The importance of both parents in minimising the exposure of children to Adverse Childhood Experiences (from the recently issued 2018 “Report 1: Mental Illness – Welsh Adverse Childhood Experience (ACE) and Resilience Study”)
- The Prevalence of Fatherless Children
- Families Without Fatherhood, Dennis & Erdos, 1992
- The Big Picture – A Selection of Studies
- The Study of Radl, Salazar and Cebolla-Boado
- The Study of McLanahan, Tach and Schneider
In previous posts I have presented data on some of the mechanisms which lead to fathers becoming estranged from their children (see Marriage, and Families, Divorce, Cohabitation, and Non-Resident Parents: Where Is The Bias?). Whilst divorce is a significant contributor, it is the decline of marriage which is the dominant cause of fatherlessness. Half of children are now born outside of marriage in the UK. At any time, about one-quarter of dependent-age UK children are not living with their father. Only about half of children will live with their father continuously to the age of 16. In the USA, the Census Bureau estimates that in 2014 approximately 24% of children under age 18 lived with only their mother, 4% lived with only their father, and 4% lived with neither parent. In the USA, only about half of unmarried parents are living together even when the child is born (Fragile Families and Child Wellbeing Study).
If fatherlessness has a significant disbeneficial effect on children, the enormous scale of fatherlessness implies a similarly enormous scale of disadvantage to children.
Here I review briefly the evidence for the impact of fatherlessness on children and their adult outcomes. If fatherlessness results, even if only in part, from an empathy gap for fathers and hence a lack of societal concern for their often-involuntary estrangement from their children, then the adverse effect of fatherlessness is collateral damage to children in which society is complicit.
Worse, there is a danger that a positive feedback mechanism will propagate the adverse effects of fatherlessness to the following generation, thus driving an increasing gradient in social capital, and that this will operate differentially on males.
There is no shortage of empirical studies which show the relationship between fatherlessness and adverse outcomes for children. Fathers’ advocates will eagerly quote such sources, and attribute the blame to the exclusion of fathers from their children’s lives after parental separation. The scale at which society has been disenfranchising fathers from their children’s lives over the last 45 years, as outlined in my previous posts, lends credibility to that perspective.
Those wishing to promote more childcare by fathers, perhaps as part of a strategy to enhance women’s standing in the workplace, will also eagerly quote such sources, though they will attribute blame differently. It is men who must change, they will opine, to take some of the childcare burden from mothers. That being more involved in childcare, perhaps even the primary carer, may avail fathers little when it comes to continuing involvement with their children after separation is a fact which such advocates find convenient to ignore.
And where fathers are no longer involved in their children’s lives, the “deadbeat dad” mantra is deployed to place the blame squarely upon the fathers. For example, the USA Marriage and Religion Research Institute concludes, “society will continue to remain in a crisis until fathers accept the responsibility of caring for and protecting their children”. Perfectly good fathers who have been forcibly removed from their children’s lives by injunctions and court orders may justly react to such statements with less than complete placidity. On the other hand, some men may indeed be feckless, though if they are the second or third generation of fatherlessness, where then does the blame lie?
And then there are those who reject completely the claims that fathers bring any benefit to children at all, for example this feminist site, The Liz Library. The attempt on this site to refute the relationship between fatherlessness and adverse childhood outcomes starts with a list of famous men who did well in life but, it is claimed, were fatherless. Any such selection of individual cases is fraudulent, of course. No one is claiming that the effects are deterministic and inescapable. Such spurious arguments are essentially straw men. It did amuse me, though, that the first mentioned man was George Washington. This is odd in view of the famous, if apocryphal, story concerning Washington being unable to lie to his father about the cherry tree incident. His father died when Washington was 10. Second on the list was Thomas Jefferson. In fact, he lived with his father too, until the latter died when Jefferson was 14. We shall see below that much, perhaps most, of the benefit of father involvement relates to younger children. I confess that surprised me, but it’s what the empirical evidence suggests. So Washington and Jefferson are no refutations at all.
