I would begin by separating the tests/measures into objective and subjective data. If you then look for information on the inter and intrareliability of a specific test on the shoulder then you will perhaps have a clearer direction for searching.