Training future primary teachers in historical thinking through error-based learning and learning analytics

The paper addresses the problem that history teaching often relies on lecture-based, rote memorization, which sidelines procedural and higher-order skills and disengages students. The study proposes shifting toward competency-based instruction that promotes historical thinking (e.g., source interpretation, evidence-based reasoning, critical reflection). Error-based learning (EBL), where learners detect, correct, and reflect on deliberate errors, is presented as a promising active methodology scarcely studied in history. Additionally, traditional summative assessments inadequately capture historical thinking; technology-enabled formative assessment via learning analytics (LA) could provide ongoing insights. The research examines whether combining EBL and LA improves and formatively evaluates pre-service teachers’ historical thinking, posing two objectives: (1) test differences in academic performance in historical thinking between students using sources with errors (EBL) versus without; (2) analyse whether LA-based formative assessment predicts subsequent test performance.

The theoretical framework reviews the shift from traditional, textbook-centered, rote learning in history toward approaches fostering higher-order cognitive and disciplinary skills (historical thinking). Historical thinking is framed through the New History movement and models that emphasize concepts such as cause and consequence, historical significance, change and continuity, ethical dimensions, historical perspectives, and use of sources (Seixas and Morton). Evidence suggests active methodologies (e.g., gamification, inquiry, PBL, VR, flipped classroom) improve learning and perceptions of history but often focus on external tools rather than modifying content to foster historical thinking. EBL is introduced as an active approach rooted in constructivism, where errors are treated as learning opportunities that promote metacognition, reflection, and deeper understanding; though well-studied in STEM, its application to history is scarce. The review also covers the need to modernize assessment from summative to formative; LA enables measurement, collection, and analysis of learning data to provide immediate, ongoing feedback. While LA is recognized as a key educational trend, there is limited evidence of its use in history education. Student response systems (SRS) like Quizizz and Kahoot can operationalize LA in classrooms, improving performance and engagement and offering immediate feedback, suggesting potential to assess historical thinking processes continuously.

Design: Quantitative, quasi-experimental study. Participants: 107 pre-service Primary Education teachers enrolled in Social Science II (History and its didactics) at the Faculty of Education of Albacete, University of Castilla-La Mancha; same academic year and lecturer. Groups: Control Group (CG) n=66 used sources without deliberate errors; Intervention Group (IG) n=41 used the same sources with deliberate, controlled errors. Ethical compliance followed APA standards; informed consent was obtained; study approved by relevant ethics committee. Procedure: Six 60-minute sessions. Session 1 and session 6 administered pre-test and post-test, respectively, to measure historical thinking. In sessions 2–5, students analyzed diverse historical sources (maps, texts, reports) spanning periods from prehistory to the contemporary age, organized around themes (e.g., wars, diseases). Each session included open-ended questions mapped to the six historical thinking dimensions (Seixas and Morton). In IG, sources contained deliberate errors intended to be detected and corrected during analysis. To prevent assimilation of erroneous information and to provide formative feedback, each session concluded with an interactive multiple-choice Quizizz activity aligned to the six dimensions, identical for both groups. Instruments and measures: Pre/post ad hoc tests with closed and open-ended items linked to the six historical thinking dimensions; developed and reviewed by four university lecturers. Open responses were coded using the validated assessment scale by Sáiz and Gómez (2016). LA data comprised SRS (Quizizz) scores collected after each session, used to provide continuous feedback and for predictive analysis. Data analysis: Data exported and analyzed with SPSS v24 and R. Descriptive statistics computed by condition. Inferential analyses included one-factor ANCOVA on post-test scores with pre-test as covariate (assumptions tested: normality, homogeneity of variance, homogeneity of regression slopes). Multiple linear regression assessed the predictive potential of average LA (SRS) scores and methodology (CG vs IG) on post-test outcomes among students attending at least 75% of sessions (n=104). Effect sizes: partial eta-squared for ANCOVA; Cohen’s f² for regression, with thresholds 0.02 small, 0.15 medium, 0.35 large. Alpha set at 0.05 (95% CI).

- Descriptive outcomes (students who completed both tests): • IG (n=61): Pre-test M=3.95, SD=1.20; Post-test M=6.03, SD=1.11. • CG (n=42): Pre-test M=4.16, SD=0.59; Post-test M=5.27, SD=0.99. - ANCOVA (post-test with pre-test covariate): Significant group effect, F(1,100)=13.80, p<0.001; partial η²=0.12 (medium-large effect). Pre-test covariate not significant (p=0.173). IG outperformed CG in post-test after adjustment, evidencing EBL’s positive impact on historical thinking. - Multiple linear regression (n=104): Significant model predicting post-test from methodology and average LA (SRS) scores, F(1,101)=4.72, p=0.03, Cohen’s f²=0.12 (small-medium). Coefficients: Intercept=3.96; Methodology (0=CG, 1=IG)=+0.63; LA scores (0–10 scale)=+0.22. Both predictors were significant. Local effects: Methodology f²=0.31 (almost high), LA f²=0.22 (medium-high). This indicates EBL adds ~0.63 points and each point in LA average adds ~0.22 points to the predicted post-test score. Overall: EBL significantly improves historical thinking performance; LA-derived formative assessment data predict subsequent test performance and are performance-enhancing across methodologies.

Findings support moving beyond lecture-based, memorization-focused history teaching toward active, competency-based approaches. Implementing EBL with deliberate, controlled errors improved pre-service teachers’ historical thinking, aligning with evidence that engaging with and reflecting on errors deepens understanding and promotes higher-order skills. The study demonstrates that EBL can be effectively adapted to history education to cultivate skills like interpretation, critical reflection, and evidence-based argumentation. In parallel, integrating LA via SRS provided ongoing, actionable formative feedback and predicted summative outcomes, addressing a key challenge in assessing historical thinking continuously. Together, EBL and LA align with contemporary pedagogical principles by centering students in the learning process, focusing on processes as well as outcomes, and enabling timely instructional adjustments. The results indicate practical, scalable ways to strengthen historical thinking training and assessment in pre-service teacher education.

The study shows that incorporating error-based learning into history instruction for pre-service primary teachers significantly enhances historical thinking, with a medium-large effect size after controlling for prior knowledge. It also validates learning analytics through SRS as a feasible, effective means of formative assessment that predicts later test performance in historical thinking. Combined, EBL and LA offer a robust approach to both fostering and continuously evaluating this competency in history education. Future research should examine student motivation and satisfaction with these methods, explore potential differences by variables such as gender, extend interventions across more sessions and different educational levels (e.g., primary education), and replicate with larger, more diverse samples to strengthen generalizability.

- The study focused primarily on learning gains and evaluation; it did not measure students’ interest or motivation toward EBL and LA. - Sample limited to pre-service primary teachers from a single university context; generalizability is constrained. - Short intervention (six sessions); effects over longer durations or at other educational levels were not assessed. - Some analyses involved only students who completed/attended requisite sessions, potentially limiting representativeness. - While groups used identical instruments except for the presence of errors in sources, the quasi-experimental context may still entail uncontrolled factors beyond the random assignment of group condition.