What Men Want in a Woman: Personality Is More Important than Academic Record or Athleticism ()
1. Introduction
There is an extensive literature on the topic of mate preferences and selection ( Buss, 1989, 1994; Buss & Schmitt, 1993; Botwin, Buss, & Shackelford, 1997 ; Prokosch, Coss, Scheib, & Blozis, 2009; Shackelford, Schmitt, & Buss, 2005; Schwarz & Hassebrauck, 2012 ). Much of the recent literature has been driven by debates on the power of the body mass index (BMI) over waist-to-hip ratio (WHR) in an attempt to determine the universality of male mate preferences (Swami & Furnham, 2008) .
There are various types of studies on the topic of mate choice: content analysis of “lonely hearts” columns (Greenlees & McGrew, 1994) ; post “commercial date” speed interactions ratings (Kurzban & Weeden, 2005) ; on-line daters research (Gebauer, Leary, & Neberich, 2012) ; psychometric studies on hypothetical, prototypic people with known specified characteristics, as well as ratings of pictures/photographs of individuals (Wood & Brumbaugh, 2009) . Recent studies have looked at specific features of mate preferences (Fletcher, Simpson, Thomas, & Giles, 1999) ; as well as trade-offs and compromises in mate choice (Shackelford et al., 2005) ; and the effect of self-appraisal on mate choice (Kenrick, Groth, Trost, & Sadalla, 1995) . Kurzban and Weeden (2005) found the agreed-upon mate values for both sexes were related almost entirely to observable physical attributes like age, attractiveness, BMI and height and not those less observable characteristics such as education, religion, socio-sexuality or ideas about children.
Some research has examined personality factors (Wood & Brumbaugh, 2009) in mate selection. Gebauer et al. (2013) found two individual difference dimensions namely agency and warmth, which were highly valued cross-culturally. In Big Five terminology these appear to be the two traits of Extraversion and Neuroticism. Furnham (2009) found females rated intelligence, Stability, Conscientiousness, height, education, social skills and political/religious compatibility significantly higher than males, who rated good looks higher than females.
Furnham and Tsoi (2012) found that as predicted, females rated indicators of earning power significantly higher than males, who rated good looks and heredity higher. Effects of similarity attraction were shown in education and financial background, self-assessed attractiveness, values, and personality. Regressions showed that sex, personality and self-ratings (aggression and patience) were consistently related to partner preferences. More recently Neto, Pinto and Furnham (2012) replicated these findings in Brazil and Portugal.
Furnham and McClelland (2015) in a paper methodologically similar to this, found females preferred younger, taller males from their own racial group and within a higher social class. Women showed a strong preference for men their own age, ethnicity and especially height.
This study examines four factors, often ignored with respect to men’s preference: one cognitive/ability factor, one physiological factor and two personality factors. The first factor was ability/intelligence as operationalised by “academic performance”. Whilst there is considerable evidence that women seek out cognitively and emotionally intelligent men there is also evidence that males can be inhibited by women more intelligent than themselves (Szymanowicz & Furnham, 2011) . Academic ability is also in part determined by personality factors namely Openness and Conscientiousness but more of the variance is determined by intelligence (Poropat, 2009) . Hence it was hypothesized that men would favour women of average, over those with superior, intelligence as indicated by an average academic record (H1).
The second factor was fitness operationalized as a “sporty/athletic” dimension. This was taken as a proxy for the degree of health and attractiveness. Whilst most sports require physical fitness and a low BMI though sometimes higher WHR this is not always the case. Considerable previous literature suggests that men are attracted to fit, healthy, fertile young women (Swami & Furnham, 2008) therefore it was predicted athletic/sporty women would be preferred over less or average sporty/athletic women (H2).
The third factor was trait Extraversion. Both males and females seek socially confident, warm optimistic mates and would therefore favour signs of Extraversion over Introversion (H3).
The fourth variable was trait Neuroticism which concerns anxiety, depression, moodiness and proneness to hypochondriasis. There is considerable evidence to suggest that Neuroticism is linked to poor academic and work performance as well as mental and physical illness hence it was predicted that participants would chose stable over unstable mates (H4).
It was also predicted that there would be significant interactions between Extraversion and Neuroticism with a strong preference for Sanguine types. Of particularly interest in this study were standardized effect sizes measures, which it was hoped would provide an index of the relative strength of some characteristics over others.
