The IQ Controversy

There are a number of reasons why the measurement and interpretation of a ‘Quotient Intelligence (IQ)’ is controversial. Some believe that the measurement of IQ is unreliable in that it does not necessarily represent the true nature of human intelligence. Others for reasons that might be described as political correctness believe that it may fuel racism that could ultimately lead to the dark idea of eugenics.

Note: At the start of this discussion, it will be stated that while an IQ score may give an indication of general intelligence, it does not necessarily provide an accurate or complete measure of the complexity of all human thought processes. Without being too detailed at this stage, there are aspects of intelligence related to short-term memory, reasoning skills and verbal ability, which IQ tests may not accurately assess. Likewise, we might cite other abilities, such as creativity, emotional sensitivity or even social awareness, which are also not included in most IQ assessments – see The Theory of Multiple Intelligence for more details.

While this introduction may appear to throw doubt on the very concept of IQ as a measure of intelligence, we know from everyday experience that we can often make a general assessment of a person’s intelligence, which while possibly subjective and not very scientific, turns out not to be so wide of the mark. However, the accuracy of such judgement implies some probability, e.g. we are usually 60% right and 40% wrong.

Note: The issue of probability needs some understanding of mathematical statistics, although this discussion will not directly address this complication. However, the interested reader might wish to review an earlier discussion entitled Introduction to Statistics, which also considers the issue of Distribution and Standard Deviations plus Correlations and Regression. The reader might also want to review the discussion entitled ‘Social Evolution’ by way of initial background to the IQ debate to follow.

The idea of ‘correlation’ can be particularly important as it can often help quantify an association between two observations. For example, any measure of intelligence may not necessarily be correlated to whether a person has a ‘kind personality’, but may have a stronger correlation to whether a person will be ‘successful in life’. Of course, at a general level, we realise that when we infer a statistical probability, this is not certainty, especially when discussing the probability curve of an entire population. By way of an initial example, the following chart suggests a correlation between job types and cognitive intelligence, which while not implying certainty in any individual case may be generally accurate as a statistical probability.

Note: While there are any number of different types of distribution curve, the curve shown above corresponds to a Gaussian function, which is sometimes called a ‘bell curve’, because of its symmetrical shape. In this type of curve, the majority of the population sample is centralised around the peak, where IQ is lower on the left and higher on the right.

So, while this discussion has highlighted a certain amount of caution about reading too much into an IQ score without context, the scientific field of psychometrics has focused much attention on assessing skills, knowledge, abilities, attitudes and personality traits and correlating these factors against educational achievement. In the context of this research, the idea of a general intelligence factor, or g-factor, was first proposed by psychologist Charles Spearman in the early 20th century. This idea was originally based on the performance of children across a broad range of subjects and indicated a positive correlation to a general assessment of mental ability. Later research then suggested that the g-factor also appeared to show a positive correlation with an individual's performance across most cognitive tasks.

Note: The terminology of IQ, general intelligence, general cognitive ability, general mental ability or simply intelligence are often used interchangeably, although the g-factor targets a more specific measure of general intelligence. Empirical research on the nature of the g-factor has also involved much research into cognitive psychology, brain physiology, molecular genetics and primate evolution. Today, the g-factor is generally assumed to be uncontroversial in the field of science, irrespective of whether controversy exists in the social science when viewed in terms of political correctness.

Today, much of the controversy surrounding the issue of intelligence is often characterised in terms of the ‘nature versus nurture’ debate. Of course, while we might reasonably assume that most skills have to be learnt, such that they might be categorised as nurture, we might also recognise that different individuals may have genetically inherited a higher cognitive ability that allows them to learn more quickly.

Note: At this point, the previous introduction has not really touched on anything that might be considered too controversial. Clearly, some individuals are smarter than others in the same way as some are taller or smaller. However, the nature of the controversy starts to become more obvious when we attempt to apply a probability distribution to an entire population, although the level of this controversy often depends on the trait in question. For example, nobody really questions that some populations may have differences in physiology, i.e. taller or shorter, but extending this idea to suggest that some populations may be statistically more intelligent is an entirely different matter. Of course, the primary issue of debate is whether these population differences are caused by nature or nurture.

We might now attempt to characterise the controversy in terms of two publications, which were both heavily based on the use of statistical analysis but were quickly perceived to have far wider political and economic consequences that many simply dismissed as politically incorrect, if not racist in their conclusions.

