Hedging : The IQ Paradox Resolved? A Critical Analysis

The IQ Paradox Resolved? A Critical Analysis

The Flynn effect, or IQ paradox, is simply the rise in mean IQ scores during the 20th century. The paper first goes into some detail on what intelligence is and how it is measured. Then a critical review and analysis of the Flynn effect and the Dickens-Flynn (2001) models is discussed. The basic idea of these models is that environment, with hereditary, influences IQ and IQ in return affects that environment. The conclusion is that the model may be appropriate for explaining the rise in IQ scores in future years but there is still major work that must be done on it.
The IQ paradox is simply the large gains over time in intelligence test scores (Flynn, 1999). The Flynn effect is Flynn's explanation of the rise in mean IQ scores during the 20th century (Rowe & Rodgers, 2002). He and his colleagues were the first to notice this rise and it is now fairly accepted that there is one. It has been shown and is universally accepted that word knowledge has risen significantly in the last 20 years by about 5 verbal IQ points (Nettelbeck & Wilson, 2004).
The Dickens-Flynn models try to explain this rise and attribute it to a model incorporating both heritability and environment. The models are recursive models of IQ growth in which phenotypes and their supportive environments are correlated (Dickens & Flynn, 2001). The basic idea is that a person's environment affects their IQ and that IQ affects their environment. This suggests that small environmental influences produce large changes in IQ and that most environmental effects are short-lived. The consequence of this is that improving childhood IQ does not necessarily mean that children will have high IQ when they are adults.
Though the Dickens-Flynn models posses a very interesting explanation and may be proven as acceptable models in the future, there is not enough evidence to suggest that it should be accepted universally as the best explanation for rise in IQ with what is known about it currently. Although the models do an adequate job at explaining the effects of environment on IQ, there are too many unknowns in the theory. The most important criticism may be that this model lacks enough experimental validation (Rowe & Rodgers, 2002). There really have not been many recent studies done, so the data it is based on is outdated. The studies that have been done have had problems of their own with keeping basic statistical assumptions true (Mahlberg, 1997). There are also some flaws with the models themselves.
With all of this in mind it is clear that a stance on the subject can be taken. The Dickens-Flynn models while well thought out and apparently noteworthy, have some major flaws. The models as they are can not be accepted for major statistical and theoretical problems. Maybe once they address these problems, they can be accepted in mainstream science as a valid explanation and some day solve the IQ paradox.
Background on IntelligenceIntelligence is defined as a general mental capability that involves the abilities to reason, comprehend complex ideas, plan, solve problems, learn quickly, think abstractly, and learn from experiences (Gottfredson, 1997). There are several issues that most scientists agree on and in 1994 the Wall Street Journal published an article, "Mainstream Science on Intelligence," which covered everything that has been universally accepted thus far, on intelligence and IQ tests. Heritability of intelligence ranges from .4 to .8 on a scale from 0 to 1. Unfortunately, there is little known about what brain processes are used in intelligence so there has been little advancement on being able to manipulate it to raise IQ permanently.
IQ is also known to lie a long a bell curve, or normal curve. The average IQ is 100, where above 130 is giftedness and below 70 is retardation. IQ scores are an accurate measure for everyone regardless of socioeconomic class or race. The averages for these groups are not the same though (Williams & Ceci, 1997).
Intelligence is a broad reflection of the capability to comprehend a person's surroundings and in this form it can be measured (Neisser et al., 1996). The intelligence tests of today measure it very well. IQ tests measures intelligence so well, that these tests are among the most accurate, reliable and valid, in psychology. There are two kinds of intelligence tests, word and numbers, or shapes and designs. The words and numbers test is culture bound and expects certain knowledge of the culture such as vocabulary. On the other hand there is the shape and design test which only requires knowledge of universal concepts such as open vs. closed. Both of these tests measure the same thing though; what is commonly known as the 'g factor.'
