Improving the Interpretation of Random Effects Regression Results

Mummolo and Peterson improve the use and interpretation of fixed-effects models by pointing out that unit intercepts fundamentally reduce the amount of variation of variables in fixed-effects models. Along a similar vein, we make two claims in the context of random effects models. First, we show that potentially large reductions in variation, in this case caused by quasi-demeaning, also occur in models using random effects. Second, in many instances, what authors claim to be a random effects model is actually a pooled model after the quasi-demeaning process, affecting how we should interpret the model. A literature review of random effects models in top journals suggests that both points are currently not well understood. To better help users interested in improving their interpretation of random effects models, we provide Stata and R programs to easily obtain post-estimation quasi-demeaned variables.

Adolescent Violent Delinquency Associated With Increased Emergency Department Usage in Young Adulthood

Limited research has examined the extent to which adolescent delinquency predicts healthcare usage in young adulthood, including emergency department (ED) visits. This study used data from 3,310 adolescents (52.05% female; mean age at Wave I = 16.04 years) from the sibling subsample of the National Longitudinal Study of Adolescent to Adult Health (Add Health). We examined whether adolescent delinquency at Wave I predicted ED visits at Wave III using sibling fixed effects models to adjust estimates for within-family unobserved heterogeneity. Increased violent, but not nonviolent, delinquency predicted a higher number of ED visits in early adulthood in the sibling fixed effects models. To our knowledge, this is the first study to examine the relationship between delinquency and ED usage using a sibling fixed effects design. Findings demonstrate that violent adolescent delinquency may increase healthcare usage and suggest the potential role of healthcare providers in improving outcomes for delinquent youth.

Association between Apolipoprotein E Polymorphism and Clinical Outcome after Ischemic Stroke, Intracerebral Hemorrhage, and Subarachnoid Hemorrhage

<b><i>Backgrounds:</i></b> Previous studies reported inconsistent results regarding associations between apolipoprotein E (<i>APOE</i>) polymorphism and clinical outcomes after ischemic stroke (IS), intracerebral hemorrhage (ICH), or subarachnoid hemorrhage (SAH). Thus, the study was designed to make a systematic review and meta-analysis regarding the association between <i>APOE</i> polymorphism and clinical outcome after IS, ICH, and SAH. <b><i>Methods:</i></b> To identify studies eligible for this meta-analysis, we searched for articles published before August 2021 in the databases (PubMed, Web of Science, and Google Scholar). We used STATA 12.0 software to compute hazard ratios (HRs) and their 95% confidence intervals (CIs) regarding <i>APOE</i> polymorphism and clinical outcome after IS, ICH, and SAH. <b><i>Results:</i></b> Meta-analysis showed no significant association between <i>APOE</i> polymorphism and functional outcome after IS with fixed effects models (ε4 carrier vs. non-ε4 carrier: HR, 1.00; 95% CI: 0.83–1.21, <i>I</i>² = 29.4%, <i>p</i> = 0.183; ε2 carrier vs. non-ε2 carrier: HR, 0.92; 95% CI: 0.72–1.16, <i>I</i>² = 15.6%, <i>p</i> = 0.307). Meta-analysis showed that ICH patients carrying ε4 allele have increased risk of poor outcome in Caucasian population with fixed effects models (ε4 carrier vs. non-ε4 carrier: HR, 1.75; 95% CI: 1.19–2.57, <i>I</i>² = 0.0%, <i>p</i> = 0.543). Meta-analysis showed no significant association between <i>APOE</i> polymorphism and functional outcomes after SAH with random effects models (ε4 carrier vs. non-ε4 carrier: HR, 1.51; 95% CI: 0.80–2.84, <i>I</i>² = 57.1%, <i>p</i> = 0.022). <b><i>Conclusions:</i></b> In conclusion, the present study demonstrated <i>APOE</i> ε4 carriers show worse functional outcomes after ICH, but not after IS or SAH. More large-scale studies were critical to explore the association between <i>APOE</i> polymorphism and clinical outcome after IS, ICH, and SAH.

