Revisiting the affective Simon effect

Abstract

We replicated the affective Simon effect found by De Houwer and Eelen (1998) in a situation in which participants had to respond by saying positive or negative depending on the grammatical category of the stimulus words while ignoring their affective connotation. Our results show that the affective Simon effect can be modulated by varying the proportion of experimental stimuli bearing a strongly polarised affective connotation. We propose that affective Simon effect depends at least in part on participants’ awareness of the correspondence between the affective connotation of the words and the responses. We also submit that this effect might not be specific to affective processing in that it is a token of a vast category of congruity effects that can be based on any kind of meaning of the stimuli, whether semantic or affective.

Acknowledgements

This work was conducted while KD was a research assistant at the National Fund for Scientific Research – Belgium, and EM a visiting student supported by an Erasmus program at the Laboratoire de Psychologie Expérimentale of the Université Libre de Bruxelles. Part of this study was financed by the Belgian Fund for Joint Basic Research (FRFC, Convention 2.4513.95).

Thanks are due to Jan De Houwer for providing his material and for his helpful comments on an earlier version of the present paper, and to Annette de Groot for helping with the testing in Amsterdam.

Notes

¹We will follow De Houwer and Eelen (1998) in using the terms of congruent and incongruent trials instead of consistent and inconsistent trials as in the terminology of Kornblum and Lee (1995).

²Also, there is a large variability in the effect. Averaged over words, the extreme values of the Simon effect were of −21 ms and +75 ms for participants. Averaged over participants, the extreme values of the effect ranged from −110 to +114 ms for words. Each word functions as its own control through the response assignment conditions, appearing for half of the participants as a congruent and for the other half as an incongruent trial.

³In the RT analysis, there were significant main effects of congruence, F(1, 19) = 4.52, MSE=1606, p<.050, and of presentation, F(1, 19) = 69.51, MSE=1338, p<.050, with a nonsignificant interaction between them. None of the effects reached significance in the analysis of the error rates. Four participants did not notice the relation between the affective connotations of the words and the responses.