Good reasons for bad testing performance: The interactional substrate of educational exams

Abstract

This paper was originally published in 1992. It arose out of a project to study how clinicians tell parents a diagnosis of a developmental disability. That specific project was part of a larger one conducted in 1985-1986 to study the delivery of bad and good news in both ordinary and medical settings (Maynard, 2003). As part of the developmental disabilities study, we also examined how testing was done as a precedent to deciding on diagnosis (Marlaire and Maynard, 1989), and the paper here about the “interactional substrate” was meant to show the orderliness of testing interactions – the basic structures that made it possible to generate valid and reliable examination scores that could lend to the official assessment. More recently (2013-2015), a research team and I returned to the same clinic with a grant from the U.S. National Science Foundation, to study more intensively the testing and diagnosis of Autism Spectrum Disorders (ASD). In the time between the 1985 study and the recent one, the prevalence of ASD had skyrocketed in the U.S. from 1 in 5000 children to the current rate of 1 in 68. A study of the micro-interactions surrounding testing and diagnosis does not explain the increase in prevalence but it does say just how testing is done and how clinicians use results and other information to diagnose children. With regard to testing, in particular, we have come back to the paper on the interactional substrate again and again because probing this substrate and the practices by which it is constituted remains as an avenue in to understanding ASD as not just a condition of the child but as something that is manifested as a feature of social interaction. There are ways in which ASD as a child’s condition is co-produced by way of (i) the orderliness of interactions between clinicians and children, and (ii) how tests constrain both the clinician and the child in terms of what is visible as “competence.” Thus, in current work, we distinguish between what we call first-order, or concrete competence, which, by way of the interactional substrate, allows testing to be done, regardless of what the official results may be, and second-order displays of abstract competence (Maynard and Turowetz, 2016). Abstract competence involves the ability to produce general answers or ones that are shorn of embodied or other contextual orientations to questions or that involve what Donaldson (1978) has called “disembedded knowledge.” The emphasis of clinical tests on measuring second-order, abstract competence may obscure various kinds of first-order, concrete competence and “autistic intelligence” a child displays (Maynard, 2005). By doing so, testing can potentially make the child seem more impaired than he or she is, or at least suppress information that could improve performance and/or be informative for how to design home and schooling environments that enhance a child’s skills and integration into these social units.

Keywords: psychological testing, conversation analysis, disability.