Sibling Fixed-effects Research Articles

Using data on biomarkers and siblings to study early-life economic determinants of type-2 diabetes.

We study the effect of economic conditions early in life on the occurrence of type-2 diabetes in adulthood using contextual economic indicators and within-sibling pair variation. We use data from Lifelines: a longitudinal cohort study and biobank including 51,270 siblings born in the Netherlands from 1950 onward. Sibling fixed-effects account for selective fertility. To identify type-2 diabetes we use biomarkers on the hemoglobin A1c concentration and fasting glucose in the blood. We find that adverse economic conditions around birth increase the probability of type-2 diabetes later in life both in males and in females. Inference based on self-reported diabetes leads to biased results, incorrectly suggesting the absence of an effect. The same applies to inference that does not account for selective fertility.

Child health and its effect on adult social capital accumulation.

Although studies have demonstrated important effects of poor health in childhood on stocks of human and health capital, little research has tested economic theories to investigate the effect of child health on social capital in adulthood. Studies on the influence of child health on adult social capital are mixed and have not used sibling fixed effects models to account for unmeasured family and genetic characteristics, that are likely to be important. Using the Add-Health sample, health in childhood was assessed as self-rated health, the occurrence of a physical health condition or mental health condition, while social capital in adulthood was measured as volunteering, religious service attendance, team sports participation, number of friends, social isolation, and social support. We used sibling fixed effects models, which attenuated several associations to non-significance. In sibling fixed effects models there was significant positive effects of greater self-rated health on participation in team sports and social support, and negative effect of mental health in childhood on social isolation in adulthood. These results suggest that children with poor health require additional supports to build and maintain their stock of social capital and highlight further potential benefits to efforts that address poor child health.

Impacts of broadband internet on adolescents’ academic outcomes: heterogeneous effects among lower secondary school students in Norway

ABSTRACT The expansion of internet is likely to influence adolescents’ academic outcomes. Yet, how internet coverage impacts students’ educational performance remains poorly understood. To address this major knowledge gap, this study uses a quasi-experimental approach with Norwegian data to causally examine how the gradual introduction of home broadband internet across municipalities impacted the academic outcomes of lower-secondary school graduates (N = 103,796). Analyses apply sibling fixed-effects models with micro-level registry data from adolescents aged 15–16, and compare differences in effects by gender, social background, migrant status, and academic achievement levels. Findings show that the introduction of broadband internet across municipalities brought moderate grade improvements concentrated among boys in the subject areas of Mathematics, Arts and Crafts, Social Sciences, and Norwegian. The positive effect of broadband internet coverage on academic performance was three times larger for boys than for girls. For boys, broadband internet coverage led to strong grade improvements among the lower-achieving and socioeconomically disadvantaged, and to moderate grade benefits among those of Norwegian background, while boys from higher-achieving and socioeconomically privileged groups reduced their grades moderately. By contrast, for girls, broadband internet coverage worsened substantially the academic performance of those from disadvantaged socioeconomic backgrounds, whereas girls of migrant background obtained higher grades with this expansion of internet. The study findings reveal how the growth of internet coverage impacts adolescent educational performance, but differently across groups, showing a complex intersection across gender, social background, migrant status, and academic achievement levels.

Childhood personality and academic performance: A sibling fixed-effects study.

This study investigated the associations between personality traits at age 8 and academic performance between ages 10 and 14, controlling for family confounds. Many studies have shown links between children's personality traits and their school performance. However, we lack evidence on whether these associations remain after genetic and environmental confounders are accounted for. Sibling data from the Norwegian Mother and Child Cohort Study (MoBa) were used (n = 9701). First, we estimated the overall associations between Big Five personality traits and academic performance, including literacy, numeracy, and foreign language. Second, we added sibling fixed effects to remove unmeasured confounders shared by siblings as well as rating bias. Openness to Experience (between-person β = 0.22 [95% CI: 0.21-0.24]) and Conscientiousness (between-person β = 0.18 [95% CI 0.16-0.20]) were most strongly related to educational performance. Agreeableness (between-person β = 0.06 [95% CI -0.08-0.04]) and Extraversion (between-person β = 0.02 [95% CI 0.00-0.04]) showed small associations with educational performance. Neuroticism had a moderate negative association (between-person β = -0.14 [95% CI -0.15-0.11]). All associations between personality and performance were robust to confounding: the within-family estimates from sibling fixed-effects models overlapped with the between-person effects. Finally, childhood personality was equally predictive of educational performance across ages and genders. Although family background is influential for academic achievement, it does not confound associations with personality. Childhood personality traits reflect unbiased and consistent individual differences in educational potential.

