A-29 Incidence of Invalid ImPACT Baseline Test Results on Initial and Follow-up Assessments among English-Speaking, Spanish-Speaking, and Bilingual Test-takers

Abstract

Abstract Objective To identify the rate of ImPACT baseline test results falling below cut-offs for invalidity indicators among English-speaking, Spanish-speaking, and bilingual-speaking, individuals, upon initial assessment and on a follow-up baseline. Method Participants were a large sample of test-takers who completed annual ImPACT baseline assessments in Greater Houston, TX, as a requirement for participation in club- or school-sponsored athletics. Participants were 12–18 years old (mean 15.3, SD = 1.5), predominantly male (69%), and assigned to independent groups on the basis of primary/secondary language: English (N = 9332), English/Spanish (N = 1380), or Spanish (N = 4720). A sub-sample of each group (English N = 815, English/Spanish N = 173, Spanish N = 63) completed 2 baseline assessments (mean 1.2 years between tests, SD = 0.55 years). Results Chi-square analysis revealed a significantly greater likelihood [X2(2) = 28.05; p < .001] of obtaining an invalid baseline for individuals with Spanish as a First (6.3%) or Second (6.9%) Language as compared to individuals speaking only English (4.8%), with an overall rate of 5.6% across all three groups. Upon re-assessment, 5.7% of English, 19% of English-Spanish, and 0% of Spanish-speakers obtained a second “invalid” assessment. The Three Letters (Total Correct < 8) and Design Memory (Learning Percentage < 50) indicators were the most common contributing indicators, across all three groups. Conclusions Primary language appears to be a contributing factor in scoring below cut-offs for invalidity indicators, especially for bilingual English-Spanish speakers. The continued rate of invalidity for the English-Spanish speakers on follow baseline assessments also warrants further review. Given the retrospective nature of this study, potential influences of acculturation and language proficiency were not available and therefore, could not be assessed.