Keywords

Since the 1960s, there has been a slow but steady rise in academic interest in the potentially harmful consequences of parental imprisonment for children (e.g. Friedman & Esselstyn, 1965; Hagan & Dinovitzer, 1999; Robins, West, & Herjanic, 1975; Wildeman & Andersen, 2017). Since then, parental imprisonment has been found to correlate with a variety of adverse intergenerational outcomes, including antisocial and delinquent behaviour, low academic attainment, and substance misuse (e.g. Murray & Farrington, 2005; Wildeman, 2014b).

Although the outcomes associated with parental imprisonment are well-established, there remains uncertainty about their cause. High levels of disadvantage in the families of prisoners make it difficult to identify whether negative outcomes are a consequence of parental imprisonment itself or a reflection of children’s greater exposure to pre-existing and concurrent risk factors (Bijleveld, 2009; Johnson & Easterling, 2012; Murray & Farrington, 2008; Wakefield, Lee, & Wildeman, 2016; Wildeman, 2014b). A global lack of longitudinal data sources with a sufficient sample size and variable range to control for selection effects has long meant that few studies could address this critical question. Most studies that investigate the direct effect of parental imprisonment on children rely on a small number of US-based data sources (Johnson & Easterling, 2012; Murray, Farrington, & Sekol, 2012a; Wildeman, 2016). A 2012 systematic review of children’s antisocial behaviour, mental health, drug use, and educational performance after parental incarceration found that of 50 research samples used to investigate these outcomes, only 14 were from outside of the USA (Murray, Loeber, & Pardini, 2012b). The body of international studies large enough to allow for rigorous causal tests is now expanding (Murray, Bijleveld, Farrington, & Loeber, 2014). This internationalization raises new questions about differences and similarities between children of prisoners in different countries.

Harms from parental imprisonment may vary across places and time periods. Differences in welfare systems, public policy, and penal policy may affect the social composition of the prison population (Hartwell, 2004). In countries where prison sentences are used as a last resort, prison populations are more likely to have other social problems, such as addiction and mental illness. In such contexts, parental imprisonment may have a less negative (or even positive) effect on families. In addition, cultural and social contexts likely determine the extent to which prisoners’ families fear and experience stigmatization (Murray et al., 2014).

Imprisonment also has different effects on different groups of children within countries. Research evidence suggests ethnicity may influence the consequences of parental imprisonment for affected children (Murray et al., 2012b; Swisher & Roettger, 2012; Wildeman, 2014b). Also, the way individual prisons regulate visitation and other forms of communication has a critical influence on children’s relationships with imprisoned parents (Comfort, 2003; Dennison, Smallbone, & Occhipinti, 2017a; Dennison & Besemer, 2018, forthcoming; Dennison, Smallbone, Stewart, Freiberg, & Teague, 2014). Such regulations and practices differ not only between countries, but also between prisons (Murray et al., 2014). Moreover, differences in welfare systems and public health care may substantially reduce the extent to which parental imprisonment introduces economic hardship and other forms of strain into children’s lives. For all of these reasons, it is likely that the consequences of imprisonment for children may differ between countries, within countries and across time periods.

There are only a few international reviews of the literature on the effects of parental imprisonment on children. Most reviews aggregate country-specific findings to identify average effects (e.g., Johnson & Easterling, 2012; Murray et al., 2012a; Wildeman, Wakefield, & Turney, 2013b). Two studies also attempted to match the samples and outcome variables between two or more international datasets so that national differences in the effects of imprisonment can be more easily identified (Besemer, van der Geest, Murray, Bijleveld, & Farrington, 2011; Murray et al., 2014). Besemer et al. (2011) found that the relationship between parental imprisonment and offspring offending differed considerably between the Netherlands and England and suggested that these might be due to major discrepancies in the penal landscape in each country. Murray et al. (2014) found differences between groups of countries. They found larger, positive relationships between parental imprisonment and male offspring offending in England, the Netherlands in the 1970s–1980s and the USA and smaller or negligible effects in the Netherlands in the 1950s–1960s and Sweden. They concluded that variations in both social and penal climates may explain these differences. The current chapter draws together and extends these reviews by considering the generalizability of international studies that look at the effects of parental imprisonment on children’s outcomes.

