Estimating Regression Models in Which the Dependent Variable Is Based on Estimates

e-mail: jblewis{at}ucla.edu (corresponding author) e-mail: dlinzer{at}ucla.edu Researchers often use as dependent variables quantities estimatedfrom auxiliary data sets. Estimated dependent variable (EDV)models arise, for example, in studies where counties or statesare the units of analysis and the dependent variable is an estimatedmean, proportion, or regression coefficient. Scholars fittingEDV models have generally recognized that variation in the samplingvariance of the observations on the dependent variable willinduce heteroscedasticity. We show that the most common approachto this problem, weighted least squares, will usually lead toinefficient estimates and underestimated standard errors. Inmany cases, OLS with White's or Efron heteroscedastic consistentstandard errors yields better results. We also suggest two simplealternative FGLS approaches that are more efficient and yieldconsistent standard error estimates. Finally, we apply the variousalternative estimators to a replication of Cohen's (2004) cross-nationalstudy of presidential approval.     

The Course and Psychosocial Correlates of Personality Disorder Symptoms in Adolescence: Erikson's Developmental Theory Revisited

Personality disorder symptoms were investigated in a community sample of young people (n = 714) to assess their relationship over time with well-being during adolescence and the emergence of intimacy in early adulthood. Drawing on Erikson's theory of psychosocial development, changes in adolescent well-being were conceptualized as indirect indicators of identity consolidation. Cluster B personality disorder symptoms (borderline, histrionic, and narcissistic symptoms) were conceptualized to represent identity diffusion—i.e. maladaptive personality traits that usually resolve during the identity crisis of adolescence. Latent growth models were used in 2 age cohorts to assess (1) interrelationships between Cluster B symptoms, well-being, and intimacy at mean ages 13.8 and 18.6 years; and (2) associations between their developmental trajectories over the next 6 years. As expected, higher personality disorder symptoms were associated with lower well-being during adolescence, and declines in personality disorder symptoms over time were associated with corresponding gains in well-being. Consistent with Erikson's developmental theory, there was an inverse relationship between Cluster B symptoms and intimacy that increased in strength as young people entered adulthood. As an indicator of successful identity consolidation, well-being was significantly associated with intimacy in female adolescents and young adults.     

Validation and implementation of the PowerPlex 16 BIO System STR multiplex for forensic casework

The PowerPlex 16 BIO multiplex short tandem repeat (STR) system contains the 13 CODIS loci (FGA, TPOX, D8S1179, vWA, D18S51, D21S11, TH01, D3S1358, CSF1PO, D16S539, D7S820, D13S317, and DS5S818), plus two pentanucleotide repeat loci (Penta D and Penta E) and the sex-identifying locus. Amelogenin. The PowerPlex 16 BIO System is optimized for use with the Hitachi FMBIO gel imaging systems. A consortium of seven independent laboratories collaborated to perform the studies defined by the FBI standards for performing a developmental validation, including the evaluation of sample concordance, percent stutter determination, nonprobative casework, precision, sensitivity, mixture determination, effect of substrates, the impact of environmental insults, and species specificity. All samples tested for concordance were consistent except for one sample from the Virginia Division of Forensic Science database that displayed discordance at D13S317, a locus whose primer sequence was altered. Stutter values were comparable to those of other STR multiplex systems, the precision was comparable to other multiplexes analyzed by gel electrophoresis, the DNA profiles were unchanged by the substrate upon which the blood samples were placed, and the nonprobative casework samples re-typed for the PowerPlex 16 BIO System were consistent with previous typing results. When greater than 0.125 ng of DNA was placed into the PowerPlex 16 BIO System amplification reaction, a full profile was generated by all laboratories. The mixture study results were comparable to those reported for other multiplex systems, the environmental study demonstrated a loss of larger molecular weight loci when samples were incubated at elevated temperatures for a prolonged period of time, and the only notable cross species hybridization was observed with primate DNA samples. This extensive validation work performed demonstrates that the PowerPlex 16 BIO System provides STR data of a quality comparable with other PowerPlex STR multiplex kits as well as other widely used STR multiplexes and is thus suitable for evidentiary casework analysis as well as database sample profiling.     