While this discussion cannot address the full scope of the Bell Curve book, we might still outline its scope without initial reference to the controversy that can polarise opinion. The book as cited above was published in 1994 and attempted to explain the variations in intelligence in US society and warn of some of the consequences. From a social research perspective, the book also attempted to forward some potential social policies that might help mitigate the worst of the consequences stemming from the IQ distribution that appear to exist in the US population. For the purposes of this discussion, we will only list the assumptions the authors claimed to be statistically relevant – see 1994 interview with Charles Murray for more details.

  1. There is a difference in the general cognitive ability of humans.
  2. All tests measure this general factor to some degree, but IQ tests designed for this purpose are most accurate.
  3.  IQ scores match, to a first degree, what people generally understand by the word ‘intelligence’.
  4. IQ scores are stable, although not perfectly so, over much of a person's life.
  5. Properly administered IQ tests are not demonstrably biased against social, economic, ethnic or racial groups.
  6. Cognitive ability is substantially heritable with a correlation no less than 40% and no more than 80%.

Note: The correlation range of inheritance is often disputed and therefore considered further later in this discussion. However, the authors also warned of inferring things about individuals based on the data presented, as they suggested that intelligence is only one of many valuable human attributes and one that might possibly be overrated.

The Bell Curve consists of 4 sections, where the first 2 might be accepted as relatively uncontroversial, but where section-3 entitled ‘The National Context’ starts to raise some very contentious issues under the 4 sub-section headings listed below:

  • Ethnic Differences in Cognitive Ability
  • Ethnic Inequalities in Relation to IQ
  • The Demography of Intelligence
  • Social Behaviour and the Prevalence of Low Cognitive Ability

Now, we might immediately perceive how these titles might give rise to controversy, irrespective of whether the statistical data was correct or not. However, at this point, we shall only quote the introduction to the first section listed above and suggest that the interested reader review the book in full before jumping to any conclusions.

“We now turn to the national scene. This means considering all races and ethnic groups, which leads to the most controversial issues we will discuss: ethnic differences in cognitive ability and social behaviour, the effects of fertility patterns on the distribution of intelligence, and the overall relationship of low cognitive ability to what has become known as the underclass. As we begin, perhaps a pact is appropriate. The facts about these topics are not only controversial but exceedingly complex. For our part, we will undertake to confront all the tough questions squarely. We ask that you read carefully.”

The second book entitled ‘IQ and the Wealth of Nations’ was published in 2002 in which the authors also forwarded the suggestion that there are different IQ distributions between national population caused by both genetic and environmental factors. They then go on to argue that the difference in IQ can result in a lower GDP, where the wealth of the nation, or lack of it, might in-turn result in lower than average IQ. At first glance, this argument does not appear to refute the argument that both nature and nurture can affect IQ, although it forwards the more controversial idea that a statistical average of IQ can not only vary between nations, but different ethnic groups. See video featuring Dr. Lynn & Dr. Rushton in 2002 for more detail.

Note: However, if we simply ignore those who might wish to promote racism or the unethical ideas often associated with eugenics, we might then be able to consider the reality of whether the combination of nature and nurture might lead to a statistical spread in the IQ distribution between nations and ethnic groups. As a general comment, political correctness does not seem to object to the observation that African ethnic groups appear to dominate many athletic sports, which we might assume has more to do with nature rather than nurture. If this is true, then the idea that IQ might vary due to both nature and nurture hardly seems controversial, such that the real issue of debate comes down to the quantifying of the correlation of IQ associated with both nature and nurture.

Having provided the framework in which this discussion wants to proceed, it will be suggested that IQ variation does exist across different populations, although its correlation in terms of both nature and nurture might still be debated. We might also have to accept that the measurement of intelligence is not an exact science, but the idea of an IQ score is not an unreasonable proxy measure, which might then be aggregated as the average IQ for some population. However, as always, it will be stressed that a statistical average cannot be applied to any specific individual within any population. In the case of ‘IQ and the Wealth of Nations’ , the authors forward a finding that suggested a statistical correlation coefficient between 0.50–0.75.

Note: For the purpose of this discussion, it will simply be stated that a correlation coefficient is a measure of the relationship between two variables, e.g. IQ and nature or IQ and nurture. The correlation coefficient is measured on a scale between [-1] and [1], where the absence of any correlation is represented by [0].

