1. Introduction

Evolutionary psychologists view jealousy as a psychological mechanism that recurrently solved an essential problem of individual reproduction in our evolutionary history: infidelity in reproductive relationships (Buss et al., 1992, Cosmides & Tooby, 1994, Daly et al., 1982, Symons, 1979). A distinctive feature of the evolutionary view is the assumption of a sex-specific jealousy mechanism (JM) because men's and women's reproductive success has been recurrently threatened by different types of infidelity. Specifically, a woman's sexual infidelity deprives her mate of a reproductive opportunity and may burden him with years of investment in a genetically unrelated child. In contrast, a man's infidelity does not burden his mate with unrelated children, but it may divert resources away from his mate's progeny. This resource threat may be signaled by his level of emotional attachment to the other female. As a consequence, men's JM is hypothesized to be particularly activated by a mate's sexual infidelity, whereas women's JM is hypothesized to be particularly activated by a mate's emotional infidelity.

The evolutionary view of a sex-specific JM spawned an impressive body of research during the past decade that has been primarily devoted to testing the hypothesis that the female JM responds with stronger emotions to a mate's emotional infidelity, whereas the male JM generates stronger emotions in response to a mate's sexual infidelity (see Harris, 2003, for a critical review; and Hofhansl, Vitouch, & Voracek, 2004, for a more recent and complete meta-analysis that supports the evolutionary view).

The present research focuses on the hypothesis derived from the evolutionary view of jealousy that the JM is a sex- and content-specific information-processing device (Schützwohl & Koch, 2004). More precisely, the male JM is hypothesized to preferentially process cues signaling a mate's sexual infidelity, whereas women's JM is predicted to preferentially process cues signaling a mate's emotional infidelity. This hypothesis has received empirical support in two recent studies. Schützwohl and Koch (2004) showed that men were better able to recall cues signaling their mates' sexual infidelity, whereas women were better able to recall cues signaling their mates' emotional infidelity. Additionally, Schützwohl (2004a) found that decisions in a forced-choice response format in favor of the evolutionary primary infidelity type (i.e., female sexual and male emotional infidelity) were made significantly faster than decisions in favor of the adaptively secondary infidelity type (i.e., female emotional and male sexual infidelity). The present study extends this line of research with respect to two aspects of the preferential processing of cues to the adaptively primary infidelity type. First, the present study examines the number of cues to sexual or emotional infidelity that men and women need to process until the first sign of feeling jealous is elicited and until the jealousy feeling becomes intolerable. Second, it investigates the efficiency of the processing of the cues indicating these two jealousy thresholds.

More specifically, depending on the condition, men and women were successively presented with cues signaling either a mate's sexual or emotional infidelity. These cues were ranked from the lowest to the highest diagnostic cue for the respective infidelity type. The participants were asked to indicate two different thresholds with respect to their jealousy feelings elicited by these cues. The first threshold concerned that cue to infidelity where they felt the first pangs of jealousy. The second threshold concerned that cue to infidelity where the intensity of the feeling of jealousy became intolerable. It was hypothesized that men would require fewer sexual infidelity cues than women would to reach the two thresholds. Conversely, women were predicted to require fewer emotional infidelity cues than men would to reach the thresholds. Additionally, as an index of processing efficiency, the decision times for the determination of the two thresholds were registered. It was hypothesized that men would more efficiently process sexual infidelity cues (i.e., need less time), whereas, conversely, women would more efficiently process emotional infidelity cues.

2. Method

3. Results

A comparison of the diagnosticity ratings for cues signaling sexual and emotional infidelity revealed considerably lower diagnosticity values for cues to sexual infidelity with low cue diagnosticity, relative to the diagnosticity values of cues to emotional infidelity (see also Appendix A). To illustrate, whereas for each sex, 11 cues to sexual infidelity received mean diagnosticity ratings lower than 4.0, for emotional infidelity, there were only two cues for men and only one cue for women below this diagnosticity value. Hence, a direct comparison of the responses to the two types of cues appears unwarranted. Therefore, the thresholds and decision times were analyzed separately for cues to sexual and emotional infidelity.

