Sunday, 25 January 2015

Davos diversity

This year, as is my custom, I did not go to Davos. I do not have any animus against the place itself, but I have never been drawn to the conference known by that name, nor have I been invited. For the avoidance of doubt, I do not object to rubbing shoulders with the Great and Good. Thompson, with a p. I assume I would enjoy it, on the basis of that famous protest banner held above a baying mob: “They promised us there would be biscuits”.

One of the things I missed by failing to attend, breathlessly reported by the unfailing BBC, was a debate on the “diversity dividend”. This is the notion that diversity brings advantages, a broader range of outlooks which translates into real world gains in productivity. For example, including Keynesians and Hayeksians in central banks, centrists and localists in governments; and fascists and communists in academia. I suppose, if diversity is in fact the key to success, that there should also be gains to be made by hiring employees of very diverse levels of intellect, lest the too clever fail to notice insights more readily available to the dull.

In fact, Davos seems to regard diversity purely in terms of sex and race, particularly which sex you are and which sex you prefer, with preferring your own sex being an indicator of diversity. I do not regard these matters as being any of my damn business, unless a friend decides to confide in me about these matters. Privacy is not what it was, and the size, shape and above all the uses to which genitalia are put are now matters of common conversation.

http://www.bbc.com/news/business-30973184

The item begins with a strong evidential claims:

There is evidence, for instance, that boards of directors with greater diversity generate more dividends. That means greater returns to shareholders and fewer bonuses paid to managers.

With respect to gender, there are numerous studies that show that adding women to the labour force increases national output, or gross domestic product

The panel referred to evidence that makes the business case that a more diverse workplace boosted the bottom line. For instance, a Massachusetts Institute of Technology study has found that changing from all-male or all-female workforces to equal numbers of both sexes could raise revenues by around 40%.

No references are given. “Trust Aunty” seems to be the mantra.

About a year ago I corresponded with people who were making such claims and I was sent one bit of evidence, suggesting that in one period stock market gains had been higher for diverse as opposed to non-diverse companies, but I did not find this compelling. Proofs of this sort are always being proposed by companies and investment groups, and generally depend on choice of time periods and comparators. That is not to say that one could not do it properly, over a longer time period, but I haven’t seen any such studies.

However, I have found the MIT study to which the BBC was referring.

Diversity, Social Goods Provision, and Performance in the Firm
Ellison, Greenbaum, Mullin. April 2011

http://economics.clemson.edu/files/diversap20.pdf

There are 4 mentions of profit, but no figures. That is to say, no data on whether the profits made by these individual offices (all members of the same company) differ according to diversity. There are 32 mentions of revenue and here is the key phrase which has been trumpeted by the BBC:

Turning to additional results in Table 5, higher levels of GendDiversity are positively
and significantly associated with office revenue in our base specification (1).
The estimated coefficient of 0.41 implies that going from an office that is either all
male or all female to an office split equally between the sexes would be associated
with a revenue gain of 41%(!).

On your behalf I bothered to look at Table 5, and the best predictor of office revenues, at a stunning 0.465, which they would no doubt describe as “boosting office revenues by 46.5% is…….. the number of years the office has been open!

Amusingly, the BBC report (see above) also makes the error of regarding revenues as “the bottom line”! No, dear reporter, revenues are the top line: the bottom line (profit) is what is left after all the costs have been removed, and this paper has nothing to say on that subject.

The authors of the paper then immediately go on to admit:

In column (4), we also report the results of a specification with office-level fixed
effects. When we control for the office fixed effects, the estimated contribution of
GendDiversity to office level revenue is no longer statistically significant.

So, the authors were right to put in their exclamation mark, knowing that the apparent effect was not a real one once local office variables were taken into account. The paper actually reports that diversity does not even boost revenues.

So, can anyone help me where the BBC cannot? Can you send me good studies showing that when the balance of a workforce is brought into line with the demographic composition of a nation it does better in terms of bottom line profits than when employees are selected solely on competence?

Finally, I may not totally follow capitalist ideology, but I had understood that the purpose of forming a company was to make profits by competing in a free market, honouring contracts drawn up freely between employer and employee, and between company and customers. Competition is the name of the game.

Suppose it was the case that having a sexually and racially diverse workforce led to larger profits, why on earth would any company admit this, and give away a useful business secret to competitors?

I hope you understand why I prefer the beach to Davos.

Friday, 23 January 2015

The Economist takes a half step forwards

There is much innocent fun to be gained from The Economist’s coy avoidance of the genetics of intelligence. They are mired in Blank-Slatism, but are cautiously tip-toeing towards admitting a few things, only to then back away again, thus taking them back to where they came from. This is not all bad: by conceding the importance of intelligence and then immediately saying it is driven by wealth they keep the Faith, whatever it is, but hint that they know more than they will let on in public.

