Longitudinal Data System Research Articles

Illuminating inequality in access: Variation in enrollment in undergraduate engineering programs across Virginia's high schools

AbstractBackgroundDetermining the root causes of persistent underrepresentation of different subpopulations in engineering remains a continued challenge. Because place‐based variation of resource distribution is not random and because school and community contexts influence high school outcomes, considering variation across those contexts should be paramount in broadening participation research.Purpose/HypothesisThis study takes a macroscopic systems view of engineering enrollments to understand variation across one state's public high school rates of engineering matriculation.Design/MethodThis study uses a dataset from the Virginia Longitudinal Data System that includes all students who completed high school from a Virginia public school from 2007 to 2014 (N = 685,429). We explore geographic variation in four‐year undergraduate engineering enrollment as a function of gender, race/ethnicity, and economically disadvantaged status. Additionally, we investigate the relationship between characteristics of the high school and community contexts and undergraduate engineering enrollment across Virginia's high schools using regression analysis.ResultsOur findings illuminate inequality in enrollment in engineering programs at four‐year institutions across high schools by gender, race, and socioeconomic status (and the intersections among those demographics). Different high schools have different engineering enrollment rates among students who attend four‐year postsecondary institutions. We show strong associations between high schools' engineering enrollment rates and four‐year institution enrollment rates as well as moderate associations for high schools' community socioeconomic status.ConclusionsStrong systemic forces need to be overcome to broaden participation in engineering. We demonstrate the insights that state longitudinal data systems can illuminate in engineering education research.

Risk factors for emergency room use and rehospitalization among opioid-exposed newborns in Massachusetts.

To determine the risk factors for receipt of emergency room (ER) care and rehospitalization among opioid-exposed newborns in Massachusetts. We analyzed two linked data sets from 2002 to 2010: (a) Massachusetts Pregnancy to Early Life Longitudinal Data System and (b) Massachusetts Bureau of Substance Abuse Services Management Information System. Generalized estimating equations were used to assess the independent association of maternal and infant characteristics with ER use and rehospitalization in the first year of life. Four thousand and five hundred and twenty-four maternal-infant dyads affected by OUD were included in the analysis. In adjusted analysis, risk factors for receipt of ER care included Hispanic ethnicity (aOR 1.63 [95% CI 1.30-2.05]), lower education levels (aOR 1.54-1.69 [95% CI 1.12-2.31]), nonprivate insurance (aOR 1.44 [95% CI 1.11-1.86]), and presence of maternal chronic conditions (aOR 1.14 [95% CI 1.01-1.29]). Risk factors for rehospitalization included prematurity (aOR 1.44 [95% CI 1.14-1.82]), low birthweight (aOR 2.02 [95% CI 1.63-2.49]), and nonprivate insurance (aOR 1.58 [95% CI 1.13-2.22]). Prolonged infant birth hospitalization was protective against both ER use (aOR 0.84 [95% CI 0.73-0.96]) and rehospitalization (aOR 0.63 [95% CI 0.53-0.75]). Clinical and sociodemographic characteristics are risk factors for receipt of ER care and rehospitalization among opioid-exposed newborns in Massachusetts.

The effects of downstream clustering in longitudinal studies

Many experimental designs in educational and behavioral research involve at least one level of clustering. Clustering affects the precision of estimators and its impact on statistics in cross-sectional studies is well known. Clustering also occurs in longitudinal designs where students that are initially grouped may be regrouped in the following year and regrouped again in subsequent years. The purpose of this article is to explore the clustering effects of multiple waves of grouping on the effect sizes of downstream outcomes; that is, outcomes that are measured after one or more waves of regrouping. We center our framework on the longitudinal cluster randomized trial in which treatments are initially assigned by clusters (such as classrooms) and students are then regrouped into different clusters in subsequent years. This article illustrates how information on the intraclass correlations in longitudinal studies can be used to correct for biases in the effect sizes and variances, and how it can also be used to adjust the significance test for the effects of longitudinal clustering. This article also provides the first empirical evidence of the magnitude of longitudinal clustering effects using data from a state longitudinal data system.

