Estimating Student Attrition in School-Based Prevention Studies: Guidance from State Longitudinal Data in Maryland.

Attrition is a critical concern for evaluating the rigor of prevention studies, and the current study provides rates of attrition for subgroups of students and schools who are often sampled for prevention science. This is the first study to provide practical guidance for expected rates of attrition using population-level statewide data; findings indicated that researchers using K-12 school-based samples should plan for attrition rates as high as 27% during middle school and 54% during elementary school. However, researchers should consider the grade levels initially sampled, the length of follow-up, and the specific student characteristics and schools available for sampling. Postsecondary attrition ranged from 45% for bachelor's degree seekers to 73% for associate degree seekers. This practical guidance can help researchers to proactively plan for attrition in the study design phase, limiting bias and increasing the validity of prevention studies.

Testing the effectiveness of the Developing Inclusive Youth program: A multisite randomized control trial.

The Developing Inclusive Youth program is a classroom-based, individually administered video tool that depicts peer-based social and racial exclusion, combined with teacher-led discussions. A multisite randomized control trial was implemented with 983 participants (502 females; 58.5% White, 41.5% Ethnic/racial minority; Mage  = 9.64 years) in 48 third-, fourth-, and fifth-grade classrooms across six schools. Children in the program were more likely to view interracial and same-race peer exclusion as wrong, associate positive traits with peers of different racial, ethnic, and gender backgrounds, and report play with peers from diverse backgrounds than were children in the control group. Many approaches are necessary to achieve antiracism in schools. This intervention is one component of this goal for developmental science.

Gender Differences in Patterns of Substance Use and Delinquency: A Latent Transition Analysis.

This study explores gender-specific patterns and transitions of adolescent substance use and delinquency in a sample of youths at ages 12, 15, and 18 (N = 803). Latent transition analysis identified "Primary Delinquent," "Delinquency and Substance Use," and "Low Risk" classes. Females were less likely to be in the "Primary Delinquent" class at age 12 than males. From 15 to 18, females were approximately equally likely to transition from "Primary Delinquent" to both other classes, whereas males were more likely to transition from "Primary Delinquent" to "Delinquency and Substance Use." These gender differences in behavior can inform services.

On the unnecessary ubiquity of hierarchical linear modeling.

In psychology and the behavioral sciences generally, the use of the hierarchical linear model (HLM) and its extensions for discrete outcomes are popular methods for modeling clustered data. HLM and its discrete outcome extensions, however, are certainly not the only methods available to model clustered data. Although other methods exist and are widely implemented in other disciplines, it seems that psychologists have yet to consider these methods in substantive studies. This article compares and contrasts HLM with alternative methods including generalized estimating equations and cluster-robust standard errors. These alternative methods do not model random effects and thus make a smaller number of assumptions and are interpreted identically to single-level methods with the benefit that estimates are adjusted to reflect clustering of observations. Situations where these alternative methods may be advantageous are discussed including research questions where random effects are and are not required, when random effects can change the interpretation of regression coefficients, challenges of modeling with random effects with discrete outcomes, and examples of published psychology articles that use HLM that may have benefitted from using alternative methods. Illustrative examples are provided and discussed to demonstrate the advantages of the alternative methods and also when HLM would be the preferred method. (PsycINFO Database Record

Modeling Clustered Data with Very Few Clusters.

Small-sample inference with clustered data has received increased attention recently in the methodological literature, with several simulation studies being presented on the small-sample behavior of many methods. However, nearly all previous studies focus on a single class of methods (e.g., only multilevel models, only corrections to sandwich estimators), and the differential performance of various methods that can be implemented to accommodate clustered data with very few clusters is largely unknown, potentially due to the rigid disciplinary preferences. Furthermore, a majority of these studies focus on scenarios with 15 or more clusters and feature unrealistically simple data-generation models with very few predictors. This article, motivated by an applied educational psychology cluster randomized trial, presents a simulation study that simultaneously addresses the extreme small sample and differential performance (estimation bias, Type I error rates, and relative power) of 12 methods to account for clustered data with a model that features a more realistic number of predictors. The motivating data are then modeled with each method, and results are compared. Results show that generalized estimating equations perform poorly; the choice of Bayesian prior distributions affects performance; and fixed effect models perform quite well. Limitations and implications for applications are also discussed.

An Evaluation of Weighting Methods Based on Propensity Scores to Reduce Selection Bias in Multilevel Observational Studies.