Here they are, saying the previously unsayable:

today’s rich increasingly pass on to their children an asset that cannot be frittered away in a few nights at a casino. It is far more useful than wealth, and invulnerable to inheritance tax. It is brains. Intellectual capital drives the knowledge economy, so those who have lots of it get a fat slice of the pie. And it is increasingly heritable. Far more than in previous generations, clever, successful men marry clever, successful women. Such “assortative mating” increases inequality by 25%, by one estimate, since two-degree households typically enjoy two large incomes. Power couples conceive bright children and bring them up in stable homes—only 9% of college-educated mothers who give birth each year are unmarried, compared with 61% of high-school dropouts. They stimulate them relentlessly: children of professionals hear 32m more words by the age of four than those of parents on welfare. They move to pricey neighbourhoods with good schools, spend a packet on flute lessons and pull strings to get junior into a top-notch college.

Notice that in The Economist’s view brighter people marrying brighter people is not seen as a positive development, but a practice which “increases inequality”. Of course, duller people marrying duller people also increased inequality. In fact, couples assort themselves on intelligence more than anything else:

http://drjamesthompson.blogspot.com/2014/12/five-gold-rings-inherited.html

They also fall for relentless stimulation and the “number of words heard” as a major causal variable.

http://drjamesthompson.blogspot.com/2013/12/does-it-really-matter-if-you-talk-to.html

Then they explain what they think needs to be done:

The moment to start is in early childhood, when the brain is most malleable and the right kind of stimulation has the largest effect. There is no substitute for parents who talk and read to their babies, but good nurseries can help, especially for the most struggling families; and America scores poorly by international standards (see article). Improving early child care in the poorest American neighbourhoods yields returns of ten to one or more; few other government investments pay off so handsomely.

So, there is “no substitute” for parents who talk and read to their babies, but child care must be improved. They do not say so, but I think they are quoting the Abecedarian study. I am a fan of that, but we still cannot be sure that it can be delivered at scale in the high quality required, and most Headstart programs have not boosted intelligence, and most no longer claim to do so.

They do not examine the usual finding that educated parents are more influential than rich parents in supposedly “boosting” intelligence. They leave out the genetic element entirely, and say it is “incomes” which are inherited. If so adoption into a rich household should have massive effects on intellect, but that is not found. If adoption cannot wipe the slate clean, what chance a kindergarten?

http://drjamesthompson.blogspot.com/2014/11/adopt-child-but-discard-illusion.html

In a related article they spell out their concerns: An hereditary meritocracy: The children of the rich and powerful are increasingly well suited to earning wealth and power themselves. That’s a problem.

America’s elite is producing children who not only get ahead, but deserve to do so: they meet the standards of meritocracy better than their peers, and are thus worthy of the status they inherit.

So, the problem seems to be that they deserve to get ahead, the bounders!

Assortative mating of this sort seems likely, on average, to reinforce the traits that bring the couple together. Though genes play a role in the variation of intelligence from person to person, this is not a crude genetic determinism. People tend to encourage in their children what they value in themselves and their partners. Thus people bought together by their education and status will typically deem such things important and do more to bring them out in their children, both deliberately and by lived example—processes in which nature and nurture are more than likely to work hand in hand.

So, genes play a role…. but then they show histograms of SAT scores by parental wealth, implying it was “wealth what done it”. They should show a structured equation of wealth, education and parental IQ if they want to be totally honest.

As for the role genes play in scholastic attainments, Plomin and his team find they account for 58% of the variance, and shared variance (family and school) surprisingly does not account for any of the variance.

For the link between intelligence and scholastic attainment, see: I. J. Deary, S. Strand, P. Smith and C. Fernandes (2007) Intelligence and educational achievement. Intelligence 35, 1, pp13-21. For private study, email the author at the University of Edinburgh and ask for a copy.

The rest of the article is about wealth, connections and influence, with a clear implication that wealth is doing the talking.

Their third article is in the same vein, but a bit more nuanced: Getting ‘em young Early education matters, but it is not everything

Andreas Schleicher, head of the OECD’s education team, says early-years investment does not “automatically produce gains in learning, unless systems transfer this to primary and secondary level”. He has just published research showing that in a worrying number of rich-world countries more than 15% of young people are “unqualified”. Those with a problem include France, the Netherlands, Norway and Denmark—all high scorers for early-years provision. A good start is not enough on its own. The system’s stamina and consistency matters just as much—and possibly more.

So, this is about “the system” and the fact that some young people are “unqualified”. Here are some of my comments on the OECD’s views on human ability and their unwillingness to countenance differences in intelligence.

http://drjamesthompson.blogspot.com/2013/10/how-illiterate-is-oecd.html

None of The Economist’s articles or the papers they quote make it clear that intelligence must be considered a driving force in economic life and, consequently, in earnings, social status and resultant wealth. Curious, isn’t it, that a magazine written by the smart fraction for the smart fraction cannot bear to mention the smart fraction in a positive light? Perhaps they fear they will be cursed by the deity, or slaughtered by the baying mob. Noblesse oblige.

