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.     

MULTIPLE IMPUTATION OF INCOMPLETE CATEGORICAL DATA USING LATENT CLASS ANALYSIS

We propose using latent class analysis as an alternative to log-linear analysis for the multiple imputation of incomplete categorical data. Similar to log-linear models, latent class models can be used to describe complex association structures between the variables used in the imputation model. However, unlike log-linear models, latent class models can be used to build large imputation models containing more than a few categorical variables. To obtain imputations reflecting uncertainty about the unknown model parameters, we use a nonparametric bootstrap procedure as an alternative to the more common full Bayesian approach. The proposed multiple imputation method, which is implemented in Latent GOLD software for latent class analysis, is illustrated with two examples. In a simulated data example, we compare the new method to well-established methods such as maximum likelihood estimation with incomplete data and multiple imputation using a saturated log-linear model. This example shows that the proposed method yields unbiased parameter estimates and standard errors. The second example concerns an application using a typical social sciences data set. It contains 79 variables that are all included in the imputation model. The proposed method is especially useful for such large data sets because standard methods for dealing with missing data in categorical variables break down when the number of variables is so large.     

Network Canvas: Key decisions in the design of an interviewer-assisted network data collection software suite

Self-reported social network analysis studies are often complex and burdensome, both during the interview process itself, and when conducting data management following the interview. Through funding obtained from the National Institute on Drug Abuse (NIDA/NIH), our team developed the Network Canvas suite of software – a set of complementary tools that are designed to simplify the collection and storage of complex social network data, with an emphasis on usability and accessibility across platforms and devices, and guided by the practical needs of researchers. The suite consists of three applications: Architect: an application for researchers to design and export interview protocols; Interviewer: a touch-optimized application for loading and administering interview protocols to study participants; and Server: an application for researchers to manage the interview deployment process and export their data for analysis. Together, they enable researchers with minimal technological expertise to access a complete research workflow, by building their own network interview protocols, deploying these protocols widely within a variety of contexts, and immediately attaining the resulting data from a secure central location. In this paper, we outline the critical decisions taken in developing this suite of tools for the network research community. We also describe the work which guides our decision-making, including prior experiences and key discovery events. We focus on key design choices, taken for theoretical, philosophical, and pragmatic reasons, and outline their strengths and limitations.     

Who Is Adopted?

ABSTRACT

There are several advantages of using large representative samples to study adoption-related issues, and existing databases are increasingly used for this purpose. Large scale surveys are often limited, however, in the number and type of questions that are asked regarding adoption. Conceptual and methodological issues in measuring adoption status are discussed, and a rich new database-the National Longitudinal Survey of Adolescent Health (Add Health)-is used to illustrate these complexities. In Add Health school survey data, adolescents were asked a direct question about whether or not they were adopted and whether they lived with biological parents. During home interviews, adolescents were not asked about adoption status per se, but they could be classified as adopted if they lived with adoptive rather than biological parents. Parents' and adolescents' reports of adoption status and living arrangements are compared in Add Health data to illustrate the complexities of measuring adoption status when using archival data. For research to be scientifically credible and to enhance the accumulation of knowledge, investigators need to be precise about the concepts and measures they use in studying adoption-related issues, particularly when using previously collected survey data.     

Investor commitment to serial entrepreneurs: A multilayer network analysis

Social networks are complex systems composed of interdependent organizations and people with diverse network structures. Understanding network dynamics, such as exchange commitment, requires a methodological toolkit that does not assume away complexity. In this study, we extend a technique for analyzing longitudinal, multilayer network data called network alignment. We introduce a novel metric – intersect proportions – for analyzing similarity between divergent graphs. We demonstrate the application of network alignment and intersect proportions to the context of investor commitment to startups and entrepreneurs. Using this technique, we are able to disentangle exchange commitment across complex networks.     

Factors Associated With Perceptions of Family Belonging Among Adolescents

Adolescents' perceptions of family belonging are associated with several well‐being indicators, yet we know little about which factors influence these perceptions or how they differ by family structure. The current study uses nationally representative data from Add Health to examine predictors of adolescents' perceptions of family belonging in two‐biological‐parent families (n = 9,686). The results are compared to a recent study using Add Health that examined family belonging in married mother–stepfather families. Findings suggest both similarities and differences across family structure in the factors associated with family belonging.     

UNTANGLING A COMPLEX MEDIA SYSTEM

This article provides empirical insights into how one online service – Twitter – was used for political purposes during three separate election campaigns in Sweden, Denmark and Norway, specifically how Twitter users, with hyperlinks, connect with other channels for political communication. Methodologically, the study employs three large sets of data on Twitter use tagged as relevant for each of the election campaigns, covering a one-month period. The approach allows for an untangling of the complex interconnections between novel online services, mainstream media, official political party websites, public information, individual blogs and social network sites. By moving beyond a study merely of the type of websites linked to, to also include classification of the actors publishing the content linked to, the article provides insights into the actual use by politicians, interest groups as well as grassroots activists of diverse Web genres.     

