谁在ACT考试中获得英语学习者测试支持？

Joann L. Moore, PhD This data byte investigates whether student characteristics predict which English learners (ELs)take the ACT®test with EL testing supports. It is a follow-up analysis to two research studiesinvestigating relationships between English-language proficiency (ELP) and performance on theACT (Moore & Schnieders, 2026; Moore et al., 2026). The two U.S. states that agreed toparticipate in this study have been administering the ACT to virtually all Grade 11 students sincebefore the 2017–2018 school year and are part of WIDA (https://wida.wisc.edu/), a consortiumof states led by the University of Wisconsin–Madison. WIDA develops resources for supportingELs and their teachers, including ACCESS, a set of summative English proficiencyassessments administered annually to K–12 students in more than 40 states. The ACCESSassessments include the four domains of Reading, Writing, Listening, and Speaking and fourcomposite scores: Oral Language (50% Listening + 50% Speaking), Literacy (50% Reading +50% Writing), Comprehension (70% Reading + 30% Listening), and Overall (35% Reading +35% Writing + 15% Listening + 15% Speaking). Each domain score and composite score isreported both as a level on a 1.0–6.0 scale and as a score on a 100–600 scale. ACT began offering EL testing supports in the fall of 2017, including (a) extended time (one andone-half time, single day), (b) translated test instructions that ACT provides, and (c) anapproved word-to-word bilingual dictionary without definitions. The supports are intended toremove construct-irrelevant variance and allow ELs to more accurately demonstrate theirknowledge and skills. All ELs are eligible to receive the supports, but not all ELs requestedthem, potentially because the ELs or their teachers and school officials were not aware of thesesupports or the benefits of using them. In two recent ACT studies (Moore & Schnieders, 2026;Moore et al., 2026), approximately 28% of the ELs in State 1 and 20% of the ELs in State 2 tookthe ACT with supports. Note that there is aprocess for requesting supports: (a) students orschool officials initiate the request, (b) school officials submit documentation to ACT that provestheir students are eligible to receive the supports, and (c) ACT approves the use of thesupports. (ACT updated this process in September 2025, described in more detail in the lastparagraph of this report.) These analyses are based on ELs who were approved to take theACT with supports, although ACT does not collect information about whether students actuallyused the supports during testing. In our analysis, we used logistic regression models to estimate which student characteristicspredicted whether a student received EL testing supports when taking the ACT. For both states,predictors included students’ ACCESS Overall score(standardized to have a mean of zero anda standard deviation of 1), free or reduced-price lunch eligibility, gender (with male as thereference group), and race/ethnicity. Race/ethnicity was classified as Black/African American,Hispanic/Latino, Asian, Native Hawaiian/other Pacific Islander, American Indian/Alaska Native,two or more races, orwhite. Because of small sample sizes in State 1, we limited the analysis of race/ethnicity to Hispanic/Latino compared with all the other racial/ethnic groups combined.For State 2, the odds of receiving supports were modeled for each racial/ethnic group withwhite as the reference group. State 2 also provided Grade 11 student transcript data, so wewere also able to include the following as predictors in the regression analysis: Grade 11 gradepoint average (GPA), number of course credits earned, and whether the student took anyadvanced coursework in Grade 11 (college-level, dual-credit, or advanced/honors versusgeneral, basic/remedial, or unspecified). Refer to the full reports for more details. In both states, ELs who tested with and without supports had similar ACCESS Englishproficiency assessment scores on average (on a 100–600 point scale), but the ELs who testedwith supports were somewhat more likely to have scores at the lowest ACCESS proficiencylevel (Level 1) compared to the ELs who tested without supports. Table 1 contains the number of ELs who took the ACT with and without testing supports,average ACCESS Overall scores, and percentages of students in each demographic categoryfor State 1. In this state, ELs who took the ACT with supports were somewhat more likely to befemale than ELs who took the ACT without supports (52% vs. 38%), more likely to bewhitethan ELs who took the ACT without supports (14% vs. 7%), and less likely to be eligible for freeor reduced-price lunch than ELs who took the ACT without supports (46% vs. 53%). Table 2contains the number ofGrade 11ELs who took the ACT with and withouttestingsupports, average ACCESS Overall scores, percentages of students in each demographiccategory, high school grades, credit hours earned, and advanced coursework taken for State 2.Demographi