Experimental strategies for research on multiple chemical sensitivity.

Skepticism about the validity of the multiple chemical sensitivity (MCS) syndrome stems in part from the lack of supporting experimental data. Performing the relevant experiments requires investigators to take account of broad variations in sensitivity and the need to establish reproducibility. The research approach best suited for MCS studies is the single-subject design. In contrast with conventional group designs, such designs emphasize repeated observations on individual subjects. Repeated observations of this kind constitute a time series in which successive measurements are serially or autocorrelated. One statistical method that bypasses the serial correlation problem is randomization tests. Explicit time series analyses take account of this aspect and can correct for it to determine the impact of an intervention such as a chemical exposure.