Wednesday, 21 January 2015

Copy number intelligence

I do not write policy papers, but much to my own surprise I had written myself an imaginary policy note to cover the eventuality that I would sent a paper reporting results on the genetics of intelligence. Would I post it up, or send it round to genetics researchers for a comment? Yesterday it dawned on me, having received such a paper from a learned psychological colleague who had found it of interest, that the best policy was to post it up with minimal comment, and ask my readers to evaluate it, or send it on to colleagues who were capable of doing so.

To my untrained eye the sample sizes look far, far too small to conclude anything. However, the small samples are drawn from larger samples precisely because they are outliers in terms of brain size, so that maximises the chances of finding a result. Furthermore, they have succeeded in replicating the finding in another sample of outliers, and have asked for others to replicate the finding on the larger samples required for a proper replication.

DUF1220 copy number is linearly associated with increased cognitive function as measured by total IQ and mathematical aptitude scores
Jonathon M. Davis · Veronica B. Searles · Nathan Anderson ·Jonathon Keeney · Armin Raznahan · L. John Horwood · David M. Fergusson · Martin A. Kennedy · Jay Giedd · James M. Sikela

Hum Genet (2015) 134:67–75 DOI 10.1007/s00439-014-1489-2

Abstract

DUF1220 protein domains exhibit the greatest human lineage-specific 
copy number expansion of any protein-coding sequence in the genome, 
and variation in DUF1220 copy number has been linked to both brain 
size in humans and brain evolution among primates. Given these 
findings, we examined associations between DUF1220 subtypes CON1 and 
CON2 and cognitive aptitude. We identified a linear association 
between CON2 copy number and cognitive function in two independent 
populations of European descent. In North American males, an increase 
in CON2 copy number corresponded with an increase in WISC IQ (R (2) = 
0.13, p = 0.02), which may be driven by males aged 6-11 (R (2) = 0.42, 
p = 0.003). We utilized ddPCR in a subset as a confirmatory 
measurement. This group had 26-33 copies of CON2 with a mean of 29, 
and each copy increase of CON2 was associated with a 3.3-point 
increase in WISC IQ (R (2) = 0.22, p = 0.045). In individuals from New 
Zealand, an increase in CON2 copy number was associated with an 
increase in math aptitude ability (R (2) = 0.10 p = 0.018). These were 
not confounded by brain size. To our knowledge, this is the first 
study to report a replicated association between copy number of a gene 
coding sequence and cognitive aptitude. Remarkably, dosage variations 
involving DUF1220 sequences have now been linked to human brain 
expansion, autism severity and cognitive aptitude, suggesting that 
such processes may be genetically and mechanistically inter-related. 
The findings presented here warrant expanded investigations in larger, 
well-characterized cohorts.

From the methods section:

We utilized two separate populations consisting of individuals of European descent: (a) from North America and (b) from New Zealand. The North American population included unrelated non-Hispanic white individuals selected for brain size extremes through a procedure identifying residual values from a regression controlling for sex and age from a population of more than 600. This included 59 individuals (41 males and 18 females) whose ages ranged from 6 to 22. To determine reproducibility of our results, we gathered array comparative genomic hybridization (arrayCGH)-based copy number data on 75 individuals from New Zealand, 51 of whom were of European decent and selected for extremes in birth head circumference from the Christchurch Health and Development Study (Fergusson and Horwood 2001) (CHDS) cohort. The CHDS is a longitudinal study of a birth cohort of 1,265 children born in the Christchurch urban area in 1977. Participants have been examined on 23 occasions through age 35, with a focus on understanding psychosocial development, health and well-being. Cognitive measures were gathered in children aged 8–13 years in the CHDS. ArrayCGH values for CON1 and CON2 were sent to blinded statisticians to identify associations with cognitive function tests. Brain and head circumference size extremes were originally selected for association studies examining DUF1220, head circumference and brain size (Dumas et al. 2012). Although selected for size extremes, these populations were not selected for IQ metrics and IQ was normally distributed in these populations.

https://drive.google.com/file/d/0B3c4TxciNeJZcHg4ek9nWDZGcWs/view?usp=sharing

Can you please wave this under the nose of a population geneticist and see if you can get a response?

Tuesday, 20 January 2015

School

 

I have always remembered the last line of Huckleberry Finn: “Aunt Sally she's going to adopt me and sivilize me, and I can't stand it.  I been there before.”

Perhaps I share Huckleberry’s dislike of schools: “WELL, three or four months run along, and it was well into the winter now. I had been to school most all the time and could spell and read and write just a little, and could say the multiplication table up to six times seven is thirty-five, and I don't reckon I could ever get any further than that if I was to live forever.  I don't take no stock in mathematics, anyway.”

However, I aim to be open to new ideas, and persons of great talent and learning assure me that going to school boosts IQ and that I should stop being churlish about such establishments.

Their arguments have to surmount a couple of hurdles. Firstly, “any fule kno” that the not so bright leave school early and the brighter ones stay on longer, so “years of schooling” is a dodgy measure which could show an apparent effect of schooling for artefactual reasons only.

