Bayesian construct validation leveraging expert knowledge for questionnaire instruments used in primary care research and practice
Bibliographic record
Abstract
BACKGROUND: Questionnaires are widely used instruments for acquiring information on latent traits, perceptions or self-reported attributes of individuals in various practical fields and research domains including education, psychology, sociology and medicine. The development of valid and reliable questionnaire instruments is a labor-intensive process requiring iterative expert input and empirical assessment of the psychometric properties of the instrument in the target population. Bayesian methods enable the incorporation of domain expert knowledge and can increase the efficiency of the development and construct validation process, potentially saving resources, time and costs. Despite numerous methodological developments in the statistical literature, Bayesian methods for questionnaire development are still underutilized in the primary care context. This is a critical gap that likely affects both practice and research in the field, as questionnaires are important instruments for day-by-day clinical decision making and research data acquisition. OBJECTIVE: The overall objective of this Ph.D. research project was to develop an effective and feasible Bayesian inference framework for questionnaire construct validation. The developed framework employs a survey approach for eliciting domain expert input to inform the required Bayesian prior distributions. METHODS: A systematic methodological review of the literature was conducted to examine the use of Bayesian methods for construct validation in the primary care context. Informed by the findings of the review, a Bayesian inference framework was developed, aiming to overcome feasibility issues of currently available methods described in the literature. The performance of the developed inference approach was assessed in comparison to standard validation approaches using an extensive Monte-Carlo simulation study. Finally, to illustrate its performance using real-world data, the developed framework was applied for the construct validation of a recently developed instrument, the McGill Empowerment Assessment – Diabetes (MEA-D) questionnaire, measuring levels of self-care in diabetes patients. RESULTS: The systematic literature review revealed that Bayesian construct validation methods are underutilized in questionnaire development studies in the primary care literature and identified prevalent shortcomings in the justification, reporting and interpretation of the respectively applied statistical validation approaches. The assessment of the newly developed Bayesian validation framework for leveraging domain expert knowledge demonstrated sound performance even under mild misspecification of expert priors. Applying the developed framework for construct validation of the MEA-D questionnaire demonstrated feasibility and consistency with the results of the standard empirical validation, yielding higher precision in estimated factor loadings.CONCLUSION: The developed Bayesian framework for leveraging domain-expert knowledge in construct validation studies enables a more inclusive and potentially more efficient (resource-saving) approach for the development of questionnaire instruments, contributing to more equitable evidence-based practice and research in primary care and beyond
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.256 | 0.547 |
| Meta-epidemiology (narrow) | 0.002 | 0.002 |
| Meta-epidemiology (broad) | 0.003 | 0.003 |
| Bibliometrics | 0.008 | 0.005 |
| Science and technology studies | 0.002 | 0.005 |
| Scholarly communication | 0.005 | 0.006 |
| Open science | 0.004 | 0.008 |
| Research integrity | 0.004 | 0.005 |
| Insufficient payload (model declined to judge) | 0.005 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; the direct Gemma label and the distilled Codex classifier agree on what is shown here.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".