An Original Comfort/Discomfort Quantification in a Bi-Variate Controlled Experiment: Application to the Discomfort Evaluation of Seated Arm Reach 2006-01-2347

Each time the subject’s response in surveys is recorded with an ordered category scales, like annoyance in noise exposures or discomfort in ergonomics, one is interested by the numerisation in order to develop and to get a numerical response’s law. A meaningful and suitable method is the “successive intervals” of Adams and Messick, originated from Thurstone psychophysical ideas, and revisited following the general and performing measurement theory (Maurin 2003a,b). In case of one objective physical variable (or geometrical), one gets a contingency table crossing the response categories in columns and some controlled values of variables in rows. Then we get a uni-variate numerical law, as for example, comfort’s law of a pedal operation in terms of the seat height, pedal resistance or pedal travel (Wang & Maurin).

Here we generalise the successive interval for bi-variate situations. The same methodology provides the meaningful numerisation of responses in case of two input objective variables. In addition to this conjoint law, some statistical fine properties make it possible to explicit the interaction between variables. We apply this to analyse the categorical discomfort ratings of arm reach motions in relation to two variables, such as target’s orientation and distance, or target’s orientation and seat’s height (Chevalot & Wang).

The bi-variate results clearly show interesting features in conjoint law and interaction, which may probably not be so well revealed using a more crude quantitative method.