2. Method
2.1. Participants
The participants in this study were two hundred and fifty eight males ranging in age from 14 to 42 years (M = 22.5, SD = 6.4). Within the sample 40.3% had A-levels (12th grade) as their highest qualification, 33.7% had a bachelor’s degree and the remainder had a higher qualification. The sample consisted primarily of Caucasians (70.4%) with 14.6% Asian, 7.3% African and 3.2% unspecified “other”. Just over half (51.4%) of the participants were single, 22.7% were with a long term partner, 16.9% with a short term partner, 6.3% were married with 2.4% reporting their status as “other” (Response rate: 98.8%). In all, 95.5% said they were heterosexual, 2.4% homosexual and 1.2% bisexual (Response rate: 95.0%). Participants were asked to state the strength of their religious beliefs on a 10 point scale (0 = Not at all; 9 = Very religious); the mean was 2.59 (SD = 2.97) (Response rate: 98.1%). They were also asked to rate their political beliefs on a 10 point scale (0 = Very Left Wing; 9 = Very Right Wing) and the mean was 3.96 (SD = 2.06) (Response rate: 95.4%). They also reported their weight (Mean in kg = 77.38; SD = 15.6) and height (Mean in cm = 179.82; SD = 13.21).
2.2. Materials
All participants completed a two-part questionnaire comprising the Great Date questionnaire and a section of demographic questions. The construction of the questionnaire was based on techniques used in previous studies to assess patient prioritisation for the allocation of scarce medical resources (e.g., Furnham, Ariffin, & McClelland, 2007 ). Participants were then presented with a short description of 16 hypothetical females, who they rated for suitability using a 9-point scale (0 = Definitely No and 8 = Definitely Yes). Each female varied on the following four characteristics: Academic record (high/average); Athleticism (high/low) and two personality variables; Extraversion (introvert/extravert) and Neuroticism (stable/neurotic). Each female represented one particular examplar of the 16 possible combinations of the four characteristics. For example:
![]()
Because this, like other studies in the area, used a within subject design some effort was made to use synonymns for some of the four major variables of interest. In order to check the “correct classification” three people were shown the 2 × 2 × 2 × 2 design and asked to place the people in the different categories. All did so correctly.
2.3. Procedure
Ethical permission was first sought and obtained. All participants were instructed to complete the questionnaire individually and anonymously. Following completion of the study, where possible, participants were fully debriefed. The task took around 10 minutes. The response rate was 97.1%.
3. Results
The mean ratings given to the 16 descriptions of the potential dates are presented in Table 1.
A repeated-measures ANOVA was conducted on the data with intelligence (as indexed by academic record), athleticism, Extraversion and Neuroticism as independent variables, and the ratings as the dependent variable. There was a significant main effect of academic record, F(1, 257) = 187.35, p < .001 (
= 42.2%) with academically stronger women (M = 4.78) being preferred to less academically able individuals (M = 3.94). This disconfirms H1. There was a weaker but still highly significant effect of athleticism, F (1, 257) = 73.56, p < .001 (
= 22.3%) with more athletic females (M = 4.66) preferred over less athletic females (M = 4.05). This confirms H2. There was a strong main effect of Extraversion, F (1, 257) = 362.54, p < .001 (
= 58.5%) with Extravert females (M = 4.97) being preferred to introvert females (M = 3.75). This confirm H3. Finally there was another strong main effect of Neuroticism, F (1, 257) = 210.90, p < .001 (
= 45.1%) with stable females (M = 4.83) preferred to neurotic females (M = 3.88). This confirms H4. Thus, three of the four hypotheses were confirmed.
There were five two-way interactions between the characteristics of the potential dates. There was a significant interaction between academic ability and athleticism, F (1, 257) = 18.34, p < .001 (
= 6.7%). Amongst less academically able females, those who were athletic (M = 4.33) were preferred over those who were not athletic (M = 3.54). The same pattern was evident for more academically able females, but the influence of athleticism was far less marked (athletic: M = 4.99; not-athletic: M = 4.57). There was also an interaction between
![]()
Table 1. Mean ratings as a function of age, social class, ethnicity and height.
Extraversion and athleticism, F (1, 257) = 26.68, p < .001 (
= 9.4%). Amongst Extraverts, athletic females (M = 5.40) were preferred to non-athletes (M = 4.54), but this effect was far weaker amongst Introverts (athletic: M = 3.92; not-athletic: M = 3.54). A very similar interaction was evident between stability and athleticism, F (1, 257) = 36.33, p < .001 (
= 12.4%). Amongst stable females, athletes (M = 5.29) were clearly preferred to non-athletes (M = 4.38), but this effect was attenuated within Neurotic females (athletic: M = 4.03; not-athletic: M = 3.73). There were two weaker (but significant) interactions between Extraversion and Neuroticism, F (1, 257) = 6.15, p < .05 (
= 2.3%), and between intelligence and Neuroticism, F (1, 257) = 4.54, p < .05 (
= 1.7%). With respect to the former, the results indicated that the negative influence of Neuroticism on attractiveness as partners within Extravert females (stable: M = 5.38; neurotic: M = 4.56) was slightly smaller than within Introverted females (stable: M = 4.28; neurotic: M = 3.21). There was a similar pattern with respect to intelligence and Neuroticism; the negative influence of Neuroticism on attractiveness within the more intelligent females (stable: M = 5.20; neurotic: M = 4.35) was slightly smaller than within the less intelligent females (stable: M = 4.47; neurotic: M = 3.41).
