Unintended Consequences: Effects of Paternal Incarcerationon Child School Readiness and Later Special EducationPlacementAnna R. HaskinsColumbia University

Abstract: Though sociologists have examined how mass incarceration affects stratification, remarkably little is knownabout how it shapes educational disparities. Analyzing the Fragile Families Study and its rich paternal incarcerationdata, I ask whether black and white children with fathers who have been incarcerated are less prepared for school bothcognitively and non-cognitively as a result, and whether racial and gendered disparities in incarceration help explain thepersistence of similar gaps in educational outcomes and trajectories. Using a variety of estimation strategies, I show thatexperiencing paternal incarceration by age five is associated with lower non-cognitive school readiness. While the maineffect of incarceration does not vary by race, boys with incarcerated fathers have substantially worse non-cognitive skillsat school entry, impacting the likelihood of special education placement at age nine. Mass incarceration facilitates theintergenerational transmission of male behavioral disadvantage, and because of the higher exposure of black children toincarceration, it also plays a role in explaining the persistently low achievement of black boys.

Keywords: educational inequality; mass incarceration; school readiness; racial disparities; gender differences; specialeducationEditor(s): Jesper Sørensen, Sarah A. Soule; Received: February 3, 2014; Accepted: Feburary 12, 2014; Published: April 21, 2014

Citation: Haskins, Anna R. 2014. “Unintended Consequences: Effects of Paternal Incarceration on Child School Readiness and Later Special EducationPlacement.” Sociological Science 1: 141-158. DOI: 10.15195/v1.a11

Copyright: c© 2014 Haskins. This open-access article has been published and distributed under a Creative Commons Attribution License, which allowsunrestricted use, distribution and reproduction, in any form, as long as the original author and source have been credited.

Over the past forty years, large gaps in aca-demic achievement and educational attain-

ment between black and white children have per-sisted (Jencks and Phillips 1998; Magnuson andWaldfogel 2008), presenting one of the most en-during problems facing a U.S. society committedto equal opportunity and racial justice. Alongsidethese racial disparities is a gender gap which is es-pecially pronounced for boys from disadvantagedbackgrounds, such that they have substantiallylower achievement trajectories throughout theireducational careers than similar female peers (En-twisle, Alexander and Olson 2007; Buchmann,DiPrete, and McDaniel 2008). In this paper,I argue that the racial and gendered dynamicsthat influence schooling trajectories for U.S. chil-dren from disadvantaged backgrounds are driven,in part, by the race- and gender-specific effectsof mass incarceration on early educational out-comes.

Incarceration has strong, negative, and last-ing effects on myriad aspects of imprisoned and

formerly imprisoned individuals’ lives (Wakefieldand Uggen 2010; Wildeman and Muller 2012),as well as on the wellbeing of families and com-munities (Clear 2007; Comfort 2008). Less wellunderstood is the impact of mass incarcerationfor child development. More than half of stateand federal inmates report having a minor child,leaving an estimated 1.5 million children with cur-rently imprisoned parents (Glaze and Maruschak2008). When paroled and recently released par-ents are included, this number rises to 3.2 millionnationwide, the largest proportions of whom areblack and have incarcerated fathers (Mumola2002; Parke and Clarke-Stewart 2003; Glaze andMaruschak 2008).

This paper brings scholarship on educationalinequalities into dialogue with the growing bodyof research on the consequences of imprisonmentfor families. Drawing on longitudinal birth-cohortdata from the Fragile Families and Child Wellbe-ing Study, I first ask whether paternal incarcera-tion diminishes young children’s school readiness.

sociological science | www.sociologicalscience.com 141 April 2014 | Volume 1

Haskins Effects of Paternal Incarceration on Child School Readiness

I then consider whether such effects vary by therace and gender of the child. Lastly, I test theextent to which school readiness mediates therelationship between paternal incarceration andspecial education placement, a particularly im-portant early schooling decision that not only hasbeen linked to later achievement and attainmenttrajectories (Morgan et al. 2010; Shifrer et al.2013), but also reflects disproportionalities byrace and gender (Hibel et al. 2010; Zhang et al.2014).

To disentangle the unique impact of paternalincarceration from the effects of preexisting disad-vantage, I condition on a wide range of covariatesthat address the process of social selection, usepropensity score matching techniques to identifyan appropriate comparison group for childrenwith incarcerated fathers in order to estimate a“treatment effect for the treated,” and conducttwo types of sensitivity analysis to assess theextent to which observed effects of paternal incar-ceration on school readiness are robust to omittedvariable bias. I find that paternal incarcerationnegatively affects children’s school readiness andthat these effects are strongest for boys and dis-proportionally experienced by black children, cre-ating lasting consequences for their educationaltrajectories. These results enhance our under-standing of how inequality is transmitted fromparent to child and the role mass incarcerationplays in the persistence of racial and gendereddisparities in educational outcomes.

Paternal Incarcerationand Children’s EducationalDevelopmentPaternal incarceration can affect children emo-tionally, developmentally, and socially through:trauma experienced as a result of parent-childseparation (e.g. Braman 2004; Comfort 2008);the isolation and shame brought on by the stigmaassociated with having a family member incarcer-ated (e.g. Goffman 1963; Murray and Farrington2008); and the social, psychological, and economicstrain imposed upon children of the incarcerateddue to financial hardship and family disruptionor dissolution (e.g. Hagan and Dinovitzer 1999;Swisher and Waller 2008). Moreover, Geller et

al. (2012) find that these harmful effects canoccur regardless of the resident status of fathersat the time of their incarceration, suggesting thata change in a father’s presence is a significant butnot necessary condition for children to be harmedby paternal incarceration.

Experiencing any of the social, emotional, ormaterial consequences of paternal incarcerationduring early childhood can be quite detrimen-tal, as this is a particularly sensitive develop-mental stage during which the foundation for achild’s cognitive and non-cognitive capacities islaid (Blair 2002; Knudsen et al. 2006). Schoolreadiness reflects this age-appropriate skill devel-opment and refers to the constellation of skillsneeded in early childhood for a successful tran-sition to formal schooling. Cognitive indicators(i.e., early numeracy, shape knowledge, languageabilities, and literacy skills) represent a child’sability to process information, apply knowledge,and engage in reasoning and problem solving.The non-cognitive dimension encompasses theattention, social, and behavioral components oflearning and includes a child’s ability to concen-trate, stay on task, cooperate, interact with peers,and exercise emotional self-regulation.

Young children’s early attention skills, socio-emotional behaviors, and cognitive knowledge arelinked to later educational achievement, educa-tional attainment, and labor market outcomes(McLeod and Kaiser 2004; Knudsen 2006; Dun-can et al. 2007). Therefore, crucial differencesin a child’s preparedness for school—his or her“school readiness”—set the course for the growthof educational disparities, and understanding thesources of these disparities early on is paramountto illuminating later-life divergent trajectories byrace and gender.

Previous Research and the Potentialfor Heterogeneity in EffectsThere is currently limited work exploring howmass paternal incarceration shapes racial and gen-dered disparities in early educational outcomes.Correspondingly, little attention has been paid tothe potential implications this has for intergen-erational transmissions of educational inequality.Existing research has tended to focus on ado-lescent outcomes (Foster and Hagan 2007; Fos-

sociological science | www.sociologicalscience.com 142 April 2014 | Volume 1

Haskins Effects of Paternal Incarceration on Child School Readiness

ter and Hagan 2009; Hagan and Foster 2012),demonstrating that paternal incarceration de-creases adolescent educational achievement andattainment, which results in the youth’s socialexclusion as an adult. However, this researchprovides little aid in understanding how pater-nal incarceration affects educational outcomes foryounger children at the onset of formal schooling.And while recent studies of younger children doemphasize impacts across individual dimensionsof behavior (e.g., externalizing or internalizing)(Wildeman 2010; Craigie 2011; Wakefield andWildeman 2011; Geller et al. 2012) or cogni-tive measures (Geller et al. 2012) that tap intoaspects of school readiness, they leave largelyunexplored any implications for education andhow the effects translate into differences in earlyschooling trajectories for black and white boysand girls.

Given the well documented racial and gen-dered disparities in both incarceration rates andschooling outcomes, a focused look at whetherpaternal incarceration has heterogeneous effectsby race and gender is needed. Wildeman’s (2010)work on physical aggression has shown impactsto be concentrated among young boys. However,Foster and Hagan (2007) find that the incarcera-tion of a resident father places adolescent daugh-ters at a unique risk of later-life homelessness andabuse. All in all, there are many possible reasonssons and daughters may respond differently whena father is incarcerated.

