The hypothesis in experiment 1 was, in accordance with Zajonc et. al.'s (1972) subliminal mere exposure effect, that previous subliminal presentation of novel material would have a positive influence on subsequent evaluation of that material.
The hypothesis in the follow-up study was that the subliminally presented material would be associated with the emotional state of the experimental situation. Thus a situation that evokes positive emotions should lead to positive evaluations of the subliminal material and vice-versa for negative emotions.
Experiment 1: Subjects (n=40, 30 female 10 male) where instructed to look at two neutral pictures in a tachistoscope (Ralph Gerbrands Co., Mass.: T-4B-1) for 3 minutes. While they where doing this, a novel symbol was flashed for 1 ms once every 0,49s. Immediately afterwards, subjects where instructed to rank 10 novel symbols (among these the presented one) from 1 to 10 based on their intuitive preferences (1 being best).
Follow-up experiment: New subjects (n=38, all male) went through the same procedure as in experiment 1, only twice. In the first cycle the visible masks were chosen to arouse a positive emotion (erotic pictures of women), and in the second cycle they were chosen to arouse a negative emotion (disfigured child and assault). The two cycles were separated by experiment 2 on gender specific influence, presented below.
Experiment 1: The symbol flashed (no. 7), was rated significantly higher (p=0,03) by the subjects in the experimental condition compared to controls. This means that the subjects in the experimental condition liked the symbol less than the controls. None of the other 9 symbols differed significantly between experimental and control groups (Se tables) (see graph)
Follow-up experiment: In the positive condition none of the ratings of the symbols differed significantly between the experimental and control groups (see tables)
In the negative condition the symbol flashed (no. 14) did not differ significantly between the two groups (p=0.13). Two other symbols differed significantly however (p=0.02 and p=0.03). One was rated higher (no. 16) and one lower (no. 13).(Se tables) (see positive and negative graph)
The hypothesis in experiment 1 was that the symbol flashed would be prioritised higher in the experimental condition than in the control condition. On the contrary it was prioritised significantly lower. This result conflicts with the subliminally mere exposure effect (Zajonc et al. 1972).
The follow-up study tried to determine whether the cause of this misdirection was the emotional content of the visible material. Watching neutral photos for 3 consecutive minutes would be perceived as boring by most people, thus maybe consequently making the target-symbol evoke the same negative emotion. The results show that this was not the case.
The results of the follow-up study are weak, as the study fails to even find any predictable influence by the subliminal exposure, thus hindering study of the emotional content's effect on the direction of this influence. Another explanation could be that of overexposure. 360 subliminal exposures in 3 minutes might be to many for 'mere exposure', in effect making subjects in experiment 1 dislike the presented symbol. In the follow-up this would
counteract any effect of the positive condition. The negative condition remains a puzzle, though. Further study is needed to clarify these issues.
The experiment found that decision making in novel situations can be influenced through subliminal presented material, where subjects lack proper decision criteria. Contrary to the "subliminal mere exposure" effect (Zajonc et al, 1972), previous exposure proved to lower subjects' preference for the presented material. A follow-up experiment found that this aversion was not caused by association with the emotional content of the visible mask, as classical conditioning could have predicted. The results show some inconsistencies, suggesting the phenomenon to be evasive with current methodology, requiring more sophisticated methods for further studies.