When asked to estimate their own intelligence, most people will say they are above average, even though this is a statistical improbability.
This is a normal, healthy cognitive bias and extends to any socially desirable trait such as honesty, driving ability and so on. This pattern is so common that it’s known as “the above-average effect”.
In a recent study, my colleagues and I explored how consistently men and women estimated their own intelligence or IQ (intelligence quotient). We also assessed measures of general self-esteem and masculine and feminine personality traits.
We found the strongest predictors of overestimating IQ were biological sex and then psychological gender. Being born male and having strong masculine traits (both men and women) were associated with an inflated intellectual self-image.
Male hubris, female humility
Despite people’s overall tendency to overestimate their own intelligence, individuals vary. Some doubt their intellectual ability while others greatly overestimate their talents. In general, though, when asked to estimate their IQ, men think they’re significantly brighter than they are, while women’s estimates are far more modest.
Our findings are consistent with those of other studies. Psychologist Adrian Furnham has termed this effect the male hubris, female humility problem. It’s true of many cultures.
There are no gender differences in actual IQ
Psychology and intelligence researchers are unequivocal: men and women do not differ in actual IQ. There is no “smarter sex”. However, it was only with the development of objective measures of assessing intelligence that this notion was invalidated.
Historically, women were believed to be intellectually inferior as they had slightly smaller skulls. By the same logic, an elephant’s intelligence dwarfs ours! Bigger is not necessarily better when it comes to brain size.
In the past century, gender stereotypes have changed greatly. Today, when asked explicitly, most people will agree men and women are equally intelligent. Overt endorsements of gender stereotypes about intelligence are rare in most countries.
But there is quite a difference in implicit beliefs about gender and intellect. Covert and indirect endorsement can still be widely seen.
In a classic social psychology study, researchers asked parents to estimate the intelligence of their children. Sons were rated significantly more intelligent than daughters. This finding has been replicated across the world.
Parental expectations may be particularly important in influencing their children’s intellectual self-image, and are also predictive of later academic achievement.
Gender differences in self-esteem might also be an important factor, as people with higher self-esteem tend to see all aspects of their life (including intellectual ability) more positively. Girls and women rate their general self-esteem significantly lower than boys and men. This difference emerges early in adolescence.
What did our study find?
In our study, we asked participants to estimate their IQ after briefing them on how intelligence is scored. The average score is 100 points. We showed participants that two-thirds (66%) of people score in the range between 85 and 115 points to give them a frame of reference for estimates.
Where our study differed is that we told participants they would complete an IQ test after estimating their own IQ. This would help counter false bragging and inflated estimates, and allow us to test the accuracy of the male and female self-estimates.
Participants also completed a measure of general self-esteem, and the Bem Sex-Role Inventory, which measures masculine and feminine personality traits. We had a hypothesis that psychological gender (specifically masculinity) would be a better predictor of self-estimates than biological sex (male or female at birth).
First, we examined the accuracy of their judgments, as one possibility might simply be that one gender (males or females) had completely unrealistic estimates of ability. Looking at the lines plotting self-estimated IQ against actual IQ, we can see men and women in our sample were fairly consistent in their accuracy. The difference was that male scores (in blue) were more more often overestimates (above the line) and females scores (in green) were more often underestimates (below the line).
After statistically controlling for the effects of actual measured IQ, we next examined the strongest predictors of self-estimated intelligence. The results showed biological sex remained the strongest factor: males rated their intelligence as higher than females. However, psychological gender was also a very strong predictor, with highly masculine subjects rating their intelligence higher (importantly, there was no association with femininity).
There was also a strong contribution of general self-esteem to participants’ intellectual self-image. As noted above, males report higher self-esteem than females.
Why does all this matter?
Educational psychologists pay attention to intellectual self-image because it’s often a self-fulfilling prophecy: if you think you can’t, you won’t.
When girls undervalue their intelligence in school, they tend to choose less challenging course content – especially in science, technology, engineering and mathematics (the STEM subjects). These decisions limit their education and career choices after school.
We need to lift girls’ aspirations if they are to go on to solve the complex problems our society faces, while achieving equal pay. It starts early with gendered parental expectations of intelligence, and differences in self-esteem between boys and girls.
Wouldn’t it be nice if, as parents, educators and a society, we could build the confidence of girls and young women to a level where they believe in themselves and are free of those doubts?