There were also a number of three-way interactions between the independent variables. There was a strong interaction between Extraversion, athleticism and Neuroticism, F (1, 257) = 134.90, p < .001 (
= 34.4%). Within Extraverted females, there was an interaction between athleticism and Neuroticism, F (1, 257) = 16.01, p < .001 (
= 5.9%) such that the difference in attractiveness between stable athletic individuals (M = 5.68) and stable non-athletic individuals (M = 5.08) was considerably smaller than the difference between neurotic athletes (M = 5.12) and neurotic non-athletes (M = 3.99). Within Introverted females, the interaction between athleticism and Neuroticism was much stronger, F (1, 257) = 142.02, p < .001 (
= 35.6%) and was of a different pattern. Here there was a much larger difference in perceived attractiveness between stable athletic (M = 4.89) and non-athletic females (M = 3.68) than between neurotic athletes (M = 2.94) and non-athletes (M = 3.47). There was also a significant interaction between academic record, athleticism and Extraversion, F (1, 257) = 18.14, p < .001 (
= 6.6%). Amongst Extravert females, there was an interaction between intelligence and athleticism, F (1, 257) = 33.94, p < .001 (
= 11.7%). Within intelligent females, there was only a small difference in the rated attractiveness as long-term partners between athletic (M = 5.68) and non-athletic (M = 5.14) individuals, but a much larger difference for females of only average intelligence (athletic: M = 5.19, non-athletic: 3.93). However, amongst Introverted females, there was no interaction between intelligence and athleticism, F < 1. Thus the influence of athleticism is moderated by intelligence amongst Extraverted―but not Introverted―females. A three-way interaction was also evident between intelligence, Neuroticism and Extraversion, F (1, 257) = 34.60, p < .001 (![]()
= 11.9%). Within extravert females there was a significant interaction between intelligence and Neuroticism, F (1, 257) = 38.55, p < .001 (![]()
= 13.0%). There was a smaller difference in attractiveness between intelligent stable (M = 5.62) and neurotic (M = 5.15) individuals than between less intelligent stable (M = 5.15) and neurotic (M = 3.97) individuals. Again, intelligence acted as a moderator with respect to the Neuroticism variable amongst Extraverted females. Within Introverted females the interaction between intelligence and Neuroticism was significant but weak, F (1, 257) = 4.52, p < .05 (![]()
= 1.7%). Here the difference in rated attractiveness of the females as long term partners between stable (M = 4.78) and neurotic (M = 3.56) intelligent individuals was marginally larger than the difference between stable (M = 3.78) and neurotic (M = 2.86) individuals with lower intelligence―but this effect was weak.
4. Discussion
The results of this study confirmed three of the four hypotheses. The lowest preference ratings were for an unstable, athletic, Introvert of average academic ability, followed by an unstable, non-athletic Extravert of average academic ability. Interestingly the effect sizes were greater for personality than intelligence or athleticism. There was a strong preference for Sanguine people (Stable Extraverts) who are classed as easygoing, responsive, and lively, and an avoidance of Melancholic individuals―unstable Introverts classified as anxious, rigid and reserved.
This certainly makes sense in terms of evolutionary theory. Nettle (2006) considered the positive benefits and negative costs of the Big Five personality types. Thus Extraverts are attractive because they have big social networks and are good at initiating, but not always maintaining relationships. They tend to be happy, though somewhat impulsive. There are few positive benefits of being Neurotic save social sensitivity and hyper- vigilance. There are however many costs associated with Neuroticism: anxiety, depression, poor mental and physical health and stress sensitivity.
The study showed the relative power of personality variables over ability or health variables through the way these were conceptualised for these results. It is possible that some males, particularly if unfit themselves would be less attractive to a fit and possibly competitive female. This hypothesis warrants investigation. It also implies an obvious interaction between the self-assessed characteristics of the male participant/potential mate and the characteristics of the female he prefers.
In this study (and in contrast to H1) males preferred more rather than less intelligent females as a possible mate. Whilst previous studies have suggested females have a strong preference for more intelligent men, who would presumably have higher paying occupations, some data suggest men can be intimidated by clever women (Szymanowicz & Furnham, 2011) . Again this variable may interact with participant self-estimated intelligence.
This study had limitations. Participants were not making real but hypothetical decisions, and it would have been desirable to have details on the athleticism, intelligence and personality of the male participants in order to examine possible matching effects. Further, the information presented to each participant about each hypothetical female is minimal and confined to just four variables. It is quite possible that there is other information that the participants might have preferentially sought. One important problem associated with a within subject design is to devise people descriptions which are sufficiently different from one another yet realistic. Therefore in this study bubbly and outgoing were used as indicators of extraversion, shy for introvert and moody and unstable indicators of neuroticism. Similarly athletic, sporty and fit were all considered equivalent which may not have always been the case. Despite the fact that we found that people could correctly categorise the 16 people according to our design we recognise that some descriptors may have not been equivalent: i.e., sensitive, touchy, neurotic. Therefore replication is encouraged, perhaps using a between subject design.