Moreover, effects of paternal incarcerationmay also be experienced differently dependingon the race of the child. The double jeopardyhypothesis (Dowd and Bengtson 1978) suggeststhat being associated with multiple marginalizedstatuses (in this case, the dual stigma of raceand criminality; Goffman 1963) could result instronger or more negative outcomes for black chil-dren with incarcerated fathers. Alternatively, theresilience hypothesis suggests that forms of dis-ruption or environmental shocks are less stressfulwhen the experience is less unexpected and alter-native support systems are in place (Mineka andKihlstrom 1978). Because paternal incarcerationis more common among recent cohorts of AfricanAmerican children, the effects of paternal incar-ceration could potentially be less negative at theindividual level for black children compared to

white, while still having lasting effects on racialinequality at a population level.

Selection into IncarcerationComplicating any work in this area is the factthat incarceration does not happen at random,and many of the same factors that predict incar-ceration also predict child school readiness. Theincarcerated are disproportionally poor, AfricanAmerican, and poorly educated (Western andBeckett 1999). Thus, children of the incarceratedand formerly incarcerated are likely to suffer fromforms of socio-structural disadvantage indepen-dent of their parent’s incarceration. Second, con-trolling for demographics, fathers who becomeincarcerated exhibit higher levels of antisocialand deviant behavior (e.g., domestic violence, im-pulsivity, and substance abuse) (Murray, Loeberand Pardini 2012), and these behaviors also haveconsequences for child outcomes. Because thesedifferences likely impact child outcomes outsideof the father’s incarceration, it is necessary forresearchers to contend with the possibility thatboth types of preexisting differences account formuch of the disadvantage these children experi-ence.

In this paper, I attempt to address the ubiq-uitous problem of separating the causal effectof paternal incarceration from differential selec-tion into incarceration by using the restrictedFragile Families Study data, which allows for theinclusion of measures that account for economicconstraints, demographic and household charac-teristics, neighborhood context, and a numberof paternal psycho-social and deviant behaviors(all measured prior to the father’s incarceration)that might drive the association between pater-nal incarceration and child schooling outcomes.Moreover, I employ quasi-experimental matchingmethods in the form of propensity score matchingmodels that allow me to make appropriate schoolreadiness comparisons between children with dif-fering paternal incarceration experiences but withsimilar characteristics of preexisting disadvantage.Finally, I conduct two types of sensitivity analy-sis to assess the extent to which observed effectsof paternal incarceration on school readiness arerobust to omitted variable bias.

sociological science | www.sociologicalscience.com 143 April 2014 | Volume 1

Haskins Effects of Paternal Incarceration on Child School Readiness

In the following sections I describe my data,outline the measures and methods used, and re-port the results. I hypothesize that paternal in-carceration has a negative impact on child schoolreadiness and later special education placement,with the strongest deleterious effects observed forblack children and boys.

Data, Measures, and Methods

DataThe Fragile Families and Child Wellbeing Study(FFS) is a longitudinal birth-cohort study thatfollows 4,898 focal children and their parents. In20 large U.S. cities between the years of 1998and 2000, marital and non-marital births wererandomly sampled within hospitals that werestratified by labor market conditions and pol-icy environments (for a complete description seeReichman et al. 2001). The FFS provides sev-eral benefits for directly studying the effects of afather’s incarceration on their child’s school readi-ness because it: follows both parents over timeas their child grows; comprises a large and di-verse sample of children with sufficient variationin paternal incarceration experiences to exploreeffects; and collects information on a variety ofbackground, demographic, environmental, behav-ioral, health, and economic indicators to allowfor a rich set of controls.

The data, which include interviews of bothparents and in-home assessments of children andtheir home environments, are collected from moth-ers and fathers separately. This study attributehelps to validate the reliability of the child andparent measures, in addition to providing infor-mation (via maternal reports) about disadvan-taged fathers otherwise unavailable given theirfrequent underrepresentation in surveys (Hernan-dez and Brandon 2002). The initial baseline waveof interviews was conducted for mothers in thehospital within 48 hours after the birth of thefocal child and for fathers soon thereafter. Fourfollow-up waves of interviews of both mothersand fathers were conducted by phone approx-imately one, three, five, and nine years afterthe focal child’s birth. In-home assessments oc-curred at the three-, five-, and nine-year follow-up waves. The baseline response rate for the

nationally representative sample of mothers is 86percent, while for fathers it is slightly lower at79 percent. Follow-up interview response ratesfor both parents across waves can be found inAppendix A of the supplementary materials.

Given my interest in the effects of paternalincarceration and of black-white racial disparitiesin both the criminal justice system and in edu-cational outcomes, I exclude from my analyticsample Latino children (n =1,224), as well as anychildren who experience maternal incarceration(n =283).1 Additionally, any children with miss-ing information on the school readiness outcomemeasures were dropped, providing final analyticsamples of N=2,602 (for non-cognitive readiness)and N=1,709 (for cognitive readiness). AppendixB in the supplementary materials presents a de-scriptive snapshot of this analytical sample bypaternal incarceration status. I preserve missingcovariates by producing five multiply imputeddatasets and averaging across them (Rubin 1987).

MeasuresSchool Readiness. The two outcomes of interest—cognitive and non-cognitive school readiness—together comprise the behavioral, social, andcognitive dimensions of learning. To conceptu-ally encompass the non-cognitive component ofschool readiness, a single index was created us-ing 17 parent-reported items from the reducedChild Behavior Checklist (CBCL) (Achenbachand Rescorla 2000). The items were reported onat the five-year follow-up wave, when the focalchild was approximately age five.2 The CBCL isone of the most widely used standardized mea-sures in child psychology for evaluating maladap-tive behavioral and emotional problems. The di-mensions measured in the non-cognitive readinessscale include internalizing (anxiety, depression,social withdrawal), externalizing (aggression, dis-obedience), and attention (hyperactivity, concen-tration) behaviors. To compute non-cognitivereadiness scores, responses to each of the 17items (0=not true; 1=somewhat/sometimes true;

1 Children of mothers with an incarceration historyare excluded in order to isolate the effect of paternalincarceration. Earlier analyses were run using a samplethat included Latinos, but for the purposes of clarity andparsimony they have currently been excluded.

2 See Appendix C of the supplementary materials forthe list of questions that comprise this scaled measure.

sociological science | www.sociologicalscience.com 144 April 2014 | Volume 1

Haskins Effects of Paternal Incarceration on Child School Readiness

2=very true/often true) were summed, averaged,and then reversed scaled so that higher numberswould indicate higher non-cognitive school readi-ness. The Cronbach alpha for this compositemeasure is 0.835.

The second outcome of interest, cognitivereadiness, is represented by the child’s standardscore on the Peabody Picture Vocabulary Test-III(PPVT). The PPVT, which is often used as an in-dicator of cognitive knowledge for preschool-agedchildren, is an individually administered, norm-referenced assessment used in early-childhoodscreenings to evaluate Standard American En-glish language development (Dunn and Dunn1997). It measures receptive vocabulary—an indi-vidual’s listening comprehension of spoken words—which is distinct from expressive vocabulary ora measure of cognitive aptitude such as an IQtest. Thus, although the PPVT isn’t the ideal sin-gular measure for representing a child’s broadercognitive ability or developmental potential, it isamong one of the best-established indicators ofchildhood verbal intelligence and general scholas-tic aptitude (Tenenbaum et al. 2007). PPVTassessments were administered to the subset offocal children whose parents agreed to participatein the five-year “in-home” portion of survey/datacollection. For ease of interpretation, the scoresfor both cognitive and non-cognitive readiness arestandardized with coefficients reported in stan-dard deviation units.

Special Education Placement. This secondaryoutcome of interest is a binary indictor repre-senting primary caregiver reports of whether thefocal child was receiving any special education orrelated services at their elementary school duringthe day. The responses were collected at the nine-year follow-up, when the focal children were onaverage nine years old and in third grade, withabout 18 percent of the full study sample indicat-ing parent-reported special education receipt.

