Gender differences in self estimation of IQ

Abstract. This research sought to establish whether females still self estimated their IQ’s at a lower level than males. 20 subjects (10 males, 10 females) were asked to estimate their own IQ. Analysis of group data, using Mann Whitney ‘U’ test, showed no significant difference. Males self estimated their IQ higher than females and their estimations covered a wide range. Speculation about the causality of this difference was inconclusive.


Over the past few decades, gender differences in self-estimation of IQ (Intelligence Quotient) have been the subject of debate and question. It was a commonly accepted belief that men were more intelligent than women and because there was no actual evidence to support this stereotype, it attained its credibility. Thus, research was conducted in Louisiana and Tennessee by sociologist Hogan (1973-76 cited in The Psychologist, July, 1992 p.309-311) to find out whether this was factual or not. He originated a study asking over 1800 high school & college students, as well as non-student adults, ‘by the comparison with the national average score of 100, what do you estimate your IQ would be…’ What he found was females underestimated their IQ while men overestimated their IQ.

In the experiment Higgins (1987 cited from The Guardian, 10th February, 1987) conducted at Chester College, women’s average IQ estimations were 111 while the average of men’s remained 118. Five years later, Beloff (1992) followed the design of Hogan and Higgins (1992 cited in The Psychologist, July 1992). Her findings demonstrate, that as a group, males in her class estimated their mean IQ 6.4 higher than the mean estimate of females. Reilly and Mullhern (1994) repeated previous studies at Queen’s University in Belfast.

Students were asked to estimate their IQ and then take a digit symbol and vocabulary test from the WAIS (Wechsler Adult Intelligence Scale). The result supported the prior consistency in females underestimating their IQ. Bennet (1996) presented a study in duplication to Hogan. His findings showed, even 20 years later, males were continuing to self-estimate their IQ higher than females. Do women still self-estimate their IQ lower than men? To support the hypothesis that ‘women self estimate their IQ at a lower level than men,’ the previous studies of Higgins, Hogan, Bennet and Beloff must be challenged.


Design The design used for this study was an Independent (between participant) design. The single-tailed hypothesis is that ‘women self-estimate their IQ at a lower level than men.’ Therefore, the Independent Variable (IV), the participant’s gender, and the Dependent Variable (DV), the participants IQ estimation, is established. Employing the Hypothetico-Deductive Model, it was decided that random samples of 20 participants (P’s) would be asked to self-estimate their IQ.

Participants Having chosen to select 20 P’s (10 males, 10 females) at random, three controls were applied. Firstly, P’s had to have some understanding of what IQ was Then, P’s could not be aware of their IQ, i.e. had never taken an IQ test. Finally, P’s had to be over 16 years of age, due to ethical and consensual reasons. Materials A standardised approach and question was formulated, which was then applied to all P’s. The purpose of this was to eliminate as many extraneous variables, such as tone pace and inflection, as possible. (See appendix 4)


P’s were approached, individually, in various social settings and asked if they had a few minutes to answer some questions about IQ. Once a verbal consensus was made, P’s were then asked the standardised questions and raw data was recorded (see appendix 1). Afterwards, P’s were debriefed and informed this was a study concerning gender differences in self-estimations of IQ in partial replication of previous research by Hogan, Higgins and Beloff et al. These studies found that women self-estimated their IQ at a lower level than men and the intention of this piece of research was to see if this was still the case. P’s were then notified of confidentiality and their right to withdraw their data.

Finally, P’s were asked if they had any questions and thanked for their participation.
The descriptive statistics (see appendix 3) show, that in terms of group means, male P’s self-estimated their IQ 25.5 higher than female P’s (males =143.6, females = 118.1 See chart above). Male P’s self-estimations covered a wider range than females P’s (males=104, females =77). The most appropriate test, when comparing 2 sets of independent data, is the Mann-Whitney ‘U’ test. An exact significance of 0.0715 was determined dividing 0.143 by 2 (see appendix 2). This is more than the 0.05 level (Coolican, 1999) of significance needed to accept the hypothesis ‘Females self-estimate their IQ at a lower level than males’. Therefore, the null-hypothesis ‘ Females do not self-estimate their IQ at a lower level than men’ must be accepted.


Previous findings have, infact, supported the hypothesis that females self estimate their IQ lower than males but because women play a more prevailing and scholarly role in today’s society, it was questionable whether the results have changed or remained the same. Taking random samples from 10 females and 10 males and aiming to achieve significant results to re-establish our hypothesis, the descriptive statistics gave evidence to support that ‘women self-estimate their IQ at a lower level than men’. Males gave a mean estimate of 144 and the women of 118 IQ points. They also had a higher range deviation, median, and standard deviation than women. (See results) The Mann-Whitney ‘U’ test revealed a .0715 level of significance, which does not fall on or below the .05 needed to allow the original hypothesis.

Therefore, the null hypothesis ‘women do not self-estimate their IQ at a lower level then men’ must be accepted. Taking a critical look at study, it is possible that limitations or faults may have influenced the outcome. Some factors could be that the sample was too small, the location (Greater London) limited, or that subjects had known some reasoning or interpretation behind the study and its question. Another factor could be the standardised approach, when asking for the sample, was ineffective or confusing.

Although age was not taken into consideration when samples were taken, researchers may have unconsciously been inclined to choose participants close to their own age group and social class. This may have had an important effect on the results. Further research could be done to establish the causes of women self-estimating their IQ at a low level. These causes could lead to a better awareness of their intelligence.

Identify the aim of the research and state the experimental / alternative and null hypothesis. The aim of this study will be to see if gender effects the rate of conformity, because Jenness’s 1932 study didn’t take gender in to account …

Gender refers to the differences and disparity between males and females. Gender roles, meanwhile, refers to the social differences between males and females. Gender differences may either be innate or inherent or it can also be caused by a multitude …

This investigation was adapted from the Fagot (1985) investigation where teachers and two-year-old children were studied to discover whether there were differences in the way teachers reinforced behaviour in boys and girls, and the effects of peer group reinforcement. The …

The aim of this task was to find out whether gender affects the compliance levels in humans. Compliance is a form of conformity, when a person acts according to certain accepting standards. I will be observing peoples behaviours when crossing …

David from Healtheappointments:

Hi there, would you like to get such a paper? How about receiving a customized one? Check it out