Driving turnout: the effect of car ownership on electoral participation

Inequalities in voter participation between groups of the population pose a problem for democratic representation. We use administrative data on 6.7 million registered voters to show that a previously-ignored characteristic of voters—access to a personal automobile—creates large disparities in in-person voting rates. Lack of access to a car depresses election day voter turnout by substantively large amounts across a variety of fixed-effects models that account for other environmental and voter characteristics. Car access creates the largest hindrance to voting for those people who live farther from the polls. These effects do not appear for absentee voting, suggesting a simple policy solution to solve large disparities in political participation. This study contributes to the theoretic understanding of political participation as well as the impact of potential policy reforms to solve participatory gaps.

The Reputational Dividends of Collaborating with a Highly Reputable Agency: The Case of the FDA and Its Domestic Partner Agencies

What, if any, dividends do agencies reap from collaboration with a highly reputable agency, such as the FDA? Utilizing a dataset covering 30 U.S. federal agencies over a period of 34 years (1980–2013), we estimate the short and long-term reputational effects of interagency collaboration. Collaboration is measured by the number of memorandums of understanding (MOUs) in effect between each agency and the FDA, while agency reputation is assessed using an automated measure of media-coverage valence (positive/negative tone) for each agency-year. To account for potential reverse and reciprocal causality, we utilize cross-lagged fixed-effects models. We find evidence of moderate rises in reputation due to increased collaboration with the FDA. These effects persist significantly for two years, before decaying to null after four years. Employing similar analyses, we furthermore estimate reversed causality – of reputation on the level of consequent collaboration – finding no evidence of such effects.Research SupportMoshe Maor and Raanan Sulitzeanu-Kenan gratefully acknowledge financial support from the Israel Science Foundation under grant 1002/11.

Implementing the panel event study

Many studies estimate the impact of exposure to some quasiexperimental policy or event using a panel event study design. These models, as a generalized extension of “difference-in-differences” designs or two-way fixed-effects models, allow for dynamic leads and lags to the event of interest to be estimated, while also controlling for fixed factors (often) by area and time. In this article, we discuss the setup of the panel event study design in a range of situations and lay out several practical considerations for its estimation. We describe a command, eventdd, that allows for simple estimation, inference, and visualization of event study models in a range of circumstances. We then provide several examples to illustrate eventdd’s use and flexibility, as well as its interaction with various native Stata commands, and other relevant community-contributed commands such as reghdfe and boottest.

The Spatial Effect of Aging Population on Innovation in Entrepreneurship--Based on Empirical Analysis of Western China

This study researches the impact of an aging population on Innovation in Entrepreneurship (IE), and applying fixed effects models (FE), mediated effects models and spatial lag regression model (SAR) to panel data of Western China (excluding Tibet) from 2004 to 2019.The results showed that an aging population and IE inverted significantly U-curve, and human capital plays a significant partial mediation between the two. A theoretical perspective based on the First Law of Geography, in the western China, aging population and IE are both positive spatial correlation, and both show the characteristics of "High-High" spatial agglomeration. Under the spatial model, aging population and IE also inverted U-curve.

The effects of parenting on early adolescents’ noncognitive skills: Evidence from a sample of twins in Germany

Many theories in the social sciences assume that parenting affects child development. Previous research mostly supports the notion that parenting affects the skill development of children in early childhood. There are fewer studies testing whether parenting in early adolescence has such an influence. We estimate the effects of parenting on early adolescents’ noncognitive skills using data from the German Twin Family Panel (TwinLife). Specifically, we look at the effects of parenting styles, parental activities, and extracurricular activities on the academic self-concept, motivation, self-esteem, self-efficacy, and locus of control of 10 to 14 years old children. To control for unobserved heterogeneity and reverse causality, we employ twin fixed-effects models combined with longitudinal information. In addition, MZ twin fixed effects models also control for genetic confounding. Our findings provide no support to the notion that parenting styles, parental activities, and extracurricular activities in early adolescence affect the development of children's noncognitive skills. We conclude that our results, in combination with the majority of evidence from previous research, are in line with a model according to which parenting has larger effects on the skill development of children in early childhood than in early adolescence.