EDA was recorded from one group of subjects by using a standard constant-voltage method and from another group by using polarizable electrodes. a hypertonic electrolyte. and a constantcurrent method with ac coupled output. The robustness of stimulus-response relationships for EDA was examined by comparing the groups over (counterbalanced) blocks of 10 auditory and 10 visual stimuli of low intensity. Both raw and range-corrected EDA measures showed significant novelty and modality effects without group differences. and high between-group correlations over means (r ~ .87). These results indicate that SRR measures from field-type studies may validly represent EDA expected from a more rigorous application of SeR techniques. Respiratory pause was also examined in both groups and found to reflect only novelty changes. This lack of covariation with EDA is contrary to unitary OR theory but compatible with recent proposals of preliminary processes in OR elicitation.Phasic changes in electrodermal activity (EDA) may be derived from two different systems. The older system, employing a constant current, leads directly to the skin resistance response (SRR), while the newer system, utilizing a constant voltage, directly generates the skin conductance response (SCR). The transition between these methods, given initial impetus by the theoretical and technical arguments of Edelberg (1967Edelberg ( , 1972 and Lykken and Venables (1971), remains incomplete, and studies presenting results derived from the different methods are still appearing in the literature.The question of the comparability of such results is of continuing interest, as reflected in a recent study by Boucsein and Hoffmann (1979). These authors made a direct comparison of simultaneous SCR and SRR measures obtained (by constant-voltage and constant-current devices, respectively) from a group of subjects in response to white-noise stimuli ranging in intensity from 60 to 110 dB. They found no difference in raw amplitudes (SCRs were converted to SRR units for this comparison), but reported a significant difference between range-corrected amplitudes obtained following the Lykken and Venables (1971) procedure. This difference in range-corrected amplitudes was most surprising, particularly since there was no difference in raw scores, and it is deserving of further investigation. One variable that might be implicated in such a result is the relatively high range of stimulus intensity used in their study, so it is necessary to investigate comparability of these measures with lower intensities such as those . com-