Estimating network properties from snowball sampled data

This article addresses the estimation of topological network parameters from data obtained with a snowball sampling design. An approximate expression for the probability of a vertex to be included in the sample is derived. Based on this sampling distribution, estimators for the mean degree, the degree correlation, and the clustering coefficient are proposed. The performance of these estimators and their sensitivity with respect to the response rate are validated through Monte Carlo simulations on several test networks. Our approach has no complex computational requirements and is straightforward to apply to real-world survey data. In a snowball sample design, each respondent is typically enquired only once. Different from the widely used estimator for Respondent-Driven Sampling (RDS), which assumes sampling with replacement, the proposed approach relies on sampling without replacement and is thus also applicable for large sample fractions. From the simulation experiments, we conclude that the estimation quality decreases with increasing variance of the network degree distribution. Yet, if the degree distribution is not to broad, our approach results in good estimates for the mean degree and the clustering coefficient, which, moreover, are almost independent from the response rate. The estimates for the degree correlation are of moderated quality.     

The Relationship Between Cognitive Ability and Experiences of Vaginal,Oral, and Anal Sex in the United States

Few studies have investigated the sexual development of populations with low cognitive abilities in the United States. This article examines the relationship between cognitive ability and various sexual experiences from adolescence (ages 12 to 18) to early adulthood (ages 28 to 34). Data were from 13,845 respondents interviewed at Waves I and IV of the National Longitudinal Study of Adolescent to Adult Health (Add Health), a probability sample of adolescents in the United States followed from adolescence to adulthood. Adjusted logistic regression models were used to study relationships between cognitive ability, approximated by the Add Health Picture Vocabulary Test (AHPVT), and experiences of vaginal, oral, and anal sex. After controlling for biological sex, age, race/ethnicity, and socioeconomic status (SES), individuals in the lowest cognitive ability group had significantly lower odds of experiencing each type of sex than those in the average ability group. Although individuals in the highest cognitive ability group had significantly lower odds of experiencing vaginal intercourse than those in the average ability group, this association did not remain significant when analyses were stratified by biological sex. These differences in experiences have implications for future health and warrant further study to understand policy implications for sexual health services and education.     

The emergence of the emergence-based approach in community practice

Community practitioners are often challenged to work in complex environments and situations that naturally occur in communities. This article focuses on the theoretical and practical use of the Emergence-Based Approach that is relevant for work in complex community situations. Using this approach, the practitioner encourages the emergence of novel ideas and initiatives, and engages a decentralized network of activists, helping them develop their initiatives. Using a field study, we compare this approach to the traditional Outcomes-Based Approach. Although their mindsets are very different, and even contradict one another, we suggest practical ways to combine the two in an effective way.     

Board interlocking network and the design of executive compensation packages

The standard approach used to model interlocks in the business and management literature is to treat each interlock of a network as an independent data point. However, such an approach ignores the complex inter-dependencies among the common director interlocks. We propose that an interlocking board network is an important inter-corporate setting that has bearing on how company boards make corporate decisions. Using a sample of 725 large U.S.-based public companies over the period 2007–2010, board member information, executive compensation information, and exponential random graph modeling (ERGM) techniques for social networks, we present evidence that board interlocks are positively linked with similarities in the design of executive compensation packages in interlocked firms, particularly the proportions of the options component. We also find evidence that board interlocks are positively linked with similarities in a number of board characteristics.     

Social Background Differences in Early Family Behavior

Social background has historically been recognized as a major factor influencing family behavior, though recent work has largely emphasized racial/ethnic influences. Here we use 1994 – 1995 and 2001 – 2002 Add Health data to examine the cohabitation, first marriage, and first birth experience of young women. In a multistate life table context, hypothetical cohorts specified in terms of race and mother’s education are followed, from age 11 to age 24, as they move through 6 family‐related statuses. The results indicate that, for both Black and White women, a higher level of maternal education is generally associated with less cohabitation, less marriage, fewer first births, and a higher percentage of women who experience none of those transitions before age 24. Racial and social background differences are conceptually and empirically distinct. Because mother’s education is associated with substantially different trajectories of early family behavior for both Blacks and Whites, we argue that social background merits increased attention in research on contemporary American family patterns.     

THE ROMANTIC RELATIONSHIPS OF AFRICAN-AMERICAN AND WHITE ADOLESCENTS

Using data from the National Longitudinal Study of Adolescent Health (Add Health), we analyze African-American and white respondents' reports of the nature of their romantic experiences. We explore levels of companionate intimacy, including frequency of interaction and intimate self-disclosure, degree of romantic and sexual intimacy, as well as problem areas including relationship violence and nonexclusivity. African-American respondents report relationships of longer average duration relative to white youths, but overall responses reflect a somewhat less intense or intimate relationship style. These observed differences were not completely attenuated once relevant controls are introduced. We connect these findings to prior research on African-American adolescent family and peer relations, and call for additional research that supplements the behavioral emphasis of Add Health.     