The sheer weight of evidence discounts any denial that fatherlessness is correlated with adverse outcomes for children, or that father involvement is correlated with improved outcomes. That is not to say that all studies report such an association: they do not. In addition, one must always be careful in social science research not to confuse correlation with causality. The correlation of fatherlessness with a whole range of serious adverse outcomes for children is clear, as we will see. But is this a causal relationship? The most obvious concern is that fatherlessness is correlated also with lower socioeconomic status, and poverty is correlated with poorer child outcomes. This is just one pathway by which a relationship between fatherlessness and adverse child outcomes might arise without direct causality. Of importance, therefore, are studies which control for relevant socioeconomic variables, or other significant variables such as parental educational attainment, before drawing conclusions about the father effect. There is no shortage of these either. Of greatest interest is the thorough review by Sara McLanahan et al, who are so bold as to claim causal connection explicitly. This paper will be reviewed in detail here.
Whilst this post concentrates upon harm to children, it is important to recall throughout that this arises as a collateral consequence of the empathy gap against men which facilitates the societal equanimity at endemic fatherlessness.
I start with this classic work largely because it has some great quotes which cut across expectations based on the authors’ affiliations. Norman Dennis was a sociologist and lifelong Labour supporter and Labour councillor in Sunderland (he died in 2010). George Erdos, a psychologist at the University of Newcastle and elsewhere, styles himself an ‘ethical socialist’. Raised in Hungary, he experienced both communism and fascism at first hand before moving to the UK. Their polemic Families Without Fatherhood (Norman Dennis and George Erdos, 1992, 3rd ed 2000) was first published in 1992, it had a second edition the following year, and a third edition in 2000. It is a long work and I will restrict myself to a few extracts because, as far as data analysis is concerned, there is an embarrassment of riches in more recent studies. I wonder if one could find a currently employed sociologist and psychologist willing to publish this today…
“One of the authors was invited to prepare a paper for a seminar based upon a very extensive study of the literature on the lone-parent family. Contrary to the near-consensus among his social policy colleagues that, in its strongest version, it was reactionary nonsense to allege that the lone-parent family was inferior to the two-parent family, he could find no study which did not show clearly that, over a whole range of outcomes, children in lone-parent families suffered disabilities as compared with the average child in the stable two-parent family. In a weaker version of the consensus it was assumed that, in so far as there were differences, these were not due to lone-parenthood as such, but merely to low income. Alternatively or additionally, they were due to the fact that the non-academic public erroneously believed that lone-parenthood was to the disadvantage of children, and this erroneous belief itself, this stigma, had the effect of creating disadvantage. There was nothing a dollar and a dose of enlightenment would not fix.
Colleagues holding this view were asked for bibliographical details of the studies that supported either the stronger version or the weaker version – at that stage fully expecting that such studies were available. On one occasion this bona fide request was met, surprisingly, with a hostile, ‘You know quite well that there are no such studies!’
The success of the attack on the family was even more astonishing when the experience of children is considered; but it was even more vulnerable when the anti-family consensus began to be challenged. For the case that the family was not deteriorating only changing, so far as children were concerned not only flew full in the face of common experience. It also flew in the face of every empirical study that had ever been published on the subject that had yielded definite results on the benefits and drawbacks for children of families with fathers as compared with those households without them.”
Dismissing the above as the rantings of an antediluvian conservative is hardly possible in view of the authors’ actual political leanings and academic subjects. Dennis & Erdos summarise a mass of data analyses (by other authors). I will confine myself to one example, thus…
“All the 17,000 children born in England, Scotland and Wales in the week 3-9 March 1958 were (and are) the subject of periodic follow-up investigations by the National Child Development Study. Of these children, 600 were born outside of marriage. A principal author of the Crellin study of these (as they were called) ‘illegitimate’ children was Dr M.L. Kellmer Pringle, well-known for her publications on child development. What were the statistical associations between, on the one hand, birth taking place without the mother being married to the father (a central prohibition of the pre-1960s system) and, on the other, various aspects of the circumstances and personal characteristics of the mother and child?
In Crellin’s population, the proportions of the children of married and uncommitted fathers were spread evenly through the social classes. Whereas uncommitted fatherhood had in the past been associated with low social class, by the later-1960s this was no longer the case. Her findings confirmed that the rapid acceleration in the rise in the frequency of uncommitted fatherhood was not the result of the emergence of an ‘underclass’ which was repudiating the values of respectable society. Attitudes, law and conduct had been transformed within a decade. No underclass has at its disposal the means of moulding public opinion to achieve such a thing. It was a change throughout society of the evaluation of what respectable behaviour was. Those changes were being effected by members of society who had access to the means for communicating what is going to be accepted as true and virtuous – the effective intelligentsia.