Genetic endowments for social capital: An investigation accounting for genetic nurturing effects

Despite social capital having been shown to be important for health and well-being, relatively little research has examined genetic determinants. Genetic endowments for education have been shown to influence human, financial, and health capital, but few studies have examined social capital, and those conducted have yet to account for genetic nurturing. We used the Add-Health data to study the effect of genetic endowments on individual social capital using the education polygenic score (PGS). We used sibling fixed effects models and controlled for the family environment to account for genetic nurturing. After accounting for the family environment, we found moderately large significant associations between the education PGS and volunteering, but associations with religious service attendance and number of friends were completely attenuated in sibling fixed effects models. These findings highlight that genetic endowments play an important role in influencing volunteering and the importance of accounting for genetic nurturing.

Number of children and disability pension due to mental and musculoskeletal disorders: A longitudinal register-based study in Norway

Earlier research has documented a relationship between parity and all-cause mortality, as well as parity and cause-specific mortality (e.g. cancer and cardiovascular disease mortality). Less is known about the relationship between parity and two very common (but less deadly) types of disorder: mental and musculoskeletal. We examine the association between parity and risk of disability pensioning from all causes and due to mental or musculoskeletal disorders, using Norwegian register data. In addition to controlling for adult socio-demographic characteristics, we control for unobserved confounding from family background by estimating sibling fixed-effects models. We find a higher risk of disability pensioning among the childless and those with one child than for parents with two children, both for all causes combined and for mental disorders. Childless men and fathers with one child also experience excess risk of being pensioned due to musculoskeletal disorders. For mental disorders, we find a positive association with high parity, particularly for men.

Re-examining the relationship between education and adult mental health in the UK: A research note

Previous studies using variation in education arising from compulsory schooling laws have found no causal effects of education on mental health in the UK. We re-examine the relationship between education and mental health in the UK by taking a different approach: sibling fixed-effects with controls for polygenic scores (summary measures of genetic predisposition) for educational attainment and adult depressive symptoms. We find that higher educational attainment is associated with better adult mental health, that sibling controls reduce these associations by ∼40–70% but important associations remain and find evidence for non-monotonic effects. We also find suggestive evidence that education partially “rescues” genetic predictors of poor mental health.

Longitudinal pathways from adolescent depressive symptoms to cardiovascular disease risk in adulthood

RationaleDespite existing literature on the impact of adolescent depressive symptoms on a variety of youth outcomes, little is known about whether and how adolescent depressive symptoms are associated with physical health in adulthood. ObjectiveThis study aimed to investigate the longitudinal association between adolescent depressive symptoms and cardiovascular disease (CVD) risk in adulthood. An extensive list of potential mechanisms underlying this association, including health behaviors, substance use, academic effort and achievement, and school-based relationships, was also examined. MethodsUsing the sibling sample from the National Longitudinal Study of Adolescent to Adult Health (N = 3011), this study employed sibling fixed-effects models to address unobservable family-level confounders, such as genetic factors, parental and familial characteristics, and school and neighborhood environments. ResultsAdolescent depressive symptoms were associated with an increased risk of CVD in about 13 years later (b = 0.698). Sibling fixed-effects estimates revealed that, despite some attenuation (approximately 31%), this association is robust to adjustment for unobserved family-level heterogeneity. Substance use and academic effort/achievement explained about 35% and 16% of the association between adolescent depressive symptoms and adult CVD risk, respectively. In contrast, health behaviors and school-based relationships played little or no role in mediating the association. ConclusionThe findings of this study suggest that policymakers and practitioners may wish to develop interventions to reduce adolescent depressive symptoms and to help adolescents with depressive symptoms avoid substance use and promote effort in school. Early interventions targeting adolescents with depressive symptoms may be a cost-effective strategy for preventing adult physical health problems.

