The authors evaluated the value of the Cumulative Grade Point Average (GPAX), letters of recommendation and personal interview in predicting performance of first year residents at the Department of Pediatrics, Faculty of Medicine, Chulalongkorn University. The pre-residency data of 25 first year pediatric residents were compared to the results of the Pediatric In-training Examination (PIE) and clinical evaluation by faculty during the first 7 months of training. The Clinical Performance Rating Score (CPRS) developed by the Royal College of Pediatricians of Thailand was used for clinical performance evaluation. GPAX correlated strongly with PIE score (r = 0.69, p < 0.001) while letters or recommendation and interview score were not predictive. For the majority of residents who graduated from Chulalongkorn University, including the undergraduate program into the regression model improved this association (r = 0.90, p < 0.001). Interview score correlated moderately (r = 0.49, p = 0.02) with clinical performance measured by CPRS which entailed the evaluator to assess a resident in 6 specific areas. GPAX and letters of recommendation showed a trend toward positive correlation with clinical performance, but these associations were weak (r = 0.32-0.39). Further analysis of letters of recommendation found a significant difference between the score given by evaluators in community hospital settings compared to the score given by medical school faculty or evaluators in large regional or provincial hospitals. The authors conclude that the cognitive function of pediatric resident candidates can be assessed reliably by GPAX. Research on how to adjust for the variability of GPAX given by an individual undergraduate program will improve the residency selection process. A structured interview may correlate better with global clinical performance than GPAX or letters of recommendation but all methods are at best moderately predictive of clinical performance. Improvement can be made by continuous evaluation and adjustment of the selection process.