The Influence of Contextual Risk and Protective Factors on Youth Smoking in Rural and Urban Florida Environments

MISSING IMAGE

Material Information

Title:
The Influence of Contextual Risk and Protective Factors on Youth Smoking in Rural and Urban Florida Environments
Physical Description:
1 online resource (108 p.)
Language:
english
Creator:
Decubellis, Tracy M
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Family, Youth and Community Sciences
Committee Chair:
Fogarty, Kate
Committee Members:
Diehl, David Christopher
Johns, Tracy Lynn

Subjects

Subjects / Keywords:
florida -- rural -- smoking -- urban -- youth
Family, Youth and Community Sciences -- Dissertations, Academic -- UF
Genre:
Family, Youth and Community Sciences thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Introduction: In the State of Florida, risk and protective factors associated with cigarette use have been examined statewide; however, no information currently exists analyzing the differential influence of risk and protective factors on youth smoking among students living in urban as compared to rural populations. Purpose: The purpose of this study is to examine the contextual influence of urban as compared to rural environments on adolescent smoking behavior. This study examines the ways in which risk and protective factors interact to influence youth smoking behavior, and the ways in which these interactions may differ based on the environment in which the youth reside. Methods: This study analyzed statewide data from the 2012 Florida Youth Substance Abuse Survey (n = 33,748 high school students; 44% urban, and 56% rural). Predictor variables were chosen based on the Problem Behavior Theory and were organized into risk and protective factor categories. Results: The results indicated that for urban and rural youth, the risk factor positively associated with smoking was favorable parental attitudes toward alcohol, tobacco, and other drug use. Protective factors negatively associated with smoking were personal intolerance of deviance, and religious service attendance. Results of Fisher’s z tests suggest significant differences between the coefficients in urban and rural samples for all protective factors, and the risk factor laws favorable to drug use. Conclusion: These findings suggest contextual differences among youth living in urban and rural areas of Florida influencing protective factors associated with youth smoking behavior.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Tracy M Decubellis.
Thesis:
Thesis (M.S.)--University of Florida, 2013.
Local:
Adviser: Fogarty, Kate.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2014-02-28

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0045642:00001


This item is only available as the following downloads:


Full Text

PAGE 1

1 THE INFLUENCE OF CONTEXTUAL RISK AND PROTECTIVE FACTORS ON YOUTH SMOKING IN RURAL AND URBAN FLORIDA ENVIRONMENTS By TRACY DECUBELLIS A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFIL LMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2013

PAGE 2

2 2013 Tracy DeCubellis

PAGE 3

3 To my family thank you for your continual support

PAGE 4

4 ACKNOWLEDGMENTS There are many pe ople I would like to thank for their assistance during the process of writing this thesis. First, I would like to thank my thesis advisor, Dr. Kate Fogarty, for encouraging me to attend the University of Florida to pursue a graduate degree, and for her ins piring guidance during the entire thesis writing process. Her understanding and flexibility made it possible for me to juggle full time professional work and graduate responsibilities at the same time. Next, I would like to thank Dr. Tracy Johns for her gu idance on statistical issues and data analysis which was invaluable I would also like to thank Dr. David Diehl for his guidance, especially in considering environ mental influences that impact youth smoking behaviors. Personally, I would also like to than k my husband, my son, and my mother for their support during this process and my daughter for the inspiration to pursue a graduate degree The encouragement of my family proved to be invaluable in balancing a full time job and graduate school commitments. I would also be remiss if I neglected to thank all of the wonderful faculty who taught me valuable lessons about research, theories, and professional writing. I would specifically like to thank Dr. Mickie Swisher for her help in discussing research design and conceptualizing the problem of youth smoking. I was blessed during this process to have many professional and personal sources of knowledge and encouragement to help me successfully complete this thesis project.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 8 ABSTRACT ................................ ................................ ................................ ..................... 9 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 11 Rationale for Study ................................ ................................ ................................ 11 Understanding the Problem: Youth Smoking Behavior ................................ .... 11 The Tobacco Use Epidemic: How Youth Smoking Contributes ........................ 13 Smoking As Youth Problem Behavior ................................ .............................. 15 Motivations and factors influencing ris k according to problem behavior theory ................................ ................................ ................................ ...... 16 The importance of context for risk and protection in problem behavior t heory ................................ ................................ ................................ ...... 17 Purpose of the Study ................................ ................................ .............................. 18 Research Questions ................................ ................................ ............................... 19 2 REVIEW OF LITERATURE ................................ ................................ .................... 21 The Function of Smoking for Adolescents ................................ .............................. 21 Smoking as a Problem Behavior ................................ ................................ ...... 22 T he Influence of Social Context on the Internal Meaning of Youth Smoking .... 23 Problem Behavior Theory ................................ ................................ ....................... 25 Review of Research support ing PBT and the PBT Ecological Framework ....... 28 The Differential Influences of Risk and Pr otection on Problem Behavior in Youth from Diverse Cultures ................................ ................................ ......... 29 A multi national examination of the differential impact of risk and protection ................................ ................................ ................................ 30 The differential impact of risk and prot ection among youth in China and the United Sta tes ................................ ................................ .................... 32 The differential impact of risk and prot ection among youth of different socio economic backgrounds ................................ ................................ 36 Risk and Protectio n among Rural and Urban Youth ................................ ............... 38 Religiosity as a Protective Factor among Rural Youth ................................ ..... 40 Youth Smoking in Rural Tobacco Growing Areas ................................ ............ 42 Youth Smoking in Different Rural Environments ................................ .............. 43 Summary ................................ ................................ ................................ ................ 47

PAGE 6

6 3 METHODOLOGY ................................ ................................ ................................ ... 49 Data Collection and Research Design ................................ ................................ .... 49 Sampling ................................ ................................ ................................ ................. 53 Procedure ................................ ................................ ................................ ............... 54 Dependent Variable ................................ ................................ .......................... 55 Independent Variables ................................ ................................ ..................... 56 Protective factors ................................ ................................ ....................... 56 Risk factors ................................ ................................ ................................ 58 Data Analysis ................................ ................................ ................................ .......... 59 4 R ESULTS ................................ ................................ ................................ ............... 64 Descriptive Statistics ................................ ................................ ............................... 64 Sample Demographics ................................ ................................ ..................... 64 Varia ble Recoding ................................ ................................ ............................ 65 Cross Tabulations ................................ ................................ ................................ ... 65 Risk Factors ................................ ................................ ................................ ..... 65 Protective Factors ................................ ................................ ............................ 67 Correlations ................................ ................................ ................................ ............ 68 T tests ................................ ................................ ................................ ..................... 69 Regression ................................ ................................ ................................ .............. 71 Risk and Protective Factors ................................ ................................ .............. 71 Risk and Protective Factor Model ................................ ................................ ..... 72 Urban Ris k and Protection Model ................................ ................................ ..... 74 Rural Risk and Protection Model ................................ ................................ ...... 75 5 DISCUSSION ................................ ................................ ................................ ......... 88 Limitations ................................ ................................ ................................ ............... 95 Recommendations for Future Research ................................ ................................ 97 Conclusion ................................ ................................ ................................ .............. 98 APPENDIX LIST OF ORIGINAL VARIABLES AND RESPONSES ............................ 101 LIST OF REFERENCES ................................ ................................ ............................. 104 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 108

PAGE 7

7 L IST OF TABLES Table page 3 1 Independent variables ................................ ................................ ........................ 63 4 1 Variable one way frequencies ................................ ................................ ............ 78 4 2 Summary statistics for demographic information urban ................................ .... 79 4 3 Summary statistics for demographic information rural ................................ ...... 80 4 4 Summary statistics for demographic information statewide .............................. 80 4 5 Contingency table smoker by peer rewards for problem behavio r ................... 81 4 6 Contingency table smoker by peer models of drug use ................................ ... 81 4 7 Contingency Table smoker by favorable parental attitude toward ATOD ........ 82 4 8 Contingency table smoker by laws favorable to drug use ................................ 82 4 9 Contingency table smoker by norms favorable to drug use ............................. 82 4 10 Contingency table smoker by frequency of religious service attendance ........ 82 4 11 Contingency table smoker by intolerance of deviance ................................ ..... 82 4 12 Contingency table smoker by family involvement ................................ ............ 83 4 13 Contingency table smoker by school involvement ................................ ........... 83 4 14 Contingency table smoker by parental monitoring ................................ ........... 83 4 15 Contingency table smoker by rural ................................ ................................ .. 83 4 16 Summary of intercorrelations for variables ................................ ......................... 84 4 17 Logistic regression predicting likelihood of adolescent smoking statewide ...... 85 4 18 Logistic regression predicting likelihood of adolescent smoking urban counties ................................ ................................ ................................ .............. 86 4 19 Logistic regression predicting likeli hood of adolescent smoking rural counties ................................ ................................ ................................ .............. 87

PAGE 8

8 LIST OF FIGURES Figure page 3 1 Florida rural and urb an places ................................ ................................ ............ 63

PAGE 9

9 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for th e Degree Master of Science THE INFLUENCE OF CONTEXTUAL RISK AND PROTECTIVE FACTORS ON YOUTH SMOKING IN RURAL AND URBAN FLORIDA ENVIRONMENTS By Tracy DeCubellis August 2013 Chair: Kate Fogarty Major: Family, Youth, and Community Sciences Introduction: In the State of Florida, risk and protective factors associated with cigarette use have been examined statewid e; however, no information currently exists analyzing the differential influence of risk and protective factors on youth smoking among students living in urban as compared to rural populations. Purpose: The purpose of this study is to examine the contextu al influence of urban as compared to rural environments on adolescent smoking behavior. This study examines the ways in which risk and protective factors interact to influence youth smoking behavior, and the ways in which these interactions may differ base d on the environment in which the youth reside. Methods: This study analyzed statewide data from the 2012 Florida Youth Substance Abuse Survey (n = 33,748 high school students; 44% urban, and 56% rural). Predictor variables were chosen based on the Problem Behavior Theory and were organized into risk and protective factor categories. Results: The results indicated that for urban and rural youth, the risk factor positively associated with smoking was favorable parental attitudes toward alcohol, tobacco, and other drug use. Protective factors negatively associated with smoking

PAGE 10

10 were personal intolerance of deviance, and religious service attendance. Results of rural samples f or all protective factors, and the risk factor laws favorable to drug use. Conclusion: These findings suggest contextual differences among youth living in urban and rural areas of Florida influencing protective factors associated with youth smoking behavio r.

PAGE 11

11 CHAPTER 1 INTRODUCTION Rationale f or Study Understanding the Problem: Youth Smoking Behavior Tobacco related illness is a major public health problem in the United States and is the number one cause of preventable death. The yearly death toll from tob acco related illness is estimated to be 443,000 from diseases such as Chronic Obstructive Pulmonary Disease (COPD), heart disease, stroke, cancers, and other chronic illnesses (Centers for Disease Control and Prevention, 2011). Tobacco smoke contains the a ddictive substance, nicotine, as well as over 50 cancer causing agents which are found in secondhand smoke and sidestream smoke (U.S. Department of Health and Human Services, 2006). In addition to the high number of people who die prematurely due to tobacc o use, 8,600,000 U.S. adults suffer from chronic disease as a result of smoking cigarettes (Family Smoking Prevention and Tobacco Control Act, 2009). This places a massive burden on the health care system, and creates high economic outlays for local, state and national governments as well as incurs private costs to individuals, families, and employers. The Centers for Disease Control and Prevention (CDC) estimates that tobacco smoking is responsible for the loss of $98.6 billion in productivity in the Unit ed States each year (Centers for Diseas e Control and Prevention, 2008). points to a need for equal concern for those who are invo luntarily affected by tobacco. For example, cigar ettes are responsible for the deaths of 1,600 people each year due to house fires (Centers for Disease Control and Prevention, 2008). Many of the victims of

PAGE 12

12 house fires are the children or relatives of smokers who do not smoke themselves. Tragically, mater nal smoking is responsible for the deaths of over 2,500 infants annually due to low birth weight or a short gestational period. Maternal smoking or secondhand smoke exposure causes the deaths of over 1,300 infants through Sudden Infant Death Syndrome (SIDS ) (Centers for Disease Control and Prevention, 2008). In 2006, the United States Surgeon General published research results explaining the dangers of secondhand smoke to children, youth, and adults who are not tobacco users, but are exposed to the passive smoke of other people (U.S. Department of Health and Human part, that secondhand smoke can cause premature death in both children and adults even if they are not current smokers In 2006, the United States Surgeon General published research results explaining the dangers of secondhand smoke to children, youth, and adults who are not tobacco users, but are exposed to the passive smoke of other people (U.S. Department of Health and Human Services, 2006). Thes e reports underscore the enormous personal toll that smoking takes in the United States through tobacco related deaths each year. Moreover beyond the long term consequences of involuntary smoking, non smokers can suffer instant negative consequences when exposed to secondhand smoke. The Surgeon General reports that there is an immediate, negative impact on the cardiovascular system when som eone is exposed to passive smoke (U.S. Department of Health and Human Services, 2006). In the long term non smokers who are exposed to passive smoke, have a 25% 30% higher risk of developing some form of heart

PAGE 13

13 disease, and have an increased risk of death d ue to cardiac disease (U.S. Department of Health and Human Services, 2006). Additionally, involuntary smokers who live with a smoker have a 20% 30% increased risk of developing lung cancer due to daily smoke exposure (U.S. Department of H ealth and Human Se rvices, 2006) Children seem particularly at risk for premature death and chronic illness when exposed to secondhand smoke. Moreover, children often lack the ability to speak for themselves, or to remove themselves from environments containing secondhand s moke which truly makes them involuntary smokers. Maternal smoking during pregnancy can cause a child to have poor lung development and infant exposure to secondhand smoke after birth can lower lung functioning (U.S. Department of Health and Human Services, 2006). Passive smoke may also contribute to the development or exacerbation of asthma, ear infections, and other respiratory ailments in children (U.S. Department of Healt h and Human Services, 2006). The Tobacco Use Epidemic: How Youth S mokin g C ontributes Up to this point, smoking among adults as a concern has been discussed, as well smoking is a legal activity for adults, the majority of adult smokers likely picked up their first cigarette during adolescence. In fact, approximately 90% of adult smokers started using tobacco prior to the age of 18 (Centers for Disease Control and Prevention, 2012). Examining youth smoking rates, between the years of 1997 2003, cigarette smokin g among high school students decreased from 36.4% to 21.9%, and declined slightly to reach 19.5% in 2009 (Centers for Disease Control and Prevention, 2010). This translates into decades of exposure to the carcinogens found in tobacco smoke which is

PAGE 14

14 associat ed with a large proportion of tobacco related chronic illnesses and deaths in the United States each year. For many tobacco users, nicotine is a lifelong addiction. Smoking among youth is risky for several reasons. First, it is believed that adolescents a re more susceptible to nicotine addiction than adults, and maintain stronger addiction to nicotine than those who start smoking during adulthood (U.S. Department of Health and Human Services, 2012). Researchers estimated that over 6 million youth aged 0 17 in the year 2000, have the potential to die from smoking related illness (Hahn, Rayens, Chaloupka, Okoli, & Yang, 2002). As a result, adolescent smoking prevention programs have focused on tobacco use as a health related behavior and educating youth about the potential negative health consequences of smoking tobacco. More recently, tobacco control programs have focused on issues such as advertising aimed at youth, products designed to appeal to youth, and combating other tobacco industry tactics that are d esigned to attract adolescents to tobacco products. With regard to policy, in 2009, the Family Smoking Prevention and Tobacco Control Act (Tobacco Control Act) gave the Food and Drug Administration the jurisdiction to regulate tobacco, and also made tobac co products that appeal directly to youth, such as candy flavored cigarettes, illegal to sell in the United States (Family Smoking Prevention and Tobacco Control Act, 2009). This law has helped focus new attention on tobacco industry practices that influen ce social norms and beliefs about words carry the connotation of harm reduction. The Tobac co Control Act also created

PAGE 15

15 added taxes per pack of cigarettes in an effort to reduce youth smoking by creating a price barrier for smoking initiation (Family Smoking Prevention and Tobacco Control Act, 2009). The new law plays an important part in youth t obacco use prevention by creating regulatory oversight of tobacco products including tobacco packaging and marketing as tools which influence youth to initiate tobacco use (Family Smoking Prevention and Tobacco Control Act, 2009). While laws and regulation s assist tobacco prevention programs by encouraging tobacco free social norms and removing environmental barriers such as products that appeal to youth, they do not address the internal motivation for youth to begin smoking cigarettes. Smoking As Youth Pro blem Behavior Tobacco control and prevention programming has historically focused on adolescent tobacco use as a health related behavior and worked under the idea that one reason youth initiate smoking is to imitate what they believe is normal, adult behav ior (i.e., within social norms). Current CDC Best Practice guidelines direct tobacco control programming to work toward policy change and the shifting of social norms to make tobacco use less socially acceptable (Center for Disease Control and Prevention, 2007). However, research has revealed that many adolescents do not view tobacco use as a health related behavior; rather, smoking manifests more as an adolescent problem behavior (Turbin, Jessor & Costa, 2000). Adolescent problem behavior according to Jess Theory refers to behaviors that pose psychosocial and physical risks to healthy adolescent development which are influenced by risk factors and can be moderated by protective factors. According to the theory, problem behaviors should be viewed as a group or a lifestyle for the adolescent which are used to reach a goal or to facilitate a

PAGE 16

16 specific function, such as attaining mature status, gaining peer acceptance, or going against conventional social norms (Jessor, 1991). Problem behaviors can be predicted and studied in light of the risk and protective factors that serve as antecedents (Jessor, 1991). In recent years, the application of the problem behavior theory has become more ecological in that the social context in which the adolescent lives and is believed to interact with the domains of family, school, neighborhood and the larger political, economic and cultural environment (Jessor, 1993, pp. 122). Motivations and factors influencing risk according to problem behavior theory According to problem behavior theory, problem behaviors are believed to serve one of the following purposes or goals of an adolescent at the individual motivational level such as showing independence, maturity, or bucking against the norms of society (Jes sor, 1991). The infl uence of what adults such as parents, teachers, or neighbors think about youth engagement in problem behaviors can be considered a risk factor if those influential adults support the behaviors. Having friends who approve of, or engage i n, problem behaviors is another example of social norm based contextual risk factors within PBT. Social norm based risk factors found to be positively associated with youth cigarette use are: peer or parental models of problem behavior; laws and norms favo rable to cigarette smoking; and the perceived availability of cigarettes (Jessor et al., 2003). Protective factors are supports within a person, system, or environment that help to buffer the life compromising effects of risk factors (Jessor, 1991). Exampl es of protective factors at the person, system and environment level, respectively, that may be associated with youth smoking are personal intolerance for deviance, parental controls or restrictions, and school involvement (Jessor et al., 2003).