In remoulding attitudes and assumptions about what was factually true and ethically valid, at least as important was a part of the intelligentsia which was largely based in higher education and which made its influence felt most widely through serious newspapers and discussion programmes. Its prime commitment was to draw attention to, and remedy, the evils associated with the system of life-long monogamy. These evils included the subordinated and narrowly domestic life of the woman, the stereotyping of the roles of the son and the daughter at the cost of the latter, and of course in some cases sheer brutality and sexual abuse. Only people who were heartless, who were uncaring, would say anything that could be interpreted in any way as weakening the case for the removal of these evils.
Factual surveys like Crellin’s had shown (and continued to show for as long as they were considered worthy to be financed) that on average the life-long socially-certified monogamous family on the pre-1960s pattern was better for children than any one of a variety of alternatives practically applicable to large urban populations. These surveys undermined aspects of the arguments put by the defenders of the interests of the various problem groups. Certain of the righteousness of their respective causes, there must have been at least a few who consciously attacked what they suspected might well be true, but inconveniently and irrelevantly true.”
Allow me to paraphrase: the feminist determination to undermine the nuclear family, which no one dared contradict, was effectively propagandised by the media and the empirically clear evidence of its detrimental effects ignored. There were academics who were well aware of this 26 years ago.
The validity of the claim that there is no shortage of studies which show a significant deleterious effect on children of father absence or disengagement, or a beneficial effect of father involvement, is easily demonstrated with a few hours of Googling or reference chasing. In fact there are thousands of professional articles on the effect of fathers on child development. Not all studies identify a statistically significant disbeneficial effect of father absence. The review paper by McLanahan et al, summarised below, selected a number of studies of particular quality. About half of studies found a significant disbeneficial effect of father absence on education, whilst three-quarters of studies found a disbeneficial effect on mental health, externalizing behaviours, delinquency, substance abuse and early childbearing. The studies which did not do so merely found no statistically significant effect either way. This is significant in itself.
As far as I am aware, no studies appeared to find a beneficial effect of father absence. This is worth noting. There are many reasons why a father-effect might be hard to discern in data. For example, if the father absence occurred after a substantial period of co-residence and father involvement, does this count as fatherlessness or not? Similarly, if a study merely records divorce as the indicator, there may be no measure of the continuing involvement of a father via contact arrangements thereafter. Some studies may consider only biological father absence, without allowance for the involvement of other father figures. And even where a father is co-resident, his mere presence may not confer benefit if he is disengaged from the child. Consequently, the fact that many studies report a null result is not surprising. There will also be considerable ‘noise’ in the data. However, if those studies which report a significant disbenefit of fatherlessness were merely the result of random statistical noise, one might expect a similar number of studies spuriously reporting a beneficial effect of fatherlessness. The absence of such reverse results strongly supports the reality of the disbeneficial effects reported.
In this section I give a rather random selection of a few published works as an indication of what the literature demonstrates. I emphasise that I have merely scratched the surface of the available literature. If you can afford academic book prices, probably the best summary of the overall scene in this area is “The Role of the Father in Child Development”, 5th edition, M.E.Lamb (ed), 2010. In the following sections I shall review two studies in greater detail. I have omitted consideration of the relationship between single parent families and income or poverty, simply to keep the discussion within tolerable bounds.
The UK Fatherhood Institute summarised the position in respect of the effect of fathers on children’s educational outcomes in 2013 thus,
“Major studies across the world which follow families over time have found fathers’ involvement with their children linked with their higher educational achievement and higher educational /occupational mobility relative to their parents. For example, in the UK, fathers’ involvement with their 7 and 11 year old children is linked with their better national examination performance at age 16 and their educational attainment at age 20. This is as true for daughters as for sons, across all social classes – and whether the mother is highly involved too, or not.”
In a Finnish study, Lyytinen et al, 1998, found that the frequency of fathers’ reading to children aged 14 to 24 months was linked with children’s greater interest in books later. It is remarkable that an effect occurs at such a young age. In a low income American sample, and controlling for parent demographics and income, Ryan et al, 2006, found that the cognitive development of two and three year olds was significantly improved when either parent was supportive, and to the same extent by mothers and fathers, and that cognitive development was highest when both parents were supportive.
