Add Health Dataset Research Articles

A longitudinal examination of social connectedness and suicidal thoughts and behaviors among adolescents.

This study examines the relationship between three different types of social connectivity and suicidal thoughts and behaviors. Using the Add Health dataset, three domains of social connection were explored: parental connection, school connection, and social integration. Logistic regression was used to examine whether changes over time in connectedness predicted suicidal thoughts and behavior. Youth whose difference scores on social integration and parental connectedness increased were less likely to experience suicidal ideation. Increases in difference scores for perceived school connectedness protected youth who reported ideation from engaging in a suicide attempt. Perceptions of social connection are key factors in understanding adolescent suicidal thoughts and behaviors. It is important to consider social connection across different relationship contexts.

Do Personality Traits Mediate the Relationship Between Childhood Abuse and Migraine? An Exploration of the Relationships in Young Adults Using the Add Health Dataset

Personality traits (especially neuroticism) and childhood maltreatment have been independently related to many negative health outcomes later in life, including migraine. Studies have also shown the association between childhood maltreatment and maladaptive personality traits. The mediating role of personality traits on the relationship between childhood maltreatment and depression, psychological distress, and alcohol dependence has been extensively studied. However, this type of mediation has not been studied in the case of the development of migraine. This study investigated (1) the main effects of childhood abuse on personality traits, and of personality traits on migraine, and (2) the mediating role of neuroticism, on the relationship between childhood abuse and migraine in young adults. We analyzed retrospective, cross-sectional data from 13,493 adults aged 24-32 years in Wave 4 of the National Longitudinal Survey of Adolescent Health ("Add Health") data set. Participants were queried regarding maltreatment (emotional, physical, and sexual) during childhood, current Big Five personality traits (using mini International Personality Item Pool), current depression (using Center for Epidemiologic Studies-Depression Scale), perceived stress (Using Cohen's Perceived Stress Scale), and diagnosis of migraine by a health care provider. Linear and logistic regressions were used to assess the main effects of childhood maltreatment on the five personality traits (openness to experience, conscientiousness, extraversion, agreeableness, and neuroticism) and the main effect of the personality traits on self-reported provider diagnosis of migraine. A structural equation model (SEM) was used to examine the mediating role of neuroticism on the relationship between childhood maltreatment and migraine. Linear regression models showed that childhood abuse independently predicted increased neuroticism (β = 0.338, SE=±0.05, P < .001), and increased openness to experiences (β = 0.341, SE = ±0.06, P < .001) after adjusting for socio-demographic characteristics, current depression, and perceived stress. Logistic regression to examine the main effect of personality traits on migraine revealed that only neuroticism had a significant effect (OR = 1.07, 95%CI = 1.04-1.10) after controlling for childhood abuse, socio-demographic characteristics, current depression, and perceived stress. Our regression analyses showed that neuroticism, but not openness to experience, was a potential mediator for the relationship between childhood abuse and migraine. SEM confirmed significant mediation of the relationship between childhood abuse and migraine through neurotic personality traits (goodness of fit indices: CFI = 0.992, TLI = 0.979, RMSEA = 0.025), unadjusted for socio-demographic variables, depression, and stress. In addition to the indirect effect (β = 0.039, P < .01) of childhood abuse on migraine through neuroticism, there was also a significant direct effect (β = 0.143, P < .01) of childhood abuse on migraine. After adjusting for socio-demographic variables, other personality types, depression, and stress, both the direct effect (β = 0.127; P < .01) of childhood abuse on migraine and the indirect effect (β = 0.09; P < .01) of childhood abuse on migraine through neuroticism were attenuated, but remained significant. Childhood abuse is associated with personality and migraine. An estimated 21% of the total effect of childhood abuse on migraine could be explained by mediation through neuroticism in the unadjusted model. In the fully adjusted model, an estimated 8.7% of the total effect could be explained by mediation, although, self-reported data limit the ability to draw firm conclusions.

Estimation of social-influence-dependent peer pressure in a large network game

Research on peer effects in sociology has been focused for long on social influence power to investigate the social foundations for social interactions. This paper extends Xu(2011)'s large--network--based game model by allowing for social-influence-dependent peer effects. In a large network, we use the Katz--Bonacich centrality to measure individuals' social influences. To solve the computational burden when the data come from the equilibrium of a large network, we extend Aguirregabiria and Mira (2007)'s nested pseudo likelihood estimation (NPLE) approach to our large network game model. Using the Add Health dataset, we investigate peer effects on conducting dangerous behaviors of high school students. Our results show that peer effects are statistically significant and positive. Moreover, a student benefits more (statistically significant at the 5% level) from her conformity, or equivalently, pays more for her disobedience, in terms of peer pressures, if friends have higher social-influence status.