PAGE 17

17 The impor tance of context for risk and protection in problem behavior theory A question the proposed research aims to answer is whether the prevalence of contextual risk and protective factors are consistent in different cultures and environments. A reformulation o f PBT included the importance of social context on problem behaviors, so that the environment in which adolescents are surrounded can contribute to risk or protective factors as well (Jessor et al., 2003). Therefore, the individual as well as the social an d environmental context are important to consider when assessing risk and protective factors that may serve as behavioral antecedents or moderators for risk (Jessor et al., 2003). One environment that has received little attention is that of rural populati ons as compared to more urban populations. Currently, around 21% of the U.S. population lives in a rural environment which is defined as areas (i.e., towns or cities) with populations of less than 50,000 people (Curtis, Waters, & Brindis, 2010). Little inf ormation about risk and protective factors among youth who live in rural populations is available at this time (Curtis, Waters, & Brindis, 2010). Data collected in the State of Florida indicated a higher prevalence of smoking among adults living in non met ropolitan areas (23.4%) as compared to adults living in metropolitan areas (16.8%) of the state (Florida Department of Health, 2009). The Florida Bureau of Tobacco Prevention Program has included adults living in rural areas as one of the priority populati ons to be reached by programs to reduce tobacco related illness (Florida Department of Health, 2009). Since the vast majority of adult smokers start during adolescence, and the prevalence of adult smoking is higher among rural Florida adults than that of a dults in urban areas, it is suspected that a greater proportion of rural youth will report smoking in the 2012 Florida Youth Substance Abuse Survey. Currently, reports on the FYSAS data do not make state or county level comparisons between

PAGE 18

18 urban and rural youth on substance use. Discovering whether there is a similar disparity what risk and protective factors could be associated with youth smoking, may point to useful informa tion aiding in smoking prevention and cessation or intervention programs for youth in rural areas. Cultural differences between rural and urban youth most certainly go beyond music and fashion trends. Youth in rural areas may live in tobacco growing regio ns as some rural counties in Florida are tobacco producing counties and employ youth and adults to work on tobacco farms. Youth in rural areas may perceive tobacco use as more socially acce ptable among youth and adults. Since the smoking prevalence among r ural Florida adults is higher than urban Florida adults, it is also possible that rural youth perceive tobacco use as a rite of passage to move into maturity or adulthood (Florida Department of Health, 2009). These possibilities require further investigati on in order to move from speculation toward understanding the ways in which contextual factors may influence adolescent smoking behavior in rural areas. Purpose of the Study The purpose of this study is to examine differential influence of contextual risk and protective factors associated with cigarette smoking among adolescents in rural, as compared to urban, environments. Youth tobacco smoking will be classified in this study incorporates cont extual interacting risk and protective factors as antecedents to adolescent problem behavior. This study will s pecifically investigate whether : (1) protective and risk factor antecedents are associated with adolescent cig arette smoking, (2) protective factors moderat e the influence of risk factors, and (3) contextual

PAGE 19

19 residential differences are associated with the effects and prevalence of risk and protective factors. In other words, the study will explore the relation among these influen ces and how such influences may differ across rural, as compared to urban, settings in which youth reside. The results of this study will help explain the interaction of antecedent contextual risk and protective factors on adolescent smoking and the ways i n which risk and protective factors may differ depending on the local context in which adolescents live. Research Questions Considering current tobacco control programs tend to be geared toward changing social norms about tobacco use, and research supporti ng that tobacco use is perceived by teens as a deviant rather than health related behavior, it is important to utilize a problem behavior theoretical framework to examine the ways in which risk and protective factors influence youth smoking. Furthermore, i t is important to understand any possible interaction between the contextual environment (e.g., rural compared to urban residence) and risk and protective factors that may influence youth smoking behavior as well. This study seeks to understand the ways in which the social context of the rural environment might interact with risk and protective factors to influence adolescent problem behavior. The following research questions will be addressed in the proposed study: 1. Is cigarette smoking among Florida youth associated with other problem behaviors such as abuse of other substances and delinquent behavior? (test of risk behavior syndrome of problem behavior theory) 2. Are selected social norm based influences associated (positively associated = risk factor, negati vely associated = protective factor) with cigarette smoking

PAGE 20

20 3. Is there a difference in the rate of cigarette smoking among youth who live in rural, as compared with urban, locations in Florida? 4. Does the extent of social norm based risk and protective factors differ among youth who live in rural, as compared with urban, locations in Florida? Provided support is found for the first four research questions, the inquiry below will follow 5. Do social norm based protective factors (negativel y associated with smoking) moderate the effects of risk factors (positively associated) on smoking behavior? 6. What degree of influence do social norm based risk and protective factors have on youth cigarette smoking in rural, as compared with urban, locatio ns in Florida?

PAGE 21

21 CHAPTER 2 REVIEW OF LITERATURE The Function of Smoking for Adolescents Adolescent smoking behavior poses a health risk for adult life and can be classified as a problem behavior. Traditionally, prevention efforts from government agencies a nd schools have focused on educating youth about the harmful health effects of tobacco use. Policy efforts to decrease smoking make environmental changes such as recommendations for creating community based policies including smoke free school grounds and restricting public smoking areas to reduce youth exposure to adult smoking behaviors. Youth smoking behavior is found to be a function of social norms and reference groups. Social norms refer to expectations for behaviors within social reference groups. If a youth regards smoking tobacco as a behavior characteristic of a social reference group to which he or she belongs, then the youth is more likely to cial norms regarding smoking. Another related explanation for an act representing adult status attainment. Tobacco companies agree with youth smoking prevention programs that social norms influence smoking behavior and that the perception of attainment of adult status motivates adolescents to smoke. According to documents released to the United States Congress by Philip Morris dating from 1969, Philip Morris executives st ated: Smoking a cigarette for the beginner is a symbolic act. The smoker is telling of smoking remains a symbolic declaration of personal identity (Philip Morris, 1969).

PAGE 22

22 The consideration of the function that smoking serves for the adolescent is important in understanding the phen omenon of youth cigarette use, as well as provides significant implications for intervention and prevention efforts. As a result, much research has been done on the topic of youth smoking to answer this question. An additional explanation which stands in o pposition to the influence of social norms and motivation to attain adult status is youth smoking to take on a deviant identity. Youth out by engaging in a behavior speci fic to youth culture and contrary to social norms and expectations. In summary among individual motivations to smoke it is possible that some adolescents view smoking in terms of social norms and group belongingness, or as a means to attain adult status, o r as a way to rebel against the conventions of society. For the purpose of this study, youth smoking will be examined in light of antecedent contextual (individual, environmental) risk and protective factors Smoking as a Problem Behavior There is evidence that youth smoking is more related to other youth problem behaviors such as risky sexual behavior, delinquency, and other drug use than to health related behaviors such as healthy eating and hygiene (Turbin, Jessor, & Costa, 2000). Researchers investigate d whether a sample (n=1782) of high school students more closely associated cigarette smoking with problem behaviors such as sexual intercourse, alcohol abuse, drug use, and delinquency, or with health related behaviors such as dietary and exercise habits, dental hygiene, and seat belt usage (Turbin et al., 2000). Through using correlation and structural equation modeling analytical methods, results indicated that adolescent cigarette smoking strongly relates to other youth problem behaviors and is only sli ghtly related to risky health compromising behaviors

PAGE 23

23 (Turbin et al., 2000). According to Problem Behavior Theory (PBT) youth problem behaviors tend to occur concomitantly, with associations found between, for example, delinquency, early sexual activity ons et, and unlawful substance use. Hence, cigarette smoking will be classified herein as a problem behavior that operates according to the axioms and principles of PBT. The Influence of Social Context on the Internal Meaning of Youth Smoking With smoking reg model problem behavior is preceded by antecedents (e.g., risk and protective factors) and followed by consequences (life compromising outcomes). Also according to Jessor (1993) individual factor s, the social context, and change over time are integral components of the PBT model (pp. 120). At the individual level behaviors carry internal meaning for adolescents or help them achieve personal or social goals (Jessor, 1991). Social norms manifest at the level of social context and are likely to change over time. Moreover, the personal and social function of a problem behavior may also change over time in response to new social norms (Jessor, 1991). Because youth tobacco use is strongly related to pro blem behaviors, youth perception regarding the meaning of their tobacco use may be an important factor and provide insight into social norms. One study investigated whether the social meaning of smoking for youth had changed over the course of time. The st udy sought to discover any difference between adolescent smokers in the year 1980 (n=3495) as compared to adolescent smokers in the year 2001 (n=3166) (Chassin, Presson, Morgan Lopez, & Sherman, 2007). The researchers wanted to know if there were differenc es between the groups of adolescent smokers because the rate of cigarette experimentation, as evidenced by 12th graders who had ever tried a cigarette, fell from 76% in 1977 to 50%

PAGE 24

24 by 1996. They concluded that the meaning behind adolescent smoking may have changed over time as smoking related social norms have changed, health promotion programs have increased, and more people have negative attitudes and beliefs toward smoking in 2001 as compared to 1980 (Chassin et al., 2007). Chassin and colleagues (2007 comparing generations of adolescents and finding youth smokers in 2001 perceived their smoking behavior as more deviant than youth in 1980 when smoking was more socially acceptable. They speculated that youth s mokers in the year 2001 were actually more deviance prone than youth smokers in the year 1980 due to a different meaning assigned to their smoking behavior. Middle school students in 2001 were more likely to view youth smoking as non conventional behavior instead of as a way to appear more mature or practice adult behavior, as middle schoolers in the 1980s may have been likely t o think (Chassin et al., 2007). However, among high school students compared between 1980 and 2001, the findings were different. Th e prevalence of smoking among middle school students decreased at greater levels between the years 1980 and 2001 than did the prevalence of smoking among high school students (Chassin et al., 2007). Chassin and colleagues (2007) speculated that smoking beh avior among middle school students may indicate a is possible that in 2001, middle schoolers who were regular smokers, as compared to middle schoolers who smoked in 1980, were more likely to see their smoking behavior as non conventional, rather than behavior modeled after what they believ ed was normal, adult behavior.

PAGE 25

25 that youth who live in rural areas are surrounded by a different social context which may be embedded with different social norms regarding smoking cigarettes. Youth in urban areas are likely to be exposed to anti tobacco social norms due to higher population areas with more to bacco free zones such as business and restaurants whereas youth in rural areas live in less densely populated environments with fewer opportunities to observe tobacco free zones simply because of fewer businesses and restaurants in their area Also, youth in urban areas may tend to adopt trends, e.g., anti smoking social norms facilitated by tobacco free zones, sooner than youth in rural areas Just as the study demonstrated that younger adolescents were beginning to change their outlook on tobacco use, it i s possible that youth in urban areas are receiving either less peer pressure to smoke, or are surrounded by more peers who are modeling tobacco free lifestyles as compared to youth in rural areas. This is speculation based on the information presented in t his study, and the data to be analyzed is cross sectional rather than longitudinal and thus capable of measuring change over time. However, social norms in the form of peer, family, and community attitudes toward youth tobacco use, will be examined in this study to answer the research questions. Problem Behavior Theory The Problem Behavior Theory (PBT), developed by Richard Jessor, has been used to explain a variety of youth risk behaviors such as tobacco use, delinquency, illicit drug use, and precocious s exual activity in adolescents Problem behavior includes behaviors which have negative developmental impacts on youth in the short term and long term, as well as produce physical and socioemotional consequences. The theory provides a framework for understan ding youth risk behavior by examining different

PAGE 26

26 domains and the ways in which risk and protective factors interact and influence youth problem b ehaviors within those domains. Jessor (1991) believes that adolescents see significant meaning or function in th eir behaviors which have social and personal consequences for the actor (Jessor, 1991). For adolescents, problem behaviors such as smoking cigarettes or drinking alcohol may serve a social function to impress or gain friends, assert their own maturity and independence, or to defy social norms (Jessor, 1991). Problem behavior is viewed in terms of patterns of behavior, or lifestyle, so ewed as isolated problem behaviors (Jessor, 1991). Youth who engage in risk behavior can essentially disrupt healthy growth and development in the physical, social, emotional, and educational domains of development (Jessor, 1991). Jessor (1991, 1992) descr ibes the theoretical model of PBT as an interconnected web of social, environmental, and individual factors within a social context that influences risk behavior and are interrelated among each other. A variety of risk and protective factors fall within ea ch of t he five domains in the model: (1) biology, ( 2 ) social environment, (3) perceived environment, (4) personality and (5) behavior (Jessor, 1991; Jessor, 1992). Each of these domains contain risk factors which are known to increase the chances of poor life outcomes, and protective factors which moderate the impact of risk on adolescents and can potentially influence positive development (Jessor, 1991). Each of these domains contain risk factors which are known to increase the chances of poor life outcom es, and protective factors which moderate the impact of risk on adolescents (Jessor, 1991). It is crucial to note that high levels of risk can be

PAGE 27

27 buffered by high levels of protection, so that youth exposed to many risk factors can still experience positiv e developmental outcomes as a result of the moderating effect of protection on risk (Jessor, 1991). Protective factors can be strengthened within the personal, social, and institutional realms to reduce the impact of risk, thereby creating the likelihood o f better adolescent outcomes (Jessor, 1993). The Problem Behavior Theory has been more recently expanded to create a more ecological perspective. In this perspective the combined influence of risk and protective factors on risk behavior from each domain i s placed in an ecological framework to incorporate individual and contextual influences in the lives of adolescents (Jessor et al., 2003). This ecological framework for PBT takes into account the variety of individual and social influences in the lives of adolescents which contain embedded risk and protective factors. The ecological social contexts included in this expanded framework for PBT are: family; peers; school; and neighb orhood (Jessor et al., 2003). Protective factor categories in the PBT ecologica l framework include (1) models protection, which provides parents and peers as mode ls for conventional behavior; (2) controls protection, which includes both individual attitudes toward deviance and social influences such as parental monitoring, religious involvemen t, and school attachment; and (3) support protection, which refers to factors such as teacher involvement and support, closeness with parents, and neighborhood involvement (Jessor et al., 2003; Vazsonyi et al., 2010). Risk factor categories conta in similar constructs including (1) models risk, which refers to parents or peers as models for deviance such as substance use, or t he approval of substance use; (2) opportunity risk, which refers to the availability of cigarettes, alcohol, and other subst ances as well as access to deviant

PAGE 28

28 groups such as gangs; and ( 3 ) vulnerability risk, which largely represents individual factors such as stress, low self esteem, or depression (Jessor et al., 2003; Vazsonyi, et al., 2010). This updated model also shows the ways that risk and protective factors directly influence youth problem behaviors such as cigarette smoking and drinking alcohol as well as the moderating impact of protective factors on measures of risk (Jessor et al., 2003). Review of Research supporting PBT and the PBT Ecological Framework In the earlier work of Jessor using his PBT with risk and protective factors and the 5 domains, a longitudinal study was conducted to discover the linkage between youth problem behaviors and risk and protective factors that may impact behavior, as well as examine the possible buffers that protective factors provide. The study included a sample of urban middle schools students (n=1486) who were in grades 7 9 during the first wave of a four wave study (Jessor, Van Den Bos Vanderryn, Costa, & Turbin, 1995). The measures of protection used in the research examined the personality system which included positive orientation toward school, positive orientation toward health, and attitudinal intolerance of deviance; the environ ment system which included positive relationships with adults, perceived controls (such as family rules), and friends as models for conventional behavior; and the behavioral system which included prosocial family, community, and school activities (Jessor e t al., 1995, p. 925 926). Risk factors contained within in the personality system included low expectations for success, self esteem, and hopelessness; within the perceived environment system were friends as models for problem behavior and orientation to f riends as opposed to parents; and the measure of risk in the behavior system was grade point average (Jessor et al., 1995, p. 926).