Accuracy of Self-Estimated Intelligence
Psychological research has investigated whether people are accurate judges of their intellectual ability generally (irrespective of gender). This arises from several strands of investigation: firstly, whether people are generally sound judges of their intellectual strengths and weaknesses, and secondly, whether there are cognitive biases that affect such evaluations.
Furnham (2017) noted that an unresolved research question is whether males over-estimate their actual IQ, females under-estimate IQ, or indeed both, but writes “there are not enough good studies with both self-estimated and test-derived IQ to settle the argument,” p. 110.
This may be due to the relative ease with which self-estimates of intelligence may be obtained, but the greater difficulty, time, and expense needed to administer psychometrically valid intelligence tests. There are some examples where a proxy is used, such as a vocabulary test, to investigate the association between self-estimated IQ and intellect (r = 0.25, McCrae and Costa, 1985), while others choose to use a test of non-verbal reasoning like the Raven’s Progressive Matrices (r = 0.29, von Stumm, 2014).
In a review of studies comparing self-estimated and psychometrically assessed intelligence, Paulhus et al. (1998) note that in student subject pool samples, correlations rarely exceed r = 0.30 which is a moderately sized effect. They further note that somewhat larger correlations are found in studies that sample from the general population.
To provide a frame of reference for evaluating this, self-reports of intelligence have roughly the same predictive validity and accuracy as the situational judgment tests (SJTs) that are widely employed in organizational psychology for predicting cognitive performance (r = 0.29 in a meta-analysis by McDaniel et al., 2007). People’s impressions of their intellect are therefore grounded firmly in reality, but their accuracy is subject to distortion by cognitive biases.
One such bias noted in the literature is the “above-average effect” (Alicke, 1985; Dunning et al., 1989; Kruger, 1999; Kruger and Dunning, 1999), which holds that for socially desirable traits such as competence and intellectual ability, there is a tendency for most people to see themselves as better than the average person. The implication of this, Kruger and Dunning (1999) argue, is that such overly favorable views of their abilities mean that a large proportion of the population is “unskilled and unaware of it,” p. 1121. Such a claim stands in contrast to evidence on the general accuracy of self-estimates of intelligence reviewed above, though the number of studies empirically testing this with psychometrically valid IQ tests are few.
Another bias is the self-esteem bias (Felson, 1981), which is the tendency for people to evaluate themselves in a way that is consistent with their general self-esteem; someone who is high in self-esteem will tend to see themselves as brighter and more capable than someone lacking in self-esteem.
While self-esteem is a normally distributed trait, there are frequently observed variations for different subgroups. Gender differences in general and academic self-esteem are well documented (Eccles et al., 1993; Gentile et al., 2009), with boys and men reporting higher general and academic self-esteem than girls and women. Syzmanowicz and Furnham (2011) raised this issue in their meta-analytic review as one possible explanation for the MHFH effect. However, they reported no correlation between self-estimated intelligence and self-esteem, and it seems few studies have actually pursued this line of reasoning (Mirjalili et al., 2011).
Parental Beliefs, and Socio-Cultural Transmission of Gender Stereotypes
Environmental factors are also likely to contribute to a gender bias in self-estimated intelligence which may be an extension of existing socio-cultural gender stereotypes. Social motives (e.g., boastful pride for males or modesty for females) might explain self-estimates of intelligence. If so, when asked to estimate of other people’s intelligence the MHFH effect should not still be present.
In the original study by Hogan (1978) into self-estimates of intelligence, participants were also asked to provide an estimate of the intelligence of their mothers and fathers. Fathers were rated as more intelligent than mothers (Hogan, 1978), even though there were no gender differences in general intelligence in the community.
The effect has been replicated numerous times (Beloff, 1992; Furnham and Rawles, 1995), but should be interpreted cautiously as it might reflect the systemic educational and occupational inequalities of the time (i.e., higher male educational advancement) rather than genuinely held beliefs that men are inherently “smarter.”
Furnham and Gasson (1998) took a different approach, and instead asked parents to provide an estimation of the intelligence of their own children. Sons were rated as more intelligent than daughters (d = 0.67), and this effect has been replicated (Beloff, 1992; Furnham, 2000; Furnham et al., 2002a). Such a pattern of results suggests that environmental factors like gender stereotypes might contribute to the MHFH problem, rather than differential social desirability for intelligence between men and women.
