106


 Due to the large amount of variance found for CEMG,

 ZGL, and ZGR, log transformations were conducted. The

 findings were identical to the results obtained using the

 raw data.

 In sum, there were no significant effects between the

 shock and control trials in ipsilateral corrugator EMG or

 bilateral zygomatic EMG. There was a significant group by

 block interaction for the left-sided zygomatic EMG variable.

 Further exploration of this finding, however, revealed no

 significant differences between the groups.

 Self-assessment manikin. Since the self-report ratings

of valence, arousal, and dominance were determined using a

5-point SAM ratings, nonparametric statistics were used to

examine the differences by condition and by group. Wilcoxon

Tests for paired samples were used to analyze the

differences in ratings between the shock and no-shock

trials. Kruskal-Wallis Tests were used to analyze group

differences. The ratings at time 1 and time 2 were averaged

together for the analyses.

 For all three variables, (i.e., valence, arousal, and

dominance) there was a significant difference in the ratings

for the shock and control trials. Within the shock

condition, subjects reported less pleasant feelings during

the shock compared to the shock-control trials [Z = -5.61, P

<.0001] (shock, mean=3.32, sd=1.29; shock-control,

mean=1.29, sd=.579). The main effect for arousal [Z =