Eirini Flouri and Ann Buchanan, 2004, used longitudinal data from the National Child Development Study which tracks people born in Britain in 1958. 7,259 cohort members with valid data on both mother and father involvement at age 7 were used in the analysis. Father involvement and mother involvement at age 7 independently predicted educational attainment by age 20. The benefit was the same for sons and daughters. The benefit of father involvement was additive to the benefit of mother involvement. The author concluded that early father involvement is another protective factor in counteracting risk conditions that might lead to later low educational attainment levels.
Studies have found that a father’s interest in his child’s education has more influence on educational success than family background, the child’s personality or poverty. The next two studies summarised found low fatherly interest similarly predictive, in the other direction: a father’s low interest in his son’s education, for instance, reduces his boy’s chances of escaping poverty by up to 34%.
In 2007, Darcy Hango published, “Parental investment in childhood and educational qualifications: Can greater parental involvement mediate the effects of socioeconomic disadvantage?”. An extract from the Abstract is,
“Data are from the National Child Development Study; a longitudinal study of children born in Britain in 1958. Results suggest that parental involvement does matter, but it depends on when involvement and economic hardship are measured, as well as type of involvement and parent gender. Father interest in education reduces the impact of economic hardship on education the most, especially at age 11. Both father and mother interest in school at age 16 have the largest direct impact on education. The frequency of outings with mother at age 11 also has a larger direct impact on education than outings with father, however, neither compare with the reduction in the effect of economic hardship as a result of father interest in school.”
Under the auspices of the Department for Work and Pensions, during the Blair Labour Government, Jo Blanden published a report “Bucking the trend: What enables those who are disadvantaged in childhood to succeed later in life?” The analysis was carried out by using data from the British Cohort Study (BCS) of children born in 1970 and included all babies born in Great Britain between 4th and 11th April 1970. This was an initial sample of 18,000 individuals. One of the benefits of using the BCS is the richness of the dataset, meaning that not only is the dataset large, but that it was possible to control for many other variables so that different influences were not confounded. The results showed that the level of parental interest is extremely important; with father’s interest having a large influence on their sons, and mother’s interest most important for their daughters.
Blanden’s analysis was based on multivariate regressions. Her Table 5.1 gives the regression results where the predicted variable is the outcome for the child in terms of subsequent adult poverty status. Easily the largest negative beta coefficient (i.e., the largest cause of an adverse effect) is “father’s little or no interest in the child’s education”, indicating up to 34% greater chance of adult poverty for boys. Surprisingly, this analysis shows little effect of father’s lack of interest in education on the outcomes for girls. Conversely, a mother’s lack of interest in education adversely affects girls more than boys.
Blanden’s Table 5.2 gives the regression results where the predicted variable is children’s test scores at age 10. The results are qualitatively similar. In the case of boys, easily the largest negative beta coefficient is for “father’s little or no interest in the child’s education”. Again this affects girls test scores very little, whilst the gender effect is reversed for mother’s lack of interest in education. The key result is that it is the father’s interest, or lack thereof, which is the largest determinant of both the educational attainment of the son, and the son’s later financial success.
There are a number of further studies which consolidate the conclusion that high levels of interest by a father in his child’s schooling and education, his high expectations for their achievement and his greater direct involvement in their learning, education and schools, are associated with the child’s better exam results, greater progress at school, more constructive attitudes towards school, and higher educational expectations. And this isn’t just true for middle-class families: whatever the father’s education level his interest and participation pay off for his children.
Fathers’ higher levels of commitment to their child’s education, and their involvement with the school, are also associated with children’s better behaviour at school, including reduced risk of suspension or expulsion. Children’s school behaviour is strongly linked with their educational attainment and fathers’ influence on that behaviour is not only significant but may at times be more significant than mothers’ influence, (Lloyd et al, 2003).
Not surprisingly, if fathers’ positive involvement can be beneficial, fathers’ negative parenting can be destructive also. For instance, fathers’ harsh parenting is more strongly linked to children’s (especially boys’) aggression than is mothers’ harsh parenting (Chang et al, 2003).