Paternal Incarceration. The key explanatoryvariable used is based on a combination of mater-nal and paternal reports of the father’s current orprevious incarceration status across the first fourstudy waves. This measure combines reports ofwhether the father was “currently incarcerated”at the time of the interview and whether the fa-ther was “ever incarcerated” at any point prior tothe interview wave. Mothers are asked, througha variety of interview questions, if their child’s fa-

ther ever spent time in jail or prison, and fathersare asked if they have ever been imprisoned.3 Ifeither the mother or father answers yes, then thefather is indicated as “ever” incarcerated for thatand subsequent waves. Many of the fathers inFFS experience incarceration. Figure 1 showsthis percentage for children in the full FFS sam-ple by year. At the first follow-up wave, whenthe focal child is one year old and paternal in-carceration is first measured,4 approximately 30percent of the fathers in the study have experi-enced incarceration at some point in their lives,and this increases to nearly 42 percent by yearfive—totaling over 2,000 dads.

Previous studies using FFS data have usedvarious indicators of paternal incarceration, somewithout fully accounting for the temporal order-ing of relevant covariates (e.g., Geller et al. 2009);others only measure short-term impacts acrossa few years (e.g., Wildeman 2010). I insteaduse a measure that improves upon these earlieroperationalizations by parsing out children whoexperience paternal incarceration into two mu-tually exclusive groups. The first group of chil-dren’s fathers experienced incarceration before orby year one. Fathers in this group could haveexperienced incarceration at any point before theone-year follow-up interview. This group is thelargest in the data as well as the most difficult tomake causal claims about because incarcerationpotentially occurred before the measurement ofimportant baseline covariates. Therefore, whileimpacts for this group are still important to un-derstand, it is hard to differentiate the directionof influences, which renders any estimates of theeffect of paternal incarceration on outcomes forchildren in this first group susceptible to bias.

3 I use the constructed paternal incarceration mea-sures provided by the FFS. These indicators combineboth direct and indirect reports of fathers’ incarcerationstatus at each of the four waves. See Appendix D of thesupplementary materials for a sample of these survey ques-tions. Unfortunately, the FFS paternal incarceration datadoes not offer any information on duration or frequencyof incarceration, nor can it distinguish between stays inprison as opposed to jail or levels of severity in the crimecommitted.

4 Technically, the incarceration status of some fathers(n=182) was known at baseline, but this is only if mothers(or fathers) answered that the father was in jail or prisonat the time of the child’s birth/baseline interview. Nodirect question was asked of either parent at baselineabout past or current episodes of paternal incarceration.

sociological science | www.sociologicalscience.com 145 April 2014 | Volume 1

Haskins Effects of Paternal Incarceration on Child School Readiness

1

FIGURES Figure 1 Prevalence of Paternal Incarceration in the Fragile Families Study over Waves (in percentages)

Note: Full non-imputed study sample used, N=4898 Figure 2 Exposure to Paternal Incarceration in the Fragile Families Study by Year 5 by Race (in percentages)

Note: Imputed black-white only Study sample used, n=3674.

29.8%

38.8% 41.6%

0.0%

10.0%

20.0%

30.0%

40.0%

50.0%

60.0%

70.0%

Year 1 Year 3 Year 5

Ever

Never

Unknown

45.5% 40.1%

13.2%

69.6%

21.8%

8.6%

0.0%

10.0%

20.0%

30.0%

40.0%

50.0%

60.0%

70.0%

80.0%

Never Before or by YR1 Between YR1 and YR5

Black

White

Figure 1: Prevalence of Paternal Incarceration in the Fragile Families Study over Waves (in percentages)

Acknowledging this concern, I then createda second group of children whose fathers experi-enced first-time incarceration sometime betweenthe one- and five-year follow-up interviews, andnot earlier. This group excludes any father withpreviously indicated incarceration experience atyear one. Fathers in the between year 1 and 5group account for a smaller proportion of the in-carcerated fathers, but they are more appropriatefor estimating effects since their first-time incar-ceration occurred after the collection of baselineand year–one covariates. Finally, I use childrenwith never incarcerated fathers by year five asthe comparison group. Figure 2 shows what thesethree groups look like within the black-white sub-set of the FFS sample.

Controls. Given the wealth of information inthe restricted FFS data, the analyses control fora host of characteristics of mothers, fathers, andtheir children that likely to be associated withpaternal incarceration and child school readiness.These include demographic and household char-acteristics, measures of economic wellbeing, anindicator for interview city, a number of census-tract characteristics, and measures of paternalpsycho-social and deviant behaviors. Adjustingfor this last set of paternal behavior measures di-minishes concerns that a father’s behavior drivesboth his incarceration and impacts his child’sreadiness for formal schooling. Estimates of the

effect of paternal incarceration are only plausibleif included controls adequately address both socio-structural and deviant behavior selection. Thisrequires nuanced measures of both that precedeincarceration. In order to maintain appropriatetime ordering between the dependent, explana-tory, and control variables, I only include controlsthat were measured prior to first paternal incar-ceration or are assumed fixed characteristics. Alist of all variables, along with descriptive statis-tics by paternal incarceration status, is providedin Appendix B of the supplementary materials.

MethodsThe bulk of my analyses rely on propensity scorematching (PSM) (Rosenbaum and Rubin 1983)to estimate the relationship between paternalincarceration and child school readiness. Ordi-nary least squares (OLS) regression models areestimated to assess the overall association be-tween paternal incarceration and behavioral andcognitive school readiness for the full sample ofblack and white children in the FFS (N=2,602 fornon-cognitive readiness; N=1,709 for cognitivereadiness). However, for the reasons discussedearlier, OLS regression is limited in its ability toextract causal inferences since for some childrenpaternal incarceration occurred before baselinecovariates were measured. Therefore, I use PSM

sociological science | www.sociologicalscience.com 146 April 2014 | Volume 1

Haskins Effects of Paternal Incarceration on Child School Readiness

1

FIGURES Figure 1 Prevalence of Paternal Incarceration in the Fragile Families Study over Waves (in percentages)

Note: Full non-imputed study sample used, N=4898 Figure 2 Exposure to Paternal Incarceration in the Fragile Families Study by Year 5 by Race (in percentages)

Note: Imputed black-white only Study sample used, n=3674.

29.8%

38.8% 41.6%

0.0%

10.0%

20.0%

30.0%

40.0%

50.0%

60.0%

70.0%

Year 1 Year 3 Year 5

Ever

Never

Unknown

45.5% 40.1%

13.2%

69.6%

21.8%

8.6%

0.0%

10.0%

20.0%

30.0%

40.0%

50.0%

60.0%

70.0%

80.0%

Never Before or by YR1 Between YR1 and YR5

Black

White

Figure 2: Exposure to Paternal Incarceration in the Fragile Families Study by Year 5 by Race (inpercentages)

to focus on the relationship between paternalincarceration and child school readiness for thesecond group of children, those whose fathers’first-time incarceration occurred between years 1and 5, and their comparison counterparts whoby year five have no paternal incarceration ex-perience (N=1,780 for non-cognitive readiness;N=1,119 for cognitive readiness).5

The large number of children in the OLS mod-els who have paternal incarceration experiencesoccurring before or by year 1 are dropped fromthe PSM analyses. As Figure 2 demonstrated,this is a substantial proportion of both black(40 percent) and white (22 percent) children,and impacts of paternal incarceration on theirschool readiness are substantively important—especially when considering population-level im-pacts of mass paternal incarceration on children.However, their exclusion from the analytic samplefor PSM purposes is purely to be able to provideas unbiased estimates of the effect of paternalincarceration as possible.

5 This is done in order to avoid introducing “post-treatment bias” (Ho, Imai, King and Stuart 2007) byensuring that only pre-treatment variables are used in thematching procedure.

