Statistical errors and reporting deficiencies in clinical prosthodontic publications, 2019–2024

PURPOSE: Evidence-based dentistry relies on valid statistics, yet misuse remains common. Despite increased emphasis on statistical rigor, recent prosthodontic publications remain underexplored. We hypothesized that statistical misapplications and reporting deficiencies persist in prosthodontic research. This study evaluated the appropriateness, application, and reporting of statistical tests in high-impact prosthodontic journals. METHODS: Original clinical studies published between January 2019 and December 2024 in the Journal of Prosthodontics (JOP), International Journal of Prosthodontics (IJOP), and Journal of Prosthetic Dentistry (JPD) were screened. Eligible articles used quantitative methods with inferential statistics. A 20% stratified random sample ensured representation by journal and year. Articles were assessed using a specialty-adapted checklist covering five domains: test selection, sample size justification, assumption testing, examiner calibration, and result interpretation. A weighted scoring system categorized studies as having no major, minor, or major error. Descriptive statistics and Pearson's Chi-square test, with Monte Carlo simulation as needed, assessed associations. Twenty percent of articles (n = 24) were double-scored; consensus yielded inter-rater reliability κ = 0.94. RESULTS: A total of 119 clinical studies were reviewed (JPD 57, IJOP 37, JOP 25). Appropriate statistical tests were used in 88% of studies; notably, all 19 studies that involved a statistician demonstrated correct test selection. However, only 32% reported sample-size calculations. Assumption testing was completed in 47% of studies, and examiner calibration was described in 40%. Confidence intervals were absent in 83% of reports, with only 8% both reporting and interpreting them. Overall, 45% of studies had no major statistical error, 35% contained major errors, and 21% showed minor deficiencies. Having no major error was strongly associated with proper test selection and assumption verification (p < 0.001). CONCLUSIONS: Statistical reporting deficiencies persist in prosthodontic research, particularly in sample-size justification, assumption testing, and confidence interval reporting, highlighting the need for specialty-specific checklists, enhanced training, and stronger editorial/reviewer standards.