Learner-reported factors associated with metacognitive awareness of reading strategies among Iranian medical sciences students: a cross-sectional study

Abstract Background Metacognitive awareness of reading strategies is relevant to how medical sciences students engage with English-language academic materials. However, less is known about its concurrent associations with learner-reported cognitive, motivational, attentional, and emotion-regulation characteristics. This study examined these associations among Iranian medical sciences students enrolled in a General English course. Methods This cross-sectional analytical study examined 380 complete records from the medical sciences subgroup of a broader two-group parent study. The reviewer-driven primary analysis retained the manuscript’s observed-variable path framework but represented the multidimensional predictors by their established scores: five MCQ-30 dimensions, nine CERQ-short dimensions, and three goal-orientation dimensions. Six linked path equations were estimated, with the three goal dimensions, general self-efficacy, mindfulness, and metacognitive awareness of reading/learning strategies treated as endogenous variables. Standardised direct effects and 5,000-resample nonparametric bootstrap confidence intervals for indirect effects were estimated. The three parallel goal-orientation disturbances were allowed to covary; no other residual covariances or modification-index-driven paths were added. Results The dimension-specific path model explained 38.4% of the variance in metacognitive awareness. Direct associations with the criterion were positive for general self-efficacy (β = 0.379), mindfulness (β = 0.207), positive beliefs about worry (β = 0.201), refocusing on planning (β = 0.195), and positive reappraisal (β = 0.133), and negative for positive refocusing (β=-0.178), cognitive confidence (β=-0.167), performance orientation (β=-0.167), and uncontrollability/danger beliefs (β=-0.127). Several total indirect effects also had bootstrap intervals excluding zero. Global fit was mixed in this low-df model: χ²(3) = 8.87, p =.031, CFI=0.998, RMSEA=0.072, SRMR=0.008, and TLI=0.838. Conclusions Re-estimating the original path framework at the dimension level resolved the interpretive problem created by multidimensional total scores. Associations within MCQ-30, CERQ-short, and goal orientation were heterogeneous and sometimes opposite in direction. The cross-sectional results support dimension-specific conditional associations and exploratory indirect-effect decompositions, not temporal or causal mediation.