Enhancing Elementary Students' Scientific Literacy through AI-Powered Inquiry Learning: A Quasi-Experimental Study

Abstract

Scientific literacy is an essential competency in elementary science education because students need to understand scientific phenomena, engage in inquiry, and interpret evidence to make reasoned decisions. However, conventional science instruction may provide limited opportunities for elementary students to engage in sustained inquiry and receive individualized learning support. This study investigated the effect of AI-powered inquiry learning on the scientific literacy of elementary school students. A quantitative approach with a quasi-experimental, non-equivalent pretest–posttest control group design was employed. The participants were 76 third-grade students at SD Negeri Anyar 1, Indonesia, comprising 38 students in the experimental class and 38 students in the control class. The experimental group participated in AI-powered inquiry learning, in which artificial intelligence was integrated into scientific questioning, investigation, evidence interpretation, feedback, and reflection, whereas the control group received conventional science instruction. Scientific literacy was measured using a researcher-developed assessment covering students’ abilities to explain scientific phenomena, evaluate scientific inquiry, and interpret scientific evidence. Data were analyzed using descriptive statistics, Shapiro–Wilk normality testing, Levene’s test, paired-samples t-tests, independent-samples t-tests, normalized gain analysis, and Cohen’s d. The results showed that the experimental group improved from a pretest mean of 58.42 (SD = 8.17) to a posttest mean of 85.76 (SD = 6.21), whereas the control group increased from 57.89 (SD = 7.94) to 72.11 (SD = 7.28). The posttest difference between groups was statistically significant (t = 8.79, p < .001), with a large effect size (Cohen’s d = 2.03). The normalized gain was also higher in the experimental group (0.66) than in the control group (0.34). These findings indicate that integrating AI-supported scaffolding into inquiry-based science instruction can substantially enhance elementary students’ scientific literacy within the context examined.