Racial and Ethnic Disparities in Birth Outcomes and Labor and Delivery Charges Among Massachusetts Women With Intellectual and Developmental Disabilities.

Understanding the pregnancy experiences of racial and ethnic minority women with intellectual and developmental disabilities (IDD) is critical to ensuring that policies can effectively support these women. This research analyzed data from the 1998-2013 Massachusetts Pregnancy to Early Life Longitudinal (PELL) data system to examine the racial and ethnic disparities in birth outcomes and labor and delivery charges of U.S. women with IDD. There was significant preterm birth disparity among non-Hispanic Black women with IDD compared to their non-Hispanic White peers. There were also significant racial and ethnic differences in associated labor and delivery-related charges. Further research, examining potential mechanisms behind the observed racial and ethnic differences in labor and delivery-related charges in Massachusetts' women with IDD is needed.

Pregnancy, Birth, and Infant Outcomes Among Women Who Are Deaf or Hard of Hearing

Being deaf or hard of hearing can be marginalizing and associated with inequitable health outcomes. Until recently, there were no U.S. population-based studies of pregnancy outcomes among deaf or hard of hearing women. In light of inconsistent findings in the limited available literature, this study sought to conduct a more rigorous study using population-based, longitudinal linked data to compare pregnancy complications, birth characteristics, and neonatal outcomes between deaf or hard of hearing and non-deaf or hard of hearing women. Researchers conducted a retrospective cohort study in 2019 using the Massachusetts Pregnancy to Early Life Longitudinal data system. This system links all Massachusetts birth certificates, fetal death reports, and delivery- and nondelivery-related hospital discharge records for all infants and their mothers. The study included women with singleton deliveries who gave birth in Massachusetts between January 1998 and December 2013. The deaf or hard of hearing women had an increased risk of chronic medical conditions and pregnancy complications including pre-existing diabetes, gestational diabetes, pre-eclampsia and eclampsia, and placental abruption. Deliveries to deaf or hard of hearing women were significantly associated with adverse birth outcomes, including preterm birth, low birth weight or very low weight, and low 1-minute Apgar score or low 5-minute Apgar score. No significant differences were found in size for gestational age, fetal distress, or stillbirth among deaf or hard of hearing women. Findings from this 2019 study indicate that deaf or hard of hearing women are at a heightened risk for chronic conditions, pregnancy-related complications, and adverse birth outcomes and underscore the need for systematic investigation of the pregnancy- and neonatal-related risks, complications, costs, mechanisms, and outcomes of deaf or hard of hearing women.

Ensuring the Life-Span Benefits of Newborn Screening.

Newborn screening is a highly successful public health program that has led to major improvements in outcomes for a variety of conditions otherwise associated with long-term disability and even death.1 In the United States, newborn screening is provided to every newborn, regardless of circumstance, leading to the identification of >13 000 newborns with a significant condition each year.1 Most of these individuals require specialized care over their life span. However, public health involvement in newborn screening typically ends once the condition has been diagnosed. This can lead to gaps in care and impede the ability to collect the data necessary for quality improvement and assess treatment effectiveness. The public health activities in newborn screening include choosing which conditions to screen for, implementing screening and monitoring quality, and ensuring follow-up testing and referrals after positive screen results. After referral for diagnosis, the public health mandate for most states ends, and the responsibility for care rests with pediatricians and other clinicians. Unfortunately, the lifelong management of the rare conditions identified through newborn screening can break down at multiple points, even for conditions that have been included in newborn screening for decades. For example, many children with sickle cell disease do not receive life-saving antibiotic prophylaxis2; the management of children with congenital hypothyroidism may be inconsistent with recommended care, undermining optimal cognitive development3; and adults with phenylketonuria often have limited access to the specialists and medical foods necessary to protect against neurocognitive impairment.4 The complexity of follow-up care has increased with the expansion … Address correspondence to Alex R. Kemper, MD, MPH, MS, Division of Ambulatory Pediatrics, Nationwide Children’s Hospital, 700 Children’s Dr, LAC5411, Columbus, OH 43205-2664. E-mail: alex.kemper{at}nationwidechildrens.org