Parental beliefs may be a particularly important mechanism in the socialization of gender stereotypes, as parental educational expectations may influence a child’s view of their own capabilities (Frome and Eccles, 1998; Jodl et al., 2001). Parental beliefs and expectations may inadvertently enhance or stifle a developing child’s intellectual self-concept and self-efficacy beliefs: raising a child that feels either bright and capable even in the face of challenges (mastery orientation) or overwhelmed and incapable of more advanced intellectual achievement (learned helplessness).
Numerous studies have demonstrated that parental beliefs about their children’s intellectual abilities predict later educational achievement in adolescence and young adulthood (Jodl et al., 2001; Phillipson and Phillipson, 2007; Gunderson et al., 2012; Pinquart and Ebeling, 2019). This may be partly through direct transmission of parental beliefs and expectations, but also because parents can provide or withhold enriching cognitive experiences which can accelerate intellectual development outside of school.
Parents are but one element in a larger ecological system that contributes to intellectual development and intellectual self-image. This system includes the role of teachers and educators in shaping the intellectual self-image of children in their care (Jussim and Harber, 2005; Kollmayer et al., 2018), as well as differential treatment of boys and girls (particularly in gender-typed courses such as mathematics and science). Children’s intellectual self-image is also shaped by media and popular culture (Solbes-Canales et al., 2020), which also plays a part in transmission of cultural gender stereotypes about intellectuality (Nosek et al., 2002; Storage et al., 2020).
Sex-Role Identification and Self-Estimated Intelligence
Another potential explanation for the MHFH effect may be the contribution of gendered personality traits, and sex-role identification. Bem (1981b) proposed gender schema theory as a cognitive account for the way that cultural prescriptions about masculinity and femininity become integrated into our self-concepts.
These self-concepts forms internalized standards for regulating our own behavior, and also evaluating that of others through the lens of a gender schema. Now, while boys and girls typically differ in their early socialization experiences (Eccles et al., 1990; Lytton and Romney, 1991), there is also considerable individual variation in the degree to which one acquires stereotypically masculine and feminine personality traits, behaviors and interests- a process termed sex-typing (Kagan, 1964; Kohlberg and Ullian, 1974).
The internalized gender schema of each individual differs and is the product of both biological and environmental factors that contribute to their sex-role identity (Tenenbaum and Leaper, 2002; Hines, 2011, 2015; Svedholm-Häkkinen et al., 2018). Highly sex-typed persons are motivated to keep their behavior and self-concept consistent with traditional gender norms of their biological sex (Maccoby, 1990; Martin and Ruble, 2004), and so implicit beliefs about gender and intellectuality could translate to higher estimates of intelligence by males and lower estimates by females.
For many people their sex-role identification is veridical with their biological sex, but others are more flexible and incorporate a healthy blend of both masculine and feminine personality traits into their self-schema. Researchers have termed this psychological androgyny (Bem, 1984; Spence, 1984; Reilly, 2019), and it has been associated with greater psychological adaptability and less rigid gender schemas. Might sex-role identification act as a better predictor of self-estimated intelligence than the social category of gender?
There are several lines of reasoning that would support such an association. Firstly, as outlined above, it has been hypothesized that self-esteem makes a strong contribution to self-estimated intelligence. While gender differences in self-esteem are frequently reported (Gentile et al., 2009), numerous studies have documented a positive association between masculinity and self-esteem in both men and women (Whitley, 1983; Burnett et al., 1995).
This, in turn, might drive higher self-estimates of intelligence. Secondly, there are links between sex-role identification and the development of cognitive ability. Nash’s (1979) sex-role mediation hypothesis proposed that both masculine and feminine sex-roles contribute to cognitive development: masculinity predicts visual-spatial performance (Reilly and Neumann, 2013), while femininity predicts verbal and language abilities (Pajares and Valiante, 2001; McGeown et al., 2011; Reilly et al., 2016).
Those higher in masculine and feminine traits may rate their abilities in those domains as higher, which may contribute to their overall impression of intellectuality. Beyer and Bowden (1997) reported the tendency for women to underestimate their performance on stereotypically masculine tasks, but that this underestimation was not found for neutral or feminine tasks. Thirdly, for those with rigid gender schema, male boastfulness and female humility may temper their self-reports and over time shape their self-concept to reflecting implicit gender stereotypes.
Several studies have tested the contribution of sex-role identification to the MHFH effect. The first by Furnham et al. (1999) recruited a small number of subject pool participants, and had them complete the Personal Attributes Questionnaire (PAQ; Spence et al., 1974) which attempts to measure masculinity and femininity as personality traits. Results were inconclusive, though the study was underpowered.
