Mating Intelligence 1
Running head: MATING INTELLIGENCE
Mating Intelligence Conceptualized as Adaptive Cross-Sex Mind-Reading Errors
Glenn Geher
StateUniversity of New York at New Paltz
Please contact the first author at
with queries regarding this work.
Abstract
A major element of Mating Intelligence (Geher & Miller, 2007) pertains to the ability to accurately read the mating-relevant judgments of potential mates. In this research, 481 young male and female heterosexual adults judged which personal ads (written by opposite-sex individuals) were most desirable as short and long-term mates. All participants then engaged in a cross-sex mind-reading task by guessing which ads were most strongly endorsed by opposite-sex individuals. Males were more accurate than females across both short and long-term judgments. A content analysis was conducted to delve into the qualitative nature of errors made. Male errors in guessing short-term desires of females tended to result from overestimating the degree to which females focused on sexual qualities in short-term mates. Female errors in guessing both short and long-term desires of males consistently showed this same tendency: overestimating the degree to which males focused on sexual qualities. Data revealed a strong, positive correlation between this overestimate sexual interest bias for males making short-term judgments about females’ desires and males’ scores on an index of general intelligence, partially supporting a model of MI as corresponding to adaptively erroneouscross-sex mind-reading.
Mating Intelligence Conceptualized as Adaptive Cross-Sex Mind-Reading Errors
Given the importance of mating psychology in the fundamental evolutionary goal of reproduction, it is reasonable to consider the role of mating as relevant across psychological domains (Miller, 2000). One major psychological area that has not seriously been considered in the light of mating psychology is intelligence. A recent content analysis of scientific journals dedicated to either intelligence or human mating revealed that nearly no intelligence researchers have addressed how intelligence relates to mating (in the journal Intelligence) and, similarly, nearly no publications in journals dealing with intimate relationships (e.g., the Journal of Social and Personal Relationships) have addressed the interface between mating and intelligence (see Geher, Miller, & Murphy, 2007). The mating intelligence construct we are developing (Geher & Miller, 2007) works to synthesize scholarship in these disparate areas. Generally, we conceive of mating intelligence as the set of cognitive abilities designed for mating-relevant purposes.
In the model of mating intelligence presented elsewhere (see Geher, Camargo, & O’Rourke, 2007), we conceive of mating intelligence as being comprised of two important classes of cognitive abilities: (a) courtship-display mechanisms: those mechanisms that function as fitness indicators in the domain of courtship (such as musical abilities, which make one attractive to potential mates) and (b) mating mechanisms: those mechanisms that are directly relevant to the mating domain (such as the ability to know if a mate is cheating in a relationship). Courtship-display mechanisms (such as displays of artistic creativity) are relevant to mating as they are effective in attracting mates (e.g., Haselton & Miller, 2006), but these mechanisms are not conceptualized as pertaining to mating directly. Mating mechanisms, on the other hand, are psychological mechanisms that deal directly with mating-relevant content.
Following on the heels of Miller’s (2000) treatise on the higher order cognitive processes as having been shaped by sexual selection forces, a new wave of research has demonstrated the predictive validity of the courtship-display component of this model of mating intelligence. Consistent with the idea of creative intelligence as having a courtship function, for instance, Haselton and Miller (2006) found women are particularly attracted to relatively creative men at or near the peak of their ovulation cycle. In a separate line of research supporting the courtship-display domain of our model of mating intelligence, Bressler, Martin, and Balshine (2006) found that heterosexual women value the ability of a potential mate to produce humor – lending support to the idea that the kind of creative intelligence needed for humor production serves a pivotal role in the domain of mating. In fact, a great deal of research is coming forward that supports this courtship-display function of human creative intelligence (Geher & Miller, 2007).
The mating mechanisms that underlie mating intelligence, on the other hand, are conceptualized as cognitive abilities that do, in fact, bear directly on mating issues. As presented elsewhere (Geher, Miller, & Murphy, 2007), these mechanisms include such abilities as accurately detecting one’s own value in a local mating market, being able to determine if a mate is engaged in infidelity, being able to separate accurate from deceptive courtship signals, etc. Once we start thinking about psychological mechanisms related to mating in terms of individual differences – by using an abilities focus– many novel questions regarding human mating suddenly become apparent (see Buss, 2007). The research summarized here focuses on a particular set of mating-relevant cognitive abilities that we conceptualize as fitting in this mating mechanisms of mating intelligence. In particular, this work addresses the ability to know the desires of potential mates; a crucial element of our psychology.
Emotional Intelligence as a Model for Studying Mating Intelligence
While research on emotional intelligence has been fraught with controversy since its inception in 1990 (Salovey & Mayer), the ability-based model for studying emotional intelligence has strongly demonstrated the emotional intelligence construct to be valid and useful in predicting behavioral outcomes (see Brackett & Mayer, 2003).