There are still many different intelligence tests (Wainwright, Wright, Geffen, Geffen, Luciano, & Martin, 2004). The most current, universal, and widely accepted test is the Wechler's Adult Intelligence Scale Revised (WAIS-R) (Kaufman, 1990). In 1939, the Wechsler-Bellevue From I (W-B I) was created by David Wechsler. This test was redesigned and reformed in 1955 and became the Wechsler Adult Intelligence Scale (WAIS). This was also edited in 1989 to become the WAIS-R. This test is now the most widely used and accepted intelligence test. Some newer and not so well known tests are the Woodcock-Johnson Psycho-Educational Battery-R (WJ), the Standford – Binet Intelligence Scale, Form IV, and the Kayfman Adolescent and Adult Intelligence Test (KAIT).
The Dickens-Flynn ModelsNow there needs to be some explanation of definitions so that the Dickens-Flynn models can be best explained. For this paper the environmental effects that are in the model are those effects that are "short-lived" (Dickens & Flynn, 2001). Most environmental effects decay over time. Attributes such as nutrition or socioeconomic class are the exception to this decay and are long term. These effects, in Flynn's eyes, rival the effects of genes and are not included under his idea of environmental effects. This means that when he argues that IQ affects environment and environment affects IQ, he is referring to the "short-lived" environment.
The Dickens – Flynn models boldly try to explain the IQ paradox and attribute it to a model with both large heritability and environmental effects. The models are recursive models of IQ growth in which phenotypes and their supportive environments are correlated (Dickens & Flynn, 2001). The idea is that environment affects IQ and that IQ affects environment. This suggests that small environmental influences produce large changes in IQ and that most environmental effects are short-lived.
There are many interesting features that the Dickens – Flynn models allow. The first and most important feature is that the models allow for a large environmental effect while still maintaining heritability. This is done by the theory that the environment has an effect on IQ and then in exchange IQ has an effect on that environment. Remember that he is referring to short-lived environment.
The second feature is that the Dickens-Flynn models can account for other phenomena. As long as the trait in question matches itself to an environment, that trait will behave like the models. The other condition for this is that the environment effect can not accumulate over time. Even with this restriction this feature is a good one statistically. It means that in the future other things regarding environmental change can already be explained statistically.
The third feature explained in Flynn's theory is why a child's IQ does not determine the adult's IQ. This is because environmental effects decay over time. In order to have a child's high IQ last, it is suggested that programs should encourage the continuance of replication of lessons long after they are done with the program (Dickens & Flynn, 2001). Basically, when a child leaves a learning program, the parents or guardian should help the child continue with the practices they learned (i.e. organizational skills or study cards).
CriticismThe criticism against the model is in many parts. The first is that there has been no explanation of the rise in IQ universally accepted (Rowe & Rodgers, 2002). The rise in IQ, the IQ paradox, is widely accepted and known, but what is causing that rise is in question. There have been several suggestions; Flynn's is just one of them (Uttl & Van Alstine, 2003). Since the publication of The Bell Curve by Herrnstein and Murray in 1994, IQ has been in question by professionals and non professionals alike. So since the IQ paradox was formed there has been some debate on what affects IQ. This in tern creates tension between the different ideas on what has caused the rise in IQ thus providing no clear agreement on the solution of the IQ paradox.
The next criticism is that so far, experiments dealing with the Dickens – Flynn model have either had statistical problems or were unable to replicate (Nettelbeck & Wilson, 2003). This could mostly be due to the fact that Dickens and Flynn used less than current data to obtain their model and is no longer representative of the current population (Rowe & Rodgers, 2002). When evaluating whether or not the models would fit outside data there were other problems. Statistically speaking this is very bad.
The results could also be attributed to other factors, which is the third criticism. Most of the other factors are the long term environmental effects. Even though the Dickens – Flynn models allow for long term environmental effects, there may not be a significant enough allowance for change over time. An example of this is nutrition. America's nutrition, and most of the world's, has been improving for the past century. More people are eating more and better than they ever did.