As regards language acquisition, to quote the Fatherhood Institute again,
“While within-gender variation is enormous, and parents’ vocabulary use is far more powerfully affected by their education level than their sex, some studies suggest that fathers’ verbal interactions with their children may differ from mothers’; and that this may sometimes be to their children’s advantage. Fathers have been found to use different and longer words with their children and also more abstract words. Topics may also vary by gender, with mothers referring more frequently to emotions (this has been found to predict children’s emotional understanding) and fathers more often using causal explanatory language, which predict their children’s theory of mind.”
References supporting those contentions include Panscofar & Vernon-Feagans, 2006, “Mother and father language input to young children: contributions to later language development” and Jennifer LaBounty et al, 2008, “Mothers’ and fathers’ use of internal state talk with their young children”
High father involvement in reading, appropriate disciplining, social activities, trips out, etc., is associated with fewer child behaviour problems, lower criminality and lower substance misuse. All these behavioural benefits impact on subsequent school achievement. Children also tend to do better and behave better when they have high levels of self-respect and self-regulation. The quality and quantity of fathers’ parenting impact strongly on these and other significant measures of adjustment. Further sources of these conclusions include the Swedish study, “Fathers’ involvement and children’s developmental outcomes: a systematic review of longitudinal studies”, by Anna Sarkadi et al, 2008; the book “Fathering & Child Outcomes” by Eirini Flouri (2005); and, of course, “The Role of the Father in Child Development”, 5th edition, M.E.Lamb (ed), 2010.
The report ‘Prisoners’ childhood and family backgrounds’, based primarily on a survey of 1,435 prisoners serving sentences between 1 month and 4 years, indicated that 53% of prisoners had not lived with both natural parents as a child, Kim Williams, Vea Papadopoulou and Natalie Booth, 2012 Many had changed living arrangements during their childhood. So, whilst 47% of prisoners reported living with both natural parents, this may not have been the case for their whole childhood. 24% of prisoners stated that they lived with foster parents or in an institution, or had been taken into care, at some point when they were a child, and hence did not enjoy the benefit of the involvement of either parent full-time. To put this in context, about 2% of the general population have spent time in care as children, FullFact 2012.
The 2013 report ‘Fractured Families’, (Centre for Social Justice), stated that their extensive research “showed that 76% of children and young people in custody had an absent father and 33% an absent mother”. The Prison Reform Trust quotes the same figures.
The Abstract of the 2003 study by Bruce Ellis et al into the impact of fatherlessness on early sexual activity is as follows,
“The impact of father absence on early sexual activity and teenage pregnancy was investigated in longitudinal studies in the United States (N = 242) and New Zealand (N = 520), in which community samples of girls were followed prospectively from early in life (5 years) to approximately age 18. Greater exposure to father absence was strongly associated with elevated risk for early sexual activity and adolescent pregnancy. This elevated risk was either not explained (in the US. study) or only partly explained (in the New Zealand study) by familial, ecological, and personal disadvantages associated with father absence. After controlling for covariates, there was stronger and more consistent evidence of effects of father absence on early sexual activity and teenage pregnancy than on other behavioral or mental health problems or academic achievement.”
The 2001 study by Kaye Wellings et al “Sexual behaviour in Britain: early heterosexual experience” was based on a survey of 4762 men and 6399 women. They found that the odds of having first intercourse before age 16 was roughly doubled for children raised with only one, or neither, parent compared with children raised with both parents (x2.29 for boys, x1.65 for girls). A similar increase in odds applied to the likelihood of pregnancy before age 18.
A Swedish study reported in The Lancet in 2003, Weitoft et al, used medical and mortality data from 65,085 children with single parents and 921,257 children with two parents. Analysis of risk was conducted using Poisson regression, adjusted for factors that might be presumed to select people into single parenthood, and for other factors, mainly resulting from single parenthood, that might have affected the relation between type of parenting and risk. Children with single parents showed increased risks of psychiatric disease, suicide or suicide attempt, injury, and addiction. After adjustment for confounding factors, such as socioeconomic status and parents’ own addiction or mental illness, children in single-parent households had increased risks compared with those in two-parent households as follows,
- For psychiatric disease in childhood, relative risk for girls 2.1, and for boys 2.5;
- For suicide attempt, relative risk for girls 2.0, and for boys 2.3;
- For alcohol-related disease, relative risk for girls 2.4, and for boys 2.2;
- For narcotics-related disease, relative risk for girls 3.2, and for boys 4·0;
- Boys in single-parent families also had a raised risk of all-cause mortality.