PAGE 29

29 The results of this study indicated that as protective factors increased, problem behavior involvement decreased; and as ris k factors increased, problem behavior involvement increased (Jessor et al., 1995). Additionally, when testing for interaction between risk and protective factors, even after controlling for demographics, the data analysis showed that protective factors act ed as buffers between risk factors and youth problem behavior (Jessor et al., 1995, p. 927). The moderating effect of protective factors between risk and problem behavior were applicable across time, regardless of the gender or racial background of the par ticipants (Jessor et al., 1995). Interestingly, the most influential risk factor was friends as models for problem behavior, and the most influential protective factor was attitudinal intolerance for deviance (Jessor et al., 1995). This study adds to the u nderstanding of the way risk and protective factors interact with each other and adolescent problem behaviors. From a prevention programming perspective, the results show the importance of considering protective factors as moderators instead of focusing on ly on risk reduction techniques. The researchers indicate that having protective factors in place prior to youth involvement in problem behaviors may moderate risk factors that increase during adolescent development (Jessor et al., 1995). The Differential Influences of Risk and Protectio n on Problem Behavior in Youth f rom Diverse Cultures Examining risk and protective factors in terms of the social context in which youth are embedded is important considering the amount of time that youth spend in school set tings, with peers, in neighborhood contexts and with family members (Jessor et al., 2003). The following studies reviewed examine the application of PBT among youth of different countries, cultures, religions, socioeconomic levels, and family structures to

PAGE 30

30 cultures, as well as pointing to the ways in which social context may influence the prevalence of risk and protective factors. A multi national examination of t he different ial impact of risk and protection A study of youth, aged 14 22, in the countries of Hungary, the Netherlands, Slovenia, Spain, Switzerland, Taiwan, Turkey and the United States (n=10,310) examined how risk and protective factors are linked to problem behav ior as well as how the linkages may vary across cultures (Vazsonyi et al., 2010). The youth in this study came from different socioeconomic backgrounds, legal and governmental systems, social structures, religious backgrounds, and family struct ures (Vazson yi et al., 2010). Syndrome, measured using a Normative Deviance Scale which did not include substance abuse related items, rather items relating to vandalism, theft and assault, deviance, and school misbehavior issues. Predict or risk factors tested included (1) models of risk, which was assessed by measures that determined levels of peer deviancy an d families as models of risk; (2) opportunity risk, which was assessed by determin ing the availability of alcohol and invo lvement in gang activity; and (3) vulnerability risk, which tested levels of depression or expectations for low academic outcomes (Vazsonyi et al., 2010). Protective factors measured in the study included controls pr otection, which assess areas of parental supervision, religiosity, and level of school involvement; and support protection, which examined closeness of parental relationship with the youth and neighborhood attachment (Vazsonyi et al., 2010). Researchers f ound that many similarities exist between the way risk and protective factors are related to youth problem behaviors across countries with only a

PAGE 31

31 few observed differences. One difference occurred with the lack of a significant relationship found between su pport protection and youth problem behavior in Hungary, Spain, and Turkey (Vazsonyi et al., 2010). Additionally, there was a not a significant relationship between vulnerability risk and youth problem behavior among youth from the Netherlands, Hungary, and Switzerland (Vazsonyi et al., 2010). Additionally, there was a not a significant relationship between vulnerability risk and youth problem behavior among youth from the Netherlands, Hungary, and Switzerland (Vazsonyi et al., 2010). Across all counties; ho wever, controls protection, models risk, and opportunity risk did significantly relate to the outcome variable of youth problem behavior (Vazsonyi et al., 2010). The authors suggest that in light of the differences in some risk and protective factors acros s nations, further research should be conducted across different cultures to discover whether there are contextual factors that interact with the risk and protective factors and youth problem behaviors (Vazsonyi et al., 2010, pp. 1785). Vazsonyi and collea the cross national utility of the ecological model of Problem Behavior Theory by expanding its application to eight countries with diverse social, political, religious, governmental, and legal cont exts. The results indicate that risk and protective factors do indeed relate to youth problem behaviors across cultures and can be used to explain youth problem behaviors in a variety of settings. Moreover, the differential effects of risk and protective f actors on youth outcomes across countries pointed to the specific needs of youth in each country and where to intervene to make the most impact on reducing adolescent risk behavior.

PAGE 32

32 T he differential impact of risk and protection among youth in China and th e United States Another study was conducted among youth in grades 7 9 in China (n=1739) and the United States (n=1596) to examine the way risk and protective factors within individual and social contexts are linked to youth problem behavior in non Western countries as compared to the United States (Jessor et al., 2003). Researchers wanted to assess universal characteristics versus more local and unique characteristics of youth development and the way it impacts adolescent behavior (Jessor et al., 2003). In this study, problem behavior included youth participation in delinquent activity, tobacco, alcohol and marijuana use, other drug use, and precocious sexual involvement (Jessor et al., 2003). The model for this study sought to link risk and protective facto rs to youth problem behaviors, as well as test for moderating effects of protective factors on risk factors (Jessor et al., 2003). Protective factors for this study included models protection, which consisted of friends and parents as models of conventiona l behav ior; controls protection, which included intolerance of deviance as well as family, school, and neighborhood disapproval of deviance; and support protection which included support from the family, peers, school, and other adults in the neighborhood (Jessor et al., 2003). The risk factors included models risk, which consisted of family, friends, schools and neighbors as models of risky behavior; opportunity risk, which refers to the availability gang activity, drugs, alcohol and tobacco; and vulnerabi lity risks which examines internal factors such as stress, self esteem, and future expectations of success (Jessor et al., 2003). The explanatory model for the study showed the direct effect of protective and risk factors on youth problem behaviors as well as an impact of protective factors as

PAGE 33

33 moderators of risk factors (Jessor et al., 2003). The researchers were also interested in noting any differences in the degree or prevalence of risk and protective factors based on the differences inherent between Chi nese and American societies such as variances in the structure of the family, government, schools, economic systems and levels of religiosity (Jessor et al., 2003). For both samples the results indicated that controls protection was the strongest measure f ollowed by models risk and vulnerability risk (Jessor et al., 2003). With regard to the moderation effect of protective factors on risk factors within the two samples, controls protection was a substantial buffer for risk factors for each. In the U.S. samp le, controls protection served as a moderator for all risk factors, and for the Chinese sample it moderated both models and vulnerability risk factors (Jessor et al., 2003). In both samples, the strongest controls protection variables were intolerance for deviance and teacher support, and the strongest models of risk were peers as models of risk and school models of risk behaviors (Jessor et al., 2003). Examining the results with regard to the outcome variable of youth smoking, the study model accounted fo r 27% of the variance among the U.S. sample and 23% of the variance among the Chinese sample (Jessor et al., 2003). This study added to the body of knowledge regarding the problem of youth smoking by identifying the protective factor, intolerance of devian ce which moderated risk. It also demonstrated that similar risk and protective factors can affect youth problem behavior among adolescents from different national and cultural backgrounds. The researchers suggest that as a result of the study it may be use ful for professionals to work toward enhancing protective factors that buffer risk factors or directly influence

PAGE 34

34 problem behaviors as a part of prevention practice as opposed to using a risk based a pproach (Jessor et al., 2003). The study of risk and prote ction among youth in China and the United States was followed up by another study which intended to specifically examine the ways in which the social contexts of family, neighborhood, peers, and school can impact individual factors or other social context factors (Costa et al., 2005, pp. 80). Youth in grades 7 9 participated in the study (Costa et al., 2005). Each social context was assessed using a model of risk and protection with each context to determine if the protections existing in each context can moderate for individual levels of risk related to youth problem behavior (Costa et al., 2005, pp. 70). Furthermore, researchers wanted to know if the social contexts themselves had any impact on youth problem behaviors (Costa et al., 2005). For this study, problem behaviors were defined as cigarette smoking, delinquency, and drinking alcohol in excess (Costa et al., 2005). Protection was me asured in the following ways: (1) models pro tection was measured in the family and peer context; ( 2 ) controls protection was assessed in the family, peers, neighborhood, and school contexts; and ( 3 ) support protection was measured in the family, peers, neighborhood, and school contexts as well (Cost a et al., 2005). Risks were measured slightly differently, and in the following ways : (1) models risk was measured in the family, peers, neigh borhood, and school contexts; (2) opportunity risk was measured in the family a nd neighborhood contexts; and (3) V ulnerability risk was measured in the family, peers, and school contexts (Costa et al., 2005). Fo r the individual level context, protection and risk was measured using controls and vulnerability protection and risk

PAGE 35

35 factors because other levels of risk and protection could not be applied to this context (Costa et al., 2005, pp. 73). The results of the study indicate that protective factors within each social context had a small, significant impact on the level of individual risk for both youth from the Unite d States and China with explaining between 1% and 5% of the variance (Costa et al., 2005, pp. 81). For both countries, controls protection was the strongest protective factor among all social contexts, with the exception of the neighborhood context for you th in the U.S., and this type of protection moderated individual level risk (Costa et al., 2005). For the other categories of protection, support protection was not found to be as influential as controls protection, and models protection was not significan t as a moderator against risk for youth in either country (Costa et al., 2005, pp. 81). Differences were found between the U.S. and China samples of youth: family and peer social contexts were more influential than other social contexts in the U.S.; and fo r Chinese youth, peer and school social contexts were more influential than family and neighborhood social contexts (Costa et al., 2005). The researchers speculate that the importance of the school social context for youth in China could be due to the diff erence in overall social structure, and the important influence that Chinese schools play in the lives of youth, which is more encompassing across domains of youth development than school influence in the United States (Costa et al., 2005). Over all, the y outh in China were found to have lower levels of problem behavior as well as higher levels of protection and lower levels of risk than youth in the United States (Costa et al., 2005). The findings also supported a basic tenet of PBT that increased levels of protection moderates risk and reduces problem behavior among youth.

PAGE 36

36 of the PBT in that it was tested with two different groups, and similar results were found across cultural social context, and its unique contributions to levels of risk and protection, is taken into account. Additionally, the researchers noted that some risk and protective factors in differe nt contexts may interact, such as family levels of protection interacting with neighborhood risks (Costa et al., 2005, pp. 80). These findings inform the proposed study in that a contextual perspective on risk and protection may be important in examining t he problem of youth smoking behavior in different environmental settings. Just as some differences in the primary influences of risk and protection emerged between the U.S. and Chinese samples, it is possible that differences in risk and protection and pot ential influence on youth smoking may be found between youth living in rural and urban areas. Additionally, since Chinese youth were influenced more strongly by their educational system than youth in the U.S., it is possible that external environmental fac tors in rural and urban areas may have different levels of influence on the problem behavior of youth smoking. The differential impact of risk and protection among youth of different socio economic backgrounds In addition to testing PBT among youth of diff erent national and cultural backgrounds, researchers have also tested the influence of risk and protective factors on youth problem behaviors among groups that are not tr aditionally considered to be at risk. For example, the majority of studies about the r isk and protective factors linked to youth problem behaviors have been conducted among youth with lower or middle class socio economic status (Racz, McMahan, & Luthar, 2011). Affluent youth have are not

PAGE 37

37 considered to be at high risk for engaging in problem behaviors because it is believed they have more built in protective factors due to their socio economic status, and their families may have greater access to supports that are considered protective factors. The study concentrated on students in grades 9 1 2 (n=1147) who attended a single high school in an affluent area of the Pacific Northwest where the median income of families was more than twice that of the national level (Racz, McMahan, & Luthar, 2011). First, researchers wanted to find out if problem b ehaviors would be related in the same way among affluent youth as they are in youth who have higher levels of risk. To test for this, the outcome variables for the study were substance use, early sexual behavior, low academic achievement, and externalizing behaviors (e.g. fighting) (Racz et al., 2011). Another purpose of the study was to find out if adolescents in affluent areas who engaged in problem behaviors had negative consequences such as negative legal, health, or interpersonal problems (Racz et al., 2011). The results indicated that the proportion of affluent youth who participated in problem behaviors of all types was similar to previous studies conducted among youth living high risk communities (Racz et al., 2011). Youth were placed in the categori es, conventional or non conventional based on a cluster analysis (Racz et al., 2011). Not surprisingly, researchers found higher levels of negative consequences among youth with higher problem behavior participation within the non conventional category (Ra cz et al., 2011). These findings support PBT and extend the understanding of youth problem behavior beyond youth who are normally considered to be at risk, or hig her risk (Racz et al., 2011). In summary, the studies reviewed represent a cross cultural, be they across national or socioeconomic lines, investigation of the ways in which risk and protective

PAGE 38

38 factors interact with youth problem behaviors in diverse cultures and settings around the world. Youth who participated in these studies were embedded in di verse cultural settings, which in some studies, appeared to interact with the influence of risk and protective factors and how risk and protection impacted problem behavior. For example, the cross national study which showed difference in the ways risk and protective factors predicted problem behaviors in various countries could be a result of the interaction of the environment with the family, school, and neighborhood structures surrounding the adolescents (Jessor, 1993; Vazsonyi et al., 2010). Examining t he China and U.S. youth study results may also point to the role social context plays in interacting with risk and protection to influence youth behavior. In this study, besides personal intolerance for deviance, the strongest protective factor was relate d to school, and the strongest risk factors were related to peer and school domains in both countries (Jessor et al., 2003). Considering that youth spend the majority of their time during the week at school, surrounded by teachers and peers, the social con text in which youth are embedded could be an important consideration when examining risk and protective antecedents of youth problem behavior. Risk and Protection a mong Rural and Urban Youth Few research studies have been conducted comparing the prevalence of risk and protective factors associated with youth problem behavior between adolescents living in rural as compared to urban areas; moreover, many of those studies have samples that are small and do not represent the population adequately (Lutfiyya et a l., 2008). It is possible that a rural environment presents a distinct culture from urban locales with emergent differences in the prevalence and influence of risk and protective factors on problem behavior.

PAGE 39

39 In an effort to understand youth cigarette use i n rural areas and to add to the body of knowledge by examining youth problem behavior in rural environments, researchers conducted a study to determine the prevalence of youth smoking among youth who live in non urban environments. The sample (n=68,270) co nsisted of a diverse group of rural U.S. adolescents from the West, South, Midwest, and Northeast in grades 7 12, and was designed to be a representative sample of rural youth throughout the country (Aloise Young, Wayman, & Edwards, 2002). The sample was a lso categorized into remote, rural, and metro designations based on the population and location relative to a metropolitan area. Researchers found that youth who live in remote and rural areas smoke more than youth who live in metropolitan areas when contr olling for the variable of region (Aloise Young et al., 2002). Regionally, youth in the Southern U.S. smoked more than youth from other regions. This information expands the knowledge base regarding adolescent smoking by providing a representative sample o f rural youth that is largely underrepresented in other research literature, and determining the prevalence of smoking in rural environments as well as regional smoking prevalence among youth. However, this study did not provide information about the reas ons behind higher cigarette use in rural areas, or the prevalence of risk and protective factors that may be associated with this problem behavior. Noting that U.S. adults who live in rural locations have higher smoking prevalence than adults who live in m ore urban environments, one group of researchers wanted to learn if living in a rural environment represented a unique risk factor linked to adolescent smoking. The study was conducted using data between the years 1997

PAGE 40

40 2003 from the Youth Risk Behavior Sur veillance System (YRBSS) conducted by the CDC (Lutfiyya et al., 2008). The sample was merged from the data collected over seven years (n=60,296) from youth in grades 9 12, and included designation for youth living in urban, suburban, or rural environments in the United States (Lutfiyya et al., 2008). The results indicated that although regular cigarette smoking declined over a 7 year period, there was a higher prevalence of smoking among youth living in rural environments (37.4%) compared to youth who lived in suburban (33.9%) or urban (29.6%) areas (Lutfiyya et al., 2008). The study also analyzed smokeless tobacco use, which revealed that the prevalence of rural youth who tried smoking also had tried smokeless tobacco products when compared with adolescents from suburban and urban areas (Lutfiyya et al., 2008). Regression analysis indicated that youth were less likely to be daily smokers if they lived in urban or suburban environments as compared to youth who lived in rural environments (Lutfiyya et al., 200 8, p. 5). The researchers suggest that several factors such as availability of tobacco, lower controls protection, and lower models protection may contribute to the urban/rural difference in whether or not youth smoked and how much youth smoked. Although t he authors speculate on the reasons for higher prevalence of tobacco use by youth in rural locations, the study did not examine the reasons behind the differences between rural, suburban, and urban culture and environment that may contribute to these resul ts. Religiosity as a Protective Factor among Rural Youth In previously mentioned research, religiosity has been studied as a protective factor against youth problem behaviors (Jessor et al., 2003; Vazsonyi et al., 2010). One study explored the influence of individual spirituality and participation in organized religion as protective factors from substance abuse among youth in rural areas. A

