Threshold determination by triangle testing: effects of judgemental procedure, positional bias and incidental training

The sodium chloride (NaCl) detection threshold of 51 subjects was determined using a sequence of triangle tests in which the odd (NaCl) stimulus increased in concentration geometrically. Subjects served under two experimental conditions in which the judgemental procedure either was, or was not, specified. Judgemental procedure had no effect on discriminatory performance while threshold values obtained were lower than most reported in the literature. A significant positional-order bias was observed and performance improved significantly throughout the duration of the test. Reasons for the findings are discussed.