Sex differences in the right tail of cognitive abilities: An update and cross cultural extension

Highlights

• Sex differences in the extreme right tail of math and verbal domains in the U.S. have long been established.

• Ratios in the extreme right tail of math in the U.S. have shrunk in the last 20 years and remained relatively stable in the verbal domain.

• Similar patterns of male-female ratios in the extreme right tail were found in a sample from India.

Abstract

Male–female ability differences in the right tail (at or above the 95th percentile) have been widely discussed for their potential role in achievement and occupational differences in adults. The present study provides updated male–female ability ratios from 320,000 7th grade students in the United States in the right tail (top 5%) through the extreme right tail (top 0.01%) from 2011 to 2015 using measures of math, verbal, and science reasoning. Additionally, the present study establishes male-female ability ratios in a sample of over 7000 7th grade students in the right tail from 2011 to 2015 in India. Results indicate that ratios in the extreme right tail of math ability in the U.S. have shrunk in the last 20 years (still favoring males) and remained relatively stable in the verbal domain (still favoring females). Similar patterns of male-female ratios in the extreme right tail were found in the Indian sample.

Introduction

In a world where accessing global talent is at a premium, developing such talent is crucial (National Academy of the Sciences, 2010). Despite this importance, there have been numerous reports that half the population's math and science talent is being underdeveloped. The female half. A review of recent research yields two themes. First, representation of women is improving on many indicators. Several recent studies suggest that representation is improving and that many biases against women in Science, Technology, Engineering, and Math (STEM) fields have been greatly reduced (e.g., Ceci et al., 2014, Miller and Wai, 2015, Williams and Ceci, 2015, but cf. Cell Associates, 2010, Moss-Racusin et al., 2012). However, greater equity in societal and organizational action may not be the same as sex parity. Second, parity has not yet been achieved on all fronts. For example, women hold only about 7–16% of tenured faculty positions, and < 30% of doctorates and bachelor degrees in math-intensive fields (Ceci et al., 2014). At that level, sex parity has not been achieved.

Some, (e.g., Lynn and Irwing, 2004, Lynn and Kanazawa, 2011) have posited that there are sex differences in general intelligence with a male advantage appearing in adolescence. However, others have disputed such findings based on the exclusion of relevant data (Blinkhorn, 2005) and the types of test used (Colom et al., 2004) while others have found contradictory evidence (e.g., Keith et al., 2008). Moreover, population level analyses have also revealed greater variability in males than females (Johnson et al., 2008).

Sex differences in domains of interest may explain some of these differences in accomplishment. Women are far more likely than men to report organic interests and in working with people, whereas men are far more likely than women to report interests in inorganic things (Su et al., 2009). Relative ability in specific domains also plays a critical role in pursuing and succeeding in careers relying on those same abilities (e.g., Kell et al., 2013, Makel et al., 2016, Robertson et al., 2010). Even within select samples, higher scores in a domain are associated with higher achievements in that domain. Because of this positive relationship, sex differences in score performance can play a large role in subsequent pursuit and performance in careers in those domains.

Although sex differences at the average have largely disappeared in many different domains (Hyde, 2005, Hyde and Linn, 2008), performance differences in the extreme tails remain (Johnson et al., 2008, Wai et al., 2010). Within the education community, a longstanding question concerns how ability in specific domains manifests at different rates for males and females (e.g., Benbow and Stanley, 1983, Ceci and Williams, 2010, Wai et al., 2010, Wai et al., 2012).

Benbow and Stanley, 1980, Benbow and Stanley, 1983, relying on SAT scores from 40,000 talent search participants (all taking the SAT before turning 13), reported that the male–female ratio in the early 1980s on the SAT-Math was approximately 2 to 1 for scores ≥ 500 (top 0.5%) and roughly 13 to 1 for students scoring ≥ 700 (top 0.01%). In an update that tracked male-female ratios in talent search participants over 30 years that also added in performance ratios on tests of verbal ability, Wai et al. (2010), in an independent sample, replicated Benbow and Stanley's 13:1 male-female ratio in the extreme right tail of math ability for the early 1980s. They also reported that this ratio dropped precipitously to approximately 4:1 by 1990 where it then remained roughly stagnant for the next two decades. At the same time, Wai and colleagues found several small female advantages in performance on verbal measures of ability on the SAT and ACT. In a slight update, Wai et al. (2012) added ACT EXPLORE test (designed to be taken by 8th grade students) performance data for elementary 4th–6th grade students and found similar math and verbal sex ratio differences from younger students taking above-level tests. Using an independent similar sample, Olszewski-Kubilius and Lee (2011) found similar sex ratio differences from comparable talent search participants in the Midwestern U.S. from 2000 to 2009.

Previous comparisons of male-female performance at the extreme right-tail were based on Western, largely U.S. data. However, in an analysis of a decade's worth of PISA scores, Stoet and Geary (2013) found significant sex differences across different domains, but found no evidence that such differences were related to other gender equality indicators for each country. More recently, Stoet and colleagues also reported that PISA participating countries that had high levels of gender equality overall also showed higher levels of sex differences in math anxiety as well as parents ascribing relatively lower value on math for girls than for boys (Stoet et al., 2016). Penner (2008) reported similar data, but reported that gender equality measures, such as whether men were perceived as having higher status than women, were related to female representation in the right tail of math performance.

However, it should be noted that despite including international samples, these papers still represent a largely western and European sample. Other countries with differing cultural patterns around sex may have differing score and career trajectory patterns. For example, in India, data suggest that there is a consistent cultural bias against females. This bias is manifested as early as the birth rate, which favors males (Sen, 1992, Sen, 2003). Literacy rates in India also favor males (15–24 year old males: 88.4%; females: 74.4%), as do participation rates in secondary (males: 73.3%, females: 69.4%) and tertiary (males: 26%, females: 20.3%) education (UNESCO, 2014). In 2015, representation of females being admitted to India's prestigious Institutes of Technology for college education rose to 9% from 8% the previous year, but only 18% of students who registered to take the entrance exams were female (Rao, 2015). Similarly, Indian women appear to be participating in STEM careers at relatively low rates, but special programs have been recently developed to remove barriers (Leggon et al., 2015). However, some argue that highly educated Indian females are doing quite well in both high level STEM and business careers (Hewlett & Rashid, 2011).

The extent to which such sex differences in performance extend to the extreme right-tail of cognitive abilities among younger students in India has not been established. In the U.S., which enjoys relative sex equity in birthrate, literacy, and general education participation, there still do exist some consistent and large sex differences in the extreme right-tail of ability in various domains (e.g., Wai et al., 2010). Whether the apparent cultural differences seen in India exacerbate, diminish, or do not affect male-female ability ratios in the extreme right-tail relative to those found in the U.S. has yet to be explored. Studying sex patterns across different cultures may help unravel universal sex differences across different cultures, which can serve as a helpful benchmark when assessing underrepresentation in STEM domains.

Given that performance differences in math at an early age, even within the top 1%, have been associated with achievement differences in later high level STEM achievements such as earning PhDs, publications, patents, and university tenure (Makel et al., 2016, Park et al., 2008, Wai et al., 2005), male-female scoring differences may be linked to later STEM outcomes and should be monitored closely. These findings may also generalize to the Indian context and suggest one way performance differences in math or other measures may be of relevance to the discussion of Indian females and males in high level STEM fields.