Keeping Contact with Children: Assessing the Father/Child Post-separation Relationship from the Male Perspective


The increase in separation and divorce since the early 1970s has considerably modified the relationship between fathers and children. Following a separation, children often stop living with their father and, eventually, a sizeable proportion of them lose contact with him. Yet whether the contact is maintained or not has important consequences for children's living conditions since, as various studies have shown, a father's propensity to fulfil his financial obligations towards his children appears to be closely linked to the amount of contact he has with them (Marcil-Gratton and Le Bourdais, 1999). Unearthing the factors likely to increase the frequency of father/child contact is, therefore, essential if we wish to reduce the risk of poverty to which children of separated parents are exposed. This is the primary objective of the present research. As a first step, we attempted to measure the amount of contact separated fathers have with their children; we then tried to uncover the factors likely to increase the frequency of contact. The analyses were based on a sample of 418 Canadian children (biological or adopted), aged from 0 to 17 years; these children had been declared by 291 fathers who were living apart from the child's mother when they were contacted by the General Social Survey (GSS), in 1995.

The originality of this study lies primarily in the fact that it looks at the problem from the male perspective rather than being based uniquely on information from women, as has been the case in the past; in addition, it combines information related both to fathers and to their children. For the first time in Canada, the 1995 GSS gathered information from separated fathers themselves about the amount of time they spent with their children; this allowed us to adopt a male-centred approach, and take fathers' attitudes and perceptions of the paternal role into consideration. Moreover, applying a multi-level approach to the regression analysis meant that we could correctly model the effect of characteristics measured separately for children and fathers on the amount of father/child contact; in other words, we were able to integrate into the model the fact that fathers do not necessarily have the same behaviour with all their children.

The analysis showed that almost two-thirds of children declared by their father had regular contact with him (that is, they spent five months or more with him during the year preceding the survey); at the other extreme, almost a quarter of children had little or no contact with their father (i.e. they spent less than 7 days with him in the previous year), and one child out of six had not seen him at all. In reality, the portrait may well be more sombre than that painted here. According to the mothers interviewed at the 1995 GSS, one child out of four, rather than one out of six, had not seen his or her father in the twelve months prior to the survey, and only 17 percent (rather than 30 percent) spent five months or more with him. Over and above the fact that separated mothers and fathers probably over-estimate the time they spend with children, the difference between men and women is linked to the fact that, compared with fathers who have all but lost contact with their children, fathers who see them regularly are not only more likely to be reached by surveys like the GSS, but will also declare correctly the number of children they had in the past.

The 1995 GSS also contains information on the telephone or letter contact that fathers have with children. Analyzing this data showed that this form of communication is not a substitute for visits, used by fathers living far from their children. On the contrary, the more often men see their children, the more they are likely to communicate with them by letter or telephone frequently. Moreover, among fathers who regularly have letter or telephone contact, the majority live relatively close to their child's residence.

A number of findings emerged from the cross-tabulations and the multilevel regression analysis of the factors associated with the frequency of father/child contact. First, and as we might intuitively have expected, fathers have similar behaviour with each of their children, once child characteristics (sex, age at separation etc.) are controlled for. In other words, the multi-level regression analysis demonstrated that approximately 75 percent of the observed variation in the number of days fathers and children spend together are the result of differences between the fathers, implying that differences observed between the children of the same father are relatively small.

Second, the moment at which the parent's separation occurs in the child's life seems to be a determining factor in the amount of father/child contact. Once this variable was taken into account, the observed association between the frequency of contact and the child's age at the time of the survey on the one hand and the length of time since the separation on the other disappeared, underlining the crucial role played by the child's age at separation in determining the course their future relationship with their father is likely to take. Up to the age of 5.5 years, the number of days children spent with their father grew, as the age of the child at separation increased; it more or less levelled out subsequently until pre-adolescence when the amount of contact started to rise again sharply.

Third, the geographic distance separating the separated parents' households is negatively linked to the amount of time fathers spend with their children. Thus, children who lived 50 kilometres or more from their father's home saw him much less often than those living less than 10 kilometres away, and the impact of this variable remained significant when the socio-demographic characteristics of fathers and children were controlled for. With the introduction into the equation of the fathers' attitude and perception variables, the coefficients associated with distance were considerably lower, losing their statistical significance. This implies that a part of the negative effect previously attributed to distance came from the fact that fathers living far from their children are less likely to report that they are happy to have children and to express more dissatisfaction with the time they spent with them.