PAGE 41

41 sample of youth in grades 7 12, and GED students, (n=414) from rural New Mexico participated in the study to determine whether reporting higher levels of spirituality or religious participation would be associated with abstinence from substance use (Hodge, Cardenas, & Montoya, 2001). The participants were 84% Hispanic, 6%, Native American, and 4% non Hispanic white, with 75% identifying themselves as Roman Catholics use (Hodge et al., 2001). Spirituality was measured using the Index of Core Spiritual Experiences (INSPIRIT); however, this survey instrument was found to have a low level of reliability (Hodge et al., 2001). Religiosity was measured by asking participants about the frequency of church attendance. The researchers reported that participation in organized religious activities was related to a higher probability of alcohol abstinence, but was not related to mariju ana use or illicit drug use (Hodge et al., 2001). Spirituality was related to a higher probability of abstinence from marijuana and illicit drugs, but was not related to alcohol use abstinence (Hodge et al., 2001). The researchers noted that alcohol might be a gateway to other substance use, so demonstrating the level of protection that spirituality and religiosity play might be predictive of protection against other types of substance use and problem behaviors (Hodge et al., 2001). While the aim of Hodge between the protection of religiosity and spirituality against alcohol and drug use, a major limitation was the low reliability of the spirituality measurement of protection. Another limitation of the stu dy is the non representative make up of the sample with a disproportionately high percentage of Hispanic youth participating as compared to the national population of rural youth. However, the findings shed new light on the way that

PAGE 42

42 religion, specifically Roman Catholicism, influences rural Hispanic youth as a protective factor against alcohol use. Youth Smoking in Rural Tobacco Growing Areas Some researchers have conducted studies in an effort to learn more about the causes of youth smoking in rural areas. One such study was conducted in a rural, tobacco growing county to determine whether: tobacco use serves as a gateway to alcohol use; alcohol use serves as a gateway to tobacco use; or other independent risk factors are linked to adolescent tobacco and al cohol use. The sample (n=630) was taken from almost all of the high school students in one tobacco growing rural county with a population of about 16,000 people (Ritchey, Reid, & Hasse, 2001). The outcome variables of the study were being a current smoker or drinker as defined by having used either cigarettes or alcohol at least once in the past 30 days (Ritchey et al., 2001). The independent variables were not listed as risk and protective factors; however, the majority of the variables were similar to ris k and protective factors, including social norm based ones, described in previous studies. Predictor variables included: believing that smoking is offensive; both mother and father living at home; friends who drink alcohol, approval of drinking alcohol; pe rceptions of adult alcohol and tobacco use; peer pressure to use alcohol; grade point average; gender; and school (Ritchey et al., 2001). Factors that increased the risk of smoking were peer pressure to drink, having friends who drink, perceptions about adult drinking and smoking behavior, peer pressure, and attitudes accepting of youth drinking (Ritchey et al., 2001). Approval of drinking functioned as a risk factor and had the strongest association with being both a current alcohol drinker and cigarette likelihood of smoking cigarettes and drinking alcohol (Ritchey et al., 2001). Interestingly,

PAGE 43

43 peer pressure to drink alcohol, which functioned as a risk factor, also increased the likelihood of youth to bo th drink alcohol and smoke cigarettes (Ritchey et al., 2001). Certain variables functioned as protective factors and decreased the risk of youth drinking alcohol and smoking cigarettes. Not surprisingly, youth who believe that smoking was offensive were le ss likely to smoke cigarettes or drink alcohol (Ritchey et al., 2001). Additionally, students with a higher grade point average were less likely to smoke cigarettes or drink alcohol (Ritchey et al., 2001). One major finding of this study that expands the base of knowledge on youth tobacco smoking is that the study model tested the reciprocal relationship between the two outcome variables and found that smoking and drinking among youth in this study were influenced by similar risk factors (Ritchey et al., 2001). Although students who drank alcohol were much more likely to smoke cigarettes, and those who smoked cigarettes were much more likely to drink alcohol, one does not cause the other as a gateway for problem behavior (Ritchey et al., 2001). These findi ngs further support using the problem behavior model in a rural environment. A limitation of this study was that only a few questions were used to measure each variable, so it is possible that the self report responses are not as reliable as other measures which include multiple items for each variable. Youth Smoking in Different Rural Environments Another recent study investigated risk and protective factors linked to adolescent drug use among youth living in different types of rural environments versus y outh living in more urban areas. The sample was made up of 18,767 youth in grades 6, 8, 10, and 12 from non metropolitan areas (Rhe w, Hawkins, & Oesterle, 2011). Researchers used the United States Department of Agriculture (USDA) Business and Industry rura l

PAGE 44

44 definition based on a population of less than 50,000, which are also listed by the USDA as non metropolitan, to classify an environment as rural (Rhew et al., 2011). Youth response to a survey question placed them in a category as farm, country, or town dweller. The outcome variable of drug use included lifetime use and 30 day use of alcohol, cigarettes, marijuana, smokeless tobacco, inhalants, and other drugs (Rhew et al., 2011). The predictor variables included risk and protective factors present in th e domains of community, school, family, and peer/individual from which researchers created cumulative risk and protectio n indices (Rhew et al., 2011). The results indicated that farm dwelling youth experienced1.12 times greater levels of risk for drug use th an town dwelling youth (Rhew et al., 2011) Youth living on farms experienced higher levels of risk within the community and peer individual domains at a significantly higher level than youth who lived in town (Rhew et al., 2011). Interestingly, farm dwe lling youth also experienced higher levels of protection in the community area than town dwelling youth, but lower school based protection (Rhew et al., 2011). Youth living in the country had fewer differences regarding risk and protection compared to farm dwelling youth. Youth who lived in the country showed higher levels of community level risk, and lower levels of school based protective factors than town dwelling youth with no other risk or protective factors being significant influencers of drug use (R hew et al. 2011). In a study with a similar design as the present study, thesis research was conducted on the differences between cigarette smoking among youth in grades 6, 8, 10, and 12 living in rural and urban areas in the State of Virginia using Bronf (1979; 1986) ecological model as a theoretical framework (Shettler, 2005). The

PAGE 45

45 research used data from the Virginia Community Youth Survey (VCYS) which contains similar questions as those found in the Florida Youth Substance Abuse Survey (FYSAS ) which is the data collection instrument for this present study, and is based on the Communities That Care Survey which is discussed in detail in Chapter 3 of this study (Shettler, 2005). The variables studied in the thesis research included (1) the depen dent variable which measured youth cigarette smoking as having smoked at least part of one c igarette in the past 30 days; (2) individual factors which included ethnicity, age, gra de level, and rebelliousness; (3) family/parent factors which included parent ing attitudes and pr actices, and family conflict; (4) Peer factors which included friends smokin g behavior and peer pressure; (5) school factors which included academic performance and commitmen t to school and teachers; and ( 6) community factors which incl uded the perceived availability of social activities and the perceived availability of cigarettes (Shettler, 2005). These variables were analyzed to find differences or similarities between youth smoking predictors among adolescents living in rural or urba n areas. The results of this research indicated that similar risk and protective factors predict youth smoking, regardless of whether they live in a rural or urban environment g cigarettes and being thought of as cool for smoking were both significant predictors of smoking behavior for urban and rural youth. However, the degree of variance explained or influence of these two risk factors on smoking behavior was higher for rural youth. Regarding the dependent variable, the study revealed that youth in rural areas smoke more than youth in urban areas, but that youth in both areas are smoking few cigarettes

PAGE 46

46 in general (0 less than one per day) (Shettler, 2005). There was no signific ant difference between youth in rural or urban environments regarding individual factors, which largely measured demographic data, with the exception of rebelliousness measures (Shettler, 2005). The most influential predictor of youth smoking within the ru ral and urban populations was having a best friend who smoked (Shettler, 2005). Additionally, permissive parental attitudes toward smoking were positively associated with youth smoking among youth in rural and urban settings (Shettler, 2005). Some differen ces between risk and protective factors that influence youth smoking in rural and urban environments were also discovered. For example, the school based and community based factors were important for youth who lived in urban settings, but not for youth who lived in rural areas (Shettler, 2005). However, the interaction between risk and protective factors as they influence smoking behavior, was not examined in the thesis. behavi or, will be used in the proposed study, the majority of the variables selected for the proposed study will not be of focus in the present study. The same dependent variable will be tested; however, the measure of youth smoking will be based on youth smokin g at least one cigarette in the past 30 days. Similar peer factors will be included where one behavior is not isolated from other problem behaviors; therefore, questions ab out alcohol and marijuana use will be included in this study (Jessor, 1991) to confirm that smoking can be feasibly classified as a problem behavior. The variables measured outh

PAGE 47

47 smoking, which may be why the results indicated little difference between rural and urban youth smoking behaviors, other than in degree of influence. Although the variables for the present study are discussed in greater detail in Chapter 3, it is wor th noting that variables are being selected for the purpose of determining the ways in which the social structure may interact with factors such as family, school, and neighborhoods to influence adolescent attitudes toward smoking, and adolescent smoking b ehavior (Jessor, 1993). Also, the proposed study is working (1979; 1986) ecological theory of human development. Th erefore, this study yield s (2005) thesis research as predicted and indicated in the hypothese s listed in Chapter 2 Summary Researchers have supported that among adolescents, cigarette smoking is a function of youth problem behavior rather than a health related behavior. Similar ris k and protective factors are linked with adolescent smoking behavior as are other youth problem behaviors such as early sexual behavior, drug use, and delinquency (indicating ive factors cross culturally has helped researchers determine that PBT is applicable to youth from different countries, religions, and cultural backgrounds. Despite the exploration of measures of risk and protection among youth in countries around the worl d, there is currently a paucity of information regarding the way risk and protective factors interact and influence adolescent smoking in rural environments, as compared to urban envi ronments, in the United States. It is believed that the rural U.S. consti tutes a unique cultural area which deserves the attention of researchers in order to determine

PAGE 48

48 first whether a greater rate of cigarette smoking is endemic to adolescents in rural communities in Florida and how smoking behavior may be linked to different r isk and protective factors than those found influencing youth sm oking behavior in urban areas. With the main research questions (whether rate of youth smoking behavior is different in rural versus urban areas of Florida, and, if so, to what degree do selec ted contextual, social norm based, risk and protective factors influence smoking behavior across rural and urban contexts?) the following hypotheses will be tested using data from the Florida Youth Substance Abuse Survey of 2012: H1. Smoking behavior is po sitively associated with other youth problem behaviors such as alcohol use and ATOD. (test of smoking as part of the risk behavior syndrome) (Research Question #1) H2. Social norm based risk and protective factors will be negatively and positively associat (Research Question #2) H3. The proportion of rural youth who smoke will be higher than the proportion of urban youth in Florida. (Research Question #3) H4. The relationship between social norm based risk factors for smoking among youth who live in rural, as compared with urban, locations in Florida will be stronger. (Research Question #4) H5. The relationship between social norm based protective factors for smoking among youth who live in urban as compared with rural, locations in Florida will be stronger. (Research Question #4) H6. Social norm based protective factors will have a moderating effect on risk factors associated with smoking behavior. (Research Question #5) H7. Risk fa ctors will ha ve a greater extent of influence on youth cigarette smoking in rural, as compared to urban, locations in Florida. (Research Question #6) H8. Protective fa ctors will have a greater extent of influence on youth cigarette smoking in rural, as compared to urba n, locations in Florida. (Research Question #6)

PAGE 49

49 CHAPTER 3 METHODOLOGY Data Collection and Research Design The data set used for this project comes from student responses to the 2012 Florida Youth Substance Abuse Survey (FYSAS) which was administered in pu blic middle and high schools throughout Florida as part of the Florida Youth Survey in cooperation with the Florida Department of Children and Families, the Florida Department of Health, Bureau of Epidemiology, and the Florida Department of Education. The FYSAS portion of the Florida Youth Survey is administered concurrently with the Florida Youth Tobacco Survey (FYTS) with students randomly receiving either a FYSAS questionnaire or a FYTS questionnaire. The data were obtained through a data request submitt ed to the Florida Department of Children and Families. The FYSAS is a cross sectional design, self report study. The design includes data collection at a specific point in time, measuring differences between groups at the point of survey administration as opposed to measuring change following intervention (deVaus, 2010). The Florida Youth Substance Abuse Survey is administered yearly, but county specific data are collected on a biennial basis. County level data are collected on even numbered years and state wide data are collected and reported on odd numbered years. Data are collected yearly in a cross sectional design that is repeated frequently to account for the lack of a time dimension that is inherent in this type of research design (deVaus, 2010). For t he purposes of this study, only data for the year 2012 survey administration will be examined as it contains the most recent county data. As a result of the cross sectional design, the information learned will be a result of any differences

PAGE 50

50 between countie s as they exist at the specific point in time that the survey was given, and will not be longitudinal (deVaus, 2010). The Florida Youth Substance Abuse Survey is based on the Communities That Care (CTC) survey (Florida Department of Children and Families, 2011). The Social Developmental Model (SDM) is the theoretical underpinning for the survey which is built on previous theories which were developed to explain criminal behavior (Catalano & Hawkins, 1996). According to SDM, an individual participates in ant isocial or prosocial behavior depending upon the influence of socialization, and based on sets of risk and protective factors related to individual, family, and external domains (Catalano & Hawkins, 1996). Social influence on an individual, either prosoci al or antisocial, is Hawkins, 1996). The Social Developmental model states that as individuals move through the developmental cycle, they also bond with other people in society; therefore, anti social or prosocial behaviors are a result of social bonds which are influenced by risk and protective factors and the extent to which those with whom a person is bonded comply or deviate from laws and norms (Catalano & Hawkins, 19 96). Problem Behavior Theory and its capacity to incorporate ecological contexts for youth development, rather than SDM, will remain the guiding theoretical approach to the analyses in the study. As a foundation of the proposed study, PBT is better suited than SDM to explain: the classification of adolescent smoking as problem behavior; the influences of and interactions between contextual risk and protective factors on smoking; and the differential effects of urban and rural settings on contextual risk and protective factors and smoking behavior. The Problem Behavior Theory has been tested

PAGE 51

51 cross culturally in many different ecological contexts that influence youth development in the United States and in other nations throughout the world (Costa et al., 200 5; Jessor et al., 2003; Vazsonyi et al 2010). Previous studies using PBT have been able to identify ways in which the culture and environment interact with risk and protection to influence youth participation in problem youth behaviors (Costa et al., 200 5; Jessor et al., 2003; Vazsonyi et al 2010). The State of Florida has been using the CTC as a basis for the FYSAS since the year 2000, and uses the data from FYSAS for prevention and intervention program planning. The following is a description of the C TC according to the Florida Department of Children and Families: The Communities That Care Youth Survey was developed from research funded by the Center for Substance Abuse Prevention of the U.S. Department of Health and Human Services. This student survey measures the following items: T he prevalence and frequency of drug use, T he prevalence and frequency of other antisocial behaviors, and T he degree to which risk and protective factors exist that can predict ATOD [alcohol, tobacco, and other drug] use, delinquency, gang involvement and other pro blem behaviors in adolescents ( pp. 1). A meta analysis of risk factors that appeared to be similar to those found in CTC was conducted using over 3,000 longitudinal studies to determine if the risk factors found i n CTC are actually correlated with adolescent alcohol, tobacco, and marijuana use (Derzon, 2000). The majority of studies were completed in the United States, but several were from other countries throughout the world such as Canada, Scandinavia, Great Bri tain and Australia (Derzon, 2000). The majority of the youth participating in the studies that were part of the meta analysis were born prior to 1980 and were 14 years

PAGE 52

52 old at the time of data collection (Derzon, 2000). The results of the analysis indicated that many of the CTC variables were not statistically correlated to alcohol, tobacco use, or marijuana use; and those that did correlate showed predominantly weak correlations. Tobacco use appeared to have different contributing risk factors than alcohol and marijuana use, which could mean there are different antecedents for tobacco use than other substances (Derzon, 2000). For tobacco use, the strongest CTC variables were positive attitudes toward substance use, poor family management, and family attitud es favorable toward substance use (Derzon, 2000). Out of 35 risk factors measured by CTC, 6 of them did correlate with adolescent substance use (Derzon, 2000). grounded in the S ocial Developmental Model as predictors of adolescent substance use. This may be, in part, because a portion of people do not form healthy attachment with their parents, and as a result, they do not bond with societal units as SDM indicates. In fact, a stu dy with a nationally representative sample of youth and adults aged 15 54 (n=65,244) found that only 59% of older adolescents and adults have secure attachment styles, 25.2% of participants showed insecure avoidant attachment, 11.3% of participants showed insecure anxious attachment styles, and the remainder could not be classified (Mickelson, Kessler & Shaver, 1997). A majority of older adolescents and adults have secure attachments, and should be able to form social bonds as SDM theorizes; however, a sign ificant number do not have secure attachment styles, and may not be able to form the types of social bonds discussed in SDM. While attachment styles can change through the lifespan, the fact that approximately 40% of older adolescents and adults in a natio nally representative study did not have secure

PAGE 53

53 attachment raises concerns about the appropriateness of using SDM as a theory to understand adolescent smoking. While social influences are still important in the lives of people without secure attachments, th e social bond may not be the best way to view the antecedents for their behavior. These results demonstrate that SDM may not be the proper theoretical framework for understanding youth smoking behavior; therefore, this study will designate variables as def ined by the Problem Behavior Theory. The risk and protective factors designated in PBT have been demonstrated to be related to youth problem behavior which has been shown to predict adolescent problem behaviors in youth from different national, cultural, a nd socioeconomic backgrounds. At this time, no studies have been conducted using CTC based surveys from the PBT perspective. Further, the Florida Department of Children and Families has eliminated risk and protective factors from FYSAS that were considere d to be less important for planning prevention services, and focuses questions on 15 risk factors and 6 protective factors among Florida youth (Florida Department of Children and Families, 2010). These risk and protective factors are organized in the domai ns of school, community, family, peer, and individual areas which is consistent with the PBT framework (Florida Department of Children and Families, 2010). Sampling The Florida Youth Substance Abuse Survey 2012 State Report and corresponding data were rele ased in early 2013 by the Department of Children and Families (DCF). According to the Florida Youth Substance Abuse Survey 2012 State Report provided as a supplement to the data set, the sampling process for the FYSAS is a stratified, two stage cluster sam ple made up of all public high schools and middle

