Amelia M. Haviland

Combining propensity score matching and group-based trajectory analysis in an observational study

In a nonrandomized or observational study, propensity scores may be used to balance observed covariates and trajectory groups may be used to control baseline or pretreatment measures of outcome. The trajectory groups also aid in characterizing classes of subjects for whom no good matches are available and to define substantively interesting groups between which treatment effects may vary. These and related methods are illustrated using data from a Montreal-based study. The effects on subsequent violence of gang joining at age 14 are studied while controlling for measured characteristics of boys prior to age 14. The boys are divided into trajectory groups based on violence from ages 11 to 13. Within trajectory group, joiners are optimally matched to a variable number of controls using propensity scores, Mahalanobis distances, and a combinatorial optimization algorithm. Use of variable ratio matching results in greater efficiency than pair matching and also greater bias reduction than matching at a fixed ratio. The possible impact of failing to adjust for an important but unmeasured covariate is examined using sensitivity analysis.

Gender Wage Disparities among the Highly Educated.

We examine gender wage disparities for four groups of college-educated women—black, Hispanic, Asian, and non-Hispanic white—using the National Survey of College Graduates. Raw log wage gaps, relative to non-Hispanic white male counterparts, generally exceed –0.30. Estimated gaps decline to between –0.08 and –0.19 in nonparametric analyses that (1) restrict attention to individuals who speak English at home and (2) match individuals on age, highest degree, and major. Among women with work experience comparable to men’s, these estimated gaps are smaller yet—between –0.004 and –0.13. Importantly, we find that inferences from familiar regression-based decompositions can be quite misleading.

Why do minority men earn less? A study of wage differentials among the highly educated

We estimate wage gaps using nonparametric matching methods and detailed measures of field of study for university graduates. We find a modest portion of the wage gap is the consequence of measurement error in the Census education measure. For Hispanic and Asian men, the remaining gap is attributable to premarket factorsprimarily differences in formal education and English language proficiency. For black men, only about one-quarter of the wage gap is explained by these same factors. For a subsample of black men born outside the South to parents with some college education, these factors do account for the entire wage gap.

Combining group-based trajectory modeling and propensity score matching for causal inferences in nonexperimental longitudinal data.

A central theme of research on human development and psychopathology is whether a therapeutic intervention or a turning-point event, such as a family break-up, alters the trajectory of the behavior under study. This article describes and applies a method for using observational longitudinal data to make more transparent causal inferences about the impact of such events on developmental trajectories. The method combines 2 distinct lines of research: work on the use of finite mixture modeling to analyze developmental trajectories and work on propensity score matching. The propensity scores are used to balance observed covariates and the trajectory groups are used to control pretreatment measures of response. The trajectory groups also aid in characterizing classes of subjects for which no good matches are available. The approach is demonstrated with an analysis of the impact of gang membership on violent delinquency based on data from a large longitudinal study conducted in Montréal, Canada.

Understanding Nonresponse to the 2007 Medicare CAHPS Survey

PURPOSE The Medicare Consumer Assessments of Healthcare Providers and Systems (MCAHPS) survey, a primarily English-language mail survey with English and Spanish telephone follow-up, is the primary means of assessing the health care experiences of American seniors. We examine unit (whole survey) and item nonresponse for this survey to explore issues regarding surveying seniors about their health care. DESIGN AND METHODS We describe overall rates and analyze predictors of unit and item nonresponse for the 695,197 Medicare beneficiaries selected for the 2007 MCAHPS survey (335,249 unit respondents, 49% overall response rate). RESULTS Asians, African Americans, and Hispanics responded at adjusted response rates 7-17 percentage points lower than non-Hispanic Whites (p < .001 for each). Among seniors, response rates dropped beyond age 75. Asians and older beneficiaries were especially likely to respond by mail, and African Americans and Hispanics by phone. Breakoff from telephone surveys was most common among African Americans and older respondents. Among respondents, older age was the strongest predictor of item missingness (e.g., those 85 years and older failed to answer items at twice the rate of those aged 65-74 years, p < .001). Non-Hispanic Whites had lower rates of item missingness than other racial/ethnic groups (p < .001 for each; one-third lower than African Americans). IMPLICATIONS Survey research on older adults, especially regarding racial/ethnic disparities in health care, could benefit from improved response rates. These results suggest that targeted prenotification materials and campaigns, tailored follow-up, targeted Spanish mailings, Chinese translations/calls, and adjustments to telephone protocols may improve representation and response.

Group-Based Trajectory Modeling Extended to Account for Nonrandom Participant Attrition

This article reports on an extension of group-based trajectory modeling to address nonrandom participant attrition or truncation due to death that varies across trajectory groups. The effects of the model extension are explored in both simulated and real data. The analyses of simulated data establish that estimates of trajectory group size as measured by group membership probabilities can be badly biased by differential attrition rates across groups if the groups are initially not well separated. Differential attrition rates also imply that group sizes will change over time, which in turn has important implications for using the model parameter estimates to make population-level projections. Analyses of longitudinal data on disability levels in a sample of very elderly individuals support both of these conclusions.

