Electrodermal activity was registered during examination with the Defe
nse Mechanism Test of 21 patients diagnosed with anxiety disorder, aff
ective disorder, or schizophrenic disorder. The test can be interprete
d as a model situation of how a person defends himself against a threa
t to avoid anxiety. We used Andersson's modified version of the test a
nd tested the hypothesis that electrodermal activity should increase w
hen there were responses categorised as Anxiety and decrease when ther
e were responses categorised as defences or when the threat was correc
tly identified. We found significant increase in all electrodermal var
iables in connection with responses categorised as Anxiety. After expo
sures with responses categorised as Isolation, the maximal skin conduc
tance level and the magnitude of late nonspecific responses were signi
ficantly decreased. After exposures when the threat was identified and
thus no longer subliminal, the electrodermal activity was significant
ly decreased. All these findings support our hypothesis. After exposur
es with responses categorised as Denial all electrodermal variables we
re significantly increased. Similarly in responses categorised as Repr
ession, Introaggression, and Disavowal or denial of hero's sex the fre
quency of late nonspecific responses were significantly increased. The
increased electrodermal activity could be due to insufficient defence
strategies as categorised in the Defense Mechanism Test.