PAGE 54

54 schools in the State of Florida, except for special education, adult education, and correctional schools (Florida Department of Children and Families, 2013). The first stage of sampling was conducted by str atifying the sample by county and by school level (middle and high school) after which schools were randomly selected; however, the probability of a school being selected for participation in the survey was related to the size of the student population (Fl orida Department of Children and Families, 2013). In counties with fewer than eight schools, every school was chosen to participate during the first stage of the sampling process. The second stage of the sampling process occurred at the school level when s urvey administrators randomly selected classes from within the school to participate in the survey. All classes had an equal chance of being selected, except special education and English classes for speakers of other languages (Florida Department of Child ren and Families, 2013). The researcher also has worked in the data collection process at the county level as a survey administrator for a school district helping to create the school level selection list and to facilitate county wide survey administration for each FYSAS between the years 2008 2012. It should be noted that one county, Osceola, did not participate in the 2012 FYSAS administration so data from this area will not be available for study. For the 2012 FYSAS administration, the school sample was 417 middle schools and 329 high schools, and the student sample totaled 38,742 middle school students and 36,154 high school students (Florida Department of Children and Families, 2013). Procedure Prior to survey administration notification was sent home t o parents informing them of the survey in counties with passive permission policies. Any parent who did not give permission for their child to participate in the survey could return a signed form to

PAGE 55

55 the school requesting student exemption from the survey a dministration. Only a few school districts have active permission policies where parent signed permission slips are required for students to participate in the survey. Second period was recommended as the administration time to maximize participation rates and minimize absences due to students who may have late busses or arrive tardy to school. All teachers in classrooms chosen to participate were given packets containing the surveys, administration directions, and scripts. All participants in the school co mpleted the survey during the same class period. Raw data from the 2012 Florida Youth Substance Abuse Survey were obtained through request by the researcher from the Department of Children and Families in December 2012. Dependent Variable Smoked cigarettes within the last 30 days: This dependent variable was measured according to the CDC definition of current cigarette smoking which is determined by smoking at least one cigarette during the last 30 days (Centers for Disease Control and Prevention, 2010). Th e response categories for number of cigarettes smoked within the last 30 days were: none at all; less than one cigarette per day; one to five cigarettes per day; about one half pack per day; and about one pack per day. Each response category was coded, cor respondingly, from 0 to 6. For the smoking at least one cigarette per day or more.

PAGE 56

56 Indepen dent Variables The independent variables fall into three categories environments (urban or rural), contextual or social norm based, and risk factors and protective factors. Risk and protective factors were selected based on PBT. This study seeks to under stand the extent of contextual risk and protective factors among youth from rural environments as compared to youth from more urban areas, as well as discover any influence that risk and protective factors have on youth smoking. The designation of urban wa s based on the U.S. Census urbanized area definition which refers to areas with populations of 50,000 of more (U.S. Department of Commerce, 2000). For the purpose of this study, counties containing urbanized areas of populations of 50,000 or more were clas sified as urban (18 counties). It is not possible to further delineate urban or rural location because the data set is organized according to data collected at the county, not school, level. Independent risk factor and protective factor variables listed in Table 3 1 have been mentioned previously, and have been used to examine problem behaviors among adolescents from different cultural backgrounds (Jessor et al., 2003; Vaszonyi et al., 2010). The independent variables were chosen from within the family, sch ool, and neighborhood settings to represent the overall social structure in which adolescent s live (Jessor, 1993). The Appendix lists items corresponding to measures of indep endent and dependent variables. Protective factors Religiosity was indicated by o

PAGE 57

57 has clear alcohol, tobacco, and drug use recoded into a dichotomous variable (0=no, 1=yes), and these items were combined together into a single variable with a scale that ranged from 0 to 6. Internal consistency n parental monitoring was .71. Family involvement was indicated by three items on the sur dichotomous variable (0=no, 1=yes) and these items were combined together into a single variable with a scale that ranged from 0 to 3. Internal consistency reliability, ial one on esponses was recoded into a dichotomous variable (0=no, 1=yes) and these items were combined together into a single variable with a scale that ranged from 0 to 3. Internal consistency Finally, How wrong do

PAGE 58

58 were recoded into dichotomous variables (0=not wrong, 1=wrong) and these items were combined together into a single variable with a scale that ranged from 0 to 3 Internal .74. Risk factors Peer models for drug use was indicated by three items on the survey, including: s) how many of your best months) how many of your best friends have tried beer, wine or hard liquor when their t friends. In the past year (12 responses was coded 0 to 4 (0= no friends, 4=4 friends), and these items were combined together into a single variable with a scale that ranged from 0 to 12. Internal behavior was .82. Peer rewards for problem behavior was indicated by three items on if you smoked these responses was coded 0 to 4 (0= no chance, 4=very good chance), and these items were combined together into a single variable with a scale that ranged from 0 12. Internal consistency lem behavior was .84. Favorable parental attitudes toward ATOD was indicated by three items on the

PAGE 59

59 parents feel it would dichotomous variable (0=not wrong, 1=wrong) and these items were combin ed together into a single variable with a scale that ranged from 0 to 3. Internal consistency reliability, Laws favorable to drug use was indicated by two items on the surv drank some beer, wine or hard liquor (for example, vodka, whiskey or gin) in our ese responses was recoded into a dichotomous variable (0=no, 1=yes) and these items were combined together into a single variable with a scale that ranged from 0 to 3. Internal consistency g use, was .81. Norms was for kids your age to dichotomous variable (0=not wrong, 1=wrong) and these items were combined together into a single variable with a scale that ranged from 0 to 3. Internal consistency reliability, Data Analysis To conduct data analysis, the first step was to run frequenc ies and measures of central tendency on the independent variables and dependent variable, and frequencies

PAGE 60

60 o n descriptive statistics (Table 4 1 through Table 4 4). Next, contingency tables were created between each independent variable and the de pendent va riable ( Table 4 5 through Table 4 15). Pearson correlations were then run between the independent variables and the dependent variable. These analyses showed selected social norm based risk and protective factors were correlated, positively and negatively, respectively, with cigarette smo Table 4 16). Further analyses were conducted including an independent samples t test on each of the risk and protective variables by urban and rural counties. Moreover, an independent samples t t est was conducted comparing urban and rural counties on the dependent variable, smoking behavior (C hapter 4 ) This test indicated higher percentages of cigarette smoking among youth living in rural as compared to urban areas. Independent samples t tests on each risk and protective factor within the groups, (rural or urban) determined whether social norm based risk and protective factors differ significantly among youth who live in rural, as compared with urban, locations in Florida. Regression was used to t est for interaction between protectiv e factors and risk factors ( Table 4 17 through Table 4 19). Finally, multiple regression (logistic) was used to assess the differential extent of influence among risk and protective factors on smoking behavior and to de termine which factors in combination serve to explain the mos t variance in smoking behavior. The following research questions were explored in this study, along with the following predictions (hypotheses) about expected results: 1. Is cigarette smoking among Florida youth associated with other problem behaviors such as abuse of other substances? (This is a test of smoking as part of a risk behavior syndrome of problem behavior theory in which Pearson

PAGE 61

61 correlations are run between whether or not youth report hav ing smoked in the past 30 days and used marijuana or alcohol in the past 30 days.) a) H1. Smoking behavior is positively associated with other youth problem behaviors such as alcohol use and ATOD (alc ohol, tobacco and other drugs). 2. Are selected social norm ba sed influences associated (positively associated = risk factor, negatively associated = protective factor, tested by Pearson a) H2. Social norm based risk and protective factors will be negatively an d 3. Is there a difference between the proportion of cigarette smoking among youth who live in rural, as compared with urban, locations in Florida? (Tested by independent sampl es t test comparing urban and rural youth on means of reported smoking in past 30 days.) a) H3. The proportion of rural youth who smoke will be higher than the proportion of youth who smoke and live in urban areas in Florida locations in Florida. 4. Does the de gree of social norm based risk and protective factors differ among youth who live in rural, as compared with urban, locations in Florida? (Tested by independent samples t test comparing urban and rural youth on means of each risk and protective factor.) a) H 4. There will be a greater degree of social norm based risk factors for smoking among youth who live in rural, as compared with urban, locations in Florida b) H5. There will be a greater degree of social norm based protective factors for smoking among youth who live in urban, as compared with rural, locations in Florida. Provided support is found for the first four research questions and corresponding hypotheses, the following questions and hypotheses will be tested. 5. Do social norm based protective factors (n egatively associated with smoking) moderate the effects of risk factors (positively associated) on smoking behavior? (Tested by logistic regression and interaction term between risk factors and protective factors regressed on smoking behavior.) a) H6. Social norm based protective factors will have a moderating effect on risk factors associated with smoking behavior. 6. What degree of influence do social norm based risk and protective factors have on youth cigarette smoking in rural, as compared with urban, locat ions in Florida? (Tested by logistic regression on urban/rural youth. Significant coefficients and

PAGE 62

62 odds ratios indicate the strength of influence of risk and protective factors on whether youth, urban compared to rural, report smoking in past 30 days.) a) H7 Risk factors will have a greater degree of influence on youth cigarette smoking in rural, as compared to urban, locations in Florida. b) H8. Protective factors will have a greater degree of influence on youth cigarette smoking in urban, as compared to rural locations in Florida.

PAGE 63

63 Table 3 1 Independent v ariables Protective Risk Controls Protection Religiosity Parental monitoring Attitudinal intolerance for substance use Models Risk Parental attitudes favorable toward ATOD use Peers as models of drug use Peer rewards for problem behavior involvement Support Protection Family involvement School involvement Opportunities Risk Laws favorable to drug use Norms favorable to drug use Figure 3 1. Florida rural and urban places ( U.S. Department of Commerce U.S. Census B ureau, 2000)

PAGE 64

64 CHAPTER 4 RESULTS The purpose of this study was to examine the differential impact of rural and urban contexts on youth smoking behavior, and to examine possible interactions between risk factors and protective factors related to youth smoking based on the rural or urban conte xt in which youth reside. C hapter 4 will review the results of the bivariate analyses, correlations, and logistic regre ssions as described Chapter 3 Descriptive Statistics Sample Demographics For the purposes of this analysis the sample from the FYSAS was restricted to high school student respondents. Analyses were conducted using urban, rural, and statewide designations to determine the distribution of grade level, gender, and racial make up of each group in 66 Florida counties (Table 4 2 through Table 4 4). A total of 18 counties were classified as urban and 48 classified as rural. The sample was divided into urban (44%, n=14,967) and rural (56%, n=18,781) groups based on population as d escribed in Chapter 3 Grade level and gender results were similar between the urban and rural youth. Females comprised 52% (n=7657) of the urban sample and 51% (n=9366) of the rural, whereas males comprised 48% (n=7075) of the urban sample and 49% (n= 9121) of the rural sample In this study, demographic information regarding race was analyzed to determine what percentage of each sample identified themselves as White, African American, or Hispanic. The racial composition of urban youth was comprised of a lower percentage of Whi te youth (52%, n=7829) as compared to the rural sample (63%, n=12,817). The percentage of African American youth (22%, n=3358) and Hispanic youth (25%, n=3667) was higher in the urban sample as

PAGE 65

65 compared to the rural sample (African American 15%, n=2749; Hi spanic 16%, n=3085). The mean age of the urban sample was 16 (SD = 1.27) and the mean age of the rural sample was also 16 (SD = 1.29). Additional summarized information about the sample demographics is detailed in Table 4 2 through Table 4 4. Variable Recod ing The independent variables were combi ned and scaled as described in C hapter 3. For all items, risk factor variables included peer models of drug use, peer rewards for problem behavior, favorable parental attitudes toward ATOD, laws favorable to drug use and norms favorable to drug use. Protective factor variables included personal intolerance of deviance, frequency of religious service attendance, parental monitoring, family involv ement, and school involvement. Due to data being skewed with values above zero being unevenly distributed, the dependent variable of smoking over the past 30 days was recoded; a response indicating that a participant smoked less than one cigarette per day in the last 30 days equaled 0, and a response indicating that a participa nt smoked one or more cigarettes per day in the last 30 days equaled 1. Cross Tabulations Cross tabulations were performed on the dependent variable by each of the independent variables. Each of the independent variables had a significant relationsh ip with the dependent variable. Risk Factors Examining the relationship between peer related risk factors and the dependent variable reveals that there is a significant relationship (chi sq=672.22, p<.0001) between peer rewards of for problem behavior and being a current smoker. Those who reported the highest levels of peer rewards for problem behavior were more likely than those with

PAGE 66

66 lowest peer rewards for problem behavior to report being a current smoker, although this relationship is fairly weak (tau c=.06). W hile just 3% of those with no peer rewards for problem behavior were smokers, 23% of those with the highest levels of peer rewards for problem behavior who were smokers ( Table 4 5). There is also a significant relationship (chi sq=4460.81, p<.0001) between peer models for drug use and being a current smoker as those who reported having the highest number of peer models of drug use were more likely to be smokers. However, this relationship is nominal (tau c=.14). While only 1% of those who friends who used d rugs were smokers, 33% of those with 12 frie nds who used drugs smokers ( Table 4 6). In examining parent and community norm related variables, the results indicate that there is a significant relationship (chi sq=1430.77, p<.0001) between favorable parental attitudes toward ATOD and being a current smoker. While only 4% of youth who reported parents with no favorable attitudes toward ATOD were smokers, 72% of those reporting the highest level of favorable attitude s toward ATOD were smokers ( Table 4 7); howev er, this relationship is nominal (tau c=.05). Additionally, there is a significant relationship (chi sq=167.72, p<.0001) between youth perception of laws favorable to drug use and being a current smoker. Youth who believed most teens would not be caught by police with alcohol or marijuana were more likely to be smokers than those who did not. However, this relationship is nominal (tau c=.03). Based on the contingency table, perceiving laws to be favorable toward drug use makes only about a 4% difference in being a current smoker ( Table 4 8). There is also a significant relationship (chi sq=837.14, p<.0001) between the perception of norms favorable to drug use and being a current smoker. While 4% of those who perceived no community

PAGE 67

67 norms favorable to drug use were smokers, 19% of youth with the highest perceptions of community norms favo ring drug use were smokers ( Table 4 9), although this relationship is nominal (tau c=.04). Protective Factors The relationship between each protective factor variable and the d ependent variable was also examined using contingency tables. There is a significant relationship (chi sq=388.89, p<.0001) between frequency of religious service attendance and being a current smoker. The relationship is nominal (tau c= .05). Based on the contingency table, about 9% of youth who currently smoke attended no religious services, while about 3% of those who attended religious services once a week or more smoke ( Table 4 10). There is also a significant relationship (chi sq=3714.18, p<.0001) betw een personal intolerance of deviance and being a current smoker. Those with the highest levels of intolerance of deviance were less likely to be current smokers, although the relationship is weak (tau c=.12). Based on the contingency table, about 33% of yo uth who smoke who had no intolerance for deviance while only about 2% of those who had the highest intoleranc e of deviance were smokers ( Table 4 11). The variable family involvement has a significant relationship (chi sq=201.14, p<.001) wi th being a curren t smoker (T able 4 12). Youth who had the highest levels of family involvement were less likely to report being a current smoker. This relationship is nominal (tau c= .04). According to the contingency table, about 9% of those having no family involvement a re smokers, while 4% of those with the highest levels of family involvement are smokers. The final protective factor, school involvement, has a significant relationship (chi sq=195.76, p<.0001) to being a current smoker ( Table 4 13).