While propensity score matching is not apanacea for selection concerns, as it can only ac-count for observed differences between treatmentand control groups, it is a valuable techniquewhen used in conjunction with a rich set of ob-served characteristics. The main purpose of thePSM model is to ensure that observed character-istics that would predict having an incarceratedfather are balanced across those children who,in reality, do and do not have incarcerated fa-thers (Augurzky and Schmidt 2001). The PSManalyses estimate the average effect of having anincarcerated father on child school readiness bysimulating “treatment” and “control” groups fromthe observational FFS data. This allows me tomake appropriate school readiness comparisonsby using a reference group of children who do notexperience paternal incarceration but are simi-larly at risk based on the observed, theoreticallyrelevant pre-treatment characteristics includedin the matching model. Moreover, the techniqueallows for the systematic judgment of whethertreatment and control groups are properly bal-anced on observables based on a balance test.6

6 Analyses were restricted to observations within the re-gion of common support. I tested that the PSM techniqueachieved covariate balance using the ‘pstest’ command

sociological science | www.sociologicalscience.com 147 April 2014 | Volume 1

Haskins Effects of Paternal Incarceration on Child School Readiness

Propensities are generated via a probit regressionmodel predicting selection into paternal incarcer-ation. I employ the kernel matching techniqueusing a Gaussian kernel and a bandwidth of 0.08to estimate the average treatment effect.7

ResultsAs discussed in Figure 1, a large number of chil-dren in the FFS experience paternal incarcerationby age five, yielding sufficient variation to exam-ine its effects on child outcomes. Looking withinthe black-white sample, Figure 2 shows thesepercentages by race for children at year five. Inthese data, early life-course racial disparities arealready present, with over 53 percent of black chil-dren experiencing some form of paternal incarcer-ation. Said differently, by the time black childrenin this sample enter formal schooling, more haveexperienced paternal incarceration than have not.For whites, the percentage of children experienc-ing paternal incarceration by year five is around30 percent. Exposure to paternal incarcerationfor both blacks and whites in the FFS sampleis high. Using age-appropriate population-levelcumulative risk estimates from Wildeman (2009)as a point of comparison, the prevalence of pater-nal incarceration in the FFS for black childrenof pre-school age is more than three times aslarge (53 percent versus 15 percent). However,the prevalence of paternal incarceration for whitepre-school age children in the FFS is twenty timeslarger (30% v. 1.5%) than what is expected forsimilar age white children on average in the U.S.With regard to gender, boys and girls experiencepaternal incarceration nearly equally.

The two panels of Table 1 present unadjustedmeans and standard deviations for both schoolreadiness measures in standardized units, weightedto be nationally representative. On average, blackchildren are less school ready than whites onboth measures; boys are less behaviorally readythan girls; and children in either of the groupsthat experience an incarcerated father by age five(before age one and between ages one and five)

in Stata. Results for the distribution of support and areporting of covariate balance statistics are available inAppendices E and F of the supplementary materials.

7 Nearest neighbor and radius matching were alsoperformed and results were similar.

have lower cognitive and non-cognitive readinessscores compared to those who do not. Not ac-counting for controls, the racial and gender gapsin non-cognitive readiness are both one-fifth ofa standard deviation (SD). For cognitive readi-ness, the racial gap is equivalent to a differenceof more than one SD, while the gender gap incomparison is quite small. Moving to the sec-ond panel, girls (and white girls in particular)with never-incarcerated fathers have the high-est non-cognitive readiness scores, while blackboys, especially those who experience either formof paternal incarceration, are consistently low-est. Moreover, as shown in Appendix B of thesupplementary materials, children experiencingpaternal incarceration are more likely to suffer awhole host of other disadvantages.

Effects of Paternal Incarceration onChild School ReadinessPropensity Score Matching Models. OLS regres-sion models (not shown) suggest that paternalincarceration is negatively associated with thenon-cognitive component of child school readi-ness for whites and boys, even after adjusting fora large set of observed baseline covariates. Thefollowing PSM models are not directly analogousbut instead allow me to focus the analyses onthe group of children with incarcerated fathersthat has the potential to produce more causalestimates. Attention can also then be paid tothe temporal ordering of relevant covariates, andI take further steps to ameliorate selection con-cerns by introducing into the matching algorithmadditional characteristics of fathers predictive ofincarceration measured at year one that were notincluded in OLS models. In these PSM models,analysis is restricted to the children whose fatherwas incarcerated for the first time between year 1and year 5 and their matched controls.

Table 2 presents PSM results for the effectof paternal incarceration on both components ofchild school readiness. Starting with Model 1aand the overall effect of paternal incarceration onnon-cognitive readiness, the standardized pointestimate of -0.143 indicates that after matching,children with fathers who experience first-timeincarceration between the ages of one and fiverank just over one-seventh of an SD lower in non-cognitive readiness compared to their matched

sociological science | www.sociologicalscience.com 148 April 2014 | Volume 1

Haskins Effects of Paternal Incarceration on Child School Readiness

Table1:W

eigh

tedMeans

andStan

dard

Deviation

sforScho

olReadiness

Variables

byRace,

Gende

r,an

dPaterna

lIncarceration

Status

ByRace

ByGen

der

ByPaterna

lIncarceration

Status

ByRace&

Gen

der

Beforeor

Between

White

Black

Overall

White

Black

Male

Female

Never

ByYR1

YR1&

YR5

Male

Female

Male

Female

Non

-cognitive

Readine

ss0.10

0.19

-0.01

0.01

0.20

0.20

-0.12

-0.13

0.09

0.32

-0.09

0.10

(NR)Stan

dardized

(0.97)

(0.94)

(0.99)

(1.04)

(0.88)

(0.93)

(0.99)

(1.12)

(1.03)

(0.81)

(1.03)

(0.94)

Cognitive

Readine

ss0.29

0.89

-0.13

0.26

0.33

0.54

-0.03

-0.27

0.84

0.96

-0.20

-0.05

(CR)Stan

dardized

(0.98)

(0.75)

(0.90)

(1.02)

(0.92)

(0.92)

(0.94)

(0.91)

(0.78)

(0.71)

(0.96)

(0.82)

NR

N=

2602

765

1837

1375

1227

1452

822

328

413

352

962

875

CR

N=

1709

424

1285

890

819

880

590

239

226

198

664

621

ByRacean

dGen

deran

dPaterna

lIncarceration

Status

Father

Incarcerated

Father

Incarcerated

Father

Never

Incarcerated

Beforeor

ByYR1

BetweenYR1an

dYR5

White

Black

White

Black

White

Black

Male

Female

Male

Female

Male

Female

Male

Female

Male

Female

Male

Female

Non

-cognitive

Readine

ss0.12

0.49

0.09

0.21

-0.10

-0.13

-0.15

-0.10

-0.15

0.33

-0.51

0.19

(NR)Stan

dardized

(1.00)

(0.80)

(0.99)

(0.86)

(1.30)

(0.72)

(0.96)

(0.97)

(0.88)

(0.80)

(1.19)

(1.03)

Cognitive

Readine

ss0.93

1.06

-0.01

0.06

0.54

0.55

-0.29

-0.13

0.23

0.61

-0.56

-0.21

(CR)Stan

dardized

(0.74)

(0.64)

(0.90)

(0.84)

(0.71)

(0.91)

(0.98)

(0.80)

(1.04)

(0.51)

(0.93)

(0.75)

NR

N=

309

257

463

423

7265

359

326

3230

140

126

CR

N=

158

138

301

283

4742

257

244

2118

106

94

Not

e:Nationa

lweigh

tsan

dim

putedblack-white

sampleda

taused

.YR1an

dYR5stan

dforYear1an

dYear5respectively.Stan

dard

deviations

arein

parenthe

ses.

sociological science | www.sociologicalscience.com 149 April 2014 | Volume 1

Haskins Effects of Paternal Incarceration on Child School Readiness

controls, and this difference is statistically sig-nificant. In the education literature, effect sizesranging between one-fourth and one-eighth aresubstantial, and these non-cognitive readiness re-sults suggest that this difference of 0.143 is equiv-alent to a loss of almost two months of schooling.

Models 2a-5a show the effect of paternal in-carceration on non-cognitive readiness within thefour race-gender pairings in the data: black boys,white boys, black girls, and white girls. Look-ing at differences in effects within these race-gender pairings allows for a direct comparison oflike with like, making additional race- or gender-motivated differences less of a concern. Resultsindicate both black boys (Model 2a) and whiteboys (Model 4a) who experience first-time pa-ternal incarceration between ages one and fiveperform significantly worse on the non-cognitivereadiness measure than matched boys who havenever had an incarcerated father. In particular,black boys perform 0.223 SD units lower, andwhite boys 0.422 SD units lower. These signifi-cant differences in non-cognitive readiness haveeducational relevance, representing a two monthloss of schooling for black boys with incarceratedfathers, and a four month loss for white boys. Atest of the difference between these point esti-mates for black and white boys shows they arenot statistically different from each other; how-ever, it is less clear whether this indicates thatpaternal incarceration has a similarly negative ef-fect on the non-cognitive readiness on both blackand white boys, or that the current sample size ofwhite boys experiencing paternal incarceration istoo small to precisely determine this relationship.I find no within-race differences in non-cognitivereadiness between treatment and control groupsfor black or white girls.