Observational studies of multilevel data to estimate treatment effects must consider both the nonrandom treatment assignment mechanism and the clustered structure of the data. We present an approach for implementation of four propensity score (PS) methods with multilevel data involving creation of weights and three types of weight scaling (normalized, cluster-normalized and effective), followed by estimation of multilevel models with the multilevel pseudo-maximum likelihood estimation method. Using a Monte Carlo simulation study, we found that the multilevel model provided unbiased estimates of the Average Treatment Effect on the Treated (ATT) and its standard error across manipulated conditions and combinations of PS model, PS method, and type of weight scaling. Estimates of between-cluster variances of the ATT were biased, but improved as cluster sizes increased. We provide a step-by-step demonstration of how to combine PS methods and multilevel modeling to estimate treatment effects using multilevel data from the Early Childhood Longitudinal Study-Kindergarten Cohort (ECLS-K).

Developmental Relations Between Alcohol and Aggressive Behavior Among Adolescents: Neighborhood and Sociodemographic Correlates.

Throughout adolescence, alcohol consumption and aggressive behaviors are associated with multiple problematic outcomes. Few studies have examined neighborhood-level predictors and individual and family-level sociodemographic variables to describe longitudinal trajectories of these problem behaviors. Therefore, this study investigated the unique contributions of neighborhood and sociodemographic factors in the shared development of aggressive behaviors and drinking in adolescents. We analyzed alcohol consumption frequency and frequency of aggressive behaviors using parallel process latent growth curve models with demographic indicators and neighborhood constructs as predictors. At all ages, alcohol use and aggression positively covaried. Male gender was associated with both aggressive episodes and alcohol use at age 12. African American ethnicity was associated with higher levels of early aggression. Higher neighborhood income was associated with lower levels of early aggression. Findings lend support to current efforts to curb early initiation of alcohol use and aggression.

Experimental pain ratings and reactivity of cortisol and soluble tumor necrosis factor-α receptor II following a trial of hypnosis: results of a randomized controlled pilot study.

OBJECTIVE: Current evidence supports the efficacy of hypnosis for reducing the pain associated with experimental stimulation and various acute and chronic conditions; however, the mechanisms explaining how hypnosis exerts its effects remain less clear. The hypothalamic-pituitary-adrenal (HPA) axis and pro-inflammatory cytokines represent potential targets for investigation given their purported roles in the perpetuation of painful conditions; yet, no clinical trials have thus far examined the influence of hypnosis on these mechanisms. DESIGN: Healthy participants, highly susceptible to the effects of hypnosis, were randomized to either a hypnosis intervention or a no-intervention control. Using a cold pressor task, assessments of pain intensity and pain unpleasantness were collected prior to the intervention (Pre) and following the intervention (Post) along with pain-provoked changes in salivary cortisol and the soluble tumor necrosis factor-α receptor II (sTNFαRII). RESULTS: Compared with the no-intervention control, data analyses revealed that hypnosis significantly reduced pain intensity and pain unpleasantness. Hypnosis was not significantly associated with suppression of cortisol or sTNFαRII reactivity to acute pain from Pre to Post; however, the effect sizes for these associations were medium-sized. CONCLUSIONS: Overall, the findings from this randomized controlled pilot study support the importance of a future large-scale study on the effects of hypnosis for modulating pain-related changes of the HPA axis and pro-inflammatory cytokines.

Enhanced cortisol increase upon awakening is associated with greater pain ratings but not salivary cortisol or soluble tumor necrosis factor-α receptor II responses to acute pain.

OBJECTIVES: The cortisol awakening response (CAR) is related to psychosocial factors and health in potentially significant ways, suggesting that it may be a distinctive marker of hypothalamic-pituitary-adrenal axis function and dysfunction. This study sought to expand upon previous work that examined the association between CAR and ratings of laboratory-evoked acute pain stimulation. In addition to evoked pain ratings, this study also tested whether CAR was prospectively related with salivary cortisol and soluble tumor necrosis factor-α receptor II responses to acute pain stimulation. METHODS: This study included 36 healthy, pain-free volunteers of both sexes recruited through posted study flyers. Prior to completion of laboratory pain testing, salivary cortisol samples were obtained at home over the course of a single morning according to the following time frame: upon awakening, and 15, 30, and 60 minute after awakening. After collection of saliva, study participants brought their home saliva samples to the laboratory for assay and subsequently completed acute experimental pain testing procedures. RESULTS: Cluster analysis of CAR revealed two distinct groups with similar patterns of cortisol response to awakening; increased and flattened. Relative to flattened CAR, increased CAR was associated with greater ratings of pain intensity and unpleasantness. Salivary cortisol was significantly increased and soluble tumor necrosis factor-α receptor II significantly decreased after pain testing, but neither of these responses differed as a function of increased versus flattened CAR. DISCUSSION: CAR may be a marker for stress sensitivity and/or the anticipation of impending stress, which could explain why the increased CAR cohort reported greater acute pain ratings.