In this chapter, we identify three specific outcomes for the children of prisoners that have been studied in more than one country: substance use, mental health problems, and adult offending. For each outcome, we discuss the extent to which studies in different national contexts show consistent results. Our main purpose is to reflect on the extent to which international evidence can be used to identify similarities or differences in the effects of parental imprisonment between countries. From this, we consider to what extent the current evidence base can be used to draw conclusions about the way imprisonment affects children globally. Finally, we discuss other ways in which the increasing internationalization of parental imprisonment research may benefit our understandings of the way imprisonment affects family members and identify gaps in the existing knowledge base.

Method

We begin by identifying outcome variables that have been investigated in more than one national context. Most studies were selected because they had been reviewed in other recent literature reviews (e.g., Johnson & Easterling, 2012; Murray et al., 2012a; Wildeman et al., 2013b). More recent work was identified through targeted searches within each topic area. We restricted the review to studies that controlled for pre-existing risk in affected children’s lives, for example, through comparison groups, fixed effect modelling, and covariate adjustment. In all studies, children were affected by imprisonment after birth. We excluded qualitative studies from our comparison, although we did consider results from these studies to inform our discussion of future directions in international family imprisonment research.

Unfortunately, many child outcomes thought to be affected by parental imprisonment have only been studied in one country. For example, effects of parental imprisonment on children’s education (e.g. Cho 2009, 2010, 2011; Dallaire, Ciccone, & Wilson, 2010; Hagan & Foster, 2012) and on physical health (e.g. Lee et al., 2014; Turney, 2014b) have, thus far, only been examined in the USA. In fact, we only found three specific outcomes for prisoners’ children that could be compared across at least two different countries: offending, substance use, and mental illness.

We reviewed the international studies both in terms of the effects they found on children’s substance use, offending or mental health, as well as for differences and similarities between the designs, locations, and samples. Specifically, we considered (a) what type of parental imprisonment was investigated; (b) whether the study population is representative of all children affected by that experience within the national population; (c) the method of causal inference (e.g., covariate adjustment, matching, or other types of analyses); and (d) the way the outcome variable was operationalized. In the following sections, we consider the effects of imprisonment on children’s later offending risk, on children’s substance use, and on children’s mental illness.

Findings

Parental Imprisonment and Intergenerational Crime and Delinquency

The association between parental imprisonment and children’s adulthood offending risk has been one of the oldest foci in the parental imprisonment literature (e.g., Farrington, Barnes, & Lambert, 1996; Glueck & Glueck, 1950; Robins et al., 1975; Wildeman & Andersen, 2017), as well as in criminology more generally (Murray et al., 2012b). Maternal and paternal imprisonment have been found to be associated with an increased adulthood offending risk in all countries in which such associations have been measured (Murray et al., 2014). However, it is not certain that these associations represent a causal effect. To date, there are only five countries in which there have been longitudinal studies that estimate direct effects of parental imprisonment on offspring offending. The characteristics of these studies are summarized in Table 6.1.

Table 6.1 Cross-national selection of studies measuring the association between parental imprisonment and offspring offending
Full size table

Wildeman and Andersen (2017) used an exogenous Danish sentencing reform as a natural experiment with which to compare the effects of parental (and specifically, paternal) imprisonment on children. This policy reform resulted in a sudden drop in the use of custodial sentences. The authors compared children ages 12–18 years at the time of their father’s sentence to custody versus a matched group of same-aged children whose fathers received non-custodial sentences. The authors found that paternal incarceration had a substantial effect on boys’ risk of criminal justice contact in the next 10 years but found no significant effect for girls (Wildeman & Andersen, 2017).