“Does Living in a Fatherless Household Compromise Educational Success? A Comparative Study of Cognitive and Non-cognitive Skills”. I have separated out this 2017 study by Jonas Radl, Leire Salazar and Hector Cebolla-Boado because of its size, involving over a quarter of a million subjects across 33 different OECD countries.
The study addresses the relationship between various family forms and the level of cognitive and non-cognitive skills among 15- to 16-year-old students. The cognitive skills were measured using standardized numeracy test scores. This data was obtained from the Programme for International Student Assessment (PISA), release for 2012, which provides data which is consistent across nations. Non-cognitive abilities were captured by a measure of “internal locus of control”.
“Locus of control” is a term describing the belief that life events are causally attributable to one’s own actions. It has been used extensively to explain differences in effort, especially among children. According to this approach, a high degree of external locus of control is the belief that fate or luck or some external “other” is the responsible factor for one’s good or bad fortune. In contrast, a high degree of internal locus of control is the belief that the driving force of success is one’s own ability and effort. Lack of internal locus of control is believed to be related to several psychological problems including depression.
The authors of the study note that there is ample evidence on the absence of one of the parents, usually the father, being negatively correlated with educational success of students. Scholars have analysed the detrimental impact on a variety of outcomes such as test scores, grades, and attitudes about school and educational aspirations. When focusing on the final level of educational achievement and/or years of schooling attained, the negative influence exerted by father absence seems to be especially pronounced.
Radl et al’s study considered only the presence or absence of the father in the household, but not the reason for his absence. Moreover, they were not able to distinguish between biological- and stepfathers. Their analysis controlled for basic sociodemographics (students’ sex, age, and social background as measured by the mother’s level of education and migrant status). The presence or otherwise of siblings and grandparents in the same household were also controlled via independent variables.
The study found that the absence of fathers from the household is associated with adverse outcomes for children in virtually all developed countries. This was generally true for both cognitive and non-cognitive skills, although the disadvantage was notably stronger in cognitive skills.
Figure 1 shows the numerical results for the regression beta coefficient associated with father absence, the two graphs showing the effect on cognitive and non-cognitive skills. The dependent variables were normalised to unit standard deviation, so the plotted beta coefficients can be regarded as fractions of a standard deviation. The central estimate for all 33 countries is negative, for both outcomes, i.e., that father absence is detrimental. However, the effect on cognitive skills is substantially larger. For almost all countries the error band is fully within the negative range for cognitive skills. In contrast, for about two-thirds of countries, the error band for non-cognitive skills is consistent with zero effect.
To put the magnitudes in Figure 1 in context, note that the beta coefficient for the dependence of numeracy on age was 0.166, on average over the 33 countries. The average beta coefficient for the dependence of numeracy on the mother’s years of education (a surrogate for socioeconomics) was 0.07.
The authors offer a possible explanation of why fathers appear to affect cognitive skill outcomes more than non-cognitive skills,
“This finding resonates with the idea developed in psychology that noncognitive skills associated with personality traits tend to be more stable over the life course. Though not immune to the biographical shocks in the family domain that are often the cause of father absence, non-cognitive characteristics show, according to our findings, more inertia than cognitive ones.”
But this is directly in contradiction to findings in the early education literature that suggest that cognitive test scores are more difficult to change than noncognitive skills and behaviours, as do the findings of McLanahan et al, below.
On the other hand, their analysis shows that sibling presence can have a positive (or negative) effect on non-cognitive skills (though sibling presence affects cognitive skills negatively). The presence of grandparents was negative for both outcomes, though this is likely to be masking a socioeconomic effect.
As a by-product, the analysis also isolates the effect of being female on these outcomes. For the cognitive skill the beta coefficient was negative (-0.136 average over countries), but this will be because numeracy was chosen as the measure. A verbal cognitive measure could be anticipated to result in a positive (and probably even larger) beta coefficient. Of greater interest is that the beta coefficient for the dependence on being female of the non-cognitive skill was also significant and negative (-0.152). This suggests that girls are more likely than boys to suffer from lack of internal locus of control, making them more vulnerable to mental illness. The greater tendency of girls to believe their fortunes are under the control of an external “other” does sound rather like feminist credo, though one must not conclude too much from a single regression analysis.