Toward a Rich Data Future for School Finance Research

This special topic is intended to advance research in school finance using two newly produced compilations of multiple longitudinal data systems, the School Finance Indicators Database ( http://schoolfinancedata.org/ ) and the Urban Institute, Education Data Explorer ( https://educationdata.urban.org/data-explorer/ ). This special topic contains two articles taking advantage of comprehensive longitudinal data on school finance. The first, by Victoria Sosina and Ericka Weathers, uses panel data from the School Finance Indicators Data System from 1999 to 2013 to evaluate whether and to what extent changes to Black-White and Latinx-White demographic differences among districts leads to greater resource disparity over time. The second article, by Knight and Mendoza, combines data from the Census Fiscal Survey with data from the California Department of Education to explore whether differences in data on and measures of school funding equity matter (i.e., lead to similar or different conclusions) when evaluating the effects of California’s 2013 adoption of the Local Control Funding Formula.

High School Completion in Context: Student- and School-Level Factors Related to On-Time Graduation

Background/Context There is an expansive body of research concerning high school graduation; however, most studies omit students who persist through four years of high school without earning a diploma. In addition, there is scant research exploring longer term outcomes among students whose academic trajectories do not fit within the traditional four-year model of high school graduation, including eventual graduation, postsecondary enrollment, or engagement in the workforce. Purpose/Objective/Research Question/Focus of Study The current study addresses the substantive gaps in research regarding high school noncompletion by examining the college and workforce outcomes of persisters—defined here as students who do not formally withdraw from high school, nor earn a regular diploma, four years after entering high school as a first-time ninth grader. Research Design The present study accessed five years of linked, longitudinal, student-level administrative data from the Maryland Longitudinal Data System. Multilevel models assessed the relationship between student- and school-level factors with the odds of students earning a high school diploma four years after beginning their freshman year. Independent variables included student-level demographic and academic indicators and school-level concentrations of student characteristics. Conclusions/Recommendations This study offers a first look into the academic and employment trajectories of an understudied and high-risk group of young adults. The multilevel examination of student- and school-level factors indicated that on-time graduation for four-year persisters should be understood as a function of students within their academic environment. Overall, persisters had less favorable college and workforce outcomes when compared with students who earned a high school diploma, suggesting the need for interventions that promote college and workforce readiness across the population of persisters. The findings presented herein suggest that the phenomenon of persisting should be considered, along with dropout, as a critical element of a more informed analysis of high school graduation. Implications for research, policy, and practice are discussed.

The Promise and Limitations of Synthetic Data as a Strategy to Expand Access to State-Level Multi-Agency Longitudinal Data

There is demand among policy-makers for the use of state education longitudinal data systems, yet laws and policies regulating data disclosure limit access to such data, and security concerns and risks remain high. Well-developed synthetic datasets that statistically mimic the relations among the variables in the data from which they were derived, but which contain no records that represent actual persons, present a viable solution to these laws, policies, concerns, and risks. We present a case study in the development of a synthetic data system and highlight potential applications of synthetic data. We begin with an overview of synthetic data, what it is, how it has been utilized thus far, and the potential benefits and concerns in its application to education data systems. We then describe our federally-funded project, proposing the steps required to synthesize a statewide longitudinal data system covering high school, postsecondary, and workforce data. Last, for use as a template for other agencies considering synthetic data, we review the challenges we have confronted in the development of our synthetic data system for research and policy evaluation purposes.

Obstetric Comorbidity and Severe Maternal Morbidity Among Massachusetts Delivery Hospitalizations, 1998-2013.

