Psychometric and Cross-Cultural Validation of the Clinical Practice Environment Scale (EAPRAC) for the Evaluation of Clinical Rotations in Latin America
Description
This dataset comes from a multicenter study designed to evaluate the psychometric validity and cross-cultural applicability of the Clinical Practice Environment Scale (EAPRAC) for assessing the teaching and learning environment during clinical rotations in Latin America. The study responds to the need to systematically monitor learning climates in residency and clinical training programs as part of educational quality assurance. EAPRAC was created to provide a practical and comparable measure of key factors that shape workplace-based learning in clinical settings, supporting training programs in identifying strengths, detecting areas for improvement, and guiding interventions that enhance both educational processes and clinical training experiences. EAPRAC is a Likert-type instrument with 39 items organized into seven domains: academic processes, clinical teachers, teaching–service agreements, well-being, academic infrastructure, clinical infrastructure, and organization and management. The repository may include de-identified item-level responses for the 39 items, and may also include derived domain scores and contextual variables describing the training setting and rotation characteristics, depending on what is shared by participating sites. The dataset is intended for psychometric analyses such as factor structure assessment and reliability evaluation, as well as for exploratory comparisons across sites, countries, training levels, or rotation types when applicable. The study uses a cross-sectional validation design implemented across multiple Latin American clinical education contexts. Sample size planning is oriented to support factor analysis with adequate respondent-to-item ratios. Analyses may include exploratory factor analysis to examine dimensionality, confirmatory factor analysis to test the proposed structure, and internal consistency estimates to evaluate reliability. All shared data are anonymized, with identifiers removed and replaced by codes; any linkage information is retained only by the primary research team. This repository is intended to facilitate reuse in medical education research and learning environment evaluation, including methodological work on scale validation and cross-context comparisons, with interpretation supported by the instrument structure and variable definitions provided in accompanying documentation (e.g., codebook/data dictionary, if included).
Files
Steps to reproduce
Download files from this repository, including: (a) the raw dataset (item-level responses), (b) the codebook/data dictionary, and (c) any analysis scripts or outputs (if provided). Verify dataset structure using the codebook: confirm the 39 EAPRAC items, domain membership, response scale coding (e.g., 1–5 or 1–7), missing value codes, and any reversed items. Clean the data: remove duplicates, check invalid values outside the Likert range, and recode missing values to standard NA format. Apply scoring rules: reverse-code the specified items (if any), then compute domain scores and (optionally) a total score exactly as described in the codebook (e.g., sum or mean). Descriptive checks: report response distributions per item, missingness per item/domain, and basic participant/site descriptors (if included). Reliability analysis: estimate internal consistency for each domain and the full scale (e.g., Cronbach’s alpha and/or omega), and inspect item–total correlations and “alpha if item deleted.” Exploratory Factor Analysis (EFA): run EFA on the item matrix (preferably with a polychoric correlation matrix for ordinal Likert items), decide the number of factors (e.g., parallel analysis), and document rotation and extraction methods. Confirmatory Factor Analysis (CFA): specify the proposed 7-domain measurement model and fit it using an estimator appropriate for ordinal data (e.g., WLSMV). Report standard fit indices (e.g., CFI, TLI, RMSEA, SRMR) and factor loadings. (Optional) Subgroup/measurement checks: if site/country/training-level variables exist, reproduce subgroup analyses (e.g., factor structure comparisons, invariance testing) following the same model specification and estimator. Reproduce tables/figures: regenerate all summary tables and model outputs using the same scoring and analysis choices, then compare results to the repository’s shared outputs (if included).