This ability-based method may serve as the basis for assessing mating intelligence. Essentially, the ability-based indices of emotional intelligence (e.g., the Emotional Accuracy Research Scale; Mayer & Geher, 1996) present items to a large group of participants, asking them to make quantifiable judgments regarding emotional stimuli (for instance, participants may read an emotionally laden vignette and then report, on a Likert scale, how happy the author of the vignette was at the time of writing. Emotional intelligence scores are then computed for each participant by increasing his or her score as a function of how much his or her judgments tended to match the group norm. In using a weighting algorithm, participants’ scores increase more for choosing items that were endorsed by a relatively large sub-sample of the group. For instance, if 15% of the group thought the author of vignette was not angry, 20% of the group thought the author was somewhat angry, and 65% of the group thought the author was very angry, a participant’s score would increase by .65 if he or she chose very angry, but only by .20 and .15 if he or she chose somewhat angry or not angry, respectively.
While this weighted-consensus scoring method has demonstrated its merits in tapping abilities that straddle the interface of cognitions and emotions, it is not without its criticisms. An important criticism of this method pertains to the fact that high scorers are essentially stereotypical thinkers (see O’Sullivan & Ekman, 2004). In other words, someone who scores high on emotional intelligence using this method may simply be someone who sees the (emotional) world as everyone else does.
While emotional intelligence researchers have developed strong empirical and theoretical arguments against this criticism (see Mayer et al., 2000), adapting this methodology to the issue of assessing cross-sex mind-reading abilities actually circumvents this issue.
In the current work, the weighted consensus scoring method was employed to examine heterosexual adults’ abilities to guess the desires of opposite-sex individuals. In short, participants read personal ads that were written by members of their own sex – and then they guessed which ad (within clusters of three) was most strongly endorsed by members of the opposite-sex. This methodology allowed for the use of the weighted consensus scoring method – by increasing participants’ scores in light of how well their guesses matched the actual choices of opposite-sex individuals – but it has the benefit of clearly tapping an actual social-intellectual skill with a correct answer. Put another way, in an emotional intelligence test, a high scorer is someone who sees emotions as others do (regardless of the correctness of their judgments). With the mating intelligence test developed for the current research, high scorers were conceptualized as those who correctly guessed the norms representing the reported desires of real opposite-sex individuals. Thus, insofar as this methodology allows for both the psychometric benefits of weighted consensus scoring (see Mayer & Geher, 1996) and the capacity to operationalize answers as correct or not, this technique represents something of an improvement over ability-based emotional intelligence measures.
The mating intelligence measures incorporated in the current research, thus, were partly based on the measurement paradigm for the ability-based emotional intelligence measures. Importantly, of course, the measures used in this research were designed to tap mating-relevant (as opposed to emotion-relevant) cognitive skills.
Based on previous work on human mating behaviors conducted by evolutionary psychologists (e.g., Buss, 2003), separate measures of mating intelligence were designed to tap the ability to know the short-term versus the long-term desires of potential mates. Further, given that heterosexual desires were examined in this research, separate tests were made for males and females. As such, four indices of mating intelligence were created in this work (males’ abilities to know the short-term desires of females, males’ abilities to know the long-term desires of females, females’ abilities to know the short-term desires of males, females’ abilities to know the long-term desires of males).
Accurate versus Adaptively Biased as “Intelligent”
In studying abilities tied to knowing the desires of opposite-sex individuals, there is a question regarding how to operationally define someone as scoring as high versus low. On first glance, it makes sense to consider a high scorer as one who is relatively accurate – one whose guesses regarding the desires of the opposite-sex tend to match their reported desires. The mating intelligence measures implemented in this research were designed with this criterion in mind. However, a growing body of literature in evolutionary social psychology suggests that judgments fraught with bias may actually be relatively adaptive and, perhaps, intelligent compared with judgments that are accurate but evolutionarily high in risk. For instance, consider research that has documented sex-specific errors in cross-sex mind-reading (see Haselton & Nettle, 2006). In such research, males have been found to consistently overestimate sexual interest on the part of females while females have been found to underestimate males’ desire for commitment to long-term relationships (i.e., females demonstrate commitment-skepticism; Haselton & Buss, 2000).
Findings documenting these systematic errors have been framed in terms of Error Management Theory (Haselton & Buss, 2000), an evolutionarily informed theory of cognitive biases which suggests that erroneous judgments that are more likely to ultimately lead to successful reproduction should be more typical of human psychology compared with accurate judgments. For instance, it may benefit males to overestimate female sexual desire in that this bias may encourage a sort of naïve optimism which may increase males’ ability to acquire short-term sexual partners. Similarly, it may benefit females to be skeptical of males’ levels of commitment; females who have an extremely conservative and skeptical screen when judging potential mates may benefit from being more likely to acquire partners who are, indeed, genuinely willing to commit.
These ideas regarding adaptive cross-sex mind-reading errors are important to consider in the development of an understanding of mating intelligence. Essentially, the error-management perspective suggests that adaptive (and, perhaps, intelligent) mating decisions may often actually be inaccurate! In measuring mating intelligence, we may actually consider high scores as less accurate, but as adaptively biased and, thus, relatively intelligent. At the very least, this adaptive-bias perspective suggests that criteria by which cross-sex mind-reading judgments are evaluated need to be examined in terms of both accuracy and a proclivity toward increasing genetic fitness.
