Orthopaedic special tests (OST) are commonly used in the assessment of the painful shoulder to assist to rule-in or rule-out specific pathology. A small number of tests with high levels of diagnostic accuracy have been identified but interexaminer reliability data is variable or lacking. The aim of this study was to determine the interexaminer reliability of a group of OST with demonstrated diagnostic accuracy at primary care level. Forty consecutive subjects with shoulder pain were recruited. Six tests were performed by two examiners (physiotherapists) on the same day. Tests included the active compression test, Hawkins-Kennedy test, drop-arm test, crank test, Kim test and belly-press test. 'Fair' reliability (kappa 0.36-0.38) was observed for the active compression test (labral pathology), Hawkins-Kennedy test and crank test. Prevalence of positive agreements was low for the active compression test (acromioclavicular joint), drop-arm test, Kim test and belly-press test. Prevalence and bias adjusted kappa (PABAK) values indicated 'substantial' reliability (0.65-0.78) for these tests. The active compression test (acromioclavicular joint), belly-press tests (observation and weakness), Kim test and drop-arm test demonstrate acceptable levels of interexaminer reliability in a group of patients with sub-acute and chronic shoulder conditions.