Diagnostic Accuracy of Sentence Recall and Past Tense Measures for Identifying Children's Language Impairments

Abstract

Purpose Measures of linguistic processing and grammatical proficiency represent strong candidates for adaptation into language screeners for early elementary students. One key barrier, however, has been the lack of consensus around the preferred reference standard for assigning affected status. Diagnostic accuracies associated with sentence recall and past tense marking index measures were examined relative to 5 different reference standards of language impairment: receipt of language services, clinically significant levels of parental concern, low performance on language measures, a composite requiring at least 2 of these indicators, and a composite requiring convergence across all indicators. Method One thousand sixty grade K–3 students participated in school-based language screenings. All students who failed the screenings and a random sampling of those who passed were invited to participate in confirmatory assessments. The community-based sample was supplemented by a clinical sample of 58 students receiving services for language impairment. Two hundred fifty-four students participated in confirmatory testing. Examiners were naive to participants' status. Results Diagnostic accuracies for the sentence recall and past tense marking index measures varied across the different reference standards (areas under receiver operating characteristic curves: .67–.95). Higher levels of convergence occurred with reference standards based on behavioral measures. When affected status was defined by receipt of services and/or parental ratings, cases presented with higher levels of performance on the language measures than when affected status was based on behavioral criteria. Conclusion These results provide additional support for the adaptation of sentence recall and past tense marking to screen for language impairments in early elementary students. Supplemental Material https://doi.org/10.23641/asha.8285786

How We Fail Children With Developmental Language Disorder

“…Plante and Vance (1995) judge screening tests to be good if their sensitivity is at least 90% and their specificity is at least 80%. Thus, in the evaluations of the Redmond (2005) sentence recall task conducted independently by Archibald and Joanisse (2009) and Redmond et al (2019) with two different large samples, specificity was good, but sensitivity was lower than desired. That does not mean that the task is not useful.…”

“…When a standard score lower than 86 on the CELF-4 was used to identify cases and performance below the 10th percentile on the screener was considered the pass/fail cutoff, the sensitivity of the sentence recall task was .846, and the specificity was .903. Redmond et al (2019) administered the Redmond (2005) sentence recall task to 1,060 students in Grades K-3; 254 participated in more extensive follow-up testing to determine cases of DLD. When a standard score lower than 80 on the CELF-4 was used to identify cases, the sensitivity of the sentence recall task was .878, and the specificity was .887 (Redmond et al, 2019).…”

“…Redmond et al (2019) administered the Redmond (2005) sentence recall task to 1,060 students in Grades K-3; 254 participated in more extensive follow-up testing to determine cases of DLD. When a standard score lower than 80 on the CELF-4 was used to identify cases, the sensitivity of the sentence recall task was .878, and the specificity was .887 (Redmond et al, 2019). Plante and Vance (1995) judge screening tests to be good if their sensitivity is at least 90% and their specificity is at least 80%.…”

Promoting Reading Achievement in Children With Developmental Language Disorders: What Can We Learn From Research on Specific Language Impairment and Dyslexia?

“…A few recent studies have evaluated researcher-administered protocols that suggest the promise of universal screening for language. For example, Redmond et al (2019) found that individually administered screens involving sentence repetition and past-tense elicitation tasks have high sensitivity and specificity against a range of reference standards for language impairment (see also Archibald & Joanisse, 2009). Whereas Redmond et al's study focused on monolingual English speakers, Pratt et al (2020) demonstrated the utility of sentence repetition tasks as a screen for language impairment in bilingual Spanish-English students (see also Gutiérrez-Clellen et al, 2006).…”

Causal Pathways for Specific Language Impairment: Lessons From Studies of Twins

“…In fact, children with SLI follow similar growth trajectories to their age peers across different dimensions of language early in childhood, followed by a leveling of rate of growth in pre-adolescence as do their age peers, leaving the children with SLI with a lower-than-expected level of language as they transition into adulthood ( Rice, 2012 , 2019 , 2020 ; Rice & Hoffman, 2015 ). A second conclusion is that certain properties of grammar (best documented for finiteness marking in morphosyntax in languages such as English and many other languages, but certainly not all) are reliable clinical markers of SLI, meeting standards of sensitivity and specificity for identification of affected children, especially well documented for the age range of 3–9 years ( Redmond et al, 2019 ; Rice, 2019 , 2020 ; Rice & Wexler, 2001 ).…”