So, a correlation between IQ and the wealth of a nation in the range 0.5-0.75 might be interpreted as ranging from moderate to strong, although it needs to be highlighted that this correlation is making no reference as to whether IQ is caused by nature or nurture. However, the authors did infer that intelligence is biased towards nature rather than nurture, which many have subsequently questioned and criticised. One of the common criticisms is that IQ tests favour western cultures, such that a specific culture or ethnicity group might be disadvantaged plus it ignores the potential impact of malnutrition and other forms of deprivation, e.g. lack of education, which might bias IQ results against impoverished nations. In this context, we might wish to question variations of IQ scores across European populations, if we make the reasonable assumption that these populations would have the same essential genetic makeup, If so, then we would have to question the genetic interpretation inferred by the authors based on the populations of West and East Germany. In this case, the IQ results for West Germany suggested IQ scores in the range 99–107, while the equivalent IQ score for East Germany, prior to unification, was estimated to be as low as 90. However, similar studies just a few years later, after reunification, suggested IQ scores for East Germany in the 97–99 range, which appear anomalous, if based solely on the genetic nature rather than the change in the nurturing environment. Therefore, at this point, we possibly need to provide some clarity on the genetic nature of IQ and the concept of heritability.

What might be inferred by the heritability of IQ?

 

The measurement of IQ in people that are related can provide some insight as to the IQ associated with nature, i.e. genetics, as opposed to nurture, i.e. environment. As such, heritability defines the genetic contribution associated with specific traits within a population, which might then be statistically quantified with a value between [0] and [1]. The study of twins, as reflected in the chart above, shows the correlation of IQ scores, not actual IQ scores, with the relationships between sibling types plus the parent/child relationship. In the first column, we see that identical twins living together have a higher IQ correlation than twins living apart. In the next column, we see that fraternal twins of the same gender have a higher IQ correlation than fraternal twins of opposite gender. In the third column, we see siblings living together have a higher IQ correlation than siblings living apart. Finally, in the last column, we see that children living with their biological parents will have a higher IQ correlation than children living with foster parents. However, the bullet points below highlight some additional comments on each variation.

  • In the case of fraternal twins that share almost identical DNA, the very high correlation of IQ seems to suggest the importance of their inherited DNA, which is slightly reduced if these twins are reared apart, which appears to suggest that nature dominates nurture. However, what is not obvious is the degree of difference between the two ‘reared apart’ environments, which conceptually range from first world to third world. While the innate intelligence of both identical twins might be similar, it would seem possible that if one twin was reared in a third world environment, the acquired knowledge and skills may be severely compromised.

  • In the case of the fraternal twins, who may have significantly different DNA, the correlation of IQ is much lower, although possibly still significant. However, this correlation is slightly further reduced, if the twins are of different gender. Here the IQ difference between fraternal twins of different gender may simply be reflecting the general IQ differences between genders – see link entitles Sex differences in Intelligence for more details on this specific issue. However, this group still suggests the role of nature.

  • In the case of normal siblings, we might question why IQ is any different to fraternal twins given that both cases imply different DNA makeup. One possible reason may be linked to the fact that fraternal twins are identical in age and therefore will probably have very similar nurturing environments, while normal siblings with a significant age gap may be subject to differing nurturing environments at home and at school, although this is just speculation at this stage. In this group, we might see a stronger argument for the role of nurture.

  • Finally, in the case of the parent child relationship, the child growing up with natural parents has to some extent inherited aspects of their DNA, while a fostered or adopted child will not. Again, the difference might be suggesting the role of nature over nurture, although we might assume that both are factors.

Again, it needs to be highlighted that correlation coefficients have to be interpreted with some care for two basic reasons. First, they only infer a statistical relationship between the variables selected and, second, high correlation may not always be linked to a direct causal relationship. In this context, aspects of the nature versus nurture debate are not necessarily conclusive, so while it might reasonably be assumed that genetics plays an important role in inherited intelligence, nurture must still have a role to play in the acquisition of knowledge and skills. This is possibly of more importance when considering the average IQ scores for different ethnic groups, where malnutrition and poor education may be significant.

Note: The Wilson Effect suggests that environmental factors can dominate the measure of IQ in early years, when genetic factors may only have a 20% influence, but grows during adolescence such that genetic effects may come to account for 80% of an IQ score in adulthood. As such, this suggests a correlation weighting biased toward the underlying genetic nature of intelligence in adulthood.

The following graph has been produced from the data listed in the website WorldData.info. It is stated that the IQ data for the 110 countries is the average of 9 international studies and compared to the average income and government expenditures on education for the years 1990 to 2010. While the average income has changed in recent years, especially in many developing countries, we might simply view the data as a snapshot by which to consider the correlation of IQ with income and education.

