ציונים יציבים, ביטוי מתפתח: ניתוק בין יישום שיפוט הערכתי לבין רפלקציה כתובה בהערכת עמיתים

The growing accessibility of knowledge through digital technologies and generative artificial intelligence intensifies the demand for transferable competencies in higher education, most notably evaluative judgment and professional reflection.Although peer assessment augmented by structured reflection has been proposed as a mechanism for cultivating these competencies, the empirical relationship between its quantitative and qualitative dimensions remains insufficiently examined. This study investigates whether a disconnect exists between peer assessment scores and the depth of written reflection across different stages of teacher training. A cross-sectional mixed-methods design was employed to analyze379 peer assessments provided by 148 pre-service teachers at an Israeli academic college of education, drawn from a first-year required course (n=81) and a year third elective course( n=67). Students evaluated peers 'pedagogical presentations using a five-criterion rubric and composed a personal reflection responding to the prompt "What did I learn from this presentation"? Quantitative data were analyzed via a multilevel linear model, while qualitative data were examined through lexical marker analysis. Findings reveal no statistically significant differences in mean scores between groups (M=49.3 vs. M=,52.3 p>05.0), yet significant differences emerged in reflection length, elaboration, and complexity favoring third-year students (p<01.0). No significant correlation was found between reflection length and assigned scores (r≈0), indicating that the two dimensions operate independently. These findings suggest a substantive disconnect between the application of evaluative judgment and its articulation in written reflection, a distinction that extends the evaluative judgment framework (Tai et al,. 2018) and underscores the importance of integrating structured reflection from the earliest stages of professional training.