The effects of self-esteem and academic performance on youth decision-making: An examination of sexual debut and illegal substance use Stephanie B. Wheeler, MPH University of North Carolina at Chapel Hill Cecil G Sheps Center for Health Services Research Background Early sexual intercourse related to: STDs Unwanted pregnancy Depression, anxiety, eating disorders Illegal use of controlled substances in youth related to: Substance dependency Motor vehicle accidents Injury from violence Criminal activity Mental health problems High school drop-out Determinants of risky behaviors among young people largely remain unknown Background Self-esteem, perceived intelligence, and academic performance may play roles in decision-making Self-esteem “Acting out” to help fill a void Confidence and self-efficacy required to say “no” Feeling invincible, immune to risk, living on the cutting edge Academic achievement Better knowledge about risks Heightened self-efficacy as result of higher perceived intelligence Having a positive outlook for one’s academic future Conceptual Framework Social Competency/Individual Deficiency theory (Norman and Turner, 1993) High risk behavior in young people associated with: Low self-esteem Inability to solve problems Difficulties with communication Inappropriate values Attempts to seek peer approval Self-esteem may be determined in part or mediated by personal, familial, and environmental factors SOCIETY Race/ethnicity COMMUNITY Gender COMPETENT Age Religion Values PERCEIVED SELF-WORTH (SELF-ESTEEM) Relationships Achievements Self-efficacy Communication Skills DEFICIENT RISKY BEHAVIOR Research Questions Does higher self-esteem at baseline Reduce the likelihood of sexual debut? Reduce the likelihood of illegal substance use? Does higher academic performance at baseline Reduce the likelihood of sexual debut? Reduce the likelihood of illegal substance use? Methods Retrospective, secondary analysis National Longitudinal Study of Adolescent Health (aka Add Health, public version) Population American youth enrolled in the 7th through 12th grades in 52 middle schools and 80 high schools Sampled originally in 1994-1995 academic years (baseline) Followed up in 1996 (wave II) and 2002-2003 (wave III) Adolescent in-home survey data only Waves I-III contain data from approximately 6500 individuals ages 11-26 Methods Face-to-face interviews and computer-assisted personal interview devices Exclusion criteria for current study: Individuals not available for sampling in all three waves Individuals who reported having had sex at baseline (sexual debut models only) Stratified multivariate logistic regression (by gender) Controlling for known confounders Key Variables Dependent Variables Sexual debut Illegal substance use Drug use Underage alcohol use Underage cigarette smoking Independent Variables Academic performance Self-esteem Ascending scale of 10 Likert-scaled survey items Range: 14-44 (median score: 33) Self Esteem Measurement Questions related to personal confidence and satisfaction with life: - “I have a lot of good qualities” - “I have a lot to be proud of” - “I like myself just the way I am” - “I feel I am doing everything just about right” - “I feel socially accepted” - “I feel I am just as good as other people” - “I enjoy life” Perceived intelligence relative to peers: - “Compared with other people your age, how intelligent are you?” Overall satisfaction with weight/size: - “How do you think of yourself in terms of weight?” - “Are you trying to lose weight, gain weight, or stay the same?” Results: Descriptive Statistics at Wave 1 Variable % or Mean SD or Range Sex by Wave I 33% .47 Drug use in Wave I 26% .44 Underage alcohol use in Wave I 52% .50 Self esteem in Wave I 32.7 15-42 A student (>/=3.5 GPA) in Wave I 43% .50 B student (>/=2.5-3.5 GPA) in Wave I 37% .48 C or lower (< 2.5 GPA) in Wave I 20% .38 PVT standardized test score Wave I 100.5 10-137 Age in Wave I 15.6 11-21 Race – White 64% .48 Race – Black 21% .41 Race – Other 15% .35 Selected Results: Adjusted Odds Ratios for Sexual Debut Females OR Males OR Wave II Wave II “A” student (defined as >=3.5 GPA) 0.49** 0.63+ “B” student (defined as >=2.5, <3.5 GPA) 0.81 0.72 Self-esteem 0.98 1.01 Age14 2.29* 1.17 Age15 3.75** 1.23 Age16 5.51** 2.25* Age17 6.16** 3.06** Age Over17 6.65** 4.68** Race African American 1.31 2.12** Influence of religion 0.90* 0.90* Having taken abstinence pledge 0.68* 0.65 Ever in romantic relationship 2.33** 2.17** Lives with one parent 1.668* 1.490 Observations 1255 974 + significant at 10%; * significant at 5%; ** significant at 1% Selected Results: Adjusted Odds Ratios for Illegal Drug Use Females OR Males OR Females OR Males OR Wave II Wave II Wave III Wave III “A” student (defined as >=3.5 GPA) 0.61** 0.89 0.92 1.01 “B” student (defined as >=2.5, <3.5 GPA) 0.57** 0.88 1.07 0.86 Self-esteem 0.96** 0.98 0.99 0.99 Prior use of substances in Wave I 11.38** 10.10** 3.44** 3.09** Age14 1.01 2.29* 0.80 0.89 Age15 1.39 1.65 0.84 0.70+ Age16 0.81 2.06* 0.71+ 0.60* Age17 0.88 2.00* 0.48** 0.48** Age Over17 0.73 2.52* 0.42** 0.40** Race African American 0.72 2.02** 0.91 0.85 Race Other 1.24 1.56* 0.93 1.08 Influence of religion 0.90** 0.92* 0.95* 0.93** Ever in romantic relationship 1.44** 1.42* 1.09 1.28* Lives in urban area 1.41* 1.11 1.18+ 0.98 Observations 1825 1517 2245 1850 + significant at 10%; * significant at 5%; ** significant at 1% Summary and Discussion Academic performance related to sexual debut and illegal substance use, but effect varies by gender and over time How to interpret significance of academic performance No strong/consistent effect of self-esteem Measurement of self-esteem Younger females appear to be more sensitive to these effects Consistent associations between age, influence of religion, relationship experience and risky behaviors Strengths Large national dataset Rich information about sensitive topics and youth behavior Estimated effect of wave I variables on wave II/III outcomes Limitations Observational study Possible omitted variable bias (risk tolerance, opportunity, sense of future, parental emphasis on avoiding sex/drugs, etc.) Implications Continued efforts to improve academic performance and selfconfidence in youth are important, particularly at young ages, and may have important long term health effects Measurement of self-esteem is complicated Non-linear, dynamic nature of self esteem Future research should explore the mechanisms through which academic achievement and self-esteem are created Peer network modeling could provide additional insight • Sally Stearns • Edward Norton • Sarah Birken • Michaela Dinan • Heather Beil Control Variables Age Race Influence or importance of religion Employment Ever in a romantic relationship Community-level socioeconomic/demographic variables Any previous psychological treatment Residence with both, one, or neither biological parent Substance use models only: • • Urban residence Prior substance use in Wave I Sexual debut models only: • • Ever taken a public abstinence pledge Height