The present study explores linguistic aspects in English reading passages from mock College Scholastic Ability Tests (CSAT) to identify differences between the 2012-2017 norm-referenced assessment and the 2018-2023 criterion-referenced assessment and differences between the CSAT June and September mock tests. The corpus was compiled of 570 reading passages from 2012 to 2023 academic years. The 2014 academic year was excluded from the analysis because test-takers were allowed to choose between two types, according to their proficiency levels only in the 2014 academic year. Coh-Metrix was used to analyze lexical aspects based on two descriptive measures, one lexical-diversity index and five word-information indices; syntactic aspects based on two descriptive measures and two syntactic-complexity indices; and textual aspects based on two readability indices, four referential-cohesion indices and one semantic-coreferentiality index. The comparison between the two grading systems indicated that the criterion-referenced assessment was significantly more difficult in both of the June and September mock tests. Moreover, from the comparison between the June and September mock tests, both of the norm- and criterion-referenced assessment showed the consistency in lexical, syntactic and textual aspects except for one syntactic-complexity index.