Secondly, people understand that school might speed up some pedestrian mental processes, but leave you at the same intrinsic level of brightness at the end of schooling that you would have achieved anyway. 17/18 years of age is the proof of the pudding, when the final educated product rolls off the production line, possibly ready to make a useful contribution to society, so we should restrict ourselves to intelligence assessments conducted at around that age.

Thirdly, a body of work asserts that it does not matter whether children start school at 5 because they do just as well when they start at 7, but if schooling boosts IQ those who start at 7 should be duller, and that should show up for ever. The counter argument might be that school starting ages do not count, but school leaving ages do, but that would go against Abecedarian type results, which indicate that massive pre-school interventions with highly vulnerable children can boost IQ by 4 points which are sustained into middle age.

Fourthly, twin studies ought to show an effect of schooling on intelligence, but tend not to, which either suggests there isn’t any, or that current schooling is so similar that there is no school-based variance within standard educational systems.

What have the learned scholars arguing for the intelligence boosting effects of schooling got to offer?

The first helping is Ceci (1991) How Much Does Schooling Influence General Intelligence and Its Cognitive Components? A Reassessment of the Evidence. Developmental Psychology 1991, Vol. 27, No. 5, 703-722.

Ceci is also known for his work on racetrack experts (Ceci & Liker, 1986a, 1986b), research which has been often quoted as meaning that IQ does not count for much in real life. Doug Detterman demolished those papers so comprehensively that I am slightly on guard when reading this particular paper, but still reading it nonetheless. (It is silly and over-dramatic to avoid a person’s publications entirely because of some errors and contested results in other papers). Among the stronger studies he quotes Harnqvist's (1968) study of Swedish men. In 1961 Harnqvist selected a 10% random sample (approximately 12,000 men and women) of the Swedish school population who had been born in 1948 and who, at the age of 13, were given IQ tests. On reaching the age of 18 (in 1966), 4,616 of the Swedish men were tested as part of their country's national military registration. Thus, the study is not vulnerable to the usual sampling biases. Harnquist was able to compare children who were comparable on IQ, SES, and school grades at age 13 and determine the impact of dropping out of school on IQ at age 18. He found that for each year of secondary school (gymnasium) not completed, there was a loss of 1.8 IQ points—up to a maximum of nearly 8 IQ points difference between 2 boys who were similar in IQ, SES, and grades at age 13 but who subsequently differed in the amount of schooling completed by up to 4 years of high school. Naturally, one would need to look at the assumptions behind the SES and school grade “corrections” before being convinced by this, but this is strong stuff.

Ceci says:  One of the best ways to document the impact of schooling on IQ is with a cohort-sequential analysis in which children of the same chronological age enter school at different times and remain there for similar lengths of time. By performing the appropriate statistical operations, it is possible to separate the effects of schooling from those of chronological age and, possibly; historical period, though the latter claim is a source of some contention.

Ceci summarises his paper’s findings thus:  Taken together, the evidence for the influence of schooling on IQ appears to be fairly pervasive. Eight types of evidence for this relationship were reviewed: (a) the link between grade attained and IQ, (b) the impact of summer vacations, (c) the relationship between intermittent attendance and IQ, (d) the effect of late school onset, (e) the effect of early school termination, (f) the equivalence of aptitude and achievement test scores, (g) the
result of cohort-related changes, and (h) historical changes in the IQ-schooling link, all involving multiple replications. Although it is not possible to estimate with any confidence the precise decrement in IQ scores that befalls each month or year of missed or delayed school (Jencks et al., 1972, hazard the estimate of 1 IQ point decrement for each year of missed school), the trend is clearly one of declining IQ scores as a function of missed school. All of the studies that were reviewed accord with this statement, and only one empirical study argues otherwise.

https://drive.google.com/file/d/0B3c4TxciNeJZV2N5QjJGaFBtRkU/view?usp=sharing

Next up are Brincha and Gallowaya (2012) Schooling in adolescence raises IQ scores. PNAS  www.pnas.org/cgi/doi/10.1073/pnas.1106077109

They say: Although some scholars maintain that education has little effect on intelligence quotient (IQ) scores, others claim that IQ scores are indeed malleable, primarily through intervention in early childhood. The causal effect of education on IQ at later ages is often difficult to uncover because analyses based on observational data are plagued by problems of reverse causation and self-selection into further education. We exploit a reform that increased compulsory schooling from 7 to 9 y in Norway in the 1960s to estimate the effect of education on IQ. We find that this schooling reform, which primarily affected education in the middle teenage years,
had a substantial effect on IQ scores measured at the age of 19 y.