PAGE 68

68 Youth with the highest levels of school involvement were less likely to be current smokers. This relationship was also nominal (tau c= .03). According to the contingency tables, about 12% of those who reported having no school involvement were smokers, while about 4% of those w ho reported having the highest levels of school involvement were smokers. In examining the relationship between parental monitoring and the dependent variable, there is a significant relationship (chi sq=517.22, p<.0001) which indicates that those who had the highest levels of parental monitoring were less likely to be current smokers. This relationship is nominal (tau c= .06), however. According to the contingency table, about 3% of those who reported having no parental monitoring were smokers, while about 4% of those who reported having the highest levels of parent al monitoring were smokers ( Table 4 14). Finally, to determine whether there is a relationship between living in a rural area and the smoking behavior, an additional cross tabulation was performe d. The results indicated that there is a significant relationship (chi sq=54.1, p<.0001) between living in a rural area and being a current smoker. Those who lived in a rural area were more likely to be smokers than those who did not. However, this relatio nship is nominal (tau b=.04). Among youth who live in urban areas, 4.5% are current smokers compared to 6.4% of y outh living in rural areas ( Table 4 15). Correlations Hypothesis 1. Smoking behavior is positively associated with other youth problem behavior s such as alcohol use and ATOD. The first Pearson correlation was performed between the dependent variable and two variables representing problem youth behaviors of being a current alcohol drinker and being a current marijuana smoker. The results indicate that being a current smoker

PAGE 69

69 is significantly and positively correlated with being both a current alcohol drinker (r=.24, p<.0001) and a current marijuana smoker (r=.31, p<.0001). Based on these results, it can be concluded that smoking behavior is positiv ely associated with other the problem youth behaviors of drinking alcohol and smoking marijuana. Hypothesis 2. Social norm based risk and protective factors will be negatively and positively associated, respectively, with cigarette smoking youth. Table 4 16 presents the correlations performed between variables. The results show that all risk factors are positively and significantly correlated to the variable current smoker, and all protective factors are negatively and significantly correla ted to the variable current smoker. Of the risk factor variables, current smoker and peer models of drug use had the strongest significant, positive (r=.28, p<.0001) relationship. Among the protective factors, intolerance of deviance had the strongest sign ificant, negative (r= .30, p<.0001) relationship with the variable current smoker. The variable rural was significantly correlated with all other variables except the protective factor variables parental monitoring, family involvement, and school involveme nt. Based on these results, it can be concluded that social norm based risk and protective factors are negatively and positively associated, respectively, with cigarette smoking among T tests Hypothesis 3. The proportion of rural youth who smoke will be higher than the proportion of urban youth who smoke in Florida. An independent samples t test was conducted to compare the means of smoking behavior between youth living in urban, as compared to rural, locations. Based on the difference of m eans test (t= 7,50, p<.0001), it can be concluded that there is a higher

PAGE 70

70 proportion of youth who are current smokers in rural areas, as compared with youth in urban locations. Hypothesis 4 There will be a higher extent of influence in social norm based r isk factors among youth who live in rural, as compared with urban, locations in Florida. A second set of independent sample t tests was conducted to determine whether there are higher levels of influence on social norm based risk factors among youth living in rural, as compared with urban, areas. When risk factor variables are examined, regarding the variable peer models of drug use, based on the difference of means test (t= 4.75, p<.0001) it can be concluded that rural youth have a higher mean of peer mode ls of drug than urban youth. For the variable favorable parental ATOD beliefs, based on the difference of means test (t= 3.99, p<.0001) rural youth have a higher mean, as compared to urban youth. Assessing the variable peer rewards for problem behavior, ba sed on the difference of means test (t= 3.81, p<.0001) rural youth have a higher mean, as compared to urban youth. For the variable laws favorable to drug use, based on the difference of means test (t= 5.15, p<.0001) rural youth have a higher mean, as comp ared to urban youth. Analyzing the variable norms favorable to drug use, based on the difference of means test (t= 2.55, p=.01) rural youth have a higher mean, as compared to urban youth. Hypothesis 5. There will be a higher extent of influence in social n orm based protective factors among youth who live in urban, as compared with rural, locations in Florida An additional set of independent sample t tests was conducted to determine whether a higher extent of influence in social norm based protective factor s among youth living in urban, as compared with rural areas is occurring. For the variable frequency of religious service attendance, based on the difference of means test (t=

PAGE 71

71 6.13, p<.0001) rural youth have a higher mean, as compared to urban youth. Based on the difference of means test (t=5.12, p<.0001) urban youth have a higher mean, as compared to rural youth. Assessing parental monitoring, based on the difference of means test (t=.13, p=.90) there is no difference in the mean of parental monitoring bet ween youth living in rural, as compared to urban, locations in Florida. Regarding the variable family involvement, based on the difference of means test (t= .42, p=.68) there is no difference in the mean between youth living in rural, as compared with urba n, areas of the state. Finally, when examining school involvement, based on the difference of means test (t=2.15, p=.03) urban youth have a higher mean scores than rural youth. Regression Risk and Protective Factors Hypothesis 6. Social norm based protecti ve factors will have a moderating effect on risk factors associated with smoking behavior. Logistic regression was used to test for interaction between protective factors and risk factors on smoking behavior. First, each protective factor was tested agains t each risk factor, then a third variable (risk by protection) was added to test the interaction between the variables in the model by comparing Max rescaled R squared values. A total of 50 logistic regression tests were performed between risk factor and p rotective factor variables. The results indicated that adding interaction between risk and protective factors did not explain any additional variance in the models. Therefore, we cannot reject the null hypothesis that there is no moderating effect of prote ctive factors on risk factors associated with smoking. Hypothesis 7. Risk factors will have a greater extent of influence in youth cigarette smoking in rural, as compared to urban, locations in Florida.

PAGE 72

72 Logistic regression of risk factors on the dependent variable was performed using risk factor variables among youth living in urban and rural areas. For youth living in urban areas, these risk variables, taken together, explain about 29% (Max rescaled R square=.29) of the variance in being a current smoker, and the likelihood is significant (chi sq=1166.46, p<.0001). For youth living in rural areas, these risk variables, taken together, explain about 30% (Max rescaled R square=.30) of the variance in being a current smoker, and the likelihood is significant ( chi sq=1941.96, p<.0001). Therefore, risk factors alone explain 1% more of the variance in smoking for rural youth compared to urban youth. Hypothesis 8. Protective factors will have a greater extent of influence on youth cigarette smoking in urban, as com pared to rural, locations in Florida. Logistic regression of protective factors on the dependent variable was performed using protective factor variables among youth living in urban and rural areas. For youth living in urban areas, these protective variabl es, taken together, explain about 23% (Max rescaled R square=.23) of the variance in being a current smoker, and the likelihood is significant (chi sq=859.10, p<.0001). For youth living in rural areas, these protective variables, taken together, also expla in about 23% (Max rescaled R square=.23) of the variance in being a current smoker, and the likelihood is significant (chi sq=1379.46, p<.0001). Therefore, we cannot reject the null hypothesis and must conclude that there is no difference in the influence of protective factors between youth living in urban, as compared to rural, locations in Florida. Risk and Protective Factor Model Table 4 17 presents the risk and protective factors logistic regression model. To further understand risk and protective facto r influence on smoking behavior, logistic

PAGE 73

73 regression of risk and protective factors on the dependent variable was performed using these variables, controlling for grade and rural location. When all independent variables equal zero, the logged odds of being a current smoker is 4.49. Living in a rural area increases the odds of being a smoker by 1.4, controlling for the other variables. So, the odds of being a smoker are 1.4 times greater for youth living in rural locations than youth living in urban locatio ns. Additionally, the odds of being a smoker increase by 1.1 for every one unit change in grade level. There were four risk factors that significantly influenced the dependent variable in this model. Among risk factors, for every one unit increase in parti cipants believing parents have a favorable attitude toward ATOD (value range of 0 3) increased the odds of being a smoker by 1.4 times, controlling for the other variables. The odds of being a smoker increased by 1.3 times for every one unit change in peer models of drug use (value range of 0 12 or 0 4 for each of 3 items added for measure), when controlling for other variables. Every one unit change in perceiving that norms are favorable to drug use (value range of 0 3) increased the odds of being a smoker by 1.1 times controlling for the other variables. Finally, the odds of being a smoker increased by 1 for every one unit change in peer rewards for problem behavior (value range of 0 12 or 0 4 for each of 3 items added for measure) controlling for the othe r variables. There were four protective factor variables that significantly influenced the dependent variable of smoking in the last 30 days. Examining protective factors, the odds of being a current smoker decreased by .92 for every one unit change in pa rental monitoring (value range of 0 6) controlling for the other variables. The odds of being a smoker decreased by .91 for every one unit change in family involvement (value range

PAGE 74

74 of 0 3) controlling for other variables. Each one unit change in the freque ncy of religious service attendance (value range of 0 to 3) decreased the odds of being a current smoker by .85 when controlling for other variables. Finally, every one unit change in intolerance of deviance (value range of 0 to 3) decreased the odds of be ing a smoker by .59 controlling for the other variables. These risk and protective factor variables taken together, explain about 34% (Max rescaled R square=.34) of the variance in being a current smoker, and the likelihood is significant (chi sq=3333.38, p<.0001). Urban Risk and Protection Model Table 4 18 presents the risk and protection logistic regression model for urban youth. In urban counties, when all independent variables equal zero, the logged odds of being a current smoker is 3.82. The odds of b eing a smoker increased by 1.1 times for every one unit change in grade level, controlling for the other variables. The variables parental monitoring and laws favorable to drug use were not significantly related to the dependent variable in the urban model There were three risk factors that significantly influenced the dependent variable in this model. Examining risk factors, the odds of being a smoker increased by 1.4 times for every one unit change in perceiving that parents have favorable attitudes towa rd ATOD, controlling for other variables. The odds of being a smoker increased by 1.3 times for every one unit change in peer models of drug use, controlling for the other variables. For the final risk factor, the odds of being a smoker increased by 1 for every one unit change in peer rewards for problem behavior. There were four protective factors that significantly influenced the dependent variable in this model. Among protective factors, the odds of being a smoker decreased by .87 times for every one uni t change in school involvement, controlling for the other

PAGE 75

75 variables. The odds of being a smoker decreased by .86 times for every one unit change in family involvement, when controlling for the other variables. Additionally, the odds of being a smoker decre ased by .81 times for every one unit change in frequency of religious attendance. Finally, the odds of being a smoker decreased by .58 times for every one unit change in intolerance of deviance. These risk and protective factor variables taken together, ex plain about 34% (Max rescaled R square=.34) of the variance in being a current smoker, and the likelihood is significant (chi sq=1250.76, p<.0001). Rural Risk and Protection Model Table 4 19 presents the risk and protection logistic regression model for ru ral youth. In rural counties, when all independent variables equal zero, the logged odds of being a current smoker is 4.52. The odds of being a smoker increased by 1.1 times for every one unit change in grade level, controlling for the other variables. Th e protective factor variables parental monitoring, family involvement, school involvement, and laws favorable to drug use did not significantly influence the dependent variable in this model. The risk factor variable peer rewards for problem behavior did n ot significantly influence the dependent variable in this model. There were three risk factors that significantly influenced the dependent variable in this model. The odds of being a smoker increased by 1.4 times for every one unit change in favorable pare ntal attitudes toward ATOD controlling for the other variables. The odds of being a smoker increased by 1.3 times for every one unit change in peer models of drug use, controlling for the other variables. Finally, the odds of being a smoker increased by 1. 1 times for every one unit change in norms favorable to drug use, when controlling for the other variables.

PAGE 76

76 T here were only two protective factor variables that significantly influenced the dependent variable in this model. The odds of being a smoker decre ased by .86 times for every one unit change in frequency of religious service attendance. Additionally, the odds of being a smoker decreased by .59 times for every one unit change in intolerance of deviance. These risk and protective factor variables taken together, explain about 34% (Max rescaled R square=.34) of the variance in being a current smoker, and the likelihood is significant (chi sq=2046.36, p<.0001). Comparing the results of the regression models for youth in urban and rural areas, the models for both areas explain about 34% of the variance in smoking behavior for high school students. Fisher s z tests were run to compare the coefficients of risk factors and protective factors on smoking behavior in urban and rural samples (urban n=11,156, rura l n= 13,921). For the risk factor laws favorable to drug use (urban r= .04, rural r .07; z= 8.67; p<.000 two tailed). Fisher s z test suggests a significant difference between coefficients on this risk factor for smoking in urban compared to rural populati ons. For the protective factor frequency of religious service attendance (urban r= .21, rural r= .15; z= 4.88; p<.000 two tailed) results suggest a significant difference between the coefficients in urban compared to rural populations. Examining the protec tive factor intolerance of deviance (urban r= .54, rural r= .52; z= 2.19; p<.030 two tailed) the results suggest a marginally significant difference between coefficients. Regarding the protective factor parental monitoring (urban r= .06, rural r= .00; z= 4 .73; p<.000 two tailed) the results suggest a significant difference between the coefficients. For the protective factor family involvement (urban r= .15, rural r= .04; z= 8.74; p<.000 two tailed) the results suggest a significant difference between the co efficients. Finally,

PAGE 77

77 for the protective factor school involvement (urban r= .14, rural r= .07; z= 5.57; p<.000 two tailed), the results suggest a significant difference between the coefficients. No other results testing the difference among urban and rural coefficients of risk and protection on smoking behavior were significant. Three risk factors and two protective factors have been found to be significantly associated with youth smoking in rural locations. Urban locations were also found to have three ris k factors that influence youth smoking; however, four protective factors that are significantly negatively associated with smoking among urban youth may point to prevention intervention. The results suggest that there are may be a greater influence of cert ain protective factors or risk factors for smoking with urban than rural populations and vice versa.

PAGE 78

78 Table 4 1. Variable one way frequencies Variable N Percent Mean S.D. Peer models of drug use 0 1 2 3 4 5 6 7 8 9 10 11 12 33,127 9463 2821 2589 230 0 2421 1777 2092 1609 2185 1524 1188 845 2313 100 29 9 8 7 7 5 6 5 7 5 4 3 7 4.15 3.96 Peer rewards for problem behavior 32,771 100 3.55 3.11 0 1 2 3 4 5 6 7 8 9 10 11 12 14,875 2094 2424 2970 2253 1702 2232 1222 1037 842 412 222 486 45 6 7 9 7 5 7 4 3 3 1 1 1 Favorable parent attitude toward ATOD 0 1 2 3 30,988 28,585 1601 471 331 100 92 5 2 1 .11 .44 Laws favorable to drug use 0 1 2 31,559 4704 3408 23,442 100 15 11 74 1.59 .73 Norms favorable to drug use 0 1 2 3 31,634 28,431 1785 638 780 100 90 6 2 2 .17 .57 Frequently attend religious services 0 1 2 3 32,955 7652 9587 4571 11,145 100 23 29 14 34 1.58 1.78 Intolerance of deviance 0 1 2 3 32,913 1738 1959 4068 25,148 100 5.3 5.9 12.4 76.4 2.60 .82

PAGE 79

79 Table 4 1. Continued Variable N Percent Mean S.D. Parental monitoring 0 1 2 3 4 5 6 29,613 833 902 1331 2752 4091 7478 12,226 100 3 3 5 9 14 25 41 4.69 1.56 Family involvement 0 1 2 3 29,976 2465 2790 7371 17,350 100 8 9 25 58 3.32 .95 School involvement 0 1 2 3 32,889 1408 5484 16,162 9835 1 00 4 17 49 30 Dependent Variable 0 1 32,650 30,836 1814 100 94 6 .06 .23 Table 4 2. Summary statistics for demographic informa tion u rban Characteristic Percentage Frequency Grade 9 30% 4540 10 27% 4040 11 24% 3613 12 19% 2774 Gender Fem ale 52% 7657 Male 48% 7075 Race White 52% 7829 African American 22% 3358 Hispanic 25% 3667

PAGE 80

80 Table 4 3. Summary statistics for demographic information r ural Characteristic Percentage Frequency Grade 9 30% 5649 10 28% 5223 11 23% 4292 12 19% 3617 Gender Female 51% 9366 Male 49% 9121 Race White 68% 12817 African American 22% 2749 Hispanic 25% 3085 Table 4 4. Summary statistics for demographic information s tatewide Characteristic Percentage Frequency Grade 9 30% 10189 10 28% 9263 11 23% 7905 12 19% 6391 Gender Female 51% 17023 Male 49% 16196 Race White 61% 20646 African American 18% 6107 Hispanic 20% 6751

PAGE 81

81 Table 4 5. Contingency table smoker by peer rewards for problem behavior Peer Models of Drug Use Smoker 0 1 2 3 4 5 6 7 8 9 10 11 12 No 97% 98% 96% 92% 94% 93% 92% 91% 92% 88% 89% 96% 77% Yes 3% 2% 4% 8% 6% 7% 8% 9% 8% 12% 11% 4% 23% Total 100% (n=14558) 100% (n=2066) 100% (n=2385) 100% (n=2912) 100% (n=2219) 100% (n=1675) 100% (n=2202) 100% (n=1207) 100% (n=1019) 100% (n=817) 100% (n=407) 100% (n=218) 100% (n=467) N=32,152 Table 4 6. Contingency table smoker by peer models of drug use Peer Models of Drug Use Smoker 0 1 2 3 4 5 6 7 8 9 10 11 12 No 99 % 99 % 99 % 98 % 97 % 98 % 96 % 95% 94% 88% 85% 81% 67% Yes 1 % 1 % 1 % 2 % 3 % 2 % 4 % 5% 6% 12% 15% 19% 33% Total 100% (n=9184 ) 100% (n=2750 ) 100% (n=2521 ) 100% (n=2237) 100% (n=2341 ) 100% (n=1722 ) 100% (n=2029 ) 100% (n=1565) 100% (n=2104) 100% (n=817) 100% (n=1157) 100% (n=825) 100% (n=2183) N=32,090

PAGE 82

82 Table 4 7. Contingency Table smoker by favorable parental attitude toward ATOD Favorable Parent Attitude Toward ATOD Smoker 0 1 2 3 No 96% 82% 73% 68% Yes 4% 18% 27% 72% Total 100% (n=28,428) 100% (n=1590) 100% (n=469) 10 0% (n=328) N=32,815 Table 4 8. Contingency table smoker by laws favorable to drug use Laws Favorable to Drug Use Smoker 0 1 2 No 98% 96% 94% Yes 2% 4% 6% Total 100% (n=4680) 100% (n=3390) 100% (n=23,324) N=31,394 Table 4 9. Contingency table smoker by norms favorable to drug use Norms Favorable to Drug Use Smoker 0 1 2 3 No 96% 84% 86% 81% Yes 4% 16% 14% 19% Total 100% (n=28291) 100% (n=1775) 100% (n=628) 100% (n=771) N=31,465 Table 4 10. Contingency table smoker by frequency of relig ious service attendance Frequency of Religious Service Attendance Smoker 0 1 2 3 No 91% 94% 95% 97% Yes 9% 6% 5% 3% Total 100% (n=7490) 100% (n=9439) 100% (n=4487) 100% (n=10,980) N=32,396 Table 4 11. Contingency table smoker by intolerance of deviance Intolerance of Deviance Smoker 0 1 2 3 No 67% 81% 91% 98% Yes 33% 19% 9% 2% Total 100% (n=1648) 100% (n=1899) 100% (n=3957) 100% (n=24,577) N=32,081