A second study by Rammstedt and Rammsayer (2002) recruited a larger sample size and instead used the Bem Sex-Role Inventory (BSRI; Bem, 1981a) which has greater psychometric validity (Choi et al., 2009). Subjects were asked both about their overall intelligence, as well as domain-specific multiple intelligences in line with Gardner’s (1999) typologies. The authors found tentative support for sex-role effects in males, with those scoring higher in masculinity rated their mathematical-logical and general reasoning higher than lower-masculinity peers.
However, the authors did not find sex-role effects for the females in their sample. Finally, a study by Storek and Furnham (2012) that recruited intellectually gifted MENSA members found a positive association between masculinity and self-estimated intelligence in both men and women. However, generalizability from such a highly-select sample is questionable. Furthermore, none of these studies included an actual measure psychometric IQ or of self-esteem to determine what role (if any) this played in the MHFH effect.
General Intelligence Versus Multiple Intelligences
Experts on human intelligence have different views on the nature and structure of intelligence to those of the everyday man and woman. Intelligence is not a unitary construct (Neisser, 1979; Halpern, 2011), and comprises a large number of distinct abilities such as verbal intelligence, mathematical/logical intelligence, emotional intelligence, and so on. Sternberg et al. (1981) examined how lay conceptions of intelligence cluster around a different set of abilities to that of intelligence experts. Sternberg (2000, p. 3) argued that understanding these implicit or lay theories of intelligence was crucial, as “implicit theories of intelligence drive the way in which people perceive and evaluate their own intelligence and that of others.”
In reference to the present topic, while gender differences in overall SEI are widely documented, we might see different estimation patterns for certain abilities, such as those stereotypically regarded as masculine or male-dominated (mathematical/analytical, spatial), and those more readily associated with femininity or that are regarded as stronger in females (e.g., verbal and emotional intelligence).
One taxonomy for considering intelligence is Gardner’s (1983, 1999) theory of multiple intelligences. Furnham (2000, 2001) first investigated whether the MHFH effect extended to Gardner’s multiple intelligences, which has since been expanded to encompass seven to nine distinct clusters of abilities depending on the definitions used (Furnham et al., 2001, 2002a,b). Even though intelligence researchers may disagree on the psychometric validity of Gardner’s multiple intelligences, student perceptions of them are important as they may guide course selection.
Subjects are typically presented with a definition of each of Gardner’s multiple intelligences, and asked to estimate their intelligence relative to others. These domains are: verbal or linguistic intelligence, logical or mathematical intelligence, spatial intelligence, musical intelligence, bodily-kinesthetic intelligence, interpersonal intelligence, intrapersonal intelligence, naturalistic intelligence, and existential/spiritual intelligence.
Research on self-estimations has revealed a complex and nuanced pattern: while gender differences were almost always found for estimates of general intelligence, they were not reliably found for all of Gardner’s multiple intelligences. Moreover, cross-cultural differences are present. For example, Yuen and Furnham (2006) found that students in Hong Kong did not exhibit significant gender differences for verbal or interpersonal intelligence (stereotypically feminine) but did for all the remaining abilities.
However, Furnham et al. (1999, Study 2) found significant gender differences with an English sample for only three of Gardner’s domains: mathematical/logical, spatial and musical intelligence. A review by Furnham (2001) on several of Furnham and colleagues’ studies noted that consistent gender differences were primarily found on stereotypically masculine intellectual abilities (mathematical/logical, and spatial), which Storek and Furnham (2012, 2014) subsequently referred to as domain-masculine intelligence (DMIQ). Furthermore, Storek and Furnham (2013) also found a moderately sized correlation between masculinity and self-estimates for DMIQ, r = 0.26, suggesting that there may be sex-role contributions to the effect.
When there are inconsistencies across studies and types of samples, the technique of meta-analysis provides a greater degree of confidence of the robustness of an effect than any single study alone. Syzmanowicz and Furnham (2011) conducted a meta-analysis on self-estimates of general intelligence and for three multiple intelligence domains, reporting moderately large gender differences favoring males for general intelligence, d = 0.37, mathematical/logical intelligence, d = 0.44, spatial intelligence, d = 0.43, and a much smaller difference for verbal intelligence, d = 0.07. However, none of the other forms of multiple intelligences were investigated. Moreover, further research is required to determine the extent of gender differences for other domains and to test potential moderators for the self-estimation effects.
reference link : https://www.frontiersin.org/articles/10.3389/fpsyg.2022.812483/full
Source: The Conversation