Figure 1: Effect of father absence on cognitive and non-cognitive outcomes from Radl et al (click to enlarge)
I have selected the 2013 study by Sara McLanahan, Laura Tach and Daniel Schneider, “
The Causal Effects of Father Absence” (McLanahan, Tach and Schneider, 2013) because it is a rare example in social science of explicitly claiming a causal effect. High quality regression analyses, which control for other implicated variables, are often highly suggestive of causality, but authors rarely make an explicit claim for causality. McLanahan et al are a rare exception, and it is surely a measure of their confidence that they did so.
McLanahan et al recognise that the literature on father absence is frequently criticized for its use of cross-sectional data and methods that fail to take account of possible omitted variable bias and reverse causality. They give a careful account of the strengths and weaknesses of a range of techniques which attempt to address causal connections. The techniques they identified were:
- Various forms of regression analysis, which means fitting the dependence of outcome variables to a set of independent variables in the attempt to isolate the dependence on different factors;
- “Lagged Dependent Variable” models, which also deploy regression, but which use longitudinal data at two time points, before and after father absence, in an attempt to cancel-out unmeasured factors;
- “Growth Curve” models which use longitudinal data at more than two points in time in order to fit both before & after effects and also rate-of-change outcome data;
- “Individual Fixed Effects” models which concentrate on changes apparent over time in individual children;
- Three model models of increasing sophistication, making 7 model types in all.
McLanahan et al identified 47 articles that make use of one or more of these methods of causal inference to examine the effects of father absence on a range of outcomes for children, including educational attainment, mental health, and relationship formation and stability.
Studies of father absence effects on education were broken down into three categories. As regards effects on test scores, of 31 analyses 14 found statistically significant disbeneficial effects of father absence on test scores, whilst 17 found no significant effect. The father absence effect on general educational attainment was more emphatic. Of nine studies examining high school graduation using multiple methodologies, only one found null effects, and this study compared siblings in blended families. The effect in the broader category of attitudes, aspirations, engagement, coursework and courses failed was more mixed. Of 18 analyses, about half showed a significant disbenefit of father absence, and half did not.
Studies of father absence effects on children’s mental health and behaviours were also broken down into three categories. As regards the effect on the mental health of children when adult, five studies out of six showed a significant negative effect of parental divorce. For the impact of father absence on children’s social-emotional problems in childhood, McLanahan et al identified 27 separate analyses that examined the association between parental divorce and some type of externalizing behaviour or delinquency. Of these, 19 analyses found a significant disbeneficial effect of divorce or father absence on problem behaviour for at least one comparison group, whereas 8 found no significant association. McLanahan et al also identified six analyses that examined children’s smoking, drug and alcohol use. The evidence for this set of outcomes was very robust, with only one analysis reporting a null effect of father absence.
The effect of father absence on their offsprings’ subsequent likelihood to marry, divorce or separate was mixed and limited to a very few studies. In respect of early childbearing, McLanahan identified only two analyses that examined the effect of father absence. Both analyses found an association between parental absence and a tendency for early childbearing, with divorce in early childhood having a stronger effect than divorce in middle childhood.
McLanahan’s key conclusions were,
- Father absence negatively affects children’s social-emotional development, particularly by increasing externalizing behaviour. These effects may be more pronounced if father absence occurs during early childhood……and they may be more pronounced for boys than for girls.
- Effects on social-emotional development persist into adolescence, for which they found strong evidence that father absence increases adolescents’ risky behaviour.
- They found strong and consistent negative effects of father absence on high school graduation…. the effects on educational attainment operate by increasing problem behaviours rather than by impairing cognitive ability.
- The longer term effects of father absence included the strongest evidence for a causal effect on adult mental health, suggesting that the psychological harms of father absence experienced during childhood persist throughout the life course.
There is a huge prevalence of fatherlessness which our society regards with equanimity, and in some cases promotes.
Fatherlessness is causally related to a range of adverse outcomes for children, many of which persist into adulthood.
Consequently, our society is complicit in widespread disadvantaging of children.