This is a substantial paper, exploiting a natural experiment, in which various educational changes were made at different rates in different municipalities, all leading to an extra 2 years of school for adolescents, and a new unified middle school for those extra years. In Norway children begin school at 7, and this late start was maintained in the reformed system. There were no IQ measures beforehand. IQ was tested at age 19 when the men entered military service. So the paper is about men not women, and might be influenced by men’s slower maturation rates, and relates to the generation born between 1950 and 1958, and contains two important simultaneous changes: length and type of education. Given that there were no pre-school or early school IQ measures, the argument is based on various assumptions about what average IQs would have looked like if there had been no changes to the educational system.

As a consequence, the analysis of the results is quite complex. I am doing my best to simplify here. The first method, a difference-in-difference (DID) analysis, is to estimate the effect of the reform on the average IQ score for Norwegian men by comparing the change in IQ scores from the pre-reform period to post-reform period for municipalities that introduced the reform in a given year with the change in IQ scores in that same period in municipalities that did not introduce the reform in that particular year.

The second approach employs instrumental variables (IV) methods in which experiencing the new schooling system is used as an instrument for educational attainment (with appropriate controls for time trends and municipality of residence). The resulting system of two simultaneous equations is then estimated by two stage
least squares (2SLS). If we assume that the exclusive mechanism by which the reform affected IQ scores is by increasing the amount of schooling, the IV/2SLS approach allows us to break down the effect of the reform into (i) the effect of the reform on educational attainment and (ii) the effect of 1 y of additional schooling on IQ scores.

The results show that the educational reform (both length and type of education) increased men’s IQ by 0.6 points, which is statistically significant in this large population sample, though small of itself. I am favour plain statistics over the other more complicated analyses, because I am a primitive sort of person. For that reason, I went straight to the supplemental data.

Looking at Table S1 in the supplement it is clear that Norwegian men’s intelligence for all reasons including the educational reform rose for those of the 1950-1958 birth period from IQ 106 to 107.5. Given that this is a population sample we need to begin with an explanation as to why the initial score as not 100. Flynn effects probably account for this, in particular post-war recovery from Nazi occupation, but in fact the secular rise is said to have been faster in other periods. The IQ test produces stanine scores, so important detail is lost, though this should not be too important in this large sample. However, if during this birth period IQ rose by 1.6 IQ points, then the higher claimed rate which the authors show is the upward deviation from the trend line which the reform created for part of the test period.

The authors say: Because we estimate the total reform effect to be 0.6 IQ points and the total Flynn effect to be roughly 1.6 IQ points, we can attribute over one-third of the Flynn effect to the direct effect of the educational reform for the population of cohorts we study.

Having a sight of the IQs over a longer period would be a reality check as to whether there was an obvious and sustained upward movement in intelligence which could be attributed to the extra two years of schooling. Long data sets are the best, because all variations can be seen in historical context. If those historical data are available, they would be extremely informative. In the mean time, please give me a hand and have a look at the results section on pages 427 onwards to see what you make of them.

This is a good paper, highly detailed and open about the assumptions made. I need to look at the further analyses in more detail to see if I understand and agree with the higher gain estimates produced by them (of an enormous 3.7 IQ points per year), but if I have to summarise, I would put the gain at 0.6 IQ points.

https://drive.google.com/file/d/0B3c4TxciNeJZcTViR0dfbmdsWnc/view?usp=sharing

It is hard to discuss any intelligence matter without making a bow towards the granite fortress of Edinburgh, and sure enough Stuart Ritchie, Timothy Bates, Geoff Der, John M. Starr, and Ian J. Deary have launched “Education is associated with higher later-life IQ scores, but not with faster cognitive processing speed”

Recent reports suggest a causal relationship between education and IQ, which has implications for cognitive development and ageing - education may improve cognitive reserve. In two longitudinal cohorts, we tested the association between education and lifetime cognitive change. We then tested whether education is linked to improved scores on processing speed variables such as Reaction Time, which are associated with both IQ and longevity. Controlling for childhood IQ score, we found that education was positively associated with IQ at ages 79 (Sample 1) and 70 (Sample 2), and more strongly for participants with lower initial IQ scores. Education, however, showed no significant association with processing speed, measured at ages 83 and 70. Increased education may enhance important later-life cognitive capacities, but does not appear to improve more fundamental aspects of cognitive processing.

In the LBC1921, each year of education was associated with a .66 IQ point advantage in IQ (95% confidence interval: .14 to 1.17 points) at age ~79, controlling for age 11 IQ, SES, age at both times of testing, and sex (Table 2, Step 3).

for individuals with lower scores on the initial IQ test, education was more strongly associated with higher later-life IQ.

This paper is particularly interesting because it has actual IQ results at age 11 and then uses exactly the same test decades later. By holding prior IQ constant it is in a position to show whether more education has an influence on later intelligence.

Correlation matrices for the two samples are shown above and below the diagonal, but are roughly the same. If we look at column 1 which relates to sample 1 then we can see that IQ at 11 is strongly related to later life IQ .66, to education .44, to age 11 social class .24 and more weakly to the reaction time and inspection time measures. Later life IQ shows the same sort of correlation with education .42 and stronger correlations with reaction time and inspection time.

