Intro
  Why Do Some Groups Score Differently?

What Should We Do About Group Differences?
  Proponents of Herrnstein and Murray

Opponents of Herrnstein and Murray 1

Opponents of Herrnstein and Murray 2

Opponents of Herrnstein and Murray 3
So Where Does This Leave Us?
Opponents of Herrnstein and Murray
Return arguments are based on four key points:

  1. Irresponsibility on the Part of the Researchers

  2. The Flynn Effect

  3. Misinterpretation of the Studies

  4. Study Choice

  5. The Nature of Dysgenesis


1) Irresponsibility on the Part of the Researchers

Gardner, in an essay in Fraser's book, notes Herrnstein and Murray "encourage the reader to draw the strongest conclusions, while allowing the authors to disavow this intention." They stop short of saying that ethnic minorities are inferior, but lay out the data and their arguments so that the reader only has to make a short jump to say it himself. Gardner calls this "scholarly brinkmanship." Thus, Herrnstein and Murray's claims of "We never said that" in response to some criticisms holds little water because so often they all but said that.


2) The Flynn Effect

The Flynn Effect is named for Flynn (Herrnstein and Murray actually coined the term), who began publishing research on it in 1984. Flynn reviewed data for the Army Intelligence Test collected over several decades. Data from 73 studies of 7500 participants showed white Americans had gained 14 IQ points between 1932 and 1978.

Jensen offered that perhaps adults reached their peak performance on IQ tests sooner, there might be some sample or testing bias, or that more g loaded tests should be used before drawing conclusions. Flynn and Jensen examined army samples of varying ages from multiple countries, and found the results unchanged. Further, more g loaded and culture fair tests showed even larger gains (the Ravens showed 7 points a decade). IQ scores increased by about 3 points points on the Wechsler tests every ten years. The tests are restandardized, Neisser (1996) explains, every 10 or so years to account for this.


While some say it is debatable as to whether the rising scores mean rising intelligence, there are three reasons for the score changes that Flynn offers:

  1. There is more intellectual flexibility and factual knowledge with each generation. This finding of The Flynn Effect is seen in minorities as well. Opponents of Herrnstein and Murray argue that if minorities simply have poor genes, why do they show IQ gains over time just like the rest of the population? (Herrnstein and Murray supporters respond that IQ scores may be increasing over time in minority groups, but they increase in time in majority groups too, and faster, so the effect is canceled out. Of course, this requires believing that IQ scores are not exact measurements of intelligence, which counters another of Herrnstein and Murray's key assumptions)


  2. Nutrition and health care have improved, leading to "healthier brains." Flynn doubts nutrition has really increased scores, since even large changes in diet and supplements produce only a small effects. He does not, however, discuss the difference between poor nutrition and poverty.


  3. It's an artifact of testing that reflects better verbal analysis and no real changes in intelligence. Flynn himself favors this interpretation, noting that while the scores may change over time, so do our ideas of intelligence, and so the change in numbers doesn't mean much. He explains this in another way. Examining the norms, 90% of the British born in 1877 would have IQs below 75 if scored in 1967. He gives the sports analogy from Jensen (a baseball fan with average IQ can understand the statistics of the game, with low average IQ can understand most of the rules, and with borderline IQ can only watch and cheer or boo when everyone else does). Flynn debates whether or not this really makes sense, pointing out that the British in 1877 were able to understand cricket. What he goes on to point out is that perhaps the g gains are "real," but even then they may not have much real world significance.

What he offers is some serious consideration to Jensen's analogy of IQ and intelligence as linked by using shadows and height. Any measurement taken repeatedly at the same place and time of day will yield very good correlations, but measuring over time would seem to indicate that people grow taller over the course of the afternoon. Thus, according to Jensen, IQ scores only predict where a group stands now, not where they will stand in the future, compared to the rest of society. The doesn't hold so well when looking at Japanese and Chinese Americans, who tend to score about 10 to 20 points below where they reach in life.

Williams, on the other hand, believes the score changes reflect changes in intellectual functioning. She offers ten reasons to explain the Flynn Effect:

  1. In the 1930s, the average years of schooling in the US was 8-9, and in the 1990s was 14 years, which should stimulate crystalized knowledge and increase intelligence test scores

  1. Computers, video games, television, and rich media children's books and products (like Happy Meals with words games on the back, place mats with mazes, and cereal boxes with "Find The Differences" pictures) all provide a kind of visual stimulation children in the 1930s would not have experienced, which should stimulate fluid intelligence and increase intelligence test scores. Flynn denies these factors have led to increases in IQ scores, as increases were already apparent before TV was popular.

  1. Reading has decreased in American children, and textbooks have been dumbed down by two grade levels in the last 30 years, and so it is unclear if that has limited scores. However, creative schooling also exposes children to art projects, multicultural projects, technology far more than a focus on the "three R's" would have provided, which should stimulate fluid intelligence and increase intelligence test scores


  2. Teachers also spend more time in class "teaching the test" to improve test taking skills and strategies, confidence and ability to manage anxiety, time management during testing, and specific factual and vocabulary knowledge for testing, which should stimulate crystalized knowledge and increase intelligence test scores


  3. Families are smaller, meaning more parental time and resources dedicated to each child, although parent-child time has decreased with dual-income families (10% of mothers worked outside the home in the 30s compared to 60% in the 90s) so it is hard to draw conclusions from this


  4. Better educated parents and teachers stress education and its importance far more, and parents become more involved in their children's education


  5. Programs for minority children, including mentoring and tutoring, and services for less gifted children also improve scores in a subsample of the population, leading to small increases in the general population average


  6. Daily life can be more complex in ways that are related to test scores (think reading bus schedules, shopping for bargains in the store…) compared to the 1930s


  7. Parents are more likely to espouse authoritarian parenting styles in this age and read parenting manuals, rather than use authoritarian "Spare the rod, spoil the child" ways, and to become involved in their children's education and social competence (one study showed specific ratings of mothers' behavior when the child was age 2 years could predict child IQ at 3 and 4 years, and achievement at 6 years)


  8. Parent health arguably is better today than in the 30s when mothers faced depression (The Great Depression and deaths after the war), as well as poorer physical health, poorer prenatal care, and less knowledge about nutrition

For what it's worth, Nisbett reports on two studies showing that on achievement testing nation-wide, blacks have increased their scores since the 1970's (indicating the increase in IQ meant something). Herrnstein and Murray say this data is useless since white people increased scores too. There's three problems with this reply though:

  • if black people have increased their scores in thirty years, and scores are mostly based on genetics, then either they changed their genes, or really achievement isn't closely tied to intelligence test scores

  • the Flynn Effect does translate to real-world gains in intelligence and achievement, and thus something other than genes is impacting scores

  • the Flynn Effect is only testing artifact, and if increases are artifact, then how can you be sure that differences between groups aren't artifacts too?

None of these possibilities are good for Herrnstein and Murray's position.

Adding to this are Hauser's (1998) criticisms. He reviewed some of their work and went back to the original sources and found five copying errors which resulted in mathematical errors, overly large standard deviations were used, and underestimates of improvement in scores when analyzing the National Assessment of Educational Progress scores. He notes while they say the Black-White achievement gap has only closed by .28 of a standard deviation, correction of these errors increases the estimate to .39, or about 5 points on an IQ test.