Abstract This study examined validity indices of the Trendelenburg test as a measure of hip abductor muscle performance (adduction of pelvis-on-femur [P-O-F]) when identifying subjects with and without hip joint osteoarthritis (OA). Muscle performance of the hip abductor muscles was obtained in standing by using the P-O-F position measured with a goniometer and in supine using a handheld dynamometer (HHD) and a manual muscle test (MMT). We studied 20 healthy adults (10 men and 10 women) and 20 adults (10 men and 10 women) with radiographically documented hip joint OA. Indices including sensitivity, specificity, and positive likelihood ratios examined values obtained from the P-O-F position and the MMT when used to identify subjects with and without hip joint OA. Sensitivity of the P-O-F position for identifying subjects with hip joint OA was 0.55, and specificity was 0.70, yielding a positive likelihood ratio of 1.83. Sensitivity of normalized hip abductor MMT strength for identifying subjects with hip joint OA was 0.35 and specificity was 0.90, yielding a positive likelihood ratio of 3.5. Based on validity information from the present study, the Trendelenburg test (P-O-F angle) is not useful in identifying subjects in the early stages of hip joint OA.