Self-Employment and Eudaimonic Well-Being: The Mediating Role of Problem- and Emotion-Focused Coping

An emerging body of research has documented that self-employed people are more likely to report higher levels of eudaimonic well-being (EWB; e.g., autonomy, competence, meaning) than their employed counterparts. In this paper, we examine why the self-employed perceive their lives as psychologically more fulfilling even though they face more complex and competing occupational demands that can expose them to more stressors. Specifically, we hypothesize that the self-employed are more likely to engage in problem-focused coping—productive and proactive behaviors and thoughts aimed to help them overcome challenges (e.g., planning and active coping)—and less likely to engage in emotion-focused coping—behaviors and thoughts to merely make them feel better (e.g., venting and denial)—which, in turn, can promote higher levels of EWB. Using data from Waves 2 and 3 of the National Study of Midlife in Development in the United States, we find supportive evidence for our theory. More importantly, we show that the well-being benefits from self-employment accrue almost entirely because the self-employed are more likely to use problem-focused coping as opposed to emotion-focused coping. In a series of robustness tests, including random-effects models, matching estimators, and twin and sibling fixed-effects, we further demonstrate the relevance of coping as a key explanatory mechanism in the relationship between self-employment and EWB.

Demystifying the link between higher education and liberal values: A within-sibship analysis of British individuals' attitudes from 1994-2020.

The link between university graduation and liberal values is well-established and often taken as evidence that higher education participation causes attitudinal change. Identification of education's causal influence in shaping individual preferences is notoriously difficult as it necessitates isolating education's effect from self-selection mechanisms. This study exploits the household structure of the Harmonized British Household Panel Study and Understanding Society data to tighten the bounds of causal inference in this area and ultimately, to provide a more robust estimate of the independent effect of university graduation on political attitudes. Results demonstrate that leveraging sibling fixed-effects to control for family-invariant pre-adult experiences reduces the size of higher education's effect on cultural attitudes by at least 70%, compared to conventional methods. Significantly, within-sibship models show that obtaining higher education qualifications only has a small direct causal effect on British individuals' adult attitudes, and that this effect is not always liberalizing. This has important implications for our understanding of the relationship between higher education and political values. Contrary to popular assumptions about education's liberalizing role, this study demonstrates that the education-political values linkage is largely spurious. It materializes predominately because those experiencing pre-adult environments conducive to the formation of particular values disproportionately enroll at universities.

Adolescent delinquent behavior and sleep deficiency: a test of multiple mechanisms using sibling comparison designs.

The aim of the study was to investigate the association between adolescent delinquency and sleep deficiency. A comprehensive set of potential mechanisms underlying this association were also examined. Using data from the National Longitudinal Study of Adolescent to Adult Health, the present study employed sibling fixed effects models to account for unobservable family-level confounders, such as genetic predisposition, parenting style, parental ability, and school and neighborhood environments. In Sobel mediation tests, the following mechanism variables were explored: substance use, school-based relationships, and parent-child relationships. An increase in delinquency (measured by the total number of types of delinquent behavior engaged) was associated with an increased risk of sleep deficiency one year later. Sibling fixed effects models with a lagged dependent variable revealed that this association is robust to adjustment for family-level heterogeneity as well as prior sleep deficiency. Substance use was the most salient pathway linking delinquency to sleep deficiency (17% for binge drinking and 26% for marijuana use), followed by student-teacher relationships (17%) and father-child relationships (16%). The results of this study suggest that policymakers and practitioners may consider developing interventions to help delinquent adolescents avoid substance use and restore disruptions of student-teacher and father-child relationships.

The Effect of Parental Loss on Social Mobility in Early Twentieth-Century Sweden.

Parents are assumed to play a crucial role in the socioeconomic attainment of children. Through investments of both time and resources, they promote the ability, human capital, networks, and motivation of their children to advance socially, or at least to maintain their social position. Consequently, losing a parent in childhood could be detrimental to adult socioeconomic outcomes. We use full-count linked census data and a comprehensive death register to study the effect of parental loss on socioeconomic outcomes in adulthood in Sweden during the first half of the twentieth century. We employ sibling fixed-effects models and the Spanish flu as an exogenous mortality shock to assess the importance of endogeneity bias in associations between parental loss and socioeconomic outcomes. Maternal death led to worse socioeconomic outcomes in adulthood in terms of occupational and class attainment, as well as for social mobility. The effects seem to be causal but the magnitudes were small. For paternal death, we find no consistent pattern, and in most models there was no effect on sons' socioeconomic outcomes. The patterns were similar for sons and daughters and do not support the theory that parental loss had important negative effects on socioeconomic outcomes in adulthood.

Association between a history of clinical depression and dementia, and the role of sociodemographic factors: population-based cohort study.

