Results of reliability studies on physical performance tests are often conflicting. The main purpose of this study was to analyse the inter-rater reliability and between-days repeatability of physical performance tests on a female working population. Twenty-two nursing aides/ assistant nurses were tested twice, with, on average, a 5-day interval. Seventeen of these were tested again, after an average of 16 days. Isometric endurance of the neck flexors, abdominal, back and shoulder muscles, standing balance and two lifting tests (lumbar and cervical Progressive Isoinertial Lifting Evaluation {PILE} tests) were evaluated. Intra-class correlation coefficient (ICC), standard error of measurement (SEM), coefficient of variation (CV) and limits of agreement (LOA and LOA[sub ratio]) were calculated. The inter-rater reliability was high for all tests. The between-days repeatability, however, differed depending on the statistical method employed. Only one of the tests - the cervical PILE test - was considered reliable between-days. Muscular endurance tests, the lumbar PILE test and standing balance tests, as performed in this study, did not give reliable between-days results. [ABSTRACT FROM AUTHOR]