The relationship between protective factors and outcomes for children investigated for maltreatment

OBJECTIVES This research explores the relationship between hypothesized protective factors and outcomes for children investigated for maltreatment. METHODS Using data from the National Survey on Child and Adolescent Well-Being (NSCAW), we ran logistic regression models to examine the relationship between hypothesized protective factors (social competence, adaptive functioning skills, and peer relationships) and outcomes (externalizing behavior, internalizing behavior, reading competence). RESULTS For each hypothesized protective factor, we found variation in individual scores and sample mean scores at the lower end of the scales, indicating that these children fare worse than most children. However, many children experienced large changes in their individual scores over time suggesting that children can and do improve on these hypothesized protective factors. In examining the relationship between hypothesized protective factors and outcomes, children with higher levels of social competence were significantly more likely to be in the normal range for both externalizing and internalizing behaviors. Children with higher mean adaptive functioning skills were more likely to be in the normal range for both externalizing behavior and reading competence. The positive nature of the childs peer relationships was also related to externalizing behavior and reading competence. CONCLUSIONS Overall, our analyses support the idea that social competence, adaptive functioning skills, and peer relationships are related to outcomes for children investigated for maltreatment. While further research is needed to establish a causal link, this work identifies three individual-level hypothesized protective factors as potential sources of variation in outcomes. PRACTICE IMPLICATIONS To prevent or alleviate the harmful consequences maltreatment, it is necessary to understand factors that help children move beyond poor outcomes. Our analyses suggest that a strong relationship exists between a childs social competence, adaptive functioning skills and positive peer relationships and select outcomes three years after being investigated for maltreatment. With these individual-level protective factors related to more positive outcomes, it suggests that intervening to increase protective factors could improve outcomes for maltreated and at-risk children.

What kinds of injuries do OSHA inspections prevent

OBJECTIVE OSHAs enforcement program is one of the major public efforts to protect American workers. We examine both the scope of injury prevention that inspections can contribute and the types of standards that contribute the most. METHODS We linked Pennsylvania Department of Labor and Industry files for lost-time injuries and employment to calculate injury rates for 1998-2005 for all single-establishment manufacturing firms. We linked these to OSHA inspection records. RESULTS Inspections with penalties did affect injury types unrelated to standards as well as those related. We also found again that citations for violations of the standard requiring personal protective equipment had the largest impact on preventing injuries. IMPACT ON INDUSTRY Programs requiring protective equipment use deserve added attention from consultants and inspectors. In addition, some inspections spur managers to undertake safety measures that go beyond compliance with standards.

Experiences of Care Among Medicare Beneficiaries With ESRD: Medicare Consumer Assessment of Healthcare Providers and Systems (CAHPS) Survey Results

BACKGROUND Patients with end-stage renal disease (ESRD) have special health needs; little is known about their care experiences. STUDY DESIGN Secondary analysis of 2009-2010 Medicare Consumer Assessment of Healthcare Providers and Systems (CAHPS) data, using representative random samples of Medicare beneficiaries. Description of Medicare beneficiaries with ESRD and investigation of differences in patient experiences by sociodemographic characteristics and coverage type. SETTING & PARTICIPANTS Data were collected from 823,564 Medicare beneficiaries (3,794 with ESRD) as part of the Medicare CAHPS survey, administered by mail with telephone follow-up of nonrespondents. PREDICTOR ESRD status, age, education, self-reported general and mental health status, race/ethnicity, sex, Medicare coverage type, state of residence, and other demographic measures. OUTCOMES 6 composite measures of patient experience in 4 care domains (access to care, physician communication, customer service, and access to prescription drugs and drug information) and 4 ratings (overall care, personal physician, specialist physician, and prescription drug plan). RESULTS Patients with ESRD reported better care experiences than non-ESRD beneficiaries for 7 of 10 measures (P < 0.05) after adjustment for patient characteristics, geography, and coverage type, although to only a small extent (adjusted mean difference, <3 points [scale, 0-100]). Black patients with ESRD and less educated patients were more likely than other patients with ESRD to report poor experiences. LIMITATIONS Inability to distinguish patient experiences of care for different treatment modalities. CONCLUSIONS On average, beneficiaries with ESRD report patient experiences that are at least as positive as non-ESRD beneficiaries. However, black and less educated patients with ESRD reported worse experiences than other ESRD patients. Stratified reporting of patient experience by race/ethnicity or education in patients with ESRD can be used to monitor this disparity. Physician choice and confidence and trust in physicians may be particularly important for patients with ESRD.

Small Businesses and Workplace Fatality Risk: An Exploratory Analysis

6 Jump down to document THE ARTS CHILD POLICY CIVIL JUSTICE EDUCATION ENERGY AND ENVIRONMENT HEALTH AND HEALTH CARE INTERNATIONAL AFFAIRS NATIONAL SECURITY POPULATION AND AGING PUBLIC SAFETY SCIENCE AND TECHNOLOGY SUBSTANCE ABUSE TERRORISM AND HOMELAND SECURITY TRANSPORTATION AND INFRASTRUCTURE WORKFORCE AND WORKPLACE The RAND Corporation is a nonprofit research organization providing objective analysis and effective solutions that address the challenges facing the public and private sectors around the world.