Cirino, P. T., Ahmed, Y., Miciak, J., Taylor, W. P., Gerst, E. H., & Barnes, M. A. (in press). A framework for executive function in the late elementary years. Neuropsychology. doi:10.1037/neu0000427
In this study, Cirino and colleagues (in press) describe the structure of executive function (EF), which they define as a domain-general control process important for managing goal-directed behavior. EF as a construct is widely represented in a number of literatures (including education-related literature, where it is often referred to as self-regulation), but it is notoriously hard to operationalize. The authors find that a bifactor model (with a common EF and five specific factors) is most parsimonious, and they discuss ways that such an understanding might affect the way EF is considered in relation to relevant outcomes, such as achievement.
EF is conceptualized in multiple ways across different literatures. The term EF is prominent in the neuropsychology literature, where it has traditionally been related to the frontal lobes of the brain. The cognitive literature on working memory shows much overlap with EF. Developmental and educational literatures often use the term self-regulation to describe a process similar to that of EF. The ways that EF and its related concepts are “operationalized,” or measured, are rather diverse. For example, complex clinical tasks (e.g., the Tower of London test), specific cognitive tasks (e.g., inhibition of a prepotent response, maintaining information in working memory while performing a secondary task, shifting between one set of rules, generating exemplars in response to a prompt), and rating scales (e.g., the BRIEF or scales of self-regulation) have each been considered to measure EF. Further, EF can be measured with tasks that differ in content or format (e.g., verbal versus nonverbal, behavioral versus performance, timed versus untimed, paper-pencil versus computer).
Three important ideas helped shape the current investigation. First, because EF is assessed in many ways, it is important to represent this diversity in measurement by selecting tasks across literatures, content, and formats. Second, what is known about the structure of EF comes predominantly from studies with young children (e.g., preschool) or with adults (e.g., college students). Interestingly, studies with young children seem to show a “unitary” EF (i.e., all the tasks appear to measure a single underlying thing), whereas studies with adults seem to show separability of factors of updating or working memory, inhibition, and shifting. There are many fewer studies with students in the between ages, such as late elementary, which is a particularly interesting age range to study because school curricula are evolving (from mastery of basic content to application) and students are undergoing developmental changes in the brain. Third, when EF is discussed, it is often to predict where students may have difficulty. For example, EF difficulties have been described in struggling readers and in students of lower socioeconomic status, making it relevant to understand the way EF is structured in such populations in particular. In addition, understanding the way that EF relates to basic information such as age and sex and economic disadvantage provides further context.
Students were in grades 3 through 5 were included in the study. The total number of students was large (N = 846), and the sample was diverse, including many fourth-grade struggling readers (because the sample was connected to a reading intervention study). Twenty-seven measures were selected in eight domains that have been associated with EF, including working memory, inhibition, shifting, planning, generative fluency, self-regulated learning, metacognition, and behavioral regulation.
The study was complex for three reasons. First, the large number of measures could not be administered to all students. Therefore, the authors used a “planned missingness” approach, which allowed for an estimate of how each measure related to all the other measures. Second, because the number of struggling readers was high, the sample was “weighted” to account for this over-representativeness. Third, the analytic method was factor analysis, which specifically allows for an estimation of how individual measures represent different domains (factors) and whether and how much these domains relate to one another. The study specifically used what is called a “bifactor” model to address the extent to which individual measures represented a general or common index of EF, versus a more specific index of EF.
The study tested whether the eight domains described above could separate from one another (it was expected that they would) and whether a bifactor representation would produce a better “match” (fit) to the data. It was also expected that EF would relate in predictable fashion with factors such as age (e.g., a positive relationship) and economic disadvantage (e.g., a negative relationship).
This study supported some of the hypotheses made, but not all of them. The bifactor structure hypothesis was supported, and there were indeed predictable relationships with demographic variables (e.g., positive with age, negative with economic disadvantage). However, the eight domains originally hypothesized were not clearly separable from one another, and instead only a reduced version of these domains remained. A reduced set of factors did fit the data well in a bifactor framework. In addition to a broad or common/general factor of EF, the study identified five separable factors (two associated with working memory and one with fluency, self-regulated learning, and metacognition).
The authors found that individual measures of EF tended to correlate poorly with one another (on average, r = .15, which is considered low), consistent with similar studies at different age ranges. The findings indicate that the individual measures are not interchangeable with one another and that when discussing EF, it is critical to differentiate between the way a specific task relates to a given outcome (such as reading or math) versus the way that the more general construct of EF relates to those same outcomes. The findings also imply that it would be dangerous to apply all of the conceptual aspects of EF to any individual (or collection of) EF measure(s). What the individual measures have in common is only a portion of any given test—the whole may be greater than the sum of its parts, but the parts contain elements that may add to or be synergistic with common EF when used to predict relevant outcomes (such as reading and/or math).
The present study related many known measures of EF to one another in a large sample of late elementary students with a significant proportion of struggling readers. A common EF factor was clearly apparent from the data, with more specific, separable components of working memory, fluency, self-regulation, and metacognition also found. The study is important because understanding the way that EF is structured can serve to help other studies operationalize EF more effectively and in a consistent manner that recognizes both the commonality that runs through measures of EF and the separability of its components. Future studies will need to evaluate the relative contributions of these components.
The following are recommendations for educators:
Gerst, E. H., Cirino, P. T., Fletcher, J. M., & Yoshida, H. (2017). Cognitive and behavioral rating measures of executive function as predictors of academic outcomes in children. Child Neuropsychology, 23(4), 381–407. doi:10.1080/09297049.2015.1120860
This recent study relates different types of measures of EF to reading and math outcomes.
Hofmann, W., Schmeichel, B. J., & Baddeley, A. D. (2012). Executive functions and self-regulation. Trends in Cognitive Sciences, 16(3), 174–180. doi:10.1016/j.tics.2012.01.006
This study shows how different conceptualizations of EF may be related to one another.
Lee, K., Bull, R., & Ho, R. M. (2013). Developmental changes in executive functioning. Child Development, 84(6), 1933–1953. doi:10.1111/cdev.12096
This is one of the few longitudinal studies of the structure of EF for children.
Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., Howerter, A., & Wager, T. D. (2000). The unity and diversity of executive functions and their contributions to complex “frontal lobe” tasks: A latent variable analysis. Cognitive Psychology, 41(1), 49–100. doi:10.1006/cogp.1999.0734
This oft-cited study demonstrates three factors of updating, shifting, and inhibition in college students.
Vaughn, S., Solís, M., Miciak, J., Taylor, W. P., & Fletcher, J. M. (2016). Effects from a randomized control trial comparing research and school-implemented treatments with fourth graders with significant reading difficulties. Journal of Research on Educational Effectiveness, 9, 23–44. doi:10.1080/19345747.2015.1126386
This report describes the results of the intervention study with which the present sample overlaps.
Wiebe, S. A., Espy, K. A., & Charak, D. (2008). Using confirmatory factor analysis to understand executive control in preschool children: I. Latent structure. Developmental Psychology, 44(2), 575–587. doi:10.1037/0012-16184.108.40.2065
This study demonstrates that a unitary executive control (similar to EF) factor is most parsimonious in preschool children.