Emotional Abuse History and Migraine Among Young Adults: A Retrospective Cross-Sectional Analysis of the Add Health Dataset.

To define and examine the relationship between self-reported childhood abuse and migraine among young adults. Headache and migraine have been linked to childhood abuse in numerous studies, but there is incomplete characterization of headache types, and limited assessment of abuse types and frequency. Only one population-based study has examined the relationship between emotional abuse and migraine. None have investigated the temporal relationship between onset of abuse and of migraine. We analyzed data from 14,356 adults aged 24-32 years in Wave 4, which is a cross-sectional subset of the longitudinal Add Health study. Participants were queried regarding abuse (emotional, physical and sexual) during childhood, diagnosis of migraine, depression and anxiety by healthcare providers, and symptoms of current depression. We used logistic regression to estimate the association between childhood abuse and migraine, controlling for socio-demographic factors, current depression, and lifetime diagnosis of anxiety and depression. About 14% (n = 2040) of respondents reported migraine. Participants with migraine (vs no migraine) reported significantly higher rates of childhood abuse overall (60.6% vs 48.9%), including emotional (57.8% vs 45.4%), sexual (8.4% vs 4.6%) and physical (22.4% vs 17.9%) abuse. Emotional abuse had a stronger association with migraine (odds ratio [OR] 1.62; 95% confidence interval [CI] 1.43-1.85) when compared with physical (OR 1.06; 95% CI 0.89-1.68) and sexual abuse (OR 1.06; 95% CI 0.93-1.68), adjusting for socio-demographic factors. The emotional abuse-migraine association remained even when controlling for lifetime diagnosis of depression and anxiety (OR 1.37; 95% CI 1.19-1.57) and for current depression (OR 1.47; 95% CI 1.30-1.67). The odds of migraine increased with increasing number of abuse types reported. There was a U-shaped distribution of odds of migraine associated with frequency of occurrences of emotional abuse, peaking at one time (OR 1.65; 95% CI 1.34-2.03) and ≥ six times (OR 1.77; 95% CI 1.49-2.10). Emotional abuse during childhood contributed more than physical or sexual abuse to the development of migraine. There is a dose-response relationship with increasing number of abuse types associated with rising odds of migraine. In addition, the relationship between the frequency of emotional abuse and the odds of migraine follows a U-shape pattern.

Exploring the behavior of juveniles and young adults raised by custodial grandmothers

The present study compares behaviors of youth raised by a custodial grandmother, biological mother, or other female caregiver from adolescence to young adulthood using data from Waves I, II, and III of the National Longitudinal Study of Adolescent Health (Add Health) data set. Caregiver type was not significantly related to antisocial behavior. However, respondents who were raised by a custodial grandmother reported being stopped by police more times than those respondents raised by their biological mothers. The effect of parenting characteristics such as warmth, control, and involvement by a custodial grandmother may operate differently than in a biological mother-headed household.

Emotional Abuse History and Migraine among Young Adults: Results from The Add Health Dataset (S26.003)

Emotional Abuse History and Migraine among Young Adults: Results from The Add Health Dataset (S26.003)

Emotional Abuse History and Migraine among Young Adults: Results from The Add Health Dataset (I15.010)

Emotional Abuse History and Migraine among Young Adults: Results from The Add Health Dataset (I15.010)

Interpersonal Neurobiology and Couple Relationship Quality: A Longitudinal Model

Interpersonal neurobiology is an interdisciplinary theory that is being used by couple and family therapists. However, more investigation is needed to determine if this theory can accurately explain family relationships over time. In this study, we used the nationally representative National Longitudinal Study of Adolescent to Adult Health (Add Health) dataset to examine key elements of interpersonal neurobiology theory. Using structural equation modeling, we explored how maternal and paternal support and self-acceptance during adolescence are related to emotional stability and adult couple relationship satisfaction 13 years later (N = 7864). Results indicate that self-acceptance and emotional stability are factors that contribute to couple relationship satisfaction. We conclude with clinical implications of this study, including how to use mindfulness practices to promote healthy relationships.

Estimation of Social-Influence-Dependent Peer Pressures in a Large Network Game

Research on peer effects in sociology has been focused for long on social influence power to investigate the social foundations for social interactions. This paper extends Xu (forthcoming)’s large–network–based game model by allowing for social–influence–dependent peer effects. In a large network, we use the Katz–Bonacich centrality to measure individuals’ social influences. To solve the computational burden when the data come from the equilibrium of a large network, we extend Aguirregabiria and Mira (2007)’s nested pseudo likelihood estimation (NPLE) approach to our large network game model. Using the Add Health dataset, we investigate peer effects on conducting dangerous behaviors of high school students. Our results show that peer effects are statistically significant and positive. Moreover, a student benefits more (statistically significant at the 5% level) from her conformity, or equivalently, pays more for her disobedience, in terms of peer pressures, if friends have higher social-influence status.

The Nitty Gritty: The Unexplored Role of Grit and Perseverance in Voter Turnout

In this paper we consider the role of general motivation, distinct from political motivation, in predicting voter turnout. Drawing on the large-scale, longitudinal Add Health dataset, we show that a measure of grit (i.e., tenacity or perseverance) measured in adolescence is a strong predictor of voter turnout in adulthood. This relationship exists independent of cognitive skills, personality measures, political interest, and many other observed and unobserved individual and contextual factors. To examine the mechanisms behind the observed relationship, we conduct a series of analyses including an original survey experiment that primes a gritty mindset. The results suggest that grittier individuals vote more because they are willing to overcome the obstacles that get in the way of voting. Compared to a control group, individuals primed to have a gritty mindset report a willingness to spend an average of 12 additional minutes waiting in line to vote and report higher levels of political efficacy.

Research note: The consequences of different methods for handling missing network data in stochastic actor based models

Although stochastic actor-based models (e.g., as implemented in the SIENA software program) are growing in popularity as a technique for estimating longitudinal network data, a relatively understudied issue is the consequence of missing network data for longitudinal analysis. We explore this issue in our research note by utilizing data from four schools in an existing dataset (the AddHealth dataset) over three time points, assessing the substantive consequences of using four different strategies for addressing missing network data. The results indicate that whereas some measures in such models are estimated relatively robustly regardless of the strategy chosen for addressing missing network data, some of the substantive conclusions will differ based on the missing data strategy chosen. These results have important implications for this burgeoning applied research area, implying that researchers should more carefully consider how they address missing data when estimating such models.

Juvenile Justice System Outcomes Among Foreign-Born and Native-Born Latinos in the United States

A growing body of research has identified a negative relationship between generational status and criminological outcomes such that foreign-born Latinos are significantly less likely to report offending, victimization, and drug use compared to their native-born counterparts. What has been explored to a lesser degree is the extent to which generational status impacts the experiences of Latino youth within the juvenile justice system. Using the Add Health data set, this article explores the prevalence of juvenile court involvement among foreign (i.e., first generation) and native-born (i.e., second generation or higher) Latino youth as well as the types of offenses for which they were adjudicated delinquent. Results suggest that significant differences exist between the foreign and native-born both in terms of juvenile court involvement and offense types. These findings are discussed relative to the extant literature and juvenile justice policy and practice.

Sports participation and suicidal behaviour: Does sport type matter?

Using the Add Health data set of 6485 adolescents aged 11-21 years, it was found that whether participation in sports was a protective factor or a risk factor for suicidal ideation depended on the particular sport, as well as the sex and ethnicity of the adolescents. In general, team sports was associated with a protective effect, while participating in a sport outside of the gender norms (for example, cheer leading/dance for boys and wrestling for girls) was a risk factor. The results were clearer for White adolescents and less clear for Black, Asian, Hispanic, and Native American adolescents.

Socioeconomic disparities in sexually transmitted infections among young adults in the United States: examining the interaction between income and race/ethnicity.

There is considerable evidence of racial/ethnic patterning of sexually transmitted infection (STI) risk in the United States. There is also evidence that poorer persons are at increased STI risk. Evidence regarding the interaction of race/ethnicity and income is limited, particularly nationally at the individual level. We examined the pattern of socioeconomic gradients in STI infection among young people in a nationwide US study and determined how these gradients varied by race/ethnicity. We estimated the cumulative diagnosis prevalence of chlamydia, gonorrhea, or trichomoniasis (via self-report or laboratory confirmation) for young adults (ages, 18-26 years old) Hispanics and non-Hispanic whites, blacks, and others across income quintiles in the Add Health data set. We ran regression models to evaluate these relationships adjusting for individual- and school-level covariates. Sexually transmitted infection diagnosis was independently associated with both racial/ethnic identity and with low income, although the racial/ethnic disparities were much larger than income-based ones. A negative gradient of STI risk with increasing income was present within all racial/ethnic categories, but was stronger for nonwhites. Both economic and racial/ethnic factors should be considered in deciding how to target STI prevention efforts in the United States. Particular focus may be warranted for poor, racial/ethnic minority women.