The rate of severe maternal morbidity in the United States increased approximately 200% during 1993-2014. Few studies have reported on the health of the entire pregnant population, including women at low risk for maternal morbidity. This information might be useful for interventions aimed at primary prevention of pregnancy complications. To better understand this, we sought to describe the distribution of comorbid risk among all delivery hospitalizations in Massachusetts and its association with the distribution of severe maternal morbidity. Using an existing algorithm, we assigned an obstetric comorbidity index (OCI) score to delivery hospitalizations contained in the Massachusetts pregnancy to early life longitudinal (PELL) data system during 1998-2013. We identified which hospitalizations included severe maternal morbidity and calculated the rate and frequency of these hospitalizations by OCI score. During 1998-2013, PELL contained 1,185,182 delivery hospitalizations; of these 5325 included severe maternal morbidity. Fifty-eight percent of delivery hospitalizations had an OCI score of zero. The mean OCI score increased from 0.60 in 1998 to 0.82 in 2013. Hospitalizations with an OCI score of zero comprised approximately one-third of all deliveries complicated by severe maternal morbidity, but had the lowest rate of severe maternal morbidity (22.8/10,000 delivery hospitalizations). The mean OCI score increased during the study period, suggesting that an overall increase in risk factors has occurred in the pregnant population in Massachusetts. Interventions that can make small decreases to the mean OCI score could have a substantial impact on the number of deliveries complicated by severe maternal morbidity. Additionally, all delivery facilities should be prepared for severe complications during low-risk deliveries.

Institute of Education Sciences: Statewide Longitudinal Data Systems (ED)

Federal Grants & ContractsVolume 43, Issue 15 p. 1-2 Grant Alerts Institute of Education Sciences: Statewide Longitudinal Data Systems (ED) First published: 02 July 2019 https://doi.org/10.1002/fgc.30511Read the full textAboutPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinkedInRedditWechat No abstract is available for this article. Volume43, Issue154 July 2019Pages 1-2 RelatedInformation

Does data matter? Statewide longitudinal data systems and student performance in the U.S.

ABSTRACTAlthough many industries have benefited from advances in data-driven technology, education is making small steps in capitalizing on the huge potential of data systems. Since 2005, the U.S. Federal Government has been making large grants to help states build statewide longitudinal data systems (SLDS) with the goal of improving programme and teacher evaluation and engage in data-driven decision-making. We analyse whether the introduction of SLDSs improved student performance, measured using test scores for math and reading tests for 4th and 8th graders, as well as high school graduation rates. We find no effects of SLDSs on student performance up to 10 years from implementation. However, we find suggestive evidence that these systems may have long-run effects, emphasizing the long-run nature of educational data collection and policy analysis.

The Impact of Inclusive STEM High Schools on Student Outcomes: a Statewide Longitudinal Evaluation of Texas STEM Academies

In recent years, the number of science, technology, engineering, and mathematics (STEM) specialized schools has been rapidly increasing internationally. In the United States, a new movement of establishing inclusive STEM high schools (ISHSs) aimed at expanding access to specialized STEM education for underrepresented and diverse student groups has begun to emerge. Due to its recent emergence, rigorous evidence on the impact of ISHSs on student outcomes is scarce and inconclusive. This study adds to this limited but growing body of empirical literature by examining whether ISHSs have an effect on student achievement, course-taking, and high school completion, and if effects vary by sociodemographic subgroups. Focusing on the Texas STEM academies—the largest cluster of ISHSs in the US, and analyzing a decade-long data from the Texas Statewide Longitudinal Data System, pooled regression analysis results showed that while ISHSs had no impact on student test scores in mathematics and science, the effects are positive on completing advanced level math courses in high school. Subgroup analyses indicated that ISHSs generally have no differential effects on student outcomes by sociodemographic subgroups, with the exception that they improve the rates of high school graduation for racial minority and low-income students.

