• A reliable study therefore: • (1) should always produce the same results from the same set of data irrespective of circumstances • (2) coders in such study must agree on the units of analysis