To compare the diagnostic powers of rheumatoid factor (RF) and anti-cyclic citrullinated peptide (CCP) in a population selected for its high statistical relevance, over a 6-month period, an informed consent to test for anti-CCP was obtained from 1,025 consecutive patients for whom RF was ordered at a University laboratory. Within 1 year, a diagnosis was obtained without informing the physician about the anti-CCP result. Extensive statistical analyses were performed. A total of 768 patients satisfied the inclusion criteria, and 132 were classified as having RA, yielding a pre-test probability of RA of 17%. The sensitivities for anti-CCP and RF were 62 and 64% (P = 0.83), and the specificities were 97 and 90% (P < 0.001), respectively. The positive predictive value (PPV) was 79% for anti-CCP and 56% for RF (P < 0.001), whereas the negative predictive value was 92% for both. The likelihood ratio (LR) was 17.9 for anti-CCP and 6.2 for RF (P < 0.005). Forty RA patients were diagnosed with RA of less than 2 years length, and the same significant statistic differences between anti-CCP and RF were observed. Placing the results of both tests together, or using different cutoff points, increased the diagnostic utility of the tests. The anti-CCP test has statistically shown significant higher specificity, PPV, and LR for RA than the RF test in a clinically diverse population. If new criteria are to be devised to help diagnose early RA, anti-CCP should be included because it has a greater diagnostic impact than RF.