The fourth is that the models do not specify a time line (Loehlin, 2002). Dickens and Flynn neglect to say if the change in IQ occurs during a person's life time or if these models are gauged to a broader population change. There are also no explanations for time gaps. An example of this is: suppose if a person learned to ride a bicycle and did so between the ages of 5 and 10. Then they stopped for 10 years. Would that person have to start over again to learn how to ride the bike or is the intelligence engraved in their mind? Dickens' and Flynn's models would say the person would have to relearn the skill, which does not seem logical.
The last and maybe the most important is that the model allows for large changes in IQ variance where historically there is no real change in variance at all (Rowe & Rodgers, 2002). This presents a very big and unwanted statistical blunder. There are two major things in statistics that most be accurate, the mean, and the variance. Since the variance increases over time the effect known as heteroscidasticity will occur. This basically means that at first the correlation between IQ and environment is constant and strong, but as time moves that correlation becomes weaker and eventually will become nonexistent. The conclusion is that the models are bad predictors and the heteroscidasticity must be solved before the models can be recognized as statistically sound.
ConclusionFor the most part, it is universally accepted that there has been a rise in IQ level. The real argument is what is causing this rise. There are arguments that range from a Freudian explanation of collective consciousness to genetic development and adaptation (Mingroni, 2004). The Dickens-Flynn models seem to be one of the most popular explanations thus far but are still a little rocky. From the analysis of the model there really seems to be too many statistical and theoretical blunders made in the development and implementation of the models. The fact that there are major statistical errors made in the models themselves is enough to discredit the theories.
Even though Flynn's theories about the IQ paradox seem appropriate on the surface, there are major theoretical issues. The first is the fact that the theory only allows for IQ to affect short-lived environmental effects. As far as that is concerned, long-term environment can be affected by IQ. An example of this is that IQ is known to correlate with socioeconomic status. Even though long term effects are in the Dickens-Flynn models it does not include it in the interaction between IQ and short-lived environment.
Even with these blunders these models could be the best explanation of the rise in IQ thus far. But like all statistical models it needs more time and studies to validate it and work out all of the kinks. If this model was reconsidered with some changes and major experiments, it may be the best theory to explain the IQ paradox yet.
Peer Commentary
Childhood Exposure to Technology May Increase IQ ScoresAllen J. HolmesRochester Institute of Technology
Flynn studied the IQ scores of separate generations of people. He discovered that, on average, IQ scores raised 3 IQ points over each decade. His studies were extensive. He analyzed data from 20 countries. Some countries displayed rapid intellectual increase of up to 20 IQ points per generation of people. This is a phenomenal increase, which suggests that the Flynn effect has probably not been occurring at all points in human history. It must be a relatively new phenomenon. If we were to assume that the Flynn effect has been occurring throughout human history, the statistical curve would suggest that the average person born only 100 years ago would be considered mentally challenged by today's standards. So what are the causes of the Flynn effect displayed in the Dickens-Flynn model?
The first cause may be food consumption. Western societies are extremely well off in food supply. The majority of people in Western culture have the choice of refusing food. This was not always true and is still not true for many non-Western peoples. Despite this fact, it seems that only severe malnutrition affects the IQ score of an individual.
Perhaps the tendency for couples to have fewer children has a direct effect on the IQ score of children. Parents with fewer children are able to spend more time with each child, thus stimulating their cognitive abilities at an early age, which is obviously beneficial to overall intelligence as well as psychological well-being. It is true that modern families have far less children than families from only 60 years in the past. I believe that this is a definite factor in the increase of IQ.
Another possibility that may increase the average IQ score over generations may be the effects of technology and the demand that technological devices impose on people. Comedians joke about how young children are able to teach their parents how to program a VCR. This is not unreasonable. Children are increasingly exposed to technology, which poses unique problems that may increase abstract problem solving ability and pattern recognition.
