Rapid HIV antibody tests (RT) now permit HIV screening in settings where laboratory personnel may not be available. This study assessed the ability of 99 individuals with no laboratory experience to conduct two RT, OraQuick and Hema-Strip; these results were compared with those generated by laboratory professionals. All participants received written instructions and one-half also received a short demonstration. Error rates ranged from 2.1% to 4.6% with or without a demonstration. However, the number of invalid tests was greatly reduced when participants received a demonstration. Appropriate RT training for non-laboratorians and continued monitoring of HIV RT performance in non-laboratory settings is recommended.