 

image

 

 

The present results suggest that education has enduring effects on IQ test performance, even controlling for childhood IQ score, and that these effects are stronger for those with lower cognitive ability in childhood. However, they also suggest that these effects work via mechanisms – perhaps those involving improvements in specific skills - that are distinct from those generating differences in more fundamental measures of processing speed.

There is a slight problem with survivor bias in both samples: the participants are somewhat brighter than average, and in my view thus more likely to have stayed in education, which might somewhat inflate the apparent contribution of education.  The first sample shows more variance in years of education than the second (2.47 vs 1.13) so in theory the effect of education should be greater in the first sample, but the second sample gains 15.24 points, the first 12.79 which is somewhat against prediction.

The paper is also interesting because it appears to reveal the nature and limitation of educational effects. It also shows greater gains for lower ability students, which would make sense if education had a mildly compensatory effect.

https://drive.google.com/file/d/0B3c4TxciNeJZNHNBVzBDbnFEaGs/view?usp=sharing

There is more in this next paper (working title: education and g) which is still under review and which is brought to you exclusively so that you can review it yourselves (thus spurring on the official reviewers):

Previous research indicates that education influences cognitive development, but it is unclear what, precisely, is being improved. Here, we tested whether education is associated with cognitive test score improvements via domain-general effects on general cognitive ability (g), or via domain-specific effects on particular cognitive skills. We conducted structural equation modeling on data from a large (n = 1,091), longitudinal sample, with a measure of intelligence at age 11 years and ten tests covering a diverse range of cognitive abilities taken at age 70. Results indicated that the association of education with improved cognitive test scores is not mediated by g, but consists of direct effects on specific cognitive skills. These results suggest a decoupling of educational gains from increases in general intellectual capacity.

The findings indicate that education’s ability to raise intelligence test scores (as shown by, e.g. Brinch & Galloway, 2012) is driven by domain-specific effects that do not show ‘far transfer’ to general cognitive ability. Such a result coheres with findings from Ritchie et al. (2013), who, in the same participants who were assessed here, showed no association of education with elementary cognitive measures such as reaction and inspection time, despite an association with improved scores on more verbal IQ subtests. Our results are also broadly consistent with recent reviews concluding that training programs targeting the specific skill of working memory can improve performance on working memory (and closely related) tasks, but this advantage does not seem to generalize to more distantly related skills such as reasoning and arithmetic (e.g. Melby-Lervåg & Hulme, 2012; though see Karbach & Verhaeghen, 2014). Finally, our results are in line with a study by Finn et al. (2014), who showed in a longitudinal sample of schoolchildren that while the quality of the school they attended had effects on tests of directly-taught subjects such as mathematics and English language, there was no relation of school quality to performance on tests of ‘fluid’ ability such as processing speed, working memory, and reasoning. These findings, along with the results of present study, point to a conceptualization of education as a training program that develops particular intellectual abilities but not more fundamental capacities such as the efficiency of cognitive operations.

https://drive.google.com/file/d/0B3c4TxciNeJZa3BnSm5xVEVWazg/view?usp=sharing

In summary, it very much looks like more years of education are associated with an increase in intelligence test scores, but not anything like as strongly to underlying general intelligence or to underlying basic processing speeds. There is some hesitancy on my part, because I need to look at the arguments and the statistical techniques in more detail, but the Lothian cohorts get round most of those problems, simply because they have IQ measures at age 11.

So, where does all this leave our hero? Huckleberry did admit: “At first I hated the school, but by and by I got so I could stand it.” Perhaps it boosted his IQ, though not his processing speed. I hope it did not stop him from drifting down the Mississippi on those sunny afternoons I still remember.

Saturday, 17 January 2015

Is your face special?

General intelligence runs through psychology like carbon through biology. Nonetheless, the ubiquity of general intelligence does not preclude the existence of some special modular skills, among which the recognition of faces is a strong candidate. Your face is special for reasons you may or may not find palatable: faces have evolved to make people recognisable. We know who you are. More important, we recognise you and remember how you have treated us. No wonder some reviewers want to remain anonymous.

So, it is interesting to find out, with a proper sample, whether face recognition is yet another g loaded task, or something which evolved on its own and is a distinct skill, and not particularly intellectual. Using that old and discredited technique, introspection, it seems to me that recognition is a simple rather than a complex task. Given time and favourable viewing circumstances we can often recognise a face, but no great intellectual insights derive from such recognition. An interesting question is whether face recognition is heritable.

To test that and whether faces are really special it is useful to compare them to the recognition of cars, which are rather recent objects, and not subject to evolutionary influences, not yet anyway. Here are the Plomin gang on the topic of faces, and Nic Shakeshaft’s powerpoint has sent me his presentation below, generously including his notes on many of the slides, which should help you navigate the content. nicholas.shakeshaft@kcl.ac.uk

The genetic specificity of face perception
Shakeshaft, N.C., Schofield, K.L., Plomin, R. King’s College London


Research Question: Specific cognitive abilities are almost invariably found to be highly heritable, and also to correlate substantially with general cognitive ability ('g'), both phenotypically and genetically. Such findings suggest that cognitive abilities in diverse domains share a common genetic aetiology to a considerable extent, falling into a single hierarchy with 'g' at its apex. Recent twin studies have suggested that the ability to recognise faces may be an exception to this pattern, being equally heritable, but phenotypically uncorrelated both with 'g' and with general object recognition. However, genetic associations cannot be determined with confidence from phenotypic correlations alone.
Method: This study presents the first investigation of face perception in a twin sample with adequate power to perform multivariate genetic analyses. More than 2000 19-20 year old UK twins from the Twins Early Development Study (TEDS) completed measures of face perception, object recognition and general (verbal and non-verbal) intelligence. TEDS is a longitudinal cohort study of twins born in England and Wales between 1994 and 1996, representative of the UK's general population. Data were subjected to twin model-fitting analyses to derive estimates
of genetic and environmental influences.
Results and conclusions: Results confirmed the substantial heritability of face perception (62%), and found it to be modestly phenotypically correlated (r < 0.2) both with general object recognition and with 'g'. Model-fitting analyses found that most of the genetic influences on face perception, accounting for most of its total variance, were not shared either with 'g' or with general object recognition.
Discussion: The results suggest that face perception is largely, but not wholly, distinct from other abilities. This makes this ability exceptional among cognitive domains, particularly in terms of its genetic uniqueness, as 'g' typically accounts for the majority of the genetic variance in each domain. Not only does this suggest that faces are 'special', but the existence of a domain largely independent from the traditional cognitive hierarchy may also shed some light on the nature - and limits - of intelligence itself.

https://drive.google.com/file/d/0B3c4TxciNeJZMWR3Zy15YzgzWkE/view?usp=sharing

There is also another paper on this topic, looking at the more general issue of social perception, and tentatively linking it with verbal intelligence.

Are general and social intelligence genetically distinct? Evidence from behavioural genetics. Schofield, K.L., Shakeshaft, N.C., Plomin, R. King’s College London 

Research Question: Studies of human cognitive ability suggest that aspects of social perception, such as the ability to process human faces, are phenotypically and genetically distinct from general cognition. Nonetheless, twin studies indicate that aspects of social perception, including face perception, are significantly heritable. The present study tests the hypotheses that 1) cognitive and self-report measures of social perception are substantially heritable; 2) social perception forms an independently heritable domain, (partially) genetically distinct from general cognitive intelligence ('g'), and more speculatively, 3) social perception is differentially related to verbal and non-verbal g, with the genetic correlation between social perception and g being explained by shared genetic influences with the verbal component specifically.
Method: This research was conducted within the Twins Early Development Study (TEDS), a longitudinal cohort study of twins born between 1994 and 1996. The participants were 1000 pairs of twins (aged 19-20) selected randomly from
the TEDS cohort, who participated via a purpose-built website.
Participants completed a cognitive test of social perception, the Cambridge Face Memory Test (CFMT), and a novel self-report measure, the Aspects of Social Perception (ASP) scale, comprising four subdomains: face identity, object
recognition, face emotion perception, and non-face emotion perception (i.e., from body language/voice). Verbal and non-verbal g, and self-report psychometric measures of social functioning, were completed at previous testing stages.

Results and conclusions: An independently heritable face-processing ability, with heritability of 62%, was confirmed. Self-reported ability in accurately perceiving emotion in faces was 30% heritable, while non-face (body language, voice) emotion perception was 18% heritable. A small (r<0.2)  but highly significant phenotypic correlation between social perception (self reported face and nonface identity and emotion recognition) and verbal g was explained almost entirely by genetic influences; no correlation was found between social perception and non-verbal g. For cognitive measures, the small phenotypic correlation with verbal g was largely genetic in origin, while that with non-verbal g was not.
Discussion: This study introduces a novel measure of social perception, demonstrates the partial independence of social perception from general cognition, and suggests differential associations with verbal and nonverbal cognitive abilities. This is the first in a series of studies intended to assess the genetic influences underlying social intelligence, and the place of this domain in the cognitive hierarchy. Future work will incorporate additional cognitive measures of
social perception. Findings are interpreted in the context of the evolutionary 'bootstrapping' hypothesis of social intelligence, and ongoing work is outlined which further explores the nature of social intelligence, and how it might be characterised in the context of verbal and nonverbal general cognition.

https://drive.google.com/file/d/0B3c4TxciNeJZXzZXblhPZHNuYXc/view?usp=sharing

Summary

Face recognition is highly heritable and weakly related to g so it probably has its own module, or a favoured set of circuits which only partly taps into the central processor power of general intelligence. So, not only is your face is special. but your specialised ability to recognise faces is even more special. Time to look in the mirror, and marvel.

Wednesday, 14 January 2015

Creativity: novelty, appropriateness, impact

 

Can creativity researchers be creative about how creativity should be measured?Davide Piffer suggests that three factors need to be understood and measured separately.