Social Networks and the Identification of Peer Effects

There is a large and growing literature on peer effects in economics. In the current article, we focus on a Manski-type linear-in-means model that has proved to be popular in empirical work. We critically examine some aspects of the statistical model that may be restrictive in empirical analyses. Specifically, we focus on three aspects. First, we examine the endogeneity of the network or peer groups. Second, we investigate simultaneously alternative definitions of links and the possibility of peer effects arising through multiple networks. Third, we highlight the representation of the traditional linear-in-means model as an autoregressive model, and contrast it with an alternative moving-average model, where the correlation between unconnected individuals who are indirectly connected is limited. Using data on friendship networks from the Add Health dataset, we illustrate the empirical relevance of these ideas.

MODELING THE ASSOCIATION BETWEEN ACADEMIC ACHIEVEMENT AND DELINQUENCY: AN APPLICATION OF INTERACTIONAL THEORY

Many studies have addressed whether delinquent behavior is associated with various aspects of schooling and academics. However, this research has been limited to examining unidirectional effects. Building on Thornberry's interactional theory, we develop a conceptual model that posits reciprocal associations among delinquent behavior, school attachment, and academic achievement. The model is tested with two waves from the Add Health data set (n = 9,381) that include measures of transcript grade point average (GPA). The results of a set of structural equation models provide evidence that academic achievement is associated with less delinquent behavior over time, as well as with higher school attachment. However, the effects of delinquency are limited to an attenuating effect on subsequent school attachment; delinquency does not directly influence academic achievement. Thus, we find only partial support for interactional theory.

Dropping out of high school: Effects of close and distant friendships

Prior research highlights the role of friends in influencing whether a student completes high school. Students who drop out tend to have fewer friends, as well as friends who are less oriented toward school success. We distinguish between close and distant friendships by developing a theoretical framework which predicts close and distant friends likely have distinct effects on dropping out. Close friendships provide valuable emotional support, and forging numerous close friendships at school should decrease one’s risk of dropping out. In contrast, the characteristics of distant friends help shape students’ social identities and beliefs about “what’s normative.” Our analyses of the Add Health data set confirm our expectations. Students with more close friendships are less likely to drop out, but close friends’ characteristics are unrelated to dropping out. Distant relationships (as measured by affect and regularity of interaction) with friends who have a high risk of dropping out significantly increase a student’s own risk of dropping out.

Mothers, Friends and Gender Identity

This paper explores a novel mechanism of gender identity formation. Specifically, we explore how the work behavior of a teenager's own mother, as well as that of her friends' mothers, affect her work decisions in adulthood. The first mechanism is commonly included in economic models. The second, which in social psychology is also emphasized as an important factor in gender identity formation, has so far been overlooked. Accordingly, our key theoretical innovation is how the utility function is modeled. It is assumed that an adult woman's work decisions are influenced by her own mother's choices as well as her friends' mothers' choices when she was a teenager, and the interaction between the two.The empirical salience of this behavioral model is tested using a network model specification together with the longitudinal structure of the AddHealth data set. We find that both intergenerational channels positively affect a woman's work hours in adulthood, but the cross effect is negative, indicating the existence of cultural substitutability. That is, the mother's role model effect is larger the more distant she is (in terms of working hours) from the friends' mothers.

Measuring Narcissism Within Add Health: The Development and Validation of a New Scale

This study reports the development of a measure of narcissism within the National Longitudinal Study of Adolescent Health (Add Health) data set. In Study 1, items were selected from Wave III to form the Add Health Narcissism Scale (AHNS). These were factor analyzed, yielding a single factor comprised of five subscales. We correlated the AHNS and its subscales with constructs within Add Health known to be associated with narcissism. Finally, we used the scale to replicate the relationship between narcissism and violence. In Study 2, a separate sample completed the AHNS, as well as other measures, including the Narcissistic Personality Inventory. These studies confirmed that the AHNS is a valid measure for use in narcissism‐related studies employing the Add Health data.

The Moderating Effects of Risk, Protection and Desistance on Violence in Female Adolescents

The major focus of this study was to test the theory driven conceptual model derived from the literature regarding risk, protective and desistance factors in relation to violent behavior of female adolescents. This study utilized The National Longitudinal Adolescent Health Study (Add Health) dataset and included 2,031 female participants. The conceptual model developed was based on life course perspective as it relates to violent behavior in adolescents. The tested SEM model suggested: (1) that each of the latent variables of risk (β = .06, t = .02, p .05), was not significantly directly associated with violent behavior, (3) there were mediating effects between risk factors, protective factors and violent behavior (β = −.12, t = .04, p .05).