In the chart above, we see that income per capita, shown left in red, and the expenditure on education, shown right in black, appears to have a very high positive correlation. We might reasonably assume that income per capita is probably reflective of the national GDP and the wealth of the nation, such that wealthier nations can afford to spend more on their education programs. Of course, the key point of interest is how these two variables are correlated with the IQ of each nation. While the actual countries are listed in the link above, the horizontal axis plots the average IQ of each nation against income in red and education in black, but appears to ‘paint’ a more complex picture of the relationship between these variables. Without attempting to be too rigorous in the analysis of this chart, the correlation of IQ in some countries with income appears weak, although there does appear to be a trend suggesting that IQ is generally higher in wealthier nations. Of course, there are undoubtedly a myriad of other factors, both nature and nurture, which influence these results that this graph is simply not taking into account, as such, the point of this graph is simply to highlight that caution in any interpretation is necessary.

So, can any sort of conclusion be reached about the IQ controversy?

Clearly, this type of general discussion cannot resolve the controversy debated by so many authoritative sources, although it might attempt to rationalise a position on the nature versus nurture debate in terms of a number of incremental assumptions. First, it is assumed that nature and nurture are both contributing factors to what we perceive as human intelligence. Second, it is assumed that IQ tests provide a reasonable first-order measurement of intelligence in a given individual. Third, an aspect of this intelligence is an innate ability based on genetic makeup. Fourth, innate intelligence without a nurturing environment cannot develop.

OK, but can the correlation coefficients for IQ against nature and nurture be accurately determined?

If the four assumptions were generally accepted, we might also accept that most of the controversy centres on the question above, because determining the statistical accuracy of these correlation coefficients are not only complex, but possibly somewhat subjective. For example, various sources suggest anywhere between 50-500 genes that might contribute to intelligence.

Note: In 2018, researchers at UK Biobank compared the DNA of more than 240,000 people. This research identified 538 potential genes linked to intellectual ability in 187 regions of the human genome associated with thinking skills. However, many of these genes are also linked to other biological processes, such that correlation is very difficult to evaluate. However, as outlined, twin studies suggest that inherited intelligence is highly dependent on genetics, possibly in the 40-80% range, although still dependent on a good nurturing environment.

While more anecdotal in scope, parents who have had more than one child quickly come to realise that their children can have very different personalities from the outset and, while having very similar nurturing environments, end up with very different intellectual abilities. While this seems to also bias the correlation factors towards nature, it is clear that nurture is important, but often difficult to quantify in terms of a correlation factor. However, this issue has been much studied from the perspective of behavioural science and amassed considerable evidence that intelligence can be negatively influenced by a poor nurturing environment, where some examples are simply listed below.

  • Malnutrition at any stage of life but particularly in early development can lead to lower IQ.
  • Exposure to toxins, trauma, stress or illness during critical developmental periods can all affect IQ.
  • Insufficient education can also lead to lower IQ has it hinders cognitive development.
  • Other factors include low socio-economic status with the additional inference of poor parenting.

While we will not necessarily pursue these issues, section-2 of the Bell Curve entitled ‘Cognitive Classes and Social Behaviour’ provided the following correlation factors to IQ that might also be seen to be reflective of the nurturing environment.

IQ <75 75–90 90–110 110–125 >125
Population distribution 5 20 50 20 5
Married by age 30 72 81 81 72 67
Frequent Unemployed 12 10 7 7 2
Divorced in 5 years 21 22 23 15 9
Live in poverty 30 16 6 3 2
Ever incarcerated 7 7 3 1 0
Welfare recipient 31 17 8 2 0
High school dropout 55 35 6 0.4 0

But what about the contentious issue of IQ distribution across different populations?

We might start by first considering a conceptual distribution of IQ across two different populations, i.e. yellow and red, which also makes reference to the training and career potential within these distributions. Clearly, the inference is that individuals in the red distribution have a higher statistical probability of getting a better paid job, which in-turn will have an impact on their socio-economic success in life. At this stage, we need make no inference about what separates these two populations, only that the IQ disparity shown does exist.

What might result from this difference in IQ distribution?

Again, we might stress that there are individuals in each population across the entire IQ range, although we have to accept that the actual numbers in each IQ range differ for each population. If so, then probability suggests that there will be a greater number of individuals in the yellow population, who may struggle for socio-economic reasons. If this greater number of lower IQ individuals then have children, heritability of IQ statistics suggest that these children may not only have a lower IQ due to genetics, but are probably more likely to suffer from the problems of a nurturing environment as outlined in the previous table.

But how might we group and separate different populations?