Results for cognitive readiness (Models 1b-5b)are less clear. Comparisons of overall treatmenteffects between non-cognitive readiness (Model1a) and cognitive readiness (Model 1b) show sim-ilarities in size and direction, but the cognitivereadiness measure does not reach statistical signif-icance. Looking within the race-gender pairings(Models 2b-5b), treatment effects for black girls,white girls, and white boys are large and negativein direction; however, the inflated standard errorsindicate that these estimates are imprecise, asthere is not enough power to determine effects.Nonetheless, the non-effect of paternal incarcera-

tion on a child’s receptive vocabulary across bothOLS and PSM models does indicate that the cog-nitive and non-cognitive components of schoolreadiness should be distinguished; non-cognitiveschool readiness appears to be more subject tolocalized family conditions, while cognitive readi-ness appears to be mainly correlated with race,and highly sensitive to poverty, maternal educa-tion, and parental cognitive abilities. Therefore,while it is unclear whether children with incar-cerated fathers are less cognitively prepared forschool than their matched counterparts, paternalincarceration does produce significant deficienciesin pre-school boys’ socio-behavioral abilities. Sim-ply put, boys with incarcerated fathers possessfewer of the non-cognitive skills needed to suc-cessfully navigate and adjust to the expectationsand interactions present at school entry.

Sensitivity Analyses and Robustness Checks.In the absence of an experiment, it is fair to as-sume that unobservable characteristics inevitablyplay some role in determining which children ex-perience paternal incarceration and which do not.To address this concern, I do two types of sen-sitivity analysis on the treatment effects foundfor boys’ non-cognitive readiness. The first is afalsification test that makes use of the recentlyreleased year nine FFS data. For this, I run ananalysis (not shown) to see if future paternal in-carceration experienced for the first time betweenyear five and year nine would predict a child’spast non-cognitive readiness at year five, account-ing for all covariates.8 I find that future paternalincarceration does not predict past non-cognitivereadiness, providing additional evidence that thetreatment effects estimated from the PSM mod-els are measuring something unique to paternalincarceration.

Secondly, I use Rosenbaum bounds (Rosen-baum 2002; DiPrete and Gangl 2004) to checkthe sensitivity of the estimated treatment effectsto omitted variable bias.9 This test assesseshow strong a hypothetical unmeasured variablerelated to selection into paternal incarcerationwould need to be to undermine the results. Theseanalyses provide support for the argument that

8 A small number of children (61 black boys and 12white boys) fall into this subgroup experiencing first-timepaternal incarceration between year five and year nine.

9 See Appendix H of the supplementary materials forthese results.

sociological science | www.sociologicalscience.com 150 April 2014 | Volume 1

Haskins Effects of Paternal Incarceration on Child School Readiness

Table 2: Results from Propensity Score Matching Predicting both types of School Readiness at AgeFive

Matched pairs

Non-Cognitive Readinessa

Difference N Treated Control

Model 1a: Paternal Incarceration −0.143∗ 1780 322 1445(0.070)

Model 2a: within Black Males −0.223∗ 603 132 464(0.123)

Model 3a: within Black Females 0.005 549 118 420(0.119)

Model 4a: within White Males −0.422∗ 341 30 253(0.231)

Model 5a: within White Females 0.148 287 23 189(0.244)

Matched pairs

Cognitive Readinessb

Difference N Treated Control

Model 1a: Paternal Incarceration −0.100 1119 229 877(0.079)

Model 2b: within Black Males 0.022 407 98 299(0.129)

Model 3b: within Black Females −0.132 377 92 282(0.142)

Model 4b: within White Males −0.299 179 15 155(0.382)

Model 5b: within White Females −0.239 156 14 138(0.297)

Note: Kernel matching model estimates shown. See Appendix B in the supplementary materialsfor a complete list of variables used in models predicting the treatment. Analyses are unweightedand done on imputed data. Standard errors are in parentheses. Significance levels are the following:†p < 0.1; ∗p < .05; ∗∗p < .01; ∗∗∗p < .001 (one-sided).a Matched pairs are average “on support” counts of treated and control cases. To help with modelconvergence, neither Models 4a or 5a include census tract characteristics in the matching algorithm.b Matched pairs are average “on support” counts of treated and control cases. To help with modelconvergence, neither Models 4b or 5b include city indicators and census tract characteristics in thematching algorithm.

paternal incarceration has a negative effect onboys’ non-cognitive school readiness. In com-parison to observed measures in the dataset, anunobserved confounding variable would need tohave an effect on predicting paternal incarcera-tion that fell in magnitude between what havinga father with only a high school diploma andhaving a father with drug and alcohol problemsdoes to undermine the reported effects. Given

the richness of the FFS data, it is difficult to iden-tify a theoretically relevant variable not alreadyincluded that has this suggested strength.

Mediation Models for Special Education. Theearlier analyses demonstrated the deleterious im-pact of paternal incarceration on the non-cognitivereadiness of five-year-old black and white boys.Given that differences found this early have thepotential to grow over time and influence later

sociological science | www.sociologicalscience.com 151 April 2014 | Volume 1

Haskins Effects of Paternal Incarceration on Child School Readiness

schooling outcomes, it is worth considering whetherconsequences of paternal incarceration for schoolreadiness at year five also present problems forboys’ educational outcomes after the start ofschooling. One way to answer this question is toask whether a portion of the negative effects wesee for later schooling outcomes (Haskins 2013)is attributable to the negative effects on schoolreadiness shown here. To address this question, Iperform Sobel-Goodman mediation tests10 to an-alyze the contribution of paternal incarceration’seffect on boys’ non-cognitive school readiness topaternal incarceration’s total effect on parent-reported special education placement, measuredat year nine, a later schooling outcome especiallymeaningful for boys (Hibel, Farkas and Morgan2010), and black males in particular (Davis 2003).

The first row of Table 3 shows the total ef-fect of first-time paternal incarceration occurringbetween years one and five on special educationplacement at year nine. Paternal incarcerationhas a statistically significant positive relationshipfor boys’ likelihood of receiving special educationservices by age nine. Secondly, school readinessmeasured four years prior mediates nearly one-fifth of the total effect, even after accounting forrelevant controls; therefore, these analyses demon-strate not only the importance of school readinessfor these young boys, but also the importance ofpaternal incarceration and its effects on specialeducation assignment, mediated through schoolreadiness. Thus, paternal incarceration’s dele-terious impact on school readiness at year fiveaccounts for a meaningful portion, though not amajority, of the negative impact it presents foryear nine educational outcomes of affected boys.

DiscussionThis study contributes to a growing body of litera-ture on the implications of mass incarceration forinequality among U.S. children. It demonstratesthe direct consequences of paternal incarcerationfor children’s educational preparedness and un-

10 I run bootstrapped Sobel-Goodman mediation testsusing the ‘sgmediation’ program in Stata 12 in order toreplicate within Stata the more appropriate Preacher andHayes (2004) bootstrapping method of mediation analysis(http://ederosia.byu.edu/blog/Eric_DeRosia/using-stata-to-perform-the-preacher-and-hayes-1994-bootstrapped-test-of-mediation/).

derscores the importance of this early schoolingoutcome for later educational decisions. More-over, it enhances our understanding of both thepersistence of racial and gendered disparities ineducational outcomes and the intergenerationaltransmission of inequality. Using multiple esti-mation strategies and attending to selection con-cerns, I found that paternal incarceration nega-tively affects boys’ non-cognitive school readinessand that these effects have lingering impacts onlater special education placement. These findingsare in line with recent work using similar dataand methods (Wildeman 2010; Wakefield andWildeman 2011; Geller et al. 2012) and presentan important extension by providing evidencethat the negative effect of paternal incarcerationon non-cognitive school readiness is one path-way through which mass incarceration createseducational disadvantages in the next generation.