Using Integrated Data to Design and Support Pay For Success Interventions

Pay for Success (PFS) interventions are increasingly being implemented in the U.S. and worldwide to assess social programs under a risk-sharing financial agreement between the public and private sectors. They seek to mitigate risk for the public sector and promote wider experimentation of programs to improve social outcomes. PFS contracts encourage coordination and alignment of goals, outcomes, and metrics across all agents involved - government, service providers, service recipients, funders and investors. Accordingly, these interventions rely heavily on access to high quality data and analysis, making integrated data systems (IDS) valuable assets to support the design, implementation, and evaluation phases of these projects. 
 The ChildHood Integrated Longitudinal Data (CHILD) System, one of the most comprehensive county-level IDS in the nation, has been used to support and inform two Pay for Success projects in Cuyahoga County (Cleveland). Partnering for Family Success is a county-level intervention in the areas of child welfare and housing instability, now into its fourth year of operation. While the intervention was implemented under a randomized controlled trial, analysis with CHILD proved instrumental to inform the project design and address challenges in program implementation. CHILD has also been used to study the feasibility of PFS as a model to expand high quality preschool, under a grant awarded to eight communities nationally. A case study of both initiatives will be presented, highlighting the role of integrated data in supporting and facilitating PFS design and analysis of outcomes, challenges encountered and lessons learned.

Antenatal Hospitalization Among U.S. Women With Intellectual and Developmental Disabilities: A Retrospective Cohort Study.

This population-based retrospective cohort study examines the prevalence of hospital utilization during pregnancy and the primary reason for antenatal hospital utilization among women with intellectual and developmental disabilities (IDD). Massachusetts residents with in-state deliveries that were ≥ 20 weeks gestational age were included via data from the 2002-2009 Massachusetts Pregnancy to Early Life Longitudinal Data System. Among women with IDD, 54.8% had at least one emergency department (ED) visit during pregnancy, compared to 23% of women without IDD. Women with IDD were more likely to have an antenatal ED visit, observational stays, and non-delivery hospital stays. This study highlights the need for further understanding of the health care needs of women with IDD during pregnancy.

Postpartum Hospital Utilization among Massachusetts Women with Intellectual and Developmental Disabilities: A Retrospective Cohort Study.

Objectives This study examined the risk of postpartum hospital admissions and emergency department (ED) visits among US women with intellectual and developmental disabilities (IDD). Methods We used the 2002-2012 Pregnancy to Early Life Longitudinal Data System and identified deliveries to women with and without IDD. Women with IDD (n = 1104) or case subjects were identified from the International Classification of Diseases and Related Health Problems 9th Revision (ICD-9 CM) codes. The study primary outcome measures were any postpartum hospital admission and any ED visit during three critical postpartum periods (1-42, 43-90, and 1-365days). We conducted unadjusted and adjusted survival analysis using Cox proportional hazard models to compare the occurrence of first hospital admission or ED visits between women with and without IDD. Results We found that women with IDD had markedly higher rates of postpartum hospital admissions and ED visits during the critical postpartum periods (within 1-42, 43-90, and 91-365days) after a childbirth. Conclusion for Practice Given the heightened risk of pregnancy complications and adverse birth outcomes and the findings of this study, there is an urgent need for clinical guidelines related to the frequency and timing of postpartum care among new mothers with IDD. Further, this study provides evidence of the need for evidence-based interventions for new mothers with IDD to provide preventive care and routine assessments that would identify and manage complications for both the mother and the infant outside of the traditional postpartum health care framework.

A New Lens on High School Dropout: Use of Correspondence Analysis and the Statewide Longitudinal Data System

ABSTRACTThe combination of log-linear models and correspondence analysis have long been used to decompose contingency tables and aid in their interpretation. Until now, this approach has not been applied to the education Statewide Longitudinal Data System (SLDS), which contains administrative school data at the student level. While some research has been conducted using the SLDS, its primary use is for state education administrative reporting. This article uses the combination of log-linear models and correspondence analysis to gain insight into high school dropouts in two discrete regions in Kentucky, Appalachia and non-Appalachia, defined by the American Community Survey. The individual student records from the SLDS were categorized into one of the two regions and a log-linear model was used to identify the interactions between the demographic characteristics and the dropout categories, push-out and pull-out. Correspondence analysis was then used to visualize the interactions with the expanded push-out categories, boredom, course selection, expulsion, failing grade, teacher conflict, and pull-out categories, employment, family problems, illness, marriage, and pregnancy to provide insights into the regional differences. In this article, we demonstrate that correspondence analysis can extend the insights gained from SDLS data and provide new perspectives on dropouts. Supplementary materials for this article are available online.