Creative solution for implementation of experiential, competency-based palliative care training for internal medicine residents.

To graduate internal medicine residents with basic competency in palliative care, we employ a two-pronged strategy targeted at both residents and attending physicians as learners. The first prong provides a knowledge foundation using web-based learning programs designed specifically for residents and clinical faculty members. The second prong is assessment of resident competency in key palliative care domains by faculty members using direct observation during clinical rotations. The faculty training program contains Competency Assessment Tools addressing 19 topics distributed amongst four broad palliative care domains designed to assist faculty members in making the clinical competency assessments. Residents are required to complete their web-based training by the end of their internship year; they must demonstrate competency in one skill from each of the four broad palliative care domains prior to graduation. Resident and faculty evaluation of the training programs is favorable. Outcome-based measures are planned to evaluate long-term program effectiveness.

Assessment of Social Information Processing in early childhood: development and initial validation of the Schultz Test of Emotion Processing-Preliminary Version.

Crick and Dodge's (Psychological Bulletin 115:74-101, 1994) social information processing model has proven very useful in guiding research focused on aggressive and peer-rejected children's social-cognitive functioning. Its application to early childhood, however, has been much more limited. The present study responds to this gap by developing and validating a video-based assessment tool appropriate for early childhood, the Schultz Test of Emotion Processing-Preliminary Version (STEP-P). One hundred twenty-five Head Start preschool children participated in the study. More socially competent children more frequently attributed sadness to the victims of provocation and labeled aggressive behaviors as both morally unacceptable and less likely to lead to positive outcomes. More socially competent girls labeled others' emotions more accurately. More disruptive children more frequently produced physically aggressive solutions to social provocations, and more disruptive boys less frequently interpreted social provocations as accidental. The STEP-P holds promise as an assessment tool that assesses knowledge structures related to the SIP model in early childhood.

Pain catastrophizing mediates the relationship between self-reported strenuous exercise involvement and pain ratings: moderating role of anxiety sensitivity.

OBJECTIVE: To investigate the cross-sectional associations among self-reported weekly strenuous exercise bouts, anxiety sensitivity, and their interaction with pain catastrophizing and pain responses to the cold pressor task (CPT) in healthy, ethnically diverse young adults (n = 79). Exercise involvement has been shown to have hypoalgesic effects and cognitive factors may partially explain this effect. Particularly, alterations in pain catastrophizing have been found to mediate the positive pain outcomes of multidisciplinary treatments incorporating exercise. Further, recent evidence suggests that exercise involvement and anxiety sensitivity may act together, as interacting factors, to exert an effect on catastrophizing and pain outcomes; however, further research is needed to clarify the nature of this interaction. METHODS: Before the CPT, participants were asked to complete the Godin Leisure-Time Exercise Questionnaire, the Beck Depression Inventory, and the Anxiety Sensitivity Index. After the CPT, participants completed a modified version of the Pain Catastrophizing Scale and the Short Form-McGill Pain Questionnaire. RESULTS: At a high level of anxiety sensitivity, controlling for depressive symptoms, CPT immersion time, and sex differences, a bias-corrected (BC), bootstrapped confidence interval revealed that pain catastrophizing significantly mediated the relationship between self-reported weekly strenuous exercise bouts and pain response (95% BC Confidence Interval = -9.558, -0.800 with 1000 resamples). At intermediate and low levels of anxiety sensitivity, no significant mediation effects were found. CONCLUSIONS: These findings support that, for pain catastrophizing to mediate the strenuous exercise-pain response relation, individuals must possess a high level of anxiety sensitivity.

The Performance of Methods to Test Upper-Level Mediation in the Presence of Nonnormal Data.

A Monte Carlo study compared the statistical performance of standard and robust multilevel mediation analysis methods to test indirect effects for a cluster randomized experimental design under various departures from normality. The performance of these methods was examined for an upper-level mediation process, where the indirect effect is a fixed effect and a group-implemented treatment is hypothesized to impact a person-level outcome via a person-level mediator. Two methods-the bias-corrected parametric percentile bootstrap and the empirical-M test-had the best overall performance. Methods designed for nonnormal score distributions exhibited elevated Type I error rates and poorer confidence interval coverage under some conditions. Although preliminary, the findings suggest that new mediation analysis methods may provide for robust tests of indirect effects.