White and black boys with incarcerated fa-thers possessed lower levels of the non-cognitiveskills necessary for successful entry into formalschooling. In the PSM models, the magnitude ofthe effect was large for both groups—0.4 standarddeviations for white boys and 0.2 for black boys,representing a two- to four-month loss of school-ing. To benchmark the relevance of this genderedeffect, paternal incarceration explains two-thirdsof the white gender gap and all of the black gen-der gap in non-cognitive school readiness. In theabsence of heterogeneous effects of paternal incar-ceration by gender, these boys would be similarlybehaviorally ready to their female counterparts atschool entry and potentially less likely to be laterplaced in special education. Moreover, togetherwith the fact that black children are dispropor-tionally exposed to paternal incarceration, thesefindings have important implications for racialeducational inequality as the rise in incarcera-tion and the excessive incarceration of AfricanAmerican males threatens to shatter what edu-cational scholars have called the “virtuous cycle”(Gamoran 2001) in which present generations ofblack children benefit from previous generations’efforts to narrow gaps in achievement and attain-ment.

These findings have three major implicationsfor sociologists interested in the intersection ofincarceration, race, gender, and educational in-equalities. First, they add to a growing body ofliterature that shows young boys appear to be

sociological science | www.sociologicalscience.com 152 April 2014 | Volume 1

Haskins Effects of Paternal Incarceration on Child School Readiness

Table 3: Results from Non-cognitive School Readiness Mediation Models for Special Education Place-ment at Year 9

Total Effect 0.08∗

(0.03)Direct Effect of Paternal Incarceration 0.07∗

(0.03)Indirect Effect mediated through Non-cognitive Readiness 0.01∗

(0.01)Confidence interval [0.001, 0.031]Proportion of Total Effect Mediated 0.18N 827

Note: Analyses are unweighted and done on imputed data. All PSM controls (see Table 2) areincluded in model. Bootstrapped standard errors in parentheses. Significance levels are the following:†p < 0.1; ∗p < .05; ∗∗p < .01; ∗∗∗p < .001 (one-sided). In order to maintain correct temporal ordering,boys who experienced first-time paternal incarceration between Year 5 and Year 9 were excluded(N = 73); an additional 44 boys were lost due to missing data on the Year 9 outcome.

more sensitive to disruptions and instability infamily structure than girls across a range of educa-tional outcomes (Cooper et al. 2011; DiPrete andBuchmann 2013). Second, they may help explainthe persistence of the black male disadvantagein educational outcomes, as black men continueto lag behind white men and white and blackwomen in high school and college completionrates (McDaniel et al. 2011). Third, as DiPreteand Jennings (2012) have recently shown, genderdifferences in non-cognitive skills present in earlyschooling grow throughout elementary school, im-pacting cognitive test scores and explaining alarge portion of the gender gap in later academicoutcomes. Socio-behavioral skills are central tochildren’s later educational success (Duckworthand Seligman 2005), and DiPrete and Jennings(2012) call for studies to think about family-levelprocesses that produce gender gaps in these non-cognitive skills. This study presents evidence ofa father’s incarceration as a family-level processable to produce such differences.

Limitations and Future DirectionsWhile informative, the results I present also raiseadditional questions and highlight areas where fu-ture work is sorely needed. Specifically, I call foradditional work on mechanisms, which I did notfully test for in these analyses. It is also worthhaving a discussion with regard to the null-effect

found for cognitive readiness. The development ofcognitive and non-cognitive skills is interrelated,and while the treatment effects found for cog-nitive readiness did not reach significance, theywere in general negatively associated with pa-ternal incarceration, and of similar magnitudeto the effects found for non-cognitive readiness.Therefore, I hesitate to conclude that paternalincarceration has no impact on cognitive skills,and I encourage future work along these lines.

The use of a parent-reported composite mea-sure of non-cognitive readiness also has potentiallimitations. A focus on the fine-grained distinc-tions between various non-cognitive skills, as doneby Geller et al. (2012), has highlighted some dif-ferences between types of child behaviors anda father’s incarceration. In exploratory factoranalyses, the 17 components used in the creationof my non-cognitive readiness measure alignedwell with sub-scales for internalizing, externaliz-ing, and attention behaviors.11 Supplementaryanalyses indicated that the components drivingthe negative effect of paternal incarceration onnon-cognitive readiness were mainly from the ex-ternalizing and attention behaviors sub-scales,similar to what has been found by Wildeman(2010) and Geller et al. (2012). Maternal re-ports of behaviors characteristic of internalizingproblems like depression, anxiety, or withdrawal

11 See Appendix I of the supplementary materials forthese results.

sociological science | www.sociologicalscience.com 153 April 2014 | Volume 1

Haskins Effects of Paternal Incarceration on Child School Readiness

seemed less affected by paternal incarceration forpre-school aged children. However, this could bedue to inaccuracy in maternal reports, as inter-nalizing behaviors are harder to identify in youngchildren. Some work comparing agreement ofchild self-reports and parent reports finds thatparents especially underreport internalizing be-haviors (Sourander, Helstela, and Helenius 1999).Future work that can bring child self-reports ofnon-cognitive skills or even reports from anotherinvolved adult into the picture would greatly addto our knowledge of the effects of paternal incar-ceration on important components of children’seducational development.

Future studies could also work to uncover ifthere is indeed a stronger relationship betweenpaternal incarceration and non-cognitive readi-ness for white boys. While this study was unableto make any conclusions on this matter, I canoffer a few different interpretations that futurestudies could explore. Given that incarceration ismore proportionally prevalent for black children,and therefore more visible in minority communi-ties, this could suggest that black families andcommunities are more resilient, supportive, orfor other reasons less sensitive to the effects ofpaternal incarceration than the white childrenand families represented in the FFS data. More-over, since the criminal justice system comes intocontact with a broader swath of the black com-munity than the white, another interpretationcould be that the white incarcerated fathers inthe FFS are in some way selectively “worse” (e.g.,committed a more severe crime or spent moretime incarcerated) than the black fathers, thuspossibly affecting their sons more severely. Athird reason could be that white families tend tobe more nuclear than black families, which canoften rely on extended kin networks (Dunifon andKowaleski-Jones 2002); thus for white childrenthe loss of a father to incarceration leaves a largerhole that in a black family could be filled by otherfamilial sources of support. A final interpretationmight be that for blacks—given residential segre-gation, spatial inequality in punishment, and thelocal concentration of incarceration in predomi-nantly black neighborhoods (Sobel 2006; Samp-son and Loeffler 2010)—paternal incarceration isa community-level as well as an individual-levelphenomenon, affecting the educational trajecto-ries of children with and without incarcerated

fathers at a more aggregate level. If this is thecase, spillover effects of incarceration would pro-duce dampened estimates of the effect of paternalincarceration on black children (but not white)when compared to same-race “control” counter-parts, as the control group would be indirectlyexposed to the treatment.

ConclusionSchool readiness represents a developmental out-come necessary for successful entry into formalschooling. Because early cognitive and non-cogni-tive abilities are central to children’s educationalsuccess, differences in school readiness affect chil-dren’s future academic and labor market trajec-tories. Moreover, in the first few years of school,key educational decisions such as placement inspecial education and ability groups are made,and this study found that a child’s behavior, at-tention, and social skills play an important role inthose decisions. As a result, early problems withnon-cognitive skills such as aggression, inatten-tiveness, disobedience, and hyperactivity can setchildren on a pathway of cumulative disadvantage(Raver 2002; McLeod and Kaiser 2004; Heckman,Stixrud, and Urzua 2006). This, coupled withevidence that effects of paternal incarceration atthis young age are strongest for boys and aredisproportionally experienced by black children,points to one potential pathway through whichpaternal incarceration facilitates the intergenera-tional transmission of disadvantage from father toson and plays a role in explaining the presence ofracial and gendered differences in early schoolingoutcomes. In short, the incarceration of a fatheraffects how ready children are for formal school-ing, and the mass incarceration of fathers in thiscountry is limiting the educational potential oftheir male children.

ReferencesAchenbach, Thomas M. and Leslie A. Rescorla.

2000. Manual for the ASEBA PreschoolForms and Profiles. Burlington, VT: Univer-sity of Vermont, Research Center for Children,Youth, and Families.

Augurzky, Boris and Christoph M. Schmidt. 2001.“The Propensity Score: A Means to an End”

sociological science | www.sociologicalscience.com 154 April 2014 | Volume 1

Haskins Effects of Paternal Incarceration on Child School Readiness

IZA Discussion Paper 271: Institute for theStudy of Labor (IZA).

Bendheim-Thoman Center for Research on ChildWellbeing. 2008. “Introduction to the FragileFamilies Public-Use Data: Baseline, One-Year,Three-Year, and Five-Year Telephone Data.”Princeton, NJ: Office of Population Research,Princeton University.

