Evaluation using artificial intelligence shows post pandemic differences in oral reading fluency between Brazilian public and private school students

The investigation into differences in academic achievement between private and public school students has long been a focal point. The aim of this study was to compare oral reading fluency between private and public schools in the post-pandemic period. A total of 1296 participants were recruited from various Brazilian cities, spanning from the 2nd to the 5th grade of elementary school. Utilizing an artificial intelligence Universal Language Model from the Azure SST platform, the audio files were analyzed reflecting the commonly spoken language in Portuguese, generating data on words read correctly per minute (WRCM), the percentage of correct words over incorrect words (PCW), average consecutive correct words (CCW), and average silence time between sentences (SBS). Statistical analysis involved linear mixed models followed by pairwise comparisons. The private school segment outperformed the public school segment in WRCM in the 4th grade and in PCW in the 3rd grade. Additionally, private schools were already reaching 5th-grade levels in WRCM during the 4th grade, and this effect was not evident within public schools. No differences were found regarding SBS and CCW. These findings highlight that differences between public and private schools concerning reading fluency may manifest during the early stages of elementary school.