be sent to you. Your participation is very important to us. The opinions of all mothers or primary care givers, whether they smoke or not, are valued and we seek an understanding from all. Please remember that your confidentiality will be maintained and the teachers are not aware of who has completed the survey and who has not. If you would like assistance filling out the survey, please leave that message on the above phone number and a time that would be convenient to call you back. If you do not wish to participate in this research, please return the empty survey in the envelope anyway. This will assist us by knowing that you just choose not to complete the survey tool. Thank you for your valuable time. Bev Temple, RN Assistant Professor Brandon University, PhD Student University of British Columbia • I wish to have another survey sent home with my child (child's name) 177 Appendix I: Reminder 3 May 26 t h, 2005 Three weeks ago I sent you a survey about your child's health and I do not have record of you returning it at this. time. If you have, please accept my thanks. If you have not returned the survey, I have attached another one to assist you to complete it today. If you are a smoking or a non-smoking mother, your information is just as important to me in gaining a full understanding. A l l of your answers will be kept confidential, there will be no identification attached to the survey, and the teacher at your child's school will not know if you have completed the survey or not. Your participation is very important to the success of my study. If you still do not wish to participate, please return the survey in the sealed envelope so that I know that you have chosen not to participate. The draw for the bicycle will be done on June 2, 2005 and your child's name will be entered as soon as you have returned the survey. Thank you for the generous donation of your time. I appreciate you completing this survey. Bev Temple, R N Assistant Professor, Brandon University PhD student, University of British Columbia (XXX) X X X - X X X X 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0092975/manifest

Comment

Related Items