Most other research in this area has controlled for selection bias through covariate-adjusted regression (though see also Murray et al., 2012b). Of these studies, the strongest intergenerational effects of offending were found in The Cambridge Study in Delinquent Development, a prospective longitudinal study of 411 boys born in 1953 in a working-class area of South London, England. Of these 411 boys, 23 boys were found to have had a father or mother imprisoned between birth and age ten. These 23 boys were more likely to engage in criminal behaviour (Murray, Janson, & Farrington, 2007) or antisocial–delinquent behaviour (Murray & Farrington, 2005) than boys affected by parental death, parental separation, and boys with parents imprisoned before their births.

Project Metropolitan in Sweden is also a prospective longitudinal survey. In a study which directly compared results from Project Metropolitan to the Cambridge study, parental incarceration in the Swedish study was found to have no significant effect on criminal convictions in adulthood after statistically controlling for the criminality of the parent. In a comparison of children exposed to parental imprisonment in childhood with children whose parents were imprisoned only before the child’s birth, both had an equal likelihood of adulthood conviction. This suggests that in Sweden, parental imprisonment was not a direct cause of children’s offending during adulthood (Murray et al., 2007).

In the Netherlands, the effect of parental imprisonment as a cause of second-generation offending has differed between studies using data from different time periods. The Criminal Career and Life Course Study (CCLS) used court information and life course data from 4615 randomly selected individuals convicted of a crime in the Netherlands in 1977. Using these data, van de Rakt, Murray, and Nieuwbeerta (2011) found that there was a significant association between fathers’ imprisonment and child convictions. When fathers’ criminal history was controlled for, the influence of paternal imprisonment became very weak and only increased risk of conviction by a factor of 1.2.

The NSCR Transfive study, also in the Netherlands, started with a group of 198 high-risk working-class boys born in 1899 (G2) (Huschek & Bijleveld, 2015). Conviction data were obtained for their children (G3), grandchildren (G4) and great-grandchildren (G5). The Dutch findings suggest that the effect of parental imprisonment varied across different historical periods. There was no significant relationship between paternal imprisonment and offspring offending in earlier generations of the study, but G5 children of incarcerated G4 parents were at a significantly increased risk of offending compared to children of criminal but never-imprisoned parents (Murray et al., 2014). This suggests that in the 1950s–1960s, parental incarceration was not a risk factor for sons’ adult crimes, but parental incarceration did influence children’s offending outcomes from the 1970s–1980s onwards, the period in which the G5 children grew up. A further study of the Transfive dataset found that for G3-G4 and G4-G5, only fathers who were incarcerated after their son was born, and before that son turned 18, had an influence over their son’s risk of being incarcerated, suggesting a causal relationship between paternal and offspring imprisonment (Dennison, Bijleveld, & van de Weijer, 2017b).

In the USA, results across different studies have largely supported a direct effect of parental imprisonment on offspring offending. Burgess-Proctor, Huebner, and Durso (2016) used the National Longitudinal Study of Adolescent Health (AddHealth), a longitudinal nationally representative sample of 20,748 respondents who were enrolled in grades 7–12 in 1994–95, 15,587 of whom had information in the first and final wave. Both maternal and paternal incarceration significantly increased the odds of adult offspring’s self-reported arrest, conviction or incarceration after age 18. Murray et al. (2012b) also reported a positive association using the Pittsburgh Youth Study. Using a combination of fixed effect models and propensity score matching, the authors found that parental imprisonment predicted increases in youth theft. However, these findings contradict an earlier study also using AddHealth data (Porter & King, 2015). In this study, delinquency measures in children in a survey wave prior to their father’s imprisonment were compared with the same measures in children who had already experienced paternal incarceration. Using this method, the authors found no significant association between paternal incarceration and offending (Porter & King, 2015).