To address this adaptive-bias perspective, two sets of mating intelligence scores were computed for participants. Half the scores represented the ability to accurately know the desires of the opposite sex, while the other half represented the proclivity toward adaptive bias. For males, this bias reflected a tendency to think that females were more interested in no-strings-attached sexual relationships than was actually the case. For females, this bias reflected commitment skepticism by tapping the tendency to think males were exclusively interested in sexual qualities in potential partners (i.e., the tendency to think that all men are pigs – the complement of commitment skepticism, in effect).
Goals of the Current Study
The goals of this study fall into two broad classes. First, the methodology allowed for an assessment of sex differences in cross-sex mind-reading abilities. Thus, separate from questions regarding individual differences in these abilities, this research was able to address questions associated with phenomenological patterns across the sexes. A second set of goals focused on mating intelligence as an individual-differences construct (more of a traditional intelligence sort of construct).
Sex-Differentiated Patterns in Cross-Sex Mind-Reading. In terms of sex-differentiated patterns, the goals were as follows:
(a) To assess overall accuracy of judgments across both type of judgment (long and short-term) and across the sexes
and
(b) To discover evidence of sex-specific adaptive biases consistent with Error Management Theory. Findings should reveal a tendency for males to overestimate females’ interest in sexual features of potential mates and, similarlythe nature of sex-specific errors in judgments that may reveal evidence for overestimating sexual interest on the part of potential mates (suggesting a she wants me bias). Further, data were predicted to show the complement of commitment skepticism in females, by showing that females overestimate males’ interest in sexualized qualities in potential mates (suggesting a men are all pigs bias).
Measuring Individual Differences in Mating Intelligence.Using a measurement scheme roughly based on the ability-based measures that exist to tap emotional intelligence (see Mayer et al., 2000), individual-difference measures were designed to tap mating intelligence for the participants in this study. The goals regarding these measures were as follows:
(a) To develop reliable and valid measures of the ability to guess the long-term desires of potential mates (for each sex)
(b) To develop reliable and valid measures of the ability to guess the short-term desires of potential mates (for each sex)
(c) To develop individual-differences indices of adaptive bias for each sex
and
(d) to examine the convergent validity of the different mating intelligence indices by addressing their inter-correlations with an index of general intelligence. As has been documented in research on emotional intelligence, for a set of abilities to be reasonably framed as an intelligence, the facets should be somewhat positively correlated with general intelligence (see Mayer et al., 1999).
Method
Participants
481 young heterosexual adults (329 females and 152 males) participated in this research. For females, the mean age was 22.17 (SD = 4.48). For males, the mean age was 24.58 (SD = 7.65). Participants were predominantly college students at SUNY New Paltz who volunteered to participate after receiving an email invitation asking them to be part of this research. Some received partial credit for their psychology classes. The web-based nature of the data collection allowed for the subject pool to go beyond the confines of New Paltz students. Additional participants were friends of New Paltz students who were invited by email to participate.
Measures
The measures used in this study included an index of general intelligence and two kinds of mating intelligence tests (for each sex). The Army Alpha Intelligence Test (Modified Vocabulary Subscale; Yerkes, 1921) was used as an index of general intelligence. This 30-item test has been found to be extremely g-loaded, correlating positively with measures of general intelligence, such as the Weschler-Bellvue test, and has been modified to serve as a proxy for general intelligence (see Mayer, Salovey, & Caruso, 1999).
For each sex, a measure of long-term mating intelligence was implemented. For this measure, participants were first presented with 10 items that included clusters of three real personal ads written by members of the opposite sex (See Tables 1 and 3). Within each cluster, they were asked to choose which ad represented the person they would most want for a long-term mate. Next, participants made cross-sex mind-reading judgments; they were presented with the long-term items that were initially given to members of the opposite sex for judgment. Participants were asked to guess which ad within each cluster was most commonly chosen by members of the opposite sex as most desirable for a long-term mate. These personal ads were collected by a team of research assistants from online dating-service sites and were modified so that demographic information such as data regarding ethnicity and religion were deleted.
Each participant also completed a sex-appropriate short-term mating intelligence test. The algorithm described in the prior section regarding the measurement of long-term mating intelligence was used to assess short-term mating intelligence with the exception that these items revolved around participants being asked to make short-term ratings (See Tables 2 and 4). Additionally, different personal ads were used in the short-term measures than in the long-term measures.
Procedure
A web-based survey was created for the purposes of data collection (using Flashlight survey software). After participants read a document providing informed-consent information, they completed the vocabulary test and the different elements of the mating intelligence measures. For the intelligence test, participants were presented with 30 words. They were asked to choose the best synonym (of four options) for each item. For instance, participants were presented with the word plenary and were instructed to choose the best synonym (of complete, candid, culpable, and cloying (the correct answer is complete)).