PAGE 83

83 Table 4 12. Contingency table smoker by family involvement Family Involvement Smoker 0 1 2 3 No 91% 92% 93% 96% Yes 9% 8% 7% 4% Total 100% (n=2409) 100% (n=2755) 100% (n=7305) 100% (n=17,230) N=29,695 Table 4 13. Contingency table smoker by school involvement School Involvement Smoker 0 1 2 3 No 89% 92% 95% 96% Yes 11% 8% 5% 4% Total 100% (n=1279) 100% (n=5266) 100% (n=15751) 100% (n=9567) N=31,863 Table 4 14. Contingency table smoker by parental monitoring Parental monitoring Smoker 0 1 2 3 4 5 6 No 96% 84% 86% 81% 92% 95% 97% Yes 4% 16% 14% 19% 8% 5% 3% Total 100% (n=28291) 100% (n=1775) 100% (n=628) 100% (n=771) 100% (n=4071) 100% (n=7446) 100% (n=12.186) N=29,464 Table 4 15. Contingency table smoker by rural Rural Smoker No Yes No 95.5% 93.6% Yes 4.5% 6.4% Total 100% (n=14457) 100% (n=18193) N=32,650

PAGE 84

84 Ta ble 4 16. Summary of i ntercorrelations for v ariables Variables 1 2 3 4 5 6 7 8 9 10 11 12 1. Current smoker 2. Peer Models of drug use .31 3. Peer Rewards for problem behavior .13 .32 4. Favorable parent ATOD beliefs .21 .19 .13 5. Laws favorable to drug use .07 .24 .17 .08 6. Norms favorable to drug use .15 .16 .10 .30 .08 7. Frequently attend religious service .11 .18 .01 ** .11 .07 .07 8. Intolerance of deviance .34 14 .22 .31 .15 .28 .16 9. Parental monitoring .13 .44 .09 .19 .20 .15 .17 .23 10. Family involvement .08 .22 .07 .08 .13 .09 .10 .14 .48 11. School involvement .07 .11 .07 .08 .10 .08 .07 .13 .19 .19 12. Rura l .04 .03 .02 .02 .03 .01 .03 .03 .00 ** .00 ** .0 p<.001, *p=.03 **not significant

PAGE 85

85 Table 4 17. Logistic regression predicting likelihood of adolescent s moking s tatewide B S.E. Wald df p Odds Ratio 95% C.I. for Odds Ratio Lower Upper Pee r models of drug use .27 .36 755.43 1 <.0001 1.31 1.28 1.34 Peer rewards for problem behavior .02 .01 4.74 1 .03 1.02 1.00 1.04 Favorable parent attitude toward ATOD .36 .04 65.55 1 <.0001 1.43 1.32 1.57 Norms favorable to drug use .11 .04 7.77 1 .005 1 .03 1 .03 1.21 Laws favorable to drug use .03 .06 .32 1 .57 1.03 .92 1.16 Frequently attend religious services .15 .03 36.15 1 <.0001 .85 .80 .89 Intolerance of deviance .52 .03 319.57 1 <.0001 .59 .56 .63 Parental monitoring .02 .02 1.20 1 .27 .98 .94 1.02 Family involvement .09 .03 6.33 1 .01 .92 .86 .98 School involvement .09 .04 6.42 1 .01 .91 .85 .98 Grade .11 .03 16.97 1 <.0001 1.12 1 .06 1.19 Rural .29 .06 30.29 1 <.0001 1.42 1.25 1.61 N=26,513; Max rescaled R square=.34

PAGE 86

86 Table 4 18 Lo gistic regression predicting l ikelihoo d of a dolescent smoking u rban c ounties B S.E. Wald df p Odds Ratio 95% C.I. for Odds Ratio Lower Upper Peer models of drug use .27 .59 268.98 1 <.0001 1.31 1.26 1.35 Peer rewards for problem behavior .03 02 4.40 1 .04 1.03 1.0 1.06 Favorable parent attitude toward ATOD .37 .07 24.68 1 <.0001 1.44 1.25 1.67 Norms favorable to drug use .12 .06 3.48 1 .06 1.13 .99 1.27 Laws favorable to drug use .04 .09 .17 1 .68 .96 .80 1.16 Frequently attend religious services .21 .05 19.73 1 <.0001 .81 .74 .89 Intolerance of deviance .54 .05 124.39 1 <.0001 .58 .53 .64 Parental monitoring .06 .03 2.76 1 .10 .95 .89 1.01 Family involvement .15 .05 8.33 1 .003 .86 .77 .95 School involvement .14 .06 5.26 1 .02 87 .77 .98 Grade .10 .05 4.57 1 .03 1.10 1 .0 1.22 N=11,659; Max rescaled R square=.3.4

PAGE 87

87 Table 4 19. Logistic regression predicting likelihoo d of adolescent smoking rural counties B S.E. Wald df p Odds Ratio 95% C.I. for Odds Ratio Lower Upper Peer models of drug use .27 .01 485.37 1 <.0001 1.31 1.28 1.34 Peer rewards for problem behavior .01 .01 1.28 1 .26 1.01 .99 1.04 Favorable parent attitude toward ATOD .36 .06 40.13 1 <.0001 1.43 1.28 1.60 Norms favorable to drug use .10 .05 4.28 1 .0 4 1.11 1.01 1.23 Laws favorable to drug use .07 .08 .94 1 .33 1.08 .93 1.25 Frequently attend religious services .15 .03 18.36 1 <.0001 .87 .80 .92 Intolerance of deviance .52 .04 197.74 1 <.0001 .60 .55 .64 Parental monitoring .00 .03 .02 1 .88 1.0 0 .94 1.05 Family involvement .04 .04 .95 1 .33 .96 .88 1.04 School involvement .07 .05 2.07 1 .15 .93 .85 1.03 Grade .12 .04 12.23 1 .0005 1.13 1.06 1.21 N=14,853; Max rescaled R square=.34

PAGE 88

88 CHAPTER 5 DISCUSSION The purpose of this study was to exa mine the influence of risk and protective factors on youth smoking behaviors within the context of rural and urban Florida environments. This study examined: the relationship between smoking and other problem youth behaviors; the relationships between risk and protective factors; the relationship between risk and protective factors and smoking; and how these relationships may vary depending on contextual influences. Context was divided into two categories based on county population areas: urban and rural. T his section will discuss the study hypotheses and how the research questions were answered. Additionally limitations of the study will be discussed as well as theoretical and practical implications of the results and possibilities for future research in th is area. The first research question asked whether smoking behavior is related to other problem youth behaviors such as drinking alcohol and smoking marijuana. I hypothesized that there would be a positive correlation between smoking and the other problem behaviors. The results corroborate my hypothesis that among Florida youth, smoking behavior is related to other problem youth behaviors. This information is important with regard to prevention practice because youth tobacco use has largely been treated as a public health issue in Florida with the emphasis on the development of policies that focus on second hand smoke exposure and tobacco marketing issues instead of conceptualizing youth tobacco use as a problem behavior that is linked to other problem behav iors as part of a syndrome (Jessor, 1991). The next research question explored whether risk factors are positively related to youth smoking, and protective factors are negatively related to youth smoking, with the

PAGE 89

89 hypothesis stating that this would be affi rmed by the results. This hypothesis was also supported by the results. It is interesting to note that the strongest relationship between a protective factor and the dependent variable of smoking for both urban and rural youth was intolerance of deviance, which was a variable comprised of questions relating to behaviors such as smoking cigarettes, drinking alcohol, and smoking marijuana. Please refer to the conclusions sectio n for further discussion of intolerance of deviance in relation to youth substance use. Of the risk factors, favorable parental attitudes about ATOD had the strongest relationship with youth smoking behavior. This finding relates the current study to the p roblem behavior theory since the relationship between models of a risk behavior and parental approval of substance use, is part of the revised problem behavior theory framework (Jessor, 2003). Even more interesting is the fact that the results of this stud y are similar to previous problem behavior research that included problem behaviors such as drinking alcohol, risky sexual behavior, drug use, and delinquency, while excluding cigarette smoking among problem behaviors (Jessor et al., 1995). In the previous research, personal intolerance for deviance and models for problem behavior were also the strongest protective and risk factors, respectively, influencing problem behaviors (Jessor et al., 1995). The next research question asked whether there is a differe nce in the extent of youth smoking among youth living in urban and rural locations. The hypothesis stated that the proportion of rural youth who smoke will be higher than the proportion of urban youth who smoke in Florida. The results indicate that there i s a greater proportion of

PAGE 90

90 youth smoking in rural areas, compared to youth in urban areas. This is important information because there has been no specific research to date comparing the proportion of youth who smoke in rural and compared to urban youth in Florida. Research question four focused on the extent of influence of risk and protective factors among youth in urban and rural locations. It was hypothesized that the influence of social norm based risk factors would be higher among youth in rural areas. Based on the difference of means test, there is evidence to support this hypothesis. The second hypothesis stated that the influence of social norm based protective factors would be higher among youth living in urban areas which was not supported. These r esults direct us to a key point of the problem behavior theory which states that higher levels of protection buffer risk, and youth who experience risk without protection are more likely to engage in problem behaviors (Jessor, 1991; Jessor et al., 1995). P reliminary results suggest that youth in rural areas have a similar number of risk factors influencing smoking as urban youth, yet experience fewer protective factors. A test of difference in coefficients indicating the strength of influence of risk and pr otection, on urban compared to rural youth smoking, was be the next step of analysis. A significant difference in coefficients was discovered among one risk factor and five protective factors. The difference in coefficients tests suggests that there are si gnificant differences in the way context influences risk f actors and protective factors. A combination of higher risk and lower protection exposes youth to the possibility of higher levels of engagement in problem behaviors (Jessor et al., 1995). These res ults may also indicate that it is important to consider the embedded culture or context in which youth live when considering risk and protective factors related to problem youth

PAGE 91

91 behavior involvement. It is possible that rural areas have unique cultural dyn amics that influence levels of risk toward and protection against risk behaviors among youth. The fifth research question asked whether social norm based risk and protective factors influence youth smoking in urban settings compared to rural settings. Hypo thesis seven stated that risk factors will have a greater influence on youth cigarette smoking in rural, as compared to urban, locations in Florida. The results supported this hypothesis, although there was a small difference in the influence that risk fac tors had on smoking behavior between youth in urban and rural settings. It should be noted that the large sample in this study can account for even small differences, making statistical significance possible when there may be little evidence of practical s ignificance in the difference between the influence that risk factors have on smoking among rural and urban youth (Mertens, 2010). To further explain the importance of practical significance, the results should be examined in terms of whether the small dif ferences between the rural and urban groups, with regard to the influence of risk factors on smoking, will have any practical value in understanding the contextual differences between rural and urban settings related to youth smoking (Rubin, 2013). Since t he size of the difference was small regarding risk among rural and urban youth, it may be more important to examine if risk factors influence smoking behavior differently among youth living in rural and urban areas of Florida. This may give a better unders tanding of the contextual difference between rural and urban settings that may influence youth behavior on a more proximal level. For both urban and rural youth, models risk was the strongest risk factor For the other categories of risk, opportunities ris k was influential among rural youth, but it was not found to be as influential as models risk. Among urban and rural

PAGE 92

92 youth, family and peer social contexts were more influential than other social contexts regarding risk factors. Overall, youth in both rura l and urban areas experienced three risk factors which were significantly related to smoking behavior. Peer models of problem behavior and favorable parental attitudes toward ATOD were risk factors that urban and rural youth had in common. For urban youth, peer rewards for problem behaviors was a unique significant influencer of smoking behavior. Among rural youth, norms favorable to drug use was a unique influencer of smoking behavior that was not significantly associated with smoking among urban youth. Th e findings point to urban youth statewide experiencing risk within the domains of peer and family, whereas youth in rural areas experienced risk within the domains of peer, family, and community. To examine the effect of protective factors on smoking behav ior, the final hypothesis stated that protective factors will have a greater extent of influence on cigarette smoking among urban youth than rural youth. This hypothesis was not supported by the results which showed no difference in the influence of protec tive factors on smoking between youth in urban and rural locations. Youth in urban areas reported a greater number of protective factors against smoking than youth in rural areas; yet the extent of influence of protection was the same for each group. For both urban and rural youth, controls protection was the strongest protective factor. For the other categories of protection, support protection was influential among urban youth, but it was not found to be as influential as controls protection. Among urban and rural youth, individual and community contexts were more influential than other contexts for protective factors. Overall, youth in urban areas experience four protective factors which are significantly negatively related to smoking behavior, while

PAGE 93

93 rur al youth experience only two protective factors. Intolerance of deviance and frequency of religious service attendance were protective factors that all Florida youth had in common as protective factors influencing smoking behavior. For urban youth, family involvement and school involvement were unique significant, negative influencers of smoking behavior. This revealed that urban youth experienced protection within school, family, community and individual domains, whereas youth in rural areas experienced pr otection within the community and individual domains only. Although youth in urban and rural areas experienced a similar number of risk factors, rural youth experienced fewer protective factors compared to their urban counterparts. According to PBT, increa sed levels of protection are needed to buffer the impact of higher levels of risk. Youth who experience higher levels of risk have a greater possibility of poorer life outcomes, especially if risk is not buffered by protective factors (Jessor, 1991). There fore, these results indicate that rural youth need more intervention, and greater emphasis on building protective factors to mitigate the risks associated with cigarette smoking (Jessor, 1993). Additionally, the results of this study did not indicate that the protective factors acted as moderators of risk regarding smoking. Therefore, risk reduction efforts may also be needed as part of prevention efforts. Based on the findings of the current study, it is possible that current tobacco prevention programming is failing to address certain areas of risk and protection that are relevant to youth in Florida today. Much of current tobacco prevention programming focuses on creating new anti tobacco policies, yet analyses in the current study showed that laws, and t he extent to which they are perceived by adolescents as being enforced in their communities (defined as likelihood of perception that they will be caught using

PAGE 94

94 ATOD), have a very weak relationship to smoking behavior, and do not significantly influence you th smoking behavior. Additionally, qualitative research conducted among law enforcement personnel, judges, and court personnel in Central Florida reveals that the underage tobacco purchase and possession law is not evenly enforced throughout Florida (Woodh ouse, Sayre & Livingood, 2001). This is due in large part to budgetary restrictions, since some judicial districts, clerk of court offices, and police departments do not have the resources to consistently follow up on tobacco use citations, fines, and lice nse suspensions as required by Florida law (Woodhouse et al., 2001). Since tobacco laws may not be enforced in an equal manner within each area of the state, youth perception of enforcement has a very weak relationship to smoking and is not a significant i nfluencer of smoking. According to the results of this study, it may be important to additionally focus prevention practices on increasing protective factors in ways that are relevant to youth in rural and urban contexts. A surprising result of the analysi s revealed that parental monitoring, within the domain of controls protection, did not function as a significant protective factor among Florida youth. Previous research showed that controls protection, including parental monitoring, was significantly rela ted to problem youth behavior across a multi national and cross cultural sample from eight different countries (Vazsonyi et al., 2010). Parental monitoring was expected to be a source of protection among Florida youth; however, the results may indicate a d awareness of their activities and actual parental enforcement or sanctions against rule infringement which may account for this result.