Looking at findings across countries, it remains difficult to draw general conclusions regarding the effects of parental imprisonment on delinquency and offending in offspring. Notably, there are few studies that include girls. For boys, the database is larger, and the effects of imprisonment on adulthood offending appear to be contextually dependent. Results differ between countries, as well as in other important ways, such as in different generations in the Netherlands. The cross-temporal differences in the Netherlands may relate to a shift in the Dutch penal climate, which became less liberal after the late 1970s. In this period, the penal climate shifted towards a greater focus on the expansion of imprisonment. At the same time, government became skeptical about the ability of prisons to rehabilitate prisoners. It is possible that over the course of this period, parental imprisonment may have become more stigmatized and perhaps more damaging to children (Murray et al., 2014). Nonetheless, it is difficult to know whether differences in results represent generational differences, national differences or differences in methodology. For example, it might seem surprising that two Scandinavian studies would find completely different effects of imprisonment on children. However, as the children in the Danish study were born roughly 30 years after the children from Stockholm and were from rural areas as well as metropolitan locations, these seemingly contrasting findings could easily be explained by differences that do not relate to the national context in which each study took place. Overall, the results do not allow for any firm conclusions about the effects of imprisonment on children across countries in terms of delinquency and crime.

Parental Imprisonment and Addictive Substance Use

There are a number of mechanisms that might link parental imprisonment to the use of addictive and/or illegal substances in offspring, including selection effects. Evidence from studies conducted within the USA shows that, within a sample of young people in mental health settings, children of prisoners were far more likely than others to have been exposed to parental alcohol or drug abuse as well as other family risk factors (Phillips, Burns, Wagner, Kramer, & Robbins, 2002). Although few authors have discussed causal mechanisms that might directly link parental imprisonment to drug use, mechanisms could include the way young people are supervised and parented as well as the psychological responses of youth to parental absence (Murray & Farrington, 2008).

Despite these potential causal connections, there have been only a few studies that investigated the connection between parental imprisonment and alcohol, drug, or other substance use in prisoners’ children. In a 2012 systematic review of effects of parental imprisonment on drug use, Murray found only eight studies that investigated the impact of parental imprisonment on either child substance use or the propensity to use. A pooled odds ratio of 1.0 across these eight studies suggested no average cross-country effect of parental imprisonment on offspring substance use. It is, however, possible that this average effect may have hidden country-specific differences (Murray et al., 2012a), or differences that depend on the type of substance misuse, and the timing at which it was measured. A selection of these and subsequent studies have been summarized in Table 6.2. For this review, we have separated studies that look at substance use in prisoners’ children during childhood, i.e., before the age of 18, and those that look at substance use in adulthood.

Table 6.2 Cross-national selection of studies measuring associations between parental imprisonment and substance use
Full size table

An Australian study by Kinner, Alati, Najman, and Williams (2007) used data from the Mater Hospital University of Queensland Study of Pregnancy (MUSP), a prospective birth cohort study of children born 1981–83 in Brisbane, Australia (N = 2399). In their study, paternal incarceration correlated significantly with children’s alcohol and tobacco use at age 14, but these associations became non-significant after controlling for family characteristics. The authors conclude that in Australia, adverse outcomes for adolescent children of imprisoned fathers (or maternal spouses) are attributable to social and familial risk factors rather than to the effects of paternal imprisonment (Kinner et al., 2007). Murray et al. (2012b) used data from the Pittsburgh Youth Study, a longitudinal study of 1009 inner-city boys, to examine within-individual change from before to after parental incarceration. They found that parental arrest and conviction had no effect on increases in youth-reported marijuana use when compared to boys with similar behaviors and family and peer environments before parental incarceration occurred (Murray et al., 2012b). As fixed effect models have the advantage of controlling for both observed and unobserved time constant variables, this lack of any effect of parental arrest and conviction on young people’s marijuana use is a strong indicator against any causal link between parental conviction and marijuana use in boys in the USA. Neither study suggests any link between (step) parental incarcerations and substance use in childhood. Effects of parental imprisonment on adulthood substance use have been more mixed, including positive, negative, and null associations.

Hayatbakhsh, Kinner, Jamrozik, Najman, and Mamun (2007) tested whether the experience of the arrest or incarceration of a mother’s partner before a child reached 14 years of age was associated with the use of cannabis in early adulthood. The study found a possible positive effect of paternal (or stepfather) imprisonment. Children who experienced the arrest of their mother’s partner before they reached the age of 14 used more cannabis at age 21 than children whose mothers’ partners were not arrested, but children whose mothers’ partners were imprisoned did not have a higher risk. The authors conclude that continued exposure to a criminal father or stepfather may pose a greater risk to children’s subsequent drug use than that person’s removal through incarceration.