Blair, Clancy. 2002. “School Readiness: In-tegrating Cognition and Emotion in a Neu-robiological conceptualization of Children’sFunctioning at School Entry.” American Psy-chologist 57:111-127. http://dx.doi.org/10.1037/0003-066X.57.2.111

Braman, Donald. 2004. Doing Time on the Out-side: Incarceration and Family Life in UrbanAmerica. Ann Arbor, MI: University of Michi-gan Press.

Buchmann, Claudia, Thomas A. DiPrete, andAnne McDaniel. 2008. “Gender Inequalitiesin Education.” Annual Review of Sociology34:319–337. http://dx.doi.org/10.1146/annurev.soc.34.040507.134719

Clear, Todd R. 2007. Imprisoning Communities.New York: Oxford University Press.

Comfort, Megan. 2008. Doing Time To-gether: Love and Family in the Shadowof the Prison. Chicago: University ofChicago Press. http://dx.doi.org/10.7208/chicago/9780226114682.001.0001

Cooper, Carey E., Cynthia A. Osborne, Au-drey N. Beck, and Sara S. McLanahan. 2011.“Partnership Instability, School Readiness, andGender Disparities.” Sociology of Education84:246-259. http://dx.doi.org/10.1177/0038040711402361

Craigie, Terry-Ann L. 2011. "The Effectof Paternal Incarceration on Early ChildBehavioral Problems: A Racial Compari-son." Journal of Ethnicity in Criminal Jus-tice 9:179-99. http://dx.doi.org/10.1080/15377938.2011.594349

Davis, James E. 2003. “Early Schooling andAcademic Achievement of African AmericanMales.” Urban Education 38:515-537. http://dx.doi.org/10.1177/0042085903256220

DiPrete, Thomas A. and Claudia Buchmann.2013. The Rise of Women: The GrowingGender Gap in Education and What It Means

for American Schools. New York: RussellSage Foundation.

DiPrete, Thomas A. and Markus Gangl. 2004.“Assessing Bias in the Estimation of CausalEffects: Rosenbaum Bounds on Matching Es-timators and Instrumental Variables Estima-tion with Imperfect Instruments.” SociologicalMethodology 34: 271-310. http://dx.doi.org/10.1111/j.0081-1750.2004.00154.x

DiPrete, Thomas A. and Jennifer L. Jen-nings. 2012. “Social and BehavioralSkills and the Gender Gap in Early Ed-ucational Achievement.” Social ScienceResearch 41:1-15. http://dx.doi.org/10.1016/j.ssresearch.2011.09.001

Dowd, James J. and Vern L. Bengtson. 1978.“Aging in Minority Populations: An Exam-ination of the Double Jeopardy Hypothe-sis.” Journal of Gerontology 33:427-36. http://dx.doi.org/10.1093/geronj/33.3.427

Duckworth, Angela L. and Martin E. P. Selig-man. 2005. “Self-Discipline OutdoesIQ in Predicting Academic Performanceof Adolescents.” Psychological Science 16:939-944. http://dx.doi.org/10.1111/j.1467-9280.2005.01641.x

Duncan, Greg J., Amy Claessens, Aletha C.Huston, Linda Pagani, Mimi Engel, HollySexton, Chantelle J. Dowsett, KatherineMagnuson, Pamela Klebanov, Leon Fein-stein, Jeanne Brooks-Gunn and KatherineDuckworth. 2007. “School Readiness andLater Achievement.” Developmental Psychol-ogy 43:1428-1446. http://dx.doi.org/10.1037/0012-1649.43.6.1428

Dunifon, Rachel and Lori Kowaleski-Jones.2002. “Who’s in the House? Race Differ-ences in Cohabitation, Single Parenthood,and Child Development.” Child Develop-ment 73:1249–1264. http://dx.doi.org/10.1111/1467-8624.00470

Dunn, Lloyd M. and Leota M. Dunn. 1997.Peabody Picture Vocabulary Test—Third Edi-tion. Circle Pines, MN: American GuidanceService.

Entwisle, Doris R., Karl L. Alexander and LindaSteffel Olson. 2007. “Early Schooling: TheHandicap of Being Poor and Male.” Sociologyof Education 80: 114-138. http://dx.doi.org/10.1177/003804070708000202

sociological science | www.sociologicalscience.com 155 April 2014 | Volume 1

Haskins Effects of Paternal Incarceration on Child School Readiness

Foster, Holly and John Hagan. 2007. “Incar-ceration and Intergenerational Social Exclu-sion.” Social Problems 54:399-433. http://dx.doi.org/10.1525/sp.2007.54.4.399

——. 2009. “The Mass Incarceration of Parents inAmerica: Issues of Race/Ethnicity, CollateralDamage to Children, and Prisoner Reentry.”Annals of the American Academy of Politicaland Social Science 623:179-194. http://dx.doi.org/10.1177/0002716208331123

Gamoran, Adam. 2001. “American Schooling andEducational Inequality: A Forecast for the21st Century.” Sociology of Education 74:135-153. http://dx.doi.org/10.2307/2673258

Geller, Amanda, Irwin Garfinkel, Carey E.Cooper and Ronald Mincy. 2009. “Parental In-carceration and Child Wellbeing: Implicationsfor Urban Families.” Social Science Quar-terly 90:1186-1202. http://dx.doi.org/10.1111/j.1540-6237.2009.00653.x

Geller, Amanda, Carey E. Cooper, IrwinGarfinkel, Ofira Schwartz-Soicher and RonaldB. Mincy. 2012. “Beyond Absenteeism: Fa-ther Incarceration and Child Development.”Demography 49:49-76. http://dx.doi.org/10.1007/s13524-011-0081-9

Glaze, Lauren E. and Laura M. Maruschak. 2008.“Parents in Prison and Their Minor Children.”Bureau of Justice Statistics Special Report.Washington, DC: Bureau of Justice Statistics.

Goffman, Erving. 1963. Stigma: Notes on theManagement of Spoiled Identity. EnglewoodCliffs, NJ: Prentice-Hall.

Hagan, John and Ronit Dinovitzer. 1999. “Collat-eral Consequences of Imprisonment for Chil-dren, Communities, and Prisoners.” Pp. 121-162 in Prisons, edited by M. Tonry and J.Petersilia. Chicago: University of ChicagoPress.

Hagan, John and Holly Foster. 2012. “Inter-generational Educational Effects of Mass Im-prisonment in America.” Sociology of Edu-cation 85:259-286. http://dx.doi.org/10.1177/0038040711431587

Haskins, Anna R. 2013. “Mass Imprisonmentand the Intergenerational Transmission of Dis-advantage: Paternal Incarceration and Chil-dren’s Cognitive Skill Development.” Fragile

Families Working Paper Series, No. WP13-15FF. Center for Research on Child Wellbeing:Princeton University.

Heckman, James J., Jora Stixrud and Ser-gio Urzua. 2006. “The Effects of Cognitiveand Noncognitive Abilities on Labor Mar-ket Outcomes and Social Behavior.” Jour-nal of Labor Economics 24:411-482. http://dx.doi.org/10.1086/504455

Hernandez, Donald, and Peter Brandon. 2002.“Who Are the Fathers of Today?” Pp. 33-62in Handbook of Father Involvement: Multi-disciplinary Perspectives, edited by C. Tamis-Lemonda and N. Cabrera. New Jersey: Erl-baum.

Hibel, Jacob, George Farkas and Paul L. Morgan.2010. “Who is Placed into Special Education?”Sociology of Education 83:312-332.http://dx.doi.org/10.1177/0038040710383518

Ho, Daniel E., Kosuke Imai, Gary King and Eliza-beth Stuart. 2007. “Matching as Nonparamet-ric Preprocessing for Reducing Model Depen-dence in Parametric Causal Inference.” Po-litical Analysis 15:199-236. http://dx.doi.org/10.1093/pan/mpl013

Jencks, Christopher and Meredith Phillips. 1998.The Black-White Test Score Gap. Washing-ton, DC: The Brookings Institute Press.

Knudsen, Eric I., James J. Heckman, Judy L.Cameron and Jack P. Shonkoff. 2006. “Eco-nomic, Neurobiological, and Behavioral Per-spectives on Building America’s Future Work-force.” Proceedings of the National Academyof Sciences 103:10155-10162. http://dx.doi.org/10.1073/pnas.0600888103

Magnuson, Katherine and Jane Waldfogel. 2008.Steady Gains and Stalled Progress: Inequalityand the Black-white Test Score Gap New York:Russell Sage Foundation.