Application of the BioMek 2000 Laboratory Automation Workstation and the DNA IQ System to the extraction of forensic casework samples

Robotic systems are commonly utilized for the extraction of database samples. However, the application of robotic extraction to forensic casework samples is a more daunting task. Such a system must be versatile enough to accommodate a wide range of samples that may contain greatly varying amounts of DNA, but it must also pose no more risk of contamination than the manual DNA extraction methods. This study demonstrates that the BioMek 2000 Laboratory Automation Workstation, used in combination with the DNA IQ System, is versatile enough to accommodate the wide range of samples typically encountered by a crime laboratory. The use of a silica coated paramagnetic resin, as with the DNA IQ System, facilitates the adaptation of an open well, hands off, robotic system to the extraction of casework samples since no filtration or centrifugation steps are needed. Moreover, the DNA remains tightly coupled to the silica coated paramagnetic resin for the entire process until the elution step. A short pre-extraction incubation step is necessary prior to loading samples onto the robot and it is at this step that most modifications are made to accommodate the different sample types and substrates commonly encountered with forensic evidentiary samples. Sexual assault (mixed stain) samples, cigarette butts, blood stains, buccal swabs, and various tissue samples were successfully extracted with the BioMek 2000 Laboratory Automation Workstation and the DNA IQ System, with no evidence of contamination throughout the extensive validation studies reported here.     

Do juvenile drug courts reduce recidivism?: Outcomes of drug court and an adolescent substance abuse program

Juvenile drug courts have emerged as “innovative” responses to juvenile drug offenders, but comparatively little is known about their operations. This paper presents results of a retrospective comparison of drug court participants to an adolescent substance abuse program (ASAP) to examine which participants fared better in terms of future recidivism. Using data collected from official case files, we compared recidivism levels for all juveniles (n = 150) terminated from drug court between 1996 and 1999 with those of a random sample of juveniles (n = 158) terminated from ASAP during 1994 and 1995. Biand multivariate analyses were conducted to identify whether significant differences existed between the groups concerning re-arrest (recidivism) over a 24-month post-release observation period. Study results highlighted by logistic regression analyses suggesting that juveniles in drug court were no more likely to recidivate than were juveniles in ASAP is a positive finding for the drug court program and is an indication that the program is working, especially given the serious nature of this juvenile offender population.     

Finding common ground in the study of child maltreatment, youth violence, and adult domestic violence

For several decades, we have witnessed a surge in public policies aimed at ending child maltreatment, youth violence, and adult domestic violence. Commensurate with this increased interest has been a growing body of research on each issue's etiology, affected population, and the public policy and prevention impacts. Even a cursory review of the literature suggests a number of commonalities across these forms of violence. As such, it seems timely and prudent to craft a research framework that facilitates the identification and dissemination of practice and policy innovations that can address all three concerns. This article articulates an overarching framework to guide researchers in better identifying common avenues of study. After summarizing the commonalities found across the three areas, the authors identify cross-cutting issues that have particular relevance for advancing our understanding of violence and its effects on personal and social interactions.     

Why triggers fail (and what to do about it): An examination of the unemployment insurance extended benefits program

During periods of high unemployment, many workers exhaust their unemployment insurance (UI) benefits before regaining employment. To help alleviate this problem, Congress created the extended benefits (EB) program, expanding the number of weeks of benefits available to UI recipients in high unemployment states. The EB program operates by “triggering on” additional weeks of benefits in states where unemployment and UI benefit receipt are above federally established thresholds. We analyze the performance of the EB program by creating a series of policy simulations using weekly UI claims and unemployment data from the program's inception in 1970 through the most recent economic expansion in 2005. Overall, we find that EB triggers, as currently constructed, fail as a policy tool for extending UI benefits. Minor adjustments to the triggers are unlikely to be effective. We develop an alternative set of “fix point” triggers that allow the EB program to trigger on and off in a more timely fashion. These triggers outperform all previously legislated triggers as well as other commonly proposed triggering mechanisms on criteria of timeliness, breadth, and duration. © 2006 by the Association for Public Policy Analysis and Management