Depression is associated with an increased dementia risk, but the nature of the association in the long-term remains unresolved, and the role of sociodemographic factors mainly unexplored. To assess whether a history of clinical depression is associated with dementia in later life, controlling for observed sociodemographic factors and unobserved factors shared by siblings, and to test whether gender, educational level and marital status modify the association. We conducted a national cohort study of 1 616 321 individuals aged 65 years or older between 2001 and 2018 using administrative healthcare data. A history of depression was ascertained from the national hospital register in the period 15-30 years prior to dementia follow-up. We used conventional and sibling fixed-effects Cox regression models to analyse the association between a history of depression, sociodemographic factors and dementia. A history of depression was related to an adjusted hazard ratio of 1.27 (95% CI 1.23-1.31) for dementia in the conventional Cox model and of 1.55 (95% CI 1.09-2.20) in the sibling fixed-effects model. Depression was related to an elevated dementia risk similarly across all levels of education (test for interaction, P = 0.84), but the association was weaker for the widowed than for the married (P = 0.003), and stronger for men than women (P = 0.006). The excess risk among men attenuated following covariate adjustment (P = 0.10). This study shows that a history of depression is consistently associated with later-life dementia risk. The results support the hypothesis that depression is an aetiological risk factor for dementia.

Pathways to Skin Color Stratification: The Role of Inherited (Dis)Advantage and Skin Color Discrimination in Labor Markets

Research has uncovered associations between skin color and myriad outcomes. What drives these associations? We develop a theoretical framework that synthesizes the multiple pathways linking skin color with life chances. Skin color stratification should be conceptualized in historical, structural terms: as the result of unequal treatment and inherited (dis)advantage, that is, unequal resources transmitted by families with different skin tones. We assess the role of two pathways— discrimination and inherited (dis)advantage—for Blacks' and Latinos' employment, earnings, and occupational prestige. We use the National Longitudinal Study of Youth 1997, which includes a visual skin color measure; multiple indicators of family background; and a sibling subsample that allows us, using fixed-effects models, to recover the effect of skin color net of family background. First, we find that darker skin tone is associated with worse labor market outcomes. Indicators of family background account for 29 to 44 percent of skin color’s associations with employment, earnings, and occupational prestige. Second, using sibling fixed-effects models, we find that darker skin tone is associated with worse labor market outcomes, but these associations are not statistically significant. In sum, our findings suggest that we pay attention to the multiple pathways linking skin color with life chances.

Genetic risks, adolescent health, and schooling attainment.

We provide new evidence on the effect of adolescent health behaviors/outcomes (obesity, depression, smoking, and attention deficit hyperactivity disorder [ADHD]) on schooling attainment using the National Longitudinal Study of Adolescent to Adult Health. We take two different approaches to deal with omitted variable bias and reverse causality. Our first approach attends to the issue of reverse causality by estimating the effect of health polygenic scores (PGSs) on schooling. Second, we estimate the effect of adolescent health using sibling fixed-effects models that control for unmeasured genetic and family factors shared by siblings. We use the PGSs as additional controls in the sibling fixed-effects models to reduce concerns about residual confounding from sibling-specific genetic differences. We find consistent evidence across both approaches that being genetically predisposed to smoking and smoking regularly in adolescence reduces schooling attainment. Estimates for depression are more imprecise, but also suggest that a high genetic risk of depression and adolescent depression reduce schooling attainment. We find mixed evidence for ADHD. Our estimates suggest that having a high genetic risk for ADHD reduces grades of schooling, but we do not find any statistically significant negative effects of ADHD. Finally, we find no consistent evidence for a detrimental effect of obesity on schooling attainment.

Child health, human capital, and adult financial behavior.

This paper provides novel evidence that child health affects adult financial behavior, that is, risky asset market participation. We do so by using a longitudinal dataset with a rich set of covariates and exploit sibling fixed-effects (FE) to control for invariant unobserved heterogeneity. We begin by proposing a new mechanism working via skill formation and portfolio choice. To be precise, we test two hypotheses. First, we expect a negative correlation between poor child health and risky asset market participation. Second, this correlation should be mostly explained by differences in skills. To test these hypotheses, we use data from the National Longitudinal Study of Youth 1979 (NLSY79). Our results show that poor child health is associated with an 11 pp decrease in adult risky market participation conditional on demographics and family background. Moreover, our results suggest that disruption in pre-labor market skill formation is a main mediator of this relationship. These results are robust to a wide range of robustness checks. Our findings have implications for the design of health policies and policies intended to increase financial literacy and asset market participation.