Collaboration Challenges and the Construction of Complex Data Systems: Lessons from the Development of the Illinois Longitudinal Data System

Over the last several years, government has invested heavily to create state longitudinal data systems. Policymakers stand poised to make use of these new systems to inform decision making in ways that were not possible in the past. Building such data systems requires the formation of complex networks involving the coordination and cooperation of large numbers of actors. Building these networks is no easy task and must be undertaken with careful planning. Employing Thomson and Perry's collaboration framework as a tool for analysis, this article examines the construction of one such data system, the Illinois Longitudinal Data System (ILDS), and offers lessons learned based on the successes and failures of its development. Findings focused on the role that negative histories and mandated formation play in network development.

Maternal Substance Use Disorders and Infant Outcomes in the First Year of Life among Massachusetts Singletons, 2003-2010

To determine the association of maternal substance use disorders (SUDs) during pregnancy with adverse neonatal outcomes and infant hospital re-admissions, observational stays, and emergency department utilization in the first year of life. We analyzed 2 linked statewide datasets from 2002 to 2010: the Massachusetts Pregnancy to Early Life Longitudinal data system and the Massachusetts Bureau of Substance Abuse Services Management Information System. Generalized estimating equations were used to assess the association of maternal SUDs and neonatal outcomes and infant hospital-based care in the first year of life, controlling for maternal and infant characteristics. Maternal SUDs increased from 19.4 per 1000 live births in 2003 to 31.1 per 1000 live births in 2009. In the adjusted analysis, exposed neonates were more likely to be born preterm (aOR 1.85; 95% CI 1.75-1.96) and low birthweight (aOR 1.94; 95% CI 1.80-2.09). After controlling for maternal characteristics and preterm birth, SUD-exposed neonates were more likely to have intrauterine growth restriction, cardiac, respiratory, neurologic, infectious, hematologic, and feeding/nutrition problems, prolonged hospital stay, and higher mortality (aOR range 1.26-3.80). Exposed infants were more likely to be rehospitalized (aOR 1.10; 95% CI 1.04-1.17) but less likely to have an observational stay (aOR 0.90; 95% CI 0.82-0.99) or use the emergency department (aOR 0.87; 95% CI 0.83-0.90) in the first year of life. Infants born to mothers with SUD are at higher risk for adverse health outcomes in the perinatal period and are also more likely to be rehospitalized in the first year of life.

Severe Maternal Morbidity at Delivery and Risk of Hospital Encounters Within 6 Weeks and 1 Year Postpartum

Little is known about the impact of severe maternal morbidity (SMM) after delivery. We examined the risk of rehospitalization in the first year postpartum among deliveries to women with and without SMM. We used the Pregnancy to Early Life Longitudinal data system, in which vital birth/fetal death records were linked with hospital delivery discharge data and subsequent nondelivery hospitalization data, including observational stays (OSs) and in-patient stays (hospital discharge [HD]) for Massachusetts residents during 2002-2011. We excluded deliveries to women with preexisting chronic conditions: hypertension, diabetes, asthma, and autoimmune conditions for a final sample of 685,228 deliveries. Multivariable log binomial regression with generalized estimating equations modeled the relative risk (RR) of hospital encounters 6 weeks and 1 year postpartum. The rate of SMM was 99 per 10,000 deliveries. In the first year postpartum, 2.8% of deliveries to women without chronic medical conditions experienced at least one HD encounter and 1.0% at least one OS encounter. The adjusted relative risk (aRR) of any HD encounter for deliveries with SMM was 2.48 (95% confidence interval [CI]: 2.20-2.80) within 6 weeks postpartum and 2.04 (95% CI: 1.87-2.23) within 1 year. For OS encounters, aRRs among deliveries with SMM at delivery were 2.47 (95% CI: 1.94-3.14) in the first 6 weeks and 1.69 (95% CI: 1.43-2.01) in 1 year. In Massachusetts, SMM increased the risk of rehospitalization in the first year postpartum among deliveries to women without chronic medical conditions.