McDaniel, Anne, Thomas A. DiPrete, ClaudiaBuchmann and Uri Shwed. 2011. “The BlackGender Gap in Educational Attainment: His-torical Trends and Racial Comparisons.” De-mography 48:889–914. http://dx.doi.org/10.1007/s13524-011-0037-0

McLeod, Jane D. and Karen Kaiser. 2004. “Child-hood Emotional and Behavioral Problems andEducational Attainment.” American Socio-

sociological science | www.sociologicalscience.com 156 April 2014 | Volume 1

Haskins Effects of Paternal Incarceration on Child School Readiness

logical Review 69:636-658. http://dx.doi.org/10.1177/000312240406900502

Mineka, S. and J. Kihlstrom. 1978. “Un-predictable and Uncontrollable Events: ANew Perspective on Experimental Neuro-sis.” Journal of Abnormal Psychology87:256-271. http://dx.doi.org/10.1037/0021-843X.87.2.256

Morgan, Paul L., Michelle Frisco, George Farkas,and Jacob Hibel. 2010. “A Propensity ScoreMatching Analysis of the Effects of SpecialEducation Services.” Journal of Special Ed-ucation 43:236–54. http://dx.doi.org/10.1177/0022466908323007

Mumola, Christopher J. 2002. Incarcerated Par-ents and their Children Washington, DC: Bu-reau of Justice Statistics.

Murray, Joseph and David Farrington. 2008.“The Effects of Parental Imprisonment onChildren.” Crime and Justice 37:133-206.http://dx.doi.org/10.1086/520070

Murray, Joseph, Rolf Loeber, and Dustin Pardini.2012. “Parental Involvement in the Crimi-nal Justice System and the Development ofYouth Theft, Marijuana Use, Depression, andPoor Academic Performance.” Criminology50:255-302. http://dx.doi.org/10.1111/j.1745-9125.2011.00257.x

Parke, Ross D. and K. Alison Clarke-Stewart.2003. “The Effects of Parental Incarcerationon Children: Perspectives, Promises, and Poli-cies.” Pp. 189-232 in Prisoners Once Re-moved: The Impact of Incarceration and Re-entry on Children, Families, and Communi-ties, edited by J. Travis and M. Waul. Wash-ington, DC: The Urban Press Institute.

Preacher, Kristopher J. and Andrew F. Hayes.2004. "SPSS and SAS Procedures for Es-timating Indirect Effects in Simple Media-tion Models." Behavior Research Methods, In-struments, & Computers 36:717-31. http://dx.doi.org/10.3758/BF03206553

Raver, C. Cybele. 2002. “Emotions Matter:Making the Case for the Role of YoungChildren’s Emotional Development for EarlySchool Readiness,” Social Policy Report 16:3-19.

Reichman, Nancy, Julian Teitler, Irwin Garfinkel,and Sara McLanahan. 2001. “Fragile Fami-

lies: Sample and Design.” Children and YouthServices Review 23:303-326.http://dx.doi.org/10.1016/S0190-7409(01)00141-4

Rosenbaum, Paul R. 2002. Observational Studies,2nd edition. New York: Springer. http://dx.doi.org/10.1007/978-1-4757-3692-2

Rosenbaum, Paul R., and Donald B. Rubin. 1983.“The Central Role of the Propensity Scorein Observational Studies for Causal Effects.”Biometrika 70:41-55. http://dx.doi.org/10.1093/biomet/70.1.41

Rubin, Donald B. 1987. Multiple Imputationfor Non-Response in Surveys. New York:John Wiley and Sons. http://dx.doi.org/10.1002/9780470316696

Sampson, Robert J. and Charles Loeffler. 2010.“Punishment’s Place: The Local Concentra-tion of Mass Incarceration.” Daedalus 139:20-31. http://dx.doi.org/10.1162/DAED_a_00020

Shifrer, Dara, Rebecca Callahan, and Chan-dra Muller. 2013. “Equity or Marginaliza-tion? The High School Course-taking ofStudents Labeled with a Learning Disabil-ity.” American Educational Research Journal50:656–82. http://dx.doi.org/10.3102/0002831213479439

Sobel, Michael E. 2006. “Spatial Concentrationand Social Stratification: Does the Clusteringof Disadvantage Beget Bad Outcomes?” Pp.204-229 in Poverty Traps, edited by S. Bowles,S. N. Durlauf, and K. Hoff. New York: RussellSage Foundation.

Sourander, Andre, Leila Helstela, and Hans He-lenius. 1999. “Parent-adolescent agreementon emotional and behavioral problems.” So-cial Psychiatry and Psychiatric Epidemiology34: 657-663. http://dx.doi.org/10.1007/s001270050189

Swisher, Raymond and Maureen Waller. 2008.“Confining Fatherhood: Incarceration and Pa-ternal Involvement among Nonresident White,African American and Latino Fathers.” Jour-nal of Family Issues 29:1067-1088. http://dx.doi.org/10.1177/0192513X08316273

Tenenbaum, Harriet, Michele Porche, CatherineSnow, Patton Tabors, and Stephanie Ross.2007. “Maternal and Child Predictors of Low-Income Children’s Educational Attainment.”

sociological science | www.sociologicalscience.com 157 April 2014 | Volume 1

Haskins Effects of Paternal Incarceration on Child School Readiness

Journal of Applied Developmental Psychology28:227-238. http://dx.doi.org/10.1016/j.appdev.2007.02.002

Wakefield, Sara and Christopher Uggen. 2010.“Incarceration and Stratification.” Annual Re-view of Sociology 36:387-406. http://dx.doi.org/10.1146/annurev.soc.012809.102551

Wakefield, Sara and Christopher Wildeman.2011. “Mass Imprisonment and RacialDisparities in Childhood Behavioral Prob-lems.” Criminology & Public Policy 10,3:793-817. http://dx.doi.org/10.1111/j.1745-9133.2011.00740.x

Western, Bruce and Katherine Beckett. 1999.“How Unregulated is The US Labor Market?The Penal System as a Labor Market Insti-tution.” The American Journal of Sociol-ogy 104:1030-1060. http://dx.doi.org/10.1086/210135

Wildeman, Christopher. 2009. “Parental Im-prisonment, the Prison Boom, and the Con-centration of Childhood Disadvantage.” De-mography 46:265-280. http://dx.doi.org/10.1353/dem.0.0052

——. 2010. “Parental Incarceration and Children’sPhysically Aggressive Behaviors: Evidencefrom the Fragile Families and Child WellbeingStudy.” Social Forces 89:285-310. http://dx.doi.org/10.1353/sof.2010.0055

Wildeman, Christopher and Christopher Muller.2012. “Mass Imprisonment and Inequal-ity in Health and Family Life. ” An-nual Review of Law and Social Science8:11-30. http://dx.doi.org/10.1146/annurev-lawsocsci-102510-105459

Zhang, Dalun, Antonis Katsiyannis, Song Juand Eric Roberts. 2014. "Minority Rep-resentation in Special Education: 5-YearTrends." Journal of Child & Family Studies23(1):118-27. http://dx.doi.org/10.1007/s10826-012-9698-6

Acknowledgements: This research was supportedby a National Science Foundation Graduate Re-search Fellowship, a Ford Foundation Disserta-tion Fellowship, the American Sociological Soci-etyâĂŹs Minority Fellowship Program, and bythe Institute of Education Sciences, US Depart-ment of Education through grant R305C050055to the University of Wisconsin-Madison. Theauthor also thanks the Eunice Kennedy ShriverNational Institute of Child Health and Hu-man Development (NICHD) through grantsR01HD36916, R01HD39135, and R01HD40421,as well as a consortium of private foundationsfor their support of the Fragile Families andChild Wellbeing Study. All errors are my own,and the contents of this paper do not neces-sarily reflect the official views of any fundingagencies. This paper has benefited from nu-merous rounds of helpful comments–too manyto fully acknowledge here. I graciously thankall those who provided feedback over the years,especially Paul Hanselman, Elizabeth Wrigley-Field and Jeffery Grigg.

Anna R. Haskins: Columbia Population Re-search Center, Columbia University. E-mail:ah3157@columbia.edu.

sociological science | www.sociologicalscience.com 158 April 2014 | Volume 1