Family Relationships and Children's School Achievement: Data from the National Longitudinal Survey of Children and Youth - October 1998

To reiterate, this study had four objectives: first, examine the relationships between children's school achievement and family processes; second investigate how family relationships affect children's achievement at different ages; third investigate possible differences between boys and girls in the way family relationships affect achievement; fourth, evaluate the adequacy with which the family-school relationships model represents the data in the NLSCY. Essentially, the data show that over the age range 6 to 11 years, no age differences (the second objective) emerge in the patterns of variables examined in this report nor were there any significant differences between boy and girls (the third objective). Consequently, the results can be understood to apply generally to children in the elementary school years.

With respect to the first objective, that of describing the way the selected variables from the NLSCY interact with each other to produce effects on children's school achievement, several useful points can be made. First, the generally accepted truism that SES is a very important determinant of a wide range of social and psychological functioning is strongly supported by the NLSCY data. These data are consistent with the possibility that school achievement may be affected by SES regardless of what families do to modify the conditions of learning within the home. In fact, SES effects are pervasive. Social support, not surprisingly, appears to be more available to parents at higher SES levels. Higher SES parents appear to experience lower levels of depression and to be marked with lower levels of hostile parenting with this latter finding being consistent with previous research (Edelman & Ladner, 1991; Hill, 1980). The data also show that, irrespective of processes within the family, the children in this sample from higher SES homes have more academic skills and focus as well as higher levels of achievement than children from lower SES homes. This finding of a direct relationship between social class and achievement related behavior is not unique nor new. Coleman, Campbell, Hobson, McPartland, Mood, Weinfield, and York (1966), among others, reported on this phenomenon decades ago although the full explanation for this correlation between class and achievement has been elusive. Kellaghan, et al. (1993), for example, observed that this relationship is probably due largely to the superior learning resources and experiences made available by parents of greater economic means. Moreover, such parents understand and are more experienced themselves in the schooling environment. The data reported here are in keeping with the suggestions by Kellaghan, et al. (1993), but they also imply that SES has an effect beyond what the family brings to the child. Certainly, higher SES families will probably associate more with other people also at higher SES levels. As a consequence, the child's social network is comprised of a greater number of people involved in intellectual and cultural activities and from whom the children will take, in part, their values and goals (Fordham & Ogbu, 1986; Steinberg, Dornbusch, & Brown, 1992). No doubt there are many more extra-familial variables and processes at work; the important point in the context of the present study is that it is probably unlikely that simply strengthening family processes themselves will be sufficient to overcome entirely an unfavourable social address.

A second general point is that the data bearing on family processes linked to achievement appear logical and consistent with the wider literature (with two exceptions to be discussed later). Greater levels of social support are associated with lower levels of depression and less family dysfunction. The effects here are highly likely to be bidirectional. Some parents who are in a dysfunctional family and who are depressed may overburden those around them with the result that their supports are fewer while others with strong and reliable supports feel less depressed and receive the kind of assistance needed to reduce the dysfunctionality in their families.

A further point concerning family processes is that higher levels of depression and greater family dysfunction are associated with more hostile-ineffective parenting. Again, these relationships are not unexpected. Parents burdened by depression simply do not have the energy to deal with many of the complexities of parenting and are pushed by circumstances or by their children into aversive strategies (Hops, Sherman, & Biglan, 1990; Patterson, Reid, & Dishion, 1992; Radke-Yarrow, 1990). The very fact that hostile interactions with their children are so frequent may make parents feel more depressed and contribute directly to overall family dysfunction beyond the parent-child relationship. Again, the direction of effects is probably bidirectional. Finally, hostile-ineffective parenting in the NLSCY sample is associated with reduced academic skill and learning effectiveness, a finding that is consistent with previous research (Glasgow, Dornbusch, Troyer, Steinberg, & Ritter, 1997; Lamborn, Mounts, Steinberg, & Dornbusch, 1991; Steinberg, Mounts, Lamborn, & Dornbusch, 1991). As before, the evidence, at this stage, is probably best interpreted bidirectionally. Children might indeed be less academically focused and effective because the parents are acting in hostile ways, but it is also easy to imagine that when a child is consistently poor in schooling skills, the parents might be driven to less effective parenting strategies. The final link in this chain of processes is the finding that better academic focus leads to better achievement although it is also possible for greater academic success to encourage the children to strengthen their academic capacities.

Two results appear anomalous. First, the positive relationship between the level of perceived social support for the parent and the level of hostile parenting is not readily explained. It is possible that parents who engage in hostile acts also seek out and receive assistance from others. If this is so, it is disturbing that the presence of the help may not be effective in moving the parent toward lower levels of hostile parenting. The second unexpected finding was that positive parenting does not lead either to better academic focus or to achievement as the literature on parenting styles would suggest (Lamborn, et al., 1991; Steinberg, et al., 1991). The failure to find a significant relationship among these variables in this study might be due to the nature of the positive parenting scale used in the NLSCY. The five items that make up this scale deal less with how the parent responds to the behaviours of the child and more with what sorts of associations exist between the parent and the child. Four items are specifically of that nature: How often do you play sports, hobbies, or games with him/her?; How often do you do something special with him/her that he/she enjoys?; How often do you and he/she laugh together?; How often do you and he/she talk or play with each other...?. The fifth item is different in that it asks about the nature of a responsive interaction: How often do you praise [child's name] by saying something like "Good for you!" or "That's good going!"? The chances are that if the survey had used a more conventional parenting scale, a significant association with academic focus would likely have emerged.

With respect to the fourth objective, the results indicate overall support for the model. The paths connecting the variables through the data are consistent with the presumption that indirect effects will predominate over direct effects. For the most part, the Level 5 variables are associated with Level 4 which associate with Level 3 which associate with Level 1 which associate with Level 0. Most of the exceptions to the assumption of adjacent-level-associations are limited to 'jumps' of one level only which means that, in relation to achievement, the effects are still chiefly indirect.

The major violation to the overall-indirect-effects assumption is the multilevel associations of the socioeconomic status variable. SES was directly associated with all variables except family dysfunction, positive parenting, and hyperactivity-attention deficit. This is evidence that the processes involved in the family system depicted in the model can never be considered a purely mediated system; theory must allow for a reasonable number of widely dispersed direct effects for variables like SES that are tremendously powerful in their consequences. Only replicated empirical work will be able to identify which variables are of this magnitude in influence.

In summary, two conclusions can be drawn. First, a consistent set of processes appears to operate to link family circumstances, family processes, and children's characteristics to children's school achievement. The same pattern of relationships appear for boys and girls at all ages between 6 and 11 years. The advantage of such consistency is that it simplifies the task of developing recommendations for social policy and for clinical or educational interventions. Second, the overall validity of the model was generally supported although the widespread impact of SES at many levels in the model suggests that some particularly powerful constructs may have many direct effects in addition to their indirect, mediated effects.