Using the Cambridge Study dataset in England, Murray, and Farrington compared drug use in children affected by parental imprisonment with four control groups: boys who did not experience separation, boys separated by hospital or death, boys separated for other reasons and boys whose parents were only imprisoned before their birth. They observed that compared to boys in these reference groups, parental imprisonment increased the odds of offspring illegal drug use between ages 32 and 48 years by a factor of 3.7 (Murray & Farrington, 2005).

Research from the USA also produced mixed results on the relationship between parental imprisonment and drug use. Roettger, Swisher, Kuhl, and Chavez (2011) used the National Longitudinal Study of Adolescent Health (AddHealth) to investigate the effects of having a biological father imprisoned on the number of days a year that young people used marijuana. Using covariate adjustment to control for confounders at the family, parental, and individual level, Roettger et al. (2011) found that for both males and females, having ever had a biological father imprisoned was associated with an increased frequency of marijuana use and increased odds of any other illegal drug use. A subsequent study, which again used AddHealth data, used propensity score matching to compare children who reported that one or both of their parents had been incarcerated with a matched sample of children with similar characteristics. Compared to this matched sample, parental incarceration significantly increased the odds of marijuana use in early as well as late adulthood, but had no significant effect on heavy alcohol use in early adolescence, and only a very small effect on heavy alcohol use in late adolescence (Mears & Siennick, 2015). However, research looking at children’s marijuana use before as well as after imprisonment did not find a net effect.

When comparing these apparently contradictory results both between and within countries, it should be emphasized that there were major differences in study designs. In the Australian Mater study, the authors operationalized paternal imprisonment as the incarceration of the mother’s spouse (Kinner et al., 2007). At age 14, it is likely that a substantial number of mothers’ spouses would have been the child’s stepfather. In AddHealth and in the Cambridge study, the imprisoned father was the child’s biological father. This is a salient difference. Stepfathers may enter and leave a child’s life and may therefore have less influence on a child’s ongoing behaviour when imprisonment results in their removal from the household. Biological fathers may be more likely to have a longer and more enduring influence on children, which could explain the difference in results. Moreover, the Mater study included tobacco and alcohol use, both of which are legal, whereas other studies examined illegal drugs. In addition, maternal imprisonment may not only have different effects from paternal imprisonment, but also select a group of children with different pre-existing problems. The extent to which there may be country-specific differences in the nature and extent of the relationship between exposure to parental imprisonment and drug use therefore remains unclear.

Mental Health

A number of studies have described the emotional distress many children experience after the imprisonment of their mother or father (Arditti, 2012; Condry, 2007; Dennison & Besemer, 2018). Such negative emotions, as well as a resulting strain on family relationships, could lead to depressive symptoms in children (Gaston, 2016). In addition, parental imprisonment may precipitate other stressful changes within a family system that may have negative impacts on the mental health of a child (Arditti, 2016). For example, material deprivation caused by increased expenses and a loss of the prisoner’s income may affect children’s ability to engage in social activities. Such deprivations may contribute to the maladaptive emotional responses of a child and contribute to the development of enduring mental health problems (Besemer & Dennison, 2017; Dennison & Besemer, 2018, forthcoming).

Some scholars have proposed that parental imprisonment in childhood could potentially be a cause of mental illness in adulthood. As mental illness is normally episodic, it is improbable that children affected by imprisonment would subsequently exhibit continuous depressive symptoms. Causal explanations of mental health problems in adult children of prisoners are therefore quite distinct from causes that may provoke depressive responses in childhood (Gaston, 2016). It is possible that children’s deprivation of shared interactions with an imprisoned parent could impair their acquisition of capabilities necessary for healthy long-term physical, socio-emotional, and cognitive development (Arditti, 2016; Dennison & Besemer, 2018), resulting in a lifelong vulnerability to mental illness. Some authors also propose that the cumulative effects of parental imprisonment on children might affect their long-term disengagement or alienation from society (Besemer & Dennison, 2017, 2018 forthcoming; Foster & Hagan, 2015). Such explanations of adulthood consequences of parental imprisonment remain speculative. However, given the likely divergent pathways between the effects of parental imprisonment during childhood and in adulthood, we discuss each outcome separately.