It is not possible to determine here to what extent the level of happiness or satisfaction expressed by fathers is the cause or the consequence of the distance separating them from their children. One thing is clear, however. There is a close association between the attitudes and perceptions of fathers and the frequency of contact with their children, a finding that constitutes the fourth highlight of our conclusion. Thus, fathers who considered that having a child made them happier, and who claimed to be satisfied with both the custody arrangements and the time they spent with their children, were also those who spent most time with them. These results present an image of fathers committed to their children that contrasts strongly with the one often presented by the media, that of absent fathers uninvolved with their children. They also underline how important it is for parents to agree about custody arrangements if separated fathers are to maintain a special relationship with their children and fulfil their financial obligations towards them; these two behaviours go hand in hand, as recent research has shown (for a review, see Le Bourdais et al., 1998).

A fifth finding is that the frequency of father/child contact is associated with the father's age and his work pattern during the year preceding the survey. The number of days spent with children grew as the age of fathers reached by the GSS increased, up to the age of about forty years, from which point it started to decline. Moreover, fathers who work part-time spent considerably less time with their children than those working regular daytime hours on a full time basis. Somewhat surprising at first, this result reflects perhaps the fact that men with regular daytime work have fixed schedules which may match better their children's timetable than part-time workers whose hours often vary; they are also likely to have higher incomes which, as we have seen, is a factor associated with more frequent father/child contact. In this sense, our findings suggest that fathers whose income does not permit them to fulfil their financial obligations towards their children will at times cut the links with their children rather than remain in a situation that they find too uncomfortable.

Sixthly, we were surprised to discover that certain variables shown in the literature to play an important role had no significant effect on the time fathers living apart from their child's mother spent with the child. In particular, we were struck by the absence of a significant relation between the conjugal situation of parents at their child's birth and the frequency of father/child contact after separation. Likewise, we had not expected to find that forming a new conjugal union, or having a baby with this new union, had no significant influence on the time that fathers spend with children.

Several elements could explain why these variables had no significant effect. First, the small sample size is undoubtedly partly responsible for the absence of statistical significance of coefficients with relatively high values. In addition, the bias in the sample of children reported by fathers certainly plays a part. Other studies (Cooksey and Craig, 1999; Seltzer, 1991) have shown that separated fathers who were never married to or never lived with the mother of their child are more likely to have little or no contact with their children. Given the greater difficulty of reaching fathers who rarely see their children, a sizeable fraction of these men may well have been excluded from the GSS sample. This sample bias could, therefore, partly explain the absence of a significant link between the frequency of father/child contact and the conjugal situation of parents at their child's birth, or with the family trajectory followed by fathers after they separate. On this subject, it should be kept in mind that the conjugal and parental life of the majority of separated fathers does not stop at the relationship surrounding the child's birth. More than half had already entered a new union between the separation and the survey; approximately one father out of eight had lived with the children of his new partner, and a similar proportion had fathered additional children within the new relationship.

Other analyses need to be carried out if we hope to reach a fuller understanding of the process set in motion by separation. For this, access to longitudinal data that follow the same individuals through the different stages of their lives is imperative. Such data are essential if we are ever to succeed in disentangling the cohort effect from that caused by the simple passage of time in the results that we observed. Our study showed, for example, that the frequency of father/child contact varies according to the age of fathers at the time of the survey and the age of children at the separation. These findings are not based on a sample of men and children followed through time, but rather on a sample of men and children interviewed only once in 1995. How should the observed relation between the frequency of contact and the father's age at the time of the survey be interpreted? Does the fact that the frequency of contact increases with the age of fathers until around forty years, when it starts to decrease, indicate a life-cycle (or passage of time) effect, meaning that fathers lose interest in their children after a certain age. Or, is it rather a cohort effect, with men from younger generations more inclined to maintain links with their children? Similarly, one wonders whether the effect of the child's age at separation is a real effect of age, or whether children were exposed to different processes at the time of separation? The National Longitudinal Survey of Children and Youth (NLSCY), which tracks a large sample of Canadian children as they grow up, will make it possible to answer some of these questions and to clarify the course of father/child contact set in motion by parental separation. This is one of the avenues of research that we intend to pursue in the future, by exploiting data from successive waves of the NLSCY.

At the same time, this research should be enhanced by analyses based on other surveys, such as the General Social Survey on the Family to be carried out by Statistics Canada in 2001. Despite the richness of the NLSCY data, it has a major weakness when it comes to explaining why fathers remain close to their children in the event of a separation: very few fathers were interviewed in the context of the NLSCY. The "person most knowledgeable" about the child was asked to reply to the questions; more than 90 percent of the time, this person was a woman, in most cases the mother of the child. The NLSCY does not, therefore, make it possible to broach the question of father/child contact directly from the man's point of view. The 2001 GSS has great research potential in other ways. First, the majority of difficulties that we experienced in the present study (such as the problem with the paths followed by the questionnaire) should be ironed out by the next survey. In addition, the sample is much larger (by two to three times) than in 1995, and should therefore permit more sophisticated analyses than those carried out here.

Date modified: