Using response time and responses of computer-administered MOCCA-College to estimate college entrance exam scores and grade point average

Abstract

Studies have found that many students are experiencing a wide range of learning difficulties (e.g., Carlson et al., 2014; McMaster et al., 2014). As students progress through higher levels of education, they transition from “learning to read to reading to learn” (Carlson et al., 2014, p. 40). Their reading and learning materials are also becoming more in-depth and complex (McMaster et al., 2012). While research has been conducted on the comprehension processes used by younger children while reading (e.g., McMaster et al., 2012), little is known about the comprehension processes used by college students. In a study conducted by ACT (2014), they found that many college students are not prepared for college-level reading or are struggling with reading comprehension (as cited in Seipel et al., 2022). This became a concern as college requires self-learning and a heavy reading load. Additionally, Clinton-Lisell and colleagues (2022) found that reading comprehension skills predict college achievement. Therefore, for college students to achieve academic success, it is essential that students are able to read and comprehend their academic texts (McMaster et al., 2014). Because reading comprehension has an impact on students' academic and ultimately professional performance, it is important to correctly identify students who may be at-risk of reading comprehension difficulties at an early stage and provide support and instructions that meet their needs.One indicator of individuals’ intelligence and educational success is their test scores. Many tests or test batteries provide more than one score, typically a total score and several subscores. In recent years, there has been an increase in interest regarding subscores, as subscores can potentially provide more detailed information about an examinee than a simple total score. However, in practice, the conventional way of creating subscores was frequently found to have little or no added value over the total score (Sinharay et al., 2007). Therefore, instead of treating subscores as the total number of items correct on subtests in a test battery or subdomains in a single test, a different approach to creating subscores must be developed. For instance, Biancarosa and colleagues (2019) used a different approach in creating subscores and found that subscores based on the types of incorrect answers are reliable, valid, and added value over and above the total number of incorrect scores. Rather than subscores based on the type of incorrect responses that students make when they make a mistake, the present research explores subscores based on speed with which a student can reach a correct answer.