As can be seen in Table 6.3, both child and adulthood mental health outcomes of parental imprisonment have been studied in more than one country. Nonetheless, as with substance use and offending, outcomes are difficult to compare. Kinner et al. (2007) found no evidence of a causal effect of prior parental imprisonment on teenage internalizing symptoms of fourteen-year-old children born in Brisbane, Australia. Murray et al. (2012b), on the other hand, found that Pittsburgh children were more likely to experience depressive symptoms within four years after a parental imprisonment that took place between ages 7 and 18. Various factors could explain this discrepancy. It may be that reactive psychological responses to parental imprisonment are immediate, rather than long-term, and were therefore not captured in the Australian study. It is also possible that imprisonment after the age of 10, which formed the majority of the Pittsburgh sample, provokes stronger psychological responses. Without further evidence, it is difficult to disentangle cross-national differences. Similarly, though results on adulthood mental health effects of parental imprisonment appear to be consistent between the UK and the USA, the large time difference between the birth years makes it difficult to be certain whether this convergence represents a true and contemporary commonality between the effects of parental imprisonment on adult offspring living in those countries.

Table 6.3 Cross-national selection of studies measuring associations between parental imprisonment and mental health problems in children or adults
Full size table

Emerging Directions for International Family Imprisonment Research

For all three subject areas reviewed in this chapter, we found that current research on the consequences of imprisonment has yielded conflicting evidence in different countries. At the same time, the work reviewed in this chapter also demonstrates that across many different locations and time periods, children with imprisoned parents are a vulnerable population. There is also enough evidence to come to the conclusion that at least in some countries, for a substantial proportion of children and families affected, parental imprisonment compounds pre-existing harms in the lives of affected children. The international evidence reviewed in this chapter thus creates a compelling argument for the need to protect children and families affected by imprisonment. However, findings are far less specific about what children and families, under what circumstances and policy climates, suffer worse outcomes. Ongoing work measuring average effects of parental imprisonment on children has therefore been less than successful in assisting policy-makers or practitioners in targeting support (see also Wakefield & Wildeman, 2013). For future studies to have a better practical application, potential reasons for heterogeneity in prisoners’ children’s outcomes need to be examined and addressed. This section will focus on two key directions such work should take and reflect on how such work could contribute to policies to protect affected children.

First, studies need to identify which children, under what circumstances are most likely to be harmed by the imprisonment of a parent (Wakefield & Wildeman, 2013). Qualitative data suggest various potential sources of variability in prisoners’ children’s outcomes, depending on a range of contextual differences between affected families and on what types of imprisonment-related experiences children are exposed (e.g. Giordano, 2010). Much of the research reviewed in this chapter uses samples of children affected by the imprisonment of a biological or social mother or father, treating these distinct experiences as a single predictor of risk. Consequently, it is still unclear whether maternal and paternal incarceration initiates distinct pathways towards negative child outcomes (Wildeman, 2014a). Similarly, few studies have been able to differentiate between children affected by varying durations of parental prison sentences (Geller, Jaeger, & Pace, 2016), by different types of criminal justice involvement or between parental imprisonments that take place at different points in children’s development (though see Murray et al., 2012b). In addition, few studies have been able to distinguish between children whose parents had more (or better quality) parenting involvement in their lives prior to their imprisonment and those whose relationship with the imprisoned parent was already impaired.

From a policy perspective, empirical studies of variability in children’s outcomes are not only important in identifying circumstances under which parental imprisonment is most likely to be harmful, but also to identify the types of families least able to provide stable support to children at stressful times (Besemer & Dennison, 2017). Studies may also help to identify circumstances in which the removal of a criminal, and potentially chaotic or violent, parent from the home may improve children’s well-being (Hissel, 2014; Jaffee, Moffitt, Caspi, & Taylor, 2003). Indeed, research from the Netherlands suggests that for children of violently criminal fathers, parental separation reduces prisoners’ children’s offending risk (van de Weijer, Thornberry, Bijleveld, & Blokland, 2015). A better understanding of such variability may also help to understand incongruities in the findings of different studies examining the effects of maternal imprisonment (Arditti, 2015; Turney & Wildeman, 2015) and paternal imprisonment (Wildeman, Wakefield, Lee, Wakefield, & Powell, 2016), including those described in this chapter.

A second direction of work pertains to the mechanisms through which imprisonment may affect children through their developmental context. Specific investigations of these mechanisms are rare (though see Murray et al., 2012b) and have received far less research attention than average effects (Auty, Farrington, & Coid, 2015). Consequently, current scholarship has yielded few theoretical foundations with which to understand the way imprisonment may impact on children and has made even less progress in testing the few theoretical mechanisms that have thus far been proposed (Auty et al., 2015).

A key reason for the lack of progress in testing and developing theories about the effects of parental imprisonment has been a lack of data. There are few large longitudinal studies that contain measures of the mechanisms that are thought to be most important in affecting children’s long-term outcomes after the imprisonment of a parent. For example, authors have proposed causal pathways relating to the effects of traumatic child-parent separation for children’s bonding and attachment; reductions in the quality of parenting, care, and supervision of children; financial hardship; and children’s development of a delinquent identity through stigma or labelling (Besemer et al., 2011; Murray, 2007; Murray & Murray, 2010; Shlafer & Poehlmann, 2010). Empirical tests of such mechanisms have remained quite limited. Some potential mechanisms, such as stigma and discrimination, are very difficult to investigate through existing longitudinal data sources due to a lack of information. A slightly larger number of studies have focused on mechanisms that may cause disruptions to children’s secure emotional support from caring adults after parental imprisonment, including effects on caregiver stress (Arditti, 2016; Chui, 2016), parenting (Turney, 2014a), and depression (Wildeman, Schnittker, & Turney, 2012). Also, a growing body of work has developed around the ways in which parental imprisonment may limit children’s ability to engage in normal social activities, potentially resulting in their long-term social exclusion (Besemer & Dennison, 2017; Dennison & Besemer, 2018; Schwartz-Soicher, Geller, & Garfinkel, 2011; Sykes & Pettit, 2015). As parental imprisonment research grows internationally, there may be greater scope for research to build and test theories about the way imprisonment impacts on children’s lives. It is essential to identify potential risk and protective factors that can be targeted through policy and for the design of successful interventions to protect and support affected children.

Finally, there is also a need to widen the scope of family imprisonment research. Until now, there has been almost no empirical evidence regarding the risks associated with the imprisonment of any household members and/or close family other than a parent (Meek, 2008; Meek, Lowe, & McPhillips, 2010; Wildeman & Wakefield, 2014). Yet from a theoretical perspective, many of the same mechanisms currently thought to affect prisoners’ children could also apply to the imprisonment of other close family members. For example, parents who are coping with stress associated with the imprisonment of their own sibling, or the imprisonment of one of their older children, may experience psychological distress that could affect their ability to parent and care for children in their household. In addition, children of non-parental incarcerated family members may similarly experience the effects of increases in household expenses associated with travel for prison visits, subsidizing prisoner phone calls and making financial contributions to prisoners’ commissary accounts for personal items. Such costs can severely limit families’ contact with the prisoner and may have a detrimental effect on households’ finances (Christian, Mellow, & Thomas, 2006; Shlafer & Poehlmann, 2010; Braman, 2004). These financial consequences may also impact on children’s ability to engage in social activities, including school activities, leisure activities or family outings (Besemer & Dennison, 2017; Dennison & Besemer, 2018, forthcoming).

A small number of studies around the world have begun to examine the effects of non-parental family imprisonments. Recent Australian evidence showed that children who experienced non-parental household imprisonment were at least as vulnerable to social exclusion as children living in families where a parent was imprisoned (Besemer & Dennison, 2018 forthcoming). In the UK, qualitative research described severe psychological distress in children affected by the imprisonment of a sibling (Meek, 2008; Meek et al., 2010; Slomkowski, Rende, Conger, Simons, & Conger, 2001). Other qualitative research in the UK showed that the stigma of imprisonment is not limited to parental incarceration, but affects other family relationships as well (Condry, 2007). A recent qualitative study in the USA, by Comfort (2016), described the cumulative disruptions to the lives of women caring for different types of family members with frequent and chronic criminal justice involvement.

Rigorous empirical studies of the direct consequences of non-parental family imprisonment have been quite rare, though limited research suggests that as with parental imprisonment, non-parental household and close family imprisonments are associated with long-term problems in children. A number of studies have confirmed that like parental imprisonment, the criminal convictions of siblings, fathers, uncles, aunts, and grandparents also predict children’s subsequent delinquency (Farrington et al., 1996, Farrington, Jolliffe, Loeber, Stouthamer-Loeber, & Kalb, 2001; Slomkowski et al., 2001). Wildeman and Wakefield (2014) found that children affected by the imprisonment of parents were also much more likely to experience the imprisonment of other family members, suggesting that the effects of parental imprisonment may be aggravated by this additional criminal justice exposure. In the Netherlands, van de Rakt, Nieuwbeerta, and Apel (2009) similarly found that non-parent family members’ offending had an additional effect on children’s own offending, net of the effect of parental imprisonment itself. However, there have been no studies examining outcomes other than offending in relation to non-parental imprisonment. If imprisonment transmits risks not only to prisoners’ own children, but also to prisoners’ siblings, grandchildren, and other family, the group of children and adults potentially affected is much larger than has previously been assumed. Even if prisoners’ family members’ risk is primarily caused by selection effects rather than by factors directly related to imprisonment, the current exclusive focus on parental imprisonment still excludes a much larger and potentially equally vulnerable group.

Conclusions

The consequences of imprisonment for children may differ substantially between countries, within countries, and across time periods. However, the nature and extent of such differences remain largely unknown. In this chapter, we attempted to find areas where outcomes studied in different countries might be sufficiently comparable to allow for a cross-national analysis of the effects of parental imprisonment on children. We found that within the area of parental imprisonment research, cross-national comparisons were difficult to make. Substantial differences between studies on parental imprisonment in different countries mean that as yet, there are only a few topics where research outcomes are similar enough to be compared. For many types of child outcomes, there were no studies that contained sufficiently similar measures across more than one country. In fact, in this study, we were only able to identify three child outcomes, substance use, offending, and mental illness, which could be compared cross-nationally.

For both substance use and offending, it was often difficult to distinguish whether conflicting cross-national findings derived from genuine differences in the social and penal contexts in which they were investigated or from differences in the research designs and methodologies (see also Murray et al., 2014). Although we found outcomes that had been studied in several countries, the studies were not easily comparable because of differences in the sample of children, the type of parental imprisonment they were affected by, the time period they grew up in and in the variables used to study each outcome. Although two previous reviews have matched international samples in order to remove some of the discrepancies in child gender, age, and child outcome between studies (Besemer et al., 2011; Murray et al., 2014), many of the differences in design and time period cannot be overcome through statistical means. To more accurately identify cross-national variations would require a harmonization of questions in the design stage of future studies of parental imprisonment, as well as in large longitudinal surveys more generally. This would make such data sources more easily comparable.

We conclude that, despite the considerable increase in international evidence on the effects of parental imprisonment on children, this evidence cannot yet be used to draw firm conclusions about the moderating effects of national context on the effect of parental imprisonment on children. While the nature and extent of country-specific effects remain unknown, researchers should be very cautious in generalizing findings regarding potential effects of imprisonment on children in one country to potential consequences for children growing up in very different national contexts. In particular, researchers should be sensitive to the possibility that findings from the USA may not be generalizable to other countries with fundamentally different penal and welfare contexts (see also Wildeman, 2016). However, although current limitations in research evidence mean that country-specific effects remain difficult to identify, a wider array of data across different parts of the world has also opened up new research opportunities to explore mechanisms through which parental imprisonment may affect children, as well as the extent to which there are broader, family-based effects associated with imprisonment.