Elderly people often have difficulty learning how to use a personal computer, whereas a young person may find computers very intuitive and easy to operate. I have a theory for this. When a person is young, he or she has an increased capacity to learn new things. This has been shown in countless studies about the benefits of head start programs sponsored by government. Perhaps early exposure to technology-oriented interfaces increases a child's ability to adapt and understand complex and abstract problems.
Flynn's studies show that there is a much greater increase in non-verbal problem solving ability than word-oriented test scores. This would support my theory that childhood interface exposure increases pattern recognition and abstract problem solving ability. Complex interfaces are a problem with which only recent generations have had to deal. This would explain the reason for the Flynn effect's recent appearance.
As the everyday use of technology increases, it is reasonable to assume that the average IQ of an individual will increase as well. This will likely level off in the coming 50 years as the increase of incorporation of technology into everyday life reaches its high water mark.
Peer Commentary
The IQ Paradox: Is There an Answer?David J. MoultonRochester Institute of Technology
This paper raised the question of the IQ paradox, which is basically the drastic increase in IQ scores during the 20th century. This work is really just a critique of Flinn's explanation for why intelligence test scores have been increasing and therefore is not an answer to the question of why IQ scores are increasing.
The Dickens-Flynn model that was critiqued had many flaws. These were discussed briefly, but the author could have used more logic and explanation. For example, after briefly researching this model myself, I feel that it does not account for the content of IQ tests. That is, what the IQ tests are actually testing should be looked at. The paper lacks any discussion of this aspect of the problem. What about the fact that we live in the information age, and that a majority of people (in the USA) own or have access to a computer? Games and other special orientated activities can be done using this new technology and may have increased a person's intelligence in this area. The paper says that the Wechsler Adult Intelligence Scale was last revised in 1989 to become the WAIS-R test. At that time, an average person may not know what a computer was; let alone what can be done with one. Computer technology has advanced at an extreme rate, and it seems as though IQ tests have not been revised enough to account for this. Since the paper says that the WAIS-R test is the most widely used one today, and its last revision was in 1989, an argument can be made from this as to why IQ has been on the rise. A lot has changed since 1989 because of the popularity of the computer. This is an environmental aspect which supports the Dickens-Flynn model, but was not discussed in the paper. As a final note on computer influence, computers have been woven into education today. Many homework and teaching techniques are facilitated by computer technology. Using a computer to do work is fairly abstract and ambiguous, but since people can come into frequent contact with they may get used to doing these abstract activities, which may translate to better performance with abstract questions on an IQ exam.
Another criticism to the Dickens-Flynn model that seems easy to make is that of an environment thousands of years ago. You could assume that our environment is much better than the environment that people lived in that long ago, yet people of that time were very intelligent. For example, the Greek civilization in view of many people was the creator of mathematics and discovered many scientific principals. Surely it requires a high level of intelligence to do these things yet, technology and other environmental factors as we know it did not exist then. How would the model explain that? Methods of education seem to have been left out of a discussion as to why IQ scores are on the rise. Could a rise in IQ be attributed to better teaching methods? Or maybe teachers are just teaching for the test itself because they know what is on it? If that were the case, then IQ test scores would increase, but this may be a false representation of real IQ since these people have been taught to answer test question and may have missed certain concepts that otherwise would have been learned through a normal method of teaching.
One last point that I disagree with is the statement that Americans are eating healthier than ever. Heart disease is still one of the most frequent causes of death in the United States, and frequently there is talk about a labeled "obesity epidemic" that is occurring. Clearly, a majority of American are not making the correct choices about what they eat. Furthermore, studies have shown that vitamins or supplements have had no impact on intelligence itself.
The IQ paradox was clearly defined and is a recognized problem. Clearly there is no clean cut answer to why IQ is increasing, but a discussion drawing from other aspects of the problem would be helpful in letting the reader realize the complexity of the question of the IQ paradox.