3.1. Thresholds

The mean number of cues until the first and second thresholds for cues to sexual and emotional infidelity are shown in Fig. 1. Mixed two-way ANOVAs, with sex as the between-subjects factor and threshold (first and second) as the within-subjects factor, were conducted separately for cues to sexual and emotional infidelity.

Fig. 1. Mean number of cues for women and men until the first and the second thresholds of jealousy as a function of cues to a mate's sexual or emotional infidelity.

With respect to sexual infidelity, the ANOVA necessarily yielded a significant main effect of the within-subjects factor number of cues to the first and second thresholds [F(1,90)=250.27, η2=.74, p<.0001], indicating more cues until the second than the first threshold. The predicted main effect of sex was not significant [F(1,90)=0]. However, the interaction between sex and number of cues until the two thresholds was significant [F(1,90)=5.37, η2=.06, p=.023]. This interaction is attributable to the fact that men and women did not differ with respect to the number of cues until the first threshold [14.1 vs. 12.9, t(90)=1.05], but there were significantly fewer cues between the first and the second thresholds for men than for women (5.9 vs. 7.9; see also Fig. 1).

The mirror image of the results emerged for the cues to emotional infidelity: The main effect of the number of cues to the first and second thresholds was, of course, also highly significant [F(1,102)=517.02, η2=.84, p<.0001], whereas the predicted main effect of sex also failed to be significant [F(1,102)=1.14]. However, the interaction effect was again significant [F(1,102)=7.24, η2=.07, p=.008]. As with the cues to sexual infidelity, men and women did not significantly differ with respect to the number of cues to the first threshold (9.2 vs. 9.3; t<1); this time, however, women processed significantly fewer cues than men did after the first threshold had been reached until the second threshold (8.4 vs. 10.6). Finally, an alternative analysis of the two thresholds, using the cumulative diagnosticity values of the cues (i.e., the sum of the diagnosticity values of the cues until the first and second thresholds) as the dependent variables, yielded virtually identical results.

3.2. Decision times

Decision times were registered in those trials in which the participants pressed the labeled keys S1 and S2, respectively, from the beginning of the presentation of the pertinent cue until the key press. Decision times exceeding the respective group means by more than three standard deviations were excluded from the following analyses. This criterion was met by five men (two in the emotional and three in the sexual infidelity group) and five women (three in the emotional and two in the sexual infidelity group). The mean decision times for the remaining men and women with respect to determining the two thresholds for cues indicating sexual and emotional infidelity are shown in Fig. 2.

Fig. 2. Women and men's mean decision times for determining the first and the second thresholds of jealousy as a function of cues to a mate's sexual or emotional infidelity.

Mixed two-way ANOVAs, with sex as the between-subjects factor and threshold (first and second) as the within-subjects factor, conducted separately for cues to sexual and emotional infidelity yielded the predicted results. Men needed considerably less time than women did for determining the two thresholds with respect to cues indicating a mate's sexual infidelity (7.9 vs. 11.8 s), as confirmed by a significant main effect for sex [F(1,85)=5.78, η2=.06, p=.018]. The remaining main effect, as well as the interaction, was not significant (Fs<1.7). Conversely, women determined the two thresholds for cues to emotional infidelity significantly more rapidly than men did [8.2 vs. 10.8 s; F(1,98)=4.04, η2=.04, p=.047]. Additionally, the main effect for the within-subjects variable threshold (first and second threshold) was also significant [F(1,98)=7.02, η2=.07, p=.009]. Both men and women were faster in determining the second than the first threshold (see also Fig. 2). The interaction between both variables was not significant (F<1.2).