It might be realised that political correctness might refute the very basis of this question, if the assumption is that there cannot be any differences. However, it is difficult to refute that people within the same population do not have different IQ, otherwise there would be no IQ distribution curve. Likewise, it is difficult to refute that there are no physical differences between different populations across the world, as statistics appear to show that African populations dominate in many fields of sport. Of course, if you accept these physical differences and the brain is a physical construct, then simply refuting the very idea that differences in IQ exist between different populations becomes problematic. However, even if we accept such assumptions, we have to return to the issue of whether any IQ difference is caused by nature or nurture, although this issue does not prevent the measurement of IQ at any point in time.

Note: What is often forgotten in the IQ controversy is that the original purpose of the IQ tests, as developed by Alfred Binet in 1904, was to facilitate a method of evaluating children who required, and would benefit from, special tutoring – see IQ Tests for more details. Likewise, much of the focus of the research in the Bell Curve was orientated towards developing more effective social policies to help different groups in the US.

Let us proceed on the assumption that individual IQ can be measured as a first-order approximation and then broadly collated to represent different population groups without necessarily knowing the correlation weights of nature or nurture. If so, then these IQ statistics would still be representative of the general IQ status in a given population, which if suffering from any of the problems identified as being affected by nurture might then be addressed. While sources and metrics for IQ differ, the general trend in the chart below might be considered generally reflective of IQ when separated into different population groups, although some may undoubtedly contest these results based on 2015 SAT scores.

Note: The chart above was presented in an article by Reeves and Halikias in 2017. They explain that, in the US, SAT scores are considered critical in identifying when a student is ready for college. From a statistical basis, these scores are representative of 1.7 million students in 2015. The mean score on the math section of the SAT for all test-takers is 511 out of 800, while average scores for different population groupings were listed as blacks (428) and Latinos (457), which were significantly lower than those of whites (534) and Asians (598).

As previously highlighted, these scores do not tell us whether they reflect nature or nurture, only that they appear statistically representative of different groups within the US population in 2015. Of course, we might reasonably assume that the students in these different groups may have grown up in very different socio-economic environments. However, the following chart also showing SAT scores, but dating back to 1995, might question whether economic income is a root cause for while SAT scores improve with income, the differentials between the various population groups appears to remain fairly consistent.

If you accept that this discussion has NOT been driven by a racist agenda and has simply tried to present some of the statistical data that is readily available from many different authoritative sources, then simply suppressing the discussion of the issues outlined will not allow the underlying problems to be addressed. In part, an aspect of this position was addressed in 2005 in a paper entitled ‘Thirty Years of Research on Race Differences in Cognitive Ability’ by Rushton and Jensen. While the link allows the paper to be reviewed in full, the following quote is representative of the closing comments under the heading ‘Conflicting Worldviews’.

A prevailing worldview throughout history has been that economic, cultural, and other environmental forces are the preeminent causes of group and individual behaviour. Modern social science has typically taken this perspective and promoted the idea that all babies are born more or less equally endowed in intelligence and learning ability. It followed therefore that inequalities were the result of social, economic, and political forces. This worldview generated many strategies for intervention in the home, the workplace, the mass media, the criminal justice system, and even the entire social–economic system. Some have been effective and are almost universally accepted, whereas others have failed and produced only shattered expectations, resentment, and inter-ethnic hostility. The major policy implication of the research reviewed here is that adopting an evolutionary genetic outlook does not undermine our dedication to democratic ideals. As E. O. Wilson aptly noted: “We are not compelled to believe in biological uniformity in order to affirm freedom and dignity”. He went on to quote the sociologist Bressler: “An ideology that tacitly appeals to biological equality as a condition for human emancipation corrupts the idea of freedom. Moreover, it encourages decent men to tremble at the prospect of ‘inconvenient’ findings that may emerge in future scientific research”. Denial of any genetic component in human variation, including between groups, is not only poor science, it is likely to be injurious both to unique individuals and to the complex structure of societies.

Unfortunately, this discussion of the IQ controversy seems to reflect a modern-day aversion towards many ‘inconvenient truths’, whether in the field of social science or climate science. In this context, we might see parallels in the last sentence above and the words of William Clifford in his essay entitled ‘Ethics of Belief’ from which the following quote is taken.

The danger to society is not merely that it should believe wrong things, though that is great enough; but that it should become credulous, and lose the habit of testing things and inquiring into them, for then it must sink back into savagery. It may matter little to me, in my cloud-castle of sweet illusions and darling lies; but it matters much to Man that I have made my neighbours ready to deceive. The credulous man is father to the liar and the cheat.