PAGE 95

95 Limitations There are some limitations inherent in using secondary data, and the current study includes several of these limitations. When administering the FYSAS, the Florida Department of Children and Families also eliminated several questions from the original Communities that Care survey, which may mean this survey d oes not give a full picture of risk and protective factors that contribute to the problem youth behaviors it attempts to measure (Florida Department of Children and Families, 2011). Although there is a large state wide sample in the present study, one of t he limitations of the study is the use of secondary data which restricts the possible responses of participants to questions, and restricts the current study to the questions asked by the previous researchers. For example, no school level data were availab le for analysis which could have assisted in more accurate categorization of whether participants resided in urban or rural locations as well as pointed to proxy measures for school receive reduced or free lunch). Additionally, some responses were originally coded as type scale which would increase variability in participant responses. Another issue with the survey is the lack of qu estions that may give a better understanding of factors that influence youth smoking. For example, in addition to asking if parents feel it is wrong for them to use cigarettes, alcohol, or drugs, the study could include specific questions on whether the pa rents actually use cigarettes, alcohol, or other drugs themselves. Parental drug use may actually be a better measure of parental modeling related to youth problem behaviors rather than the youth perception of parent attitudes. This type of question was in cluded to explore peer modeling of

PAGE 96

96 problem behavior by asking youth how many of their four best friends used cigarettes, alcohol, and drugs. A question regarding youth perception about peer rewards (being tobacco is also included in this survey; therefore, including similar types of questions regarding parent actions and attitudes may give a better explanation into the antecedents of youth problem behaviors. Another improvement in the survey would be to mat ch youth responses on amounts of cigarettes, alcohol, and drugs used with national adult studies so that the responses can be compared, or at least to allow youth to choose a specific number to provide greater accuracy. Some additional questions on the sur vey could be refined to provide more information about the ways in which school and community activities influence youth involvement questions do not appear to be an accur ate measure of that variable. Improving school involvement questions to more accurately reflect youth perceptions of connectedness to school and involvement opportunities would improve the ability to measure the impact of school influences on smoking behav ior. Questions related to parental sanctions could be added to further clarify whether youth perceptions of parental monitoring coupled with sanctions for problem behavior involvement serve as a protective factor among Florida youth. Another limitation of the study was that urban and rural locations were designated on a county wide basis, which is a less accurate way to determine whether participants lived in rural or urban areas. Although a question on the survey asked youth whether they resided on a farm, in the country not on a farm, or in a city town or suburb,

PAGE 97

97 the data were incomplete and the distinction between city, town or suburb could not be made to ascertain whether a residence was in or near a location with a population of 50,000 or more. Therefor e, this study classified entire counties as urban or rural based on population centers of more or less than 50,000 which may not have adequately captured specific rural or urban participant locations. It would be possible to refine the definition of rural and urban by choosing only the top 10 most rural and urban counties within the State of Florida to compare youth smoking behaviors. Since it is not possible to know in which specific population area within a county participants live, identifying specific c ounties with highest and lowest population density may be an alternative so that the more specific differences between rural and urban context may be discovered. Recommendations for Future Research Intolerance of deviance emerged as the protective factor w ith the strongest influence on youth smoking behaviors in both urban and rural areas. Further study should be done to determine the antecedents for intolerance of deviance to possibly identify areas of prevention intervention to strengthen this protective factor among youth. It is possible that many factors work in combination, including culture and context, to build intolerance of deviance among youth. Identifying predisposing factors to intolerance of deviance and strengthening them has great potential in reducing youth involvement in problem behaviors such as smoking. Additional research is called for to examine the way rural and urban context interacts with youth from different racial backgrounds with regard to problem behavior. Further inquiry into spe cific ways that family and peer relationships, religiosity, and perceived community norms interact within urban and rural contexts among youth of different racial backgrounds may provide information to pinpoint specific areas for future

PAGE 98

98 problem behavior pr evention services. Also including a more specific determination of rural location to better pinpoint participant population, as discussed previously, would refine the results and the understanding of the interaction of rural context with these factors. Con clusion Youth smoking is an ongoing concern in the State of Florida and the nation today. Despite continued prevention efforts, the youth smoking rate has only decreased by a few percent over the past decade. The results of the current study suggest there may be merit in re focusing efforts on the specific needs of youth, particularly addressing attention on developing and strengthening context specific (e.g., urban and rural) protective factors that influence youth problem behaviors. Since cigarette smokin g is related to a larger syndrome of problem behaviors, conceptualizing tobacco prevention as part of this syndrome, instead of treating tobacco use as a stand alone adolescent behavior, could be an important part of reducing youth smoking rates. Furthermo re, the current study indicates that high school youth in Florida are not significantly influenced by their perception of the enforcement of laws regarding problem behaviors. Current tobacco prevention efforts include an emphasis on creating new tobacco po licies, with the idea that new policies and laws will restrict current and future tobacco use, thus reducing the overall tobacco usage rate among adults and adolescents. However, new tobacco control laws are not likely to decrease adolescent smoking rates if they do not function as protective factors that buffer against smoking. Therefore, it is important that policy and prevention efforts should also include programs which target increasing the prevalence of protective factors shown to influence smoking be havior among adolescent youth, particularly in rural areas. For example, rather than

PAGE 99

99 instituting a more restrictive policy affecting the advertisement, sale or distribution of tobacco products to minors in Florida, a state legislative mandate for K 12 heal th education covering the topic of intolerance of youth tobacco use, backed by funding or incentives to schools/teachers for implementation, has greater potential to impact change. Prevention programming based youth empowerment and policy change has been s hown to have a stronger impact on youth tobacco use than policy restrictions on tobacco availability; however prevention programs based on PBT focusing on community, peer, family, and environmental domains have demonstrated longer lasting impact on youth t obacco use (Pentz, 1999). Additionally, concentrating anti tobacco advertising in rural areas has been shown to increase youth awareness of tobacco issues in those areas which are generally less served than urban areas, and may serve to support tobacco pre vention programs in rural areas (Duke et al., 2009). Intolerance of deviance has been studied using a family developmental model which allows for examination of risk and protective factors that influence youth within different context and cultural settings Research has indicated that the family domain appears to create a higher level of intolerance of deviance and boosts conventional behavior among adolescents (Brook, Brook, & Pahl, 2006). This may indicate that strengthening family involvement and parent child relationships could build increased intolerance of deviance among adolescents. Therefore, expanding the conceptualization of tobacco prevention may need to include strengthening family relationships and increasing usage of effective parenting techniq ues as long term interventions that have the potential to reduce youth involvement in problem behaviors.

PAGE 100

100 Additionally, for youth in Florida, smoking behavior is influenced by parental attitudes toward ATOD. It is unlikely that prevention program intervent ion would be successful in changing parental attitudes toward youth substance youth; however, it may be important to educate parents that their attitudes may influence problem behavior involvement in their children. For example, a father who smokes may hav e the goal for his children to be tobacco free due to health issues or problems he has experienced due to tobacco use and addiction. Tobacco prevention programs may provide interventions to help parents learn to address these issues with their children, an d communicate anti substance use messages to them. Another possible intervention in this area could be providing youth with adult models of conventional, non substance using, behavior in community settings. Increasing youth opportunities for involvement in community programs such as: 4 H, Boys and Girls Clubs, Boy Scouts, Girl Scouts, and other pro social groups, led by positive adult role models could be another way to provide protection against parental modeling of problem behavior. In addition, youth par ticipation in prosocial activities would be more likely to expose them to youth who are models of conventional behavior, which may also serve as a method of risk reduction as well. A multi dimensional approach to youth tobacco prevention including communit y, school, peer, family, and individual domains is likely to have the greatest impact on youth smoking (Ennet et al 2010).

PAGE 101

101 APPENDIX LIST OF ORIGINAL VARIABLES AND RESPONSES Dependent Variable 0 Not at all 1 Less than one cigarette per day 2 One to five cigarettes per day 3 About one half pack per day 4 About one pack per day 5 About one and one half packs per day 6 Two packs per day Independent Variables Controls Protection Reli giosity 0 Never 1 Rarely 2 1 2 times a month 3 About once a week or more Parental monitoring 0 NO! 1 no 2 yes 3 YES! Attitudinal intolerance for deviance nk beer, wine or hard liquor for example, vodka, whiskey or gin regularly? Smoke cigarettes? 0 Very wrong 1 Wrong 2 A little bit wrong 3 Not wrong at all Support Protection Family involvement 0 NO! 1 no 2 yes 3 YES!

PAGE 102

102 School involvement ool to get involved in spor ts, one on 0 NO! 1 no 2 yes 3 YES! Models Risk Parental attitude s favorable toward ATOD use Drink beer, wine or hard liquor (for example, vodka, whiskey or gin) regularly? Smoke cigarettes? 0 Very wrong 1 Wrong 2 A little bit wrong 3 Not wrong at all Peers as models of drug use year (12 months) how many of your best friends have: Smoked cigarettes? Tried beer, wine or hard liquor (for example, vodka whiskey or gin) when 0 None 1 1 2 2 3 3 4 4 Peer rewards for problem behavior involvement Smoked cigarettes? Began drinking alcoholi c beverages regularly, that is, at least once or twice a month? 0 No or very little chance 1 Little chance 2 Some chance 3 Pretty good chance 4 Very good chance Opportunities Risk Laws favorable to drug use

PAGE 103

103 arijuana in your neighborhood would he or she be caught by whiskey or gin) in our neighborhood, would he or she be caught by 0 NO! 1 no 2 yes 3 YES! Norms favorable to drug use ood think it was for kids your age: To use marijuana? To drink alcohol? 0 Very wrong 1 Wrong 2 A little bit wrong 3 Not wrong at all

PAGE 104

104 LIST OF R EFERENCES Aloise Young, P.A., Wayman, J.C., & Edwards, R.W. (2002). Prevalence of cigarette smoking among rural ado lescents in the United States. Substance Use & Misuse, 37(5 7), 613 630. Bronfenbrenner, U. (1979). The ecology of human development Cambr id ge: Harvard University Press. Bronfenbrenner, U. (1986). Ecology of the family as a context for human development: Research perspectives. Developmental Psychology, 22 723 742. Brook, J.S., Brook, D.W., Pahl, K. (2006). The developmental context for adoles cent substance abuse intervention. In Liddel, C.A. & Rowe, C.A. (Eds.), Adolescent substance abuse: Research and clinical advances (pp.25 50). Cambridge: Cambridge University Press. Catalano, R. F. & Hawkins, J. D. (1996). The social development model: A t heory of antisocial behavior. In J. D. Hawkins (Ed.), Delinquency and crime: Current theories (pp. 149 197). New York, NY: Cambridge University Press. Centers for Disease Control and Prevention. (1996). Projected smoking related deaths among youth United States. Morbidity and Mortality Weekly Report, 45(44), 971 974. Centers for Disease Control and Prevention. (2007). State and community interventions. Best Practices for Comprehensive Tobacco Control Program, Section A, 22 30. Centers for Disease Control and Prevention. (2008). Smoking attributable mortality, years of potential life lost, and productivity losses United States, 2000 2004. Morbidity and Mortality Weekly Report, 57(45), 1226 1228. Centers for Disease Control and Prevention. (2010). Cigaret te use among high school students United States, 1991 2009. Morbidity and Mortality Weekly Report, 59(26), 797 801. Centers for Disease Control and Prevention. (2011). Vital signs: Current cigarette United States, 20 05 2010. Morbidity and Mortality Weekly Report 60(35), 1207 1212. Centers for Disease Control and Prevention. (2012). Current tobacco use among middle and high school students United States, 2011, Morbidity and Mortality Weekly Report, 61(31), 581 585. Chassin, L., Presson, C., Morgan Journal of Applied Developmental Psychology, 28, 264 276. doi: 10.1016/j. appdev.2007.02.005

PAGE 105

105 Costa, F.M., Jessor, R., Turbin, M.S., Dong, Q., Zhang, H., & Wang, C. (2005). The role of social contexts in adolescence: Context protection and context risk in the United States and China. Applied Developmental Science, 9(2), 67 85. Cu rtis, A.C., Waters, C.M., & Brindis, C. (2010). Rural adolescent health: The importance of prevention services in the rural community. The Journal of Rural Health, 27, 60 71. Derzon, J.H. (2000 ) A synthesis of research on predictors of youth alcohol, toba cco, and marijuana use, I mprovi ng Prevention Effectiveness (pp.163 171). deVaus, D. (2010). Research design in social research Thousand Oaks, CA: SAGE. xposure to a tobacco preventi on media campaign in rural and low population density communities. American Journal of Public Health, 99(12), 2210 2216. Ennett, S.T., Foshie, A.F., Bauman, K.E., Hussong, A., Faris, R., Hipp, J.R., & Cai, L. (2010). A coial con textual analysis of youth cigarette smoking development. Nicotine & Tobacco Research, 12(9), 950 962. Family Smoking Prevention and Tobacco Control Act P ub. L. 111 31, div. A, 123 Stat. 1776 (2009). Florida Department of Children and Families. (201 1 ). Flo rid a Youth Substance Abuse Survey State Report Retrieved from http://www.dcf.state.fl.us/programs/samh/publications/fysas/11Survey/2011State Repo rtFinalv3.pdf Florida Department of Children and Families. (201 3 ). Florid a Youth Substance Abuse Survey State Report Retrieved from http://www.myflfamilies.com/service programs/substance abuse/fysas/2 012 Florida Department of Health, Bureau of Tobacco Prevention Program. (2009). smoking in Florida?. Retrieved from http://doh.state.fl.us/tobacco/PDF_Files/ WhosSmokingI nFlorida .pdf Hahn, E.J., Rayens, M.K., Chaloupka, F.J., Chizimuzo, O.T.C., & Yang, J. (2002). Projected smoking related deaths among U.S. youth: A 2000 update. Chicago: University of Illinois at Chicago. Retrieved from: http://www.impacteen.org/generalarea_PDFs/Hahn_researchpaper22_May2002. pdf Hodge, D.R., Cardenas, P., & Montoya, H. (2001). Substance use: Spirituality and religious participation as protective facto rs among rural youths. Social Work Research, 25(3), 153 161. doi: 10.1093/swr/25.3.153

PAGE 106

106 Jessor, R. (1991). Risk behavior in adolescence: A psychosocial framework for understanding and action. Journal of Adolescent Health, 12, 597 605. Jessor, R. (1992). REP LY: Risk behavior in adolescence : A psychosocial framework for understanding and action. Developmental Review, 12, 374 390. Jessor, R. (1993). Successful adolescent development among youth in high risk settings. American Psychologist, 48(2), 117 126. Jesso r, R., Turbin, M.S., Costa, F.M., Dong, Q., Zhang, H., & Wang, C. (2003). Adolescent problem behavior in China and the United States: A cross national study of psychosocial protective factors. Journal of Research on Adolescents, 13(3), 329 360. Jessor, R. Van Den Bos, J., Vanderryn, J., Costa, F.M. & Turbin, M.S. (1995). Protective factors in adolescent problem behavior: Moderator effects and developmental change. Developmental Psychology, 31(6), 923 933. Lutfiyya, M.N., Shah, K.K., Johnson, M., Bales, R.W ., Cha I., McGrath, C., Serpa, L., & Lipsky, M.S. (2008). Adolescent daily cigarette smoking: is rural residency a risk factor?. Rural and Remote Health, 8, 1 12. Mertens, D.M. (2010). Research and evaluation in education and psychology: Integrating diver sity with quantitative, qualitative and mixed methods Thousand Oaks, CA: SAGE. Mickelson, K.D., Kessler, R.C., & Shaver, P.R. (1997). Adult attachment in a nationally representative sample. Journal of Personality and Social Psychology, 73(5), 1092 1106. Pentz, M.A. (1999). Effective programs for tobacco use. Nicotine & Tobacco Research, 1, S99 S107. Philip M orris. (1969). Why one smokes. (Document Bates No. 1003287836) A rchives of Tobacco Documents Online. Retrieved from http://tobaccodocuments.org/landman/182914.html Racz, S.J., McMahon, R.J., & Luthar, S.S. (2011). Risky behavior in affluent youth: Examining the co occurrence and consequences of multiple problem behaviors. Journal of Child F amily Studies, 20, 120 128. doi: 10.1007/s10826 010 9385 4 Rhew, I.C., Hawkins, J.D., & Oesterle, S. (2011). Drug use and risk among youth in different rural contexts. Health & Place, 17, 775 783. doi: 10.1016/j.healthplace.2011.02003 Ritchey, P.N., Reid, G.S., & Hasse, L.A. (2001). The relative influence of smoking on drinking and drinking on smoking among high school students in a rural tobacco growing county. Journal of Adolescent Health, 29, 386 394.

PAGE 107

107 Rubin, A. (2013). Statistics for evidence based pract ice and evaluation 3 rd ed Belmont, CA: Brooks/Cole. Shettler, L.C. (2005). Risk and protective factors for adolescent smoking in rural versus urban environments Turbin, M.S., Jessor, R., & Costa, F.M. (2000). Adolescent cigarette smo king: Health related behavior or normative transgression? Prevention Science, 1(3), 115 124. U.S. Department of Commerce. U.S. Census Bureau. (2000). The urban a nd rural classifications. Geographic areas reference manual, (12) 1 24 U.S. Department of Heal th and Human Services. (2006). The health consequences of involuntary exposure to tobacco smoke: A rep ort of the Surgeon General. Atlanta, GA: U.S. Department of Health and Human Services. Retrieved from: http://www.surgeongeneral.gov/library/reports/secondhandsmoke/chapter1 U.S. Department of Health and Human Services, (2012). Preventing tobacco use among youth and young adults: A report of the Surgeon General. Retrieved from http://www.surgeongeneral.gov/library/reports/preventing youth tobacco use/index.html Woodhouse, L.D., Sayre, J.J., & Livingood, W.C. (2001). Toba cco policy and the role of law enforcement in prevention: The value of understanding c ontext Qualitative Health Research 11(5), 682 692. Vazsonyi, A.T., Chen, P., Jenkins, D.D., Burcu, E., Torrente, G., & Shue, C.J. (2010). problem behavior theo ry: Cross national evidence from Hungary, the Netherlands, Slovenia, Spain, Switzerland, Taiwan, Turkey, and the United States. Developmental Psychology, 46(6), 1779 1791. doi: 10.1037/a0020682

PAGE 108

108 BIOGRAPHICAL SKETCH Tracy DeCubellis received a Bachelor of Science degree in education from Taylor University in Upland, Indiana. Tracy came to the graduate school at Universit y of Florida to pursue a Master of S cience degree in Family, Youth, and Community Sciences after teaching public school for 10 years, a nd serving as a Tobacco Prevention Specialist for 5 years.