4. Discussion

The present experiment yielded very systematic results. Contrary to predictions, no sex-specific differences were found with respect to the number of cues processed until the first feeling of jealousy was elicited, and this was true for both cues to sexual and emotional infidelity. However, after the first threshold had been reached, men and women alike processed significantly fewer cues of the adaptively primary infidelity type until reaching the tolerance level of the jealousy feeling. Additionally, as predicted, both men's and women's decisions concerning the cues indicating the two thresholds were made more rapidly for cues of the adaptively primary infidelity type. Thus, the present findings add to the accumulating evidence that men and women systematically differ in the processing of cues to sexual and emotional infidelity, and they do so at various stages of information processing (Schützwohl, 2004a, Schützwohl & Koch, 2004). Additionally, contrary to the claim of DeSteno, Bartlett, Salovey, and Braverman (2002) and Harris (2003), these findings also demonstrate again that sex differences in jealousy are not restricted to men's and women's responses in the forced-choice format.

One possible reason for the failure to find the predicted sex-specific difference with respect to the number of cues preceding the first threshold was suggested by a reviewer. Accordingly, the JM is presumably designed to motivate and guide action. Thus, the predicted difference might only emerge at the point were action is called for. Perhaps, the cue eliciting a first feeling of jealousy does not yet require any action, but when the feeling of jealousy becomes intolerable, action is likely.

A further noteworthy aspect of the present findings concerns the dissociation between the cue diagnosticity ratings of the pretest and the number of cues processed between the two thresholds in the main experiment. Whereas the probability ratings of a mate's infidelity revealed—in agreement with the original results reported by Shackelford and Buss (1997)—only a significant sex difference for the cues to emotional infidelity, the number of cues processed between the two thresholds revealed a significant sex difference for both cues to emotional and sexual infidelity. This dissociation might be due to the fact that participants had difficulties in assigning a probability to a single cue but could easily respond with an emotion to certain cues encountered in a series of events. As noted by Cosmides and Tooby (1996, p. 15), “the ‘probability’ of a single event is intrinsically unobservable.” Applied to the requested probability judgments of a mate's infidelity, Cosmides and Tooby argue that no sense organ can discern that, if a mate's sexual behavior changes, there is a 60% probability that he/she has been sexually unfaithful. Either he/she was or was not. “As useful as a sense organ for detecting single-event probabilities might be, it is theoretically impossible to build one. No organism can evolve cognitive mechanisms designed to reason about, or receive as input, information in a format that did not regularly exist. What was available in the environment in which we evolved was the encountered frequencies of actual events” (Cosmides & Tooby, 1996, p. 15; see also Gigerenzer & Hoffrage, 1995). Additionally, frequency information that has been encoded is not transformed into single-event probabilities because of the ensuing disadvantage that they cannot be updated with each new relevant event. In contrast, frequency formats can be easily used to update one's database. These arguments suggest that the participants in the main experiment updated their frequency-based database with each single cue to infidelity capable of eliciting jealousy. This updated database, in turn, was used as input in a decision rule that determines, for each additional cue, the intensity of the jealousy feeling until the intensity is intolerable.

A new, and perhaps the most important, feature of the present data appears to be the evidence that both men and women need less processing of cues to the adaptively primary infidelity type. This efficient processing occurs in the absence of (a) any information that only cues more diagnostic of either sexual or emotional infidelity are presented and (b) any knowledge about the measurement of the time needed to determine the two requested thresholds. This sex-specific efficient processing of cues to the adaptively primary infidelity type not only fits the picture of the JM as a sex- and content-specific information processing device that preferentially processes those cues that are most relevant to the respective adaptive problem, it also supports the general view that evolved psychological mechanisms process a certain range of information according to specific procedures in a privileged manner (e.g., Buss, 1995, McDougall, 1960, Plutchik, 1980). The use of decision times as an indicator of efficient information processing might well be extendable to the investigation of the functioning of other suspected evolved psychological mechanisms in future research (e.g., Schützwohl, 2004b).