The reliability of methods of testing nerve function is important, since diagnostic decision making is a direct function of the quality of the test. Three methods of nerve function testing were investigated at the Danish Bangladesh Leprosy Mission (DBLM) in north Bangladesh, and assessed for inter-observer reliability. The three methods were 1) ballpoint pen test (BPT) for sensory function; 2) graded Semmes Weinstein monofilament test (SWM) for sensory function and 3) voluntary muscle testing (VMT) for motor function. The weighted kappa (kappa w) statistic was used to express inter-observer reliability. Using this statistic, 0 represents agreement no better than random, and 1.0 complete agreement. kappa w values of > or = 0.80 are reckoned to be adequate for monitoring and research. Fifty-three patients were tested, a Senior physiotechnician acting as 'gold standard' against whom four other staff physiotechnicians were assessed. All three testing methods were found to have minimal inter-observer variation, with the kappa w for inter-observer agreement using BPT being 0.86, the SWM 0.92, and VMT 0.94. It is concluded that in trained and experienced hands, all three methods are reliable and repeatable to a level allowing confident use of results obtained in monitoring and research.