Parental divorces and children’s educational outcomes in Senegal

This paper provides new evidence on the consequences of parental divorce for children in Africa. Using survey data that collected the detailed life histories of Senegalese women and of their children, we investigate how children’s educational outcomes are affected by their parents’ divorce. We use a sibling fixed-effects strategy that allows us to control for all the factors that are common to all children in a family, such as parental preferences regarding education or the level of education of the parents, alleviating concerns of omitted variable bias. We compare children who were old enough to have been enrolled in primary school at the time of the divorce to their younger siblings, for whom enrollment decisions had not yet been made at the time of the divorce. We find that younger siblings were more likely than their older siblings to have attended primary school. This higher level of investment does not persist in the long run: there are no differences between siblings when considering primary school completion. We find that custody and fostering decisions do not seem to mediate the positive effects on school attendance. Our findings are consistent with either an improvement of the financial situation (due to remarriage) or an increase in the decision-making power of mothers after the divorce.

Revisiting the relationship between perceived discrimination and health: Evidence from sibling models with multiple health measures

Perceived discrimination (PD) is known to be significantly related to health outcomes. This link, however, warrants further scrutiny due to the possibility of unobserved family-level confounding. Using the Add Health and sibling fixed-effect approach, we examine whether the relationship between PD and health is confounded by family background characteristics such as genetics, family environment, and childhood social context (school and neighborhood effects). While the naive OLS models exhibit significant associations between PD and both physical and psychological health outcomes, our preferred sibling fixed-effect estimates reveal that the observed associations between PD and physical health outcomes are confounded by shared family background. In contrast, the observed associations for psychological health, self-reported health, and some of health behavior outcomes are robust to adjustment for sibling fixed-effects. Furthermore, we find similar overall patterns in the link between PD and health across races/ethnicities.

Gender Differences in the Educational Penalty of Delinquent Behavior: Evidence from an Analysis of Siblings

This study examines: (a) whether the association between juvenile delinquency and educational attainment differs by gender, and (b) which factors underlie such gender differences. In order to account for the influence of unobservable family-background factors, this study applies sibling fixed-effects models on the National Longitudinal Study of Adolescent to Adult Health (Add Health). As a sensitivity check, I explore whether observed gender differences are robust to different measurements of delinquency and the potential presence of sibling spillover effects. Nearly half of the association between juvenile delinquency and educational attainment is attributable to unobservable factors related to family background. This study finds that controlling for unobserved family-level heterogeneity substantially attenuates the association between juvenile delinquency and educational attainment among females, making it no longer statistically significant. Among males, sibling fixed-effects estimates suggest that a one-standard-deviation increase in delinquent involvement is associated with a reduction in 0.23 years of schooling and a 4.6 percentage point increase in the probability of high school dropout. Supplementary analyses show that male delinquents face major disadvantages in social relationships in school settings and display lower levels of educational aspirations as well as effort. No such patterns are found among female delinquents. This study finds a negative association between delinquency and educational attainment only for males but not for females. Results suggest that failure to account for unobserved family-level heterogeneity spuriously inflates the delinquency–education association to a larger extent among females than males.

Genetic Risks, Adolescent Health and Schooling Attainment

We provide new evidence on the effect of adolescent health behaviors/outcomes (obesity, depression, smoking, and attention deficit hyperactivity disorder (ADHD)) on schooling attainment using the National Longitudinal Study of Adolescent to Adult Health. We take two different approaches to deal with omitted variable bias and reverse causality. Our first approach attends to the issue of reverse causality by using health polygenic scores (PGSs) as proxies for actual adolescent health. Second, we estimate the effect of adolescent health using sibling fixed-effects models that control for unmeasured genetic and family factors shared by siblings. We use the PGSs as additional controls in the sibling fixed-effects models to reduce concerns about residual confounding from sibling-specific genetic differences. We find consistent evidence across both approaches that being genetically predisposed to smoking and smoking regularly in adolescence reduces schooling attainment. We find mixed evidence for ADHD. Our estimates suggest that having a high genetic risk for ADHD reduces grades of schooling, but we do not find any statistically significant negative effects of ADHD on grades of schooling. Finally, results from both approaches show no consistent evidence for a detrimental effect of obesity or depression on schooling attainment.