The ACNM Certification Council is responsible for determining that the national certification examination in nurse-midwifery is not biased toward any protected group. Educational programs of nurse-midwifery have an interest in knowing that their administrasions criteria and educational design have validity. Policy makers have an interest in both of these issues. Nine demographic and preadmission factors were investigated to identify their ability to predict performance on the nurse-midwifery national certification examination. This retrospective analysis included 1,046 candidates who took any one of five criterion-referenced, modified essay forms of the examination between January 1988 and April 1994. Multiple-regression analysis revealed that five factors explained 16.2% of the variance in certification examination performance. These factors included age (as age increased, scores decreased slightly), ethnicity (white and Hispanic candidates received higher scores compared with other groups), years of nursing practice prior to nurse-midwifery education (a slight positive influence for each additional year), highest educational degree obtained (those whoose highest degree was the baccalaureate performed slightly better than candidates with no degree and candidates with graduate degrees), and academic level of the nurse-midwifery education program (certificate, master's, and doctoral level candidates all performed better than candidates from precertification programs certificate candidates performed slightly better than master's candidates). However, even though these variables contributed to a statistically significant model, the actual degree of predicted difference in scores was of little practical significance. Variables that were not predictive of performance included maritial status, U.S. versus foreign education either in nursing or in midwifery, academic level of the registered nursing educational program, and the highest academic degree obtained.