Psychometric properties of the Beck Depression Inventory for Youth in a sample of girls.

A new measure has been developed to assess depressive symptoms, the Beck Depression Inventory for Youth (BDI-Y; J. S. Beck, A. T. Beck, & J. B. Jolly, 2001). This research extends previous validation research of BDI-Y total scores by examining internal consistency and convergent and predictive validity within a school-based sample (n=859) of girls 9-13 years old by age level and for selected races or ethnic groups. Scores had high internal consistency, and there was support for using the BDI-Y to assess depressive symptoms. Reliability was slightly lower for 9-year-olds, but reliability and validity estimates did not differ by race or ethnic group. Finally, confirmatory factor analysis results provide some support for unidimensionality of scores but also point toward possible refinements.

Reading and the corpus callosum: an MRI family study of volume and area.

This study examines the corpus callosum in 68 readers nested in 24 families. Callosa were measured and controlled for whole brain volume, intelligence, and gender. The relation of corpus callosum size to the within-family variance of oral reading was investigated with various measurements: volume, midsagittal area, and anterior-to-posterior one-fifth area segmentations. Because this is the first known publication of MRI calculations of corpus callosum volume, some basic questions about bilateral symmetry and the efficacy of area versus volumetric measurements were explored. Results suggest that better readers within families have larger midsagittal areas at the midbody. Although reliably measured, volume did not contribute to oral reading but was highly correlated with area. Bilateral volumes of the corpus callosum were symmetric.

Evaluation and comparison of models of metabolic syndrome using confirmatory factor analysis.

The aim of this study was to evaluate and compare three competing models of the underlying factor structure of metabolic syndrome using confirmatory factor analysis (CFA). Data from the Insulin Resistance Atherosclerosis Study (IRAS) was used, which has previously been evaluated using principal components analysis (PCA). The three models that were evaluated consisted of oblique and orthogonal two-factor models with hypothesized underlying "metabolic" and "blood pressure" factors, and a four-factor model theorizing "insulin resistance," "obesity," "lipids," and "blood pressure" as the underlying constructs. Several CFAs were performed using EQS Multivariate Software Version 5.7b with maximum likelihood estimation. The results showed that the four-factor model yielded significantly better data-model fit than two-factor models, with a comparative fit index of 0.963, and standardized root mean square residual of 0.036. Factors exhibited good construct reliability and variance extracted estimates except for the lipids factor. We concluded that the four-factor model of metabolic syndrome was the most plausible model among the three competing models.

The Performance of Cross-Validation Indices Used to Select Among Competing Covariance Structure Models Under Multivariate Nonnormality Conditions.

Cudeck and Browne (1983) proposed using cross-validation as a model selection technique in structural equation modeling. The purpose of this study is to examine the performance of eight cross-validation indices under conditions not yet examined in the relevant literature, such as nonnormality and cross-validation design. The performance of each cross-validation index was measured in terms of true model selection rate as well as consistency of model selection. The performance of the cross-validation indices tended to improve as factor loading and sample size increased but performed less well as nonnormality increased. The double cross-validated indices outperformed their simple cross-validated counterparts in certain conditions. Recommendations are provided as to which cross-validation methods would optimally perform in a given condition.

A Comparison of Single Sample and Bootstrap Methods to Assess Mediation in Cluster Randomized Trials.

A Monte Carlo study examined the statistical performance of single sample and bootstrap methods that can be used to test and form confidence interval estimates of indirect effects in two cluster randomized experimental designs. The designs were similar in that they featured random assignment of clusters to one of two treatment conditions and included a single intervening variable and outcome, but they differed in whether the mediator was measured at the participant or site level. A bias-corrected bootstrap had the best statistical performance for each design and was closely followed by the empirical-Mtest, either of which is recommended for testing and estimating indirect effects in multilevel designs. In addition, consistent with previous research, the commonly used z test had relatively poor performance.

Core self-evaluations and job burnout: the test of alternative models.

Research on job burnout has traditionally focused on contextual antecedent conditions, although a theoretically appropriate conception implicates person-environment relationships. The authors tested several models featuring various combinations of personal and contextual influences on burnout and job satisfaction. Measures of core self-evaluations, organizational constraints, burnout, and job satisfaction were collected from 859 health care employees. Results from structural equations modeling analyses revealed an influence of core self-evaluations and perceived organizational constraints on job burnout and satisfaction, suggesting personal and contextual contributions. These results favor a broadening of current thinking about the impact of situational constraints on the expression of job burnout, as well as for the role of disposition for affective responding to effectively address occupational health problems.