Claessens, A., & Dowsett, C. (2014). Growth and change in attention problems, disruptive behavior, and achievement from kindergarten to fifth grade. Psychological Science. Advance online publication. doi:10.1177/0956797614554265
In this manuscript, Claessens and Dowsett (2014) examine relations among achievement (in both reading and math), attention, and behavioral problems. We know from prior research that there are correlations among these domains, but this study provides important new contributions. First, this study reported data over multiple time points for each domain from kindergarten through fifth grade. Second, Claessens and Dowsett were interested in what we might call the "chicken versus egg" problem—that is, whether change in one area precedes or follows change in the other area(s). Specifically, they wanted to distinguish from among four potential alternatives:
(a) Early problems with attention and behavior lead to later achievement problems
(b) Early problems with achievement lead to later attention and behavior problems
(c) Both (a) and (b) are true—the relation goes both ways
(d) The relations are due to some other factor like demographics or general knowledge
A major finding was that improvement in attention problems (but not behavior problems) during kindergarten led to subsequent improvement in math and reading achievement. The authors suggest that leveraging interventions that promote improvement in attention skills could hold benefit for later achievement.
The data came from the Early Childhood Longitudinal Study-Kindergarten and included more than 16,000 children from kindergarten to middle school from about 1998 to 2004. Using data from such a large study has some advantages. For example, the size allows researchers to detect even small effects and to control for a number of factors that could get in the way of interpretation (whether they were actually measured or not). Following students over time allows researchers to identify how changes in one factor relate to changes in another. It should be noted that analysis of large data sets is complex and requires sophisticated statistical analysis that can be interesting if interpreted carefully.
The present study measured all of the targets (reading, math, attention problems, and disruptive behavior) at five time points: beginning of kindergarten; end of kindergarten; and end of first, third, and fifth grades. These time points are all developmentally important. For example, the end of kindergarten is often the first year of formal instruction for many children. The end of third grade is often when many foundational academic achievement concepts (e.g., word reading, math facts) have been taught. At the end of fifth grade, educational requirements are more complex and require more student independence. The sample was equal in terms of sex and somewhat diverse (e.g., 57% white, 14% black; 17% Hispanic).
Reading and math were assessed by using single measures developed through a technique called item response theory, which allows differences in scores over time to be more directly compared. Attention problems were measured by using two items from the Social Rating Scale (an item for attention and another for task persistence). Disruptive behavior was assessed with four items indicating how often the child does certain things (fights, argues, gets angry, and disturbs activities). Teachers rated all items for attention and behavior on a 4-point scale (from never to very often).
The analytic technique used to evaluate the data was piecewise hierarchical linear modeling. This statistical model of growth evaluates time in specific groupings—which in this study included from the beginning to the end of kindergarten or from the end of third grade to the end of fifth grade. The analyses are complex, but what is important is that the technique allows researchers to conclude that change at some earlier time point (e.g., in attention) may invoke some later change (e.g., in achievement). The authors included statistical controls for things like sex, age, family characteristics like socioeconomic status, and general knowledge, and focused on change rather than level of performance. This part of the analysis helped determine whether the relations were real (and not "spurious," or due to other factors).
Do early attention and behavior problems lead to later achievement problems? (Table 3 in paper) The authors reported that an increase in attention problems during kindergarten leads to decreases in reading and math achievement not only during kindergarten, but also through third grade. Kindergarten attention levels were not related to achievement gains from spring of third grade through fifth grade, but later increases in attention problems were related to decreases in achievement during this time period. Put another way, improving attention during kindergarten could lead to later gains in achievement, which makes sense, given the focus on foundational skills in the early grades. Later increases in achievement appear to require attention gains closer in time. Finally, the authors reported that disruptive behavior was not consistently related to achievement, even over the same time periods.
Do early problems with achievement lead to later attention and behavior problems? (Table 4 in paper) Improvement in reading and math led to decreases in attention problems within any of the time points from fall of kindergarten to spring of fifth grade. But in math, gains in the preceding testing period were related to decreases in classroom attention problems in the following periods in kindergarten through fifth grade. For example, when math achievement gains were evident in first to third grades, there were fewer attention problems in the spring of third grade through fifth grade. The authors reported inconclusive findings related to whether early reading gains influenced later problems with attention.
Improvements in reading and math also led to decreases in disruptive behavior within the later time periods (e.g., spring of first grade to spring of third grade, spring of third grade to spring of fifth grade). But earlier achievement gains were not generally related to later changes in behavior problems (though math improvement in kindergarten led to less disruptive behavior from kindergarten to first grade).
Does the relation between achievement and attention and behavior go both ways? In math, it does seem that the relation between gains in achievement and attention goes both ways. Not only do increases in attention skills relate to later achievement, but likewise, achievement gains in earlier grades lead to increased attention later on. For reading, the relations were stronger in going from attention gains to reading gains than the reverse direction. In general, the relations between changes in disruptive behaviors and changes in achievement from one time point to the next were weak.
The overall pattern indicated the strongest support for Claessens and Dowsett's first hypothesis—that positive changes in attention during kindergarten led to improvements in reading and math through third grade. Early improvements in math also led to later improvements in attention. Therefore, some of the relations between achievement and attention were bidirectional, but moreso for math than for reading. Disruptive behavior and achievement showed weak relations, moreso when behavior was the outcome than when it was the predictor, and generally for predictions over the same time period. The results were generally consistent even though the authors controlled for a number of child and family factors that could explain relations between achievement and attention and behavior.
These findings may be used to inform professional development for early childhood teachers. Because it seems that students who exhibit greater attention during kindergarten experience improvements in reading in math through third grade, early childhood educators should be trained in instructional techniques (e.g., providing effective feedback) to maximize student attention in the early grades.
Claessens and Dowsett encourage researchers to design interventions to increase attention skills and measure their effects on academic outcomes. The most convincing demonstration of such an intervention should be conducted via randomized controlled experiments. They also encourage the replication of their findings with special populations, including struggling readers and students with disabilities.
Although the study was quite large and nationally representative and the analyses were carefully conducted, teachers and educational staff members evaluating these results might keep in mind two particular caveats. First, many large, urban school districts might not fit the overall profile of the sample considered here. For example, only 17% of students in this study were reported to be of Hispanic ethnicity. A school district with large numbers of Hispanic English language learners may want to interpret these findings with caution. Second, although the measurement properties of the academic achievement measures were very strong, they did not focus on component skills (e.g., reading decoding versus comprehension, math facts versus word problems). Despite these caveats, this study provides key data about the relation and direction between attention problems and achievement and provides other researchers and practitioners with important information and questions for further studies to address.