Peer Commentary
The Dickens-Flynn Model: A Contradictory ExplanationKyle R. SkottkeRochester Institute of Technology
One of the major areas discussed in this paper is the Dickens-Flynn model. This model is supposed to provide an explanation for the rise in IQ scores during the 20th century. It is seemingly undisputed that intelligence is rising, based on comparing results from intelligence from past decades. Although the Dickens-Flynn model is supposed to help elucidate the mechanism by which intelligence seems to be rising across the board, there appear to be major flaws in the model which makes it unacceptable in my opinion.
One key point that the author makes is that old data was used to help design and test the model. While there might have been several reasons for using older, already gathered data, such as lowering the cost of research and shortening the time required to support their theory, it seems counterproductive in this case. In designing a model which represents a current population and how it is changing currently, it seems inappropriate to use data that are out of date.
I believe that the Dickens-Flynn model does not fully appreciate the importance of our increasing food supply and nutrition. One study (Berkman, Lescano, Gilman, Lopez, & Black, 2002) found that malnourished children, and children suffering from disease causing chronic diarrhea suffer a decreased cognitive ability. In many third world countries, malnutrition and diarrhoeal diseases were a big problem, but these are being corrected by modern medicine and through the use of better nutrition, such as golden rice. Although these problems have not been completely wiped out, there has been improvement.
The improvement of childhood nutrition in third world countries, and even improved health in western nations could explain the rise in IQ. I think this is a much more plausible explanation than the short lived environmental effects that the Dickens-Flynn model emphasizes. The long term nature of environmental effects like nutrition would also help to explain why IQ scores have consistently risen, instead of more sporadic results which might be expected with short term effects such as those described in the Dickens-Flynn model.
The author mentions that the data that Dickens and Flynn relied upon was out of date and no longer relevant, which brings into question the relevance of our IQ tests. If it is possible for the data that Dickens and Flynn were using to become obsolete in current times, it also seems possible that revisions could be made to intelligence tests to make them better fit the times. The WAIS-R was revised in 1981, making it over 20 years old and in my view, a prime candidate for further modernization. This is one area that the author does not consider in the paper, but one that I feel is relevant. Although it is important to have some sort of consistency so that validity and reliability can be ensured, I also believe that IQ tests would serve as a better measure if they were updated more frequently to match an ever changing society.
Although the Dickens-Flynn model does incorporate the ability of intelligence and the environment to interact and influence each other, I believe that there are too many statistical flaws and under-representations of important factors to consider this model valid. When designing the model, Dickens and Flynn relied on out-of-date data that does not represent the current situation. In addition to this, agree with others who criticize the model for not placing enough emphasis on the effects of long term environmental factors such as malnutrition in childhood. The focus of the model is instead on short term environmental factors which pale in comparison to malnutrition as far as the effects on cognitive ability. One more criticism I have with this model, in addition to those mentioned in the paper is that the intelligence tests that are in use today are over 20 years old. When the Dickens-Flynn model can be shot down by critics because of old data, I believe that old methods of IQ quantization are also fair game for criticism. I believe that part of the reason that IQ scores have increased is because the overall nutrition of the world has increased, as well as the need for updated tests that would present the same challenge in today's society as the WISC-R did in 1981.
Author Response
Flynn and the IQ ParadoxAlexandria K. CherryRochester Institute of Technology
The IQ Paradox has been a major topic for debate and attention for some time. What is causing this dramatic rise in IQ over the past century has not been fully explained. My paper, "The IQ Paradox Resolved? A Critical Analysis," looked at one theory proposed by Flynn to see how well it held up.
The first peer commentary, from Holmes, entitled "Childhood Exposure to Technology May Increase IQ Scores," is a very interesting take on the Flynn effect but does not really comment directly on my paper. It seems that Moulton, author of "The IQ Paradox: Is There an Answer?" did understand some basic theories of the IQ paradox. There is nothing in it that suggests that this phenomenon has persisted for the entire existence of humankind.
The last commentary, by Skottke, "The Dickens-Flynn Model: A Contradictory Explanation," did an excellent job relating new information with what had previously been stated. He recognized all my problem points and did a beautiful job adding something new and interesting.