Genetic and Environmental Influences on Pubertal Timing: Results From Two National Sibling Studies

This study examined genetic and environmental effects on individual variation in pubertal timing using two national samples of siblings from the Nonshared Environment of Adolescent Development (NEAD) and the National Longitudinal Study of Adolescent Health (Add Health). In each sample, female and male siblings with different degrees of genetic relatedness, i.e., monozygotic twins, dizygotic twins, full siblings, half siblings, and unrelated siblings in blended families, were assessed. Timing of pubertal development was measured by age-adjusted self-report measures of the Pubertal Development Scale in NEAD and a four-item scale of pubertal development in the Add Health. The results indicated that both genetic and environmental influences play an important role in determining the relative timing of pubertal development for both boys and girls.     

Social Closure and Educational Attainment

This article examines how network closure among parents affects adolescents’ educational attainment. First, we introduce a distinction between informal closure and school‐based closure. Second, we investigate whether and how the effect of informal and school‐based parental network closure varies across social contexts. Findings from the National Longitudinal Study of Adolescent Health (Add Health) and multilevel models show that parental network closure modestly impacts educational outcomes. Moreover, educational benefits of informal closure in parent networks are contingent on social context. Closure only benefits educational attainment in low‐poverty schools. In high‐poverty schools, informal closure in parent networks lowers educational attainment. The social closure generated in informal connections among parents thereby contributes to the encapsulation of disadvantage in areas of concentrated poverty, which is not the case for school‐based closure.     

Scaling bias in pooled exponential random graph models

Researchers often use pooled exponential random graph models (ERGM) to analyze samples of networks. However, pooled ERGM—here, understood to include both meta-regression and combined estimation on a stacked adjacency matrix—may be biased if there is heterogeneity in the latent error variance (‘scaling’) of each lower-level model. This study explores the implications of scaling for pooled ERGM analysis. We illustrate that scaling can produce bias in pooled ERGM coefficients that is more severe than in single-network ERGM and we introduce two methods for reducing this bias. Simulations suggest that scaling bias can be large enough to alter conclusions about pooled ERGM coefficient size, significance, and direction, but can be substantially reduced by estimating the marginal effect within a block diagonal or random effects meta-regression framework. We illustrate each method in an empirical example using Add Health data on 15 in-school friendship networks. Results from the application illustrate that many substantive conclusions vary depending on choice of pooling method and interpretational quantity.     

The impact of school connectedness and teacher support on depressive symptoms in adolescents: A multilevel analysis

Using data from Waves I and II of the National Longitudinal Study of Adolescent Health (Add Health), this study examined the association between school connectedness, teacher support and depressive symptoms in a weighted sample of 11,852 adolescents from 132 schools. To account for the nested data, multilevel regression was utilized. The results indicated that higher school connectedness and getting along with teachers were significantly associated with fewer depressive symptoms. Findings offer implications for school social work practice and future research. Suggestions for future research are described and strategies to enhance school connectedness and teacher support are discussed.     

Health and the structure of adolescent social networks

Much research has explored the role of social networks in promoting health through the provision of social support. However, little work has examined how social networks themselves may be structured by health. This article investigates the link between individuals' health and the characteristics of their social network positions. We first develop theoretical predictions for how health may influence the structure of adolescent networks. We then test these predictions using longitudinal analysis of the National Longitudinal Study of Adolescent Health (Add Health). We find important relationships between the health status of adolescents and the characteristics of the social network positions within which they are embedded. Overall we find that adolescents in poor health form smaller local networks and occupy less central global positions than their healthy peers. These results also have implications for social network research, expanding the scope of factors responsible for the network positions individuals occupy.     

ADOLESCENTS' INCONSISTENCY IN SELF-REPORTED SMOKING: A COMPARISON OF REPORTS IN SCHOOL AND IN HOUSEHOLD SETTINGS

Extent and sources of inconsistency in self-reported cigarette smoking between self-administered school surveys and household interviews was examined in two longitudinal multiethnic adolescent samples, the urban Transition to Nicotine Dependence in Adolescence (TND) (N = 832) and the National Longitudinal Study of Adolescent Health (Add Health) (N = 4,414). Inconsistency was defined as a positive report of smoking in school followed by a negative report in the household. Smoking questions were ascertained with paper-and-pencil instruments (PAPI-SAQ) in school in both studies, and computer-assisted personal interviewing (CAPI) in TND but audio computer-assisted self-interviewing (ACASI) in Add Health in the household. In TND, 23.5 percent of youths who reported smoking lifetime and 20.4 percent of those who reported smoking the last 12 months in the school survey reported in the household never having smoked; in Add Health, the latter was 8.6 percent. Logistic regressions identified five common correlates of inconsistency across the two studies: younger age, ethnic minority status, lesser involvement in deviant activities, having nonsmoking parents and friends. In TND, interviewing of youth and parent by the same interviewer increased inconsistent reporting. Matching the definition of inconsistent reporting and the age, gender and race/ethnic distributions of TND on an urban Add Health subsample reduced the predicted rate of inconsistency in TND. The estimated bias attributable to CAPI compared with ACASI methodology did not reach significance in the aggregated matched samples suggesting that irrespective of administration mode, household interviews decrease reporting of smoking, especially among younger, minority and more conventional youths embedded in a social network of nonsmokers.