Can creativity be measured? An attempt to clarify the notion of creativity and general directions for future research


http://dx.doi.org/10.1016/j.tsc.2012.04.009

He says: The goal of this investigation is to demonstrate that much of the confusion regarding the measurement of creativity is caused by the insufficient clarity of its definition and to provide suggestions for an improved assessment and new possible tools of investigation (e.g. interviews).
It is shown that three dimensions of creativity (novelty, appropriateness and impact)
constitute a framework within which creativity can be defined and measured.
Further clarity to the definition of creativity is added by distinguishing between person’s and product’s creativity and providing definitions for each.
Based on this new definition, it is argued that Divergent Thinking, Remote Associates or some personality scales can be considered neither the only components of the creative process/cognition/potential nor “creativity tests”. The use of the terms “creativity test” and “measure of creative process” in the literature are criticized and it is indicated when they should be used.
It is also shown that claims to have found a general factor of creativity are based on methodological and conceptual errors.
Finally it is concluded that a person’s creativity can only be assessed indirectly (for example with self report questionnaires or official external recognition) but it cannot be measured directly.

https://drive.google.com/file/d/0B3c4TxciNeJZOGVHc3BQSG9iWXM/view?usp=sharing

Friday, 9 January 2015

Educated parents more important than rich parents

Although academics should be above such things, sometimes a particular paper becomes cherished because it seems to prove a particular point, and becomes a totem for a general world view. For many years it has been taken as axiomatic that family environments have a very strong impact on child development. How could they not? Parents are there every day to inculcate values, teach skills and attitudes, nurture or neglect, and shape their young charges into the paths of righteousness, or sling them down the slippery slope of dissolution. Furthermore, rich parents are able to buy their children educational toys, books, tutors, quiet rooms for study, good schools, proper accents and habits, and the shiny baubles of wealth indicators.

Social class is often seen as the engine room of social success, and most psychological and sociological research measures socio-economic status as a matter of course, sometimes neglecting other measures as a consequence. However, social class is far from being a pure sociological measure, most notably because most parents are bringing up their own children, so genetics and social forces are confounded. Absent a genomic analysis of parents and children (which may become standard one day) it would be at least one step forward to measure the intelligence of parents and their children. Useful, but somewhat time consuming and expensive compared with a quick assessment of social class.

As a consequence, most data sets use cheap and cheerful measures: socio-economic status of family, and educational level of parents, either as highest level achieved, or total years of schooling. Both of these last measures are pretty crude, and will under-estimate the finer detail of scholastic attainments.

The general picture which emerges from worldwide PISA results is that parental education is more influential than parental status or wealth. In the German PISA sample for Germany there were very weak correlations between wealth indicators and academic achievement: PISA sum and availability of one's own study place:
r=.09(N=32947); own desk: r=.11 (N=32983);mobile: r=−.25 (N = 32113); TV: r = −.06 (N = 32968); game computer: r = −.20 (N = 31990); and video: r = −.22 (N = 31861). By contrast, another indicator of students' learning environment, which requires relatively less expenditure than electronics, is much more and stably and
positively related to their achievement: the number of books: r = .43 (N=32568).

Having a mobile phone, a video recorder and a game computer are associated with lower scholastic ability, and the only substantial positive correlation is with the number of books, and of course the cause may not be the books themselves, but the intellect and character of the families who choose to buy books.

The frequently supposed positive effects of wealth-driven private school attendance are hardly supported by research results. According to studies done in 22 Western countries (Dronkers & Robert, 2008a,b), the “effect of private schools” on PISA scores through parental attributes is much larger than the small positive or even negative effects of private school attendance itself: “The main explanation of their gross differences in scholastic achievement is the better social composition of private
schools.” (Dronkers & Robert, 2008a, p. 541). Thus, these schools do not make students smarter, but smarter kids improve these schools, at least their measured achievement test results. Similar results were found by Bond and Saunders (1999) for Great Britain: Effects of private school attendance varied between a small positive impact on abilities (β = .07) and negative ones on later occupational status (sum
β = −.12).

Educational level is an indicator of parental cognitive ability, which can influence their educational behavior and their shaping of a benign developmental environment. This can include wealth (e.g. via books) having an environmental impact on children's cognitive development. Further, parental educational level
can be an indicator of genes underlying parents' and children's behavior.

Now Rindermann and Baumeister have taken a close look at Hart and Risely (1995) and find that as regards cognitive development, parental education is more important than socio-economic status.

Rindermann, H. & Baumeister, A. E. E. (2015). Parents’ SES vs. parental
educational behavior and children’s development: A reanalysis of the
Hart and Risley study. Learning and Individual Differences, 37, 133-138.

In their seminal study “Meaningful differences in the everyday  experience of young American children”, Hart and Risley (1995) showed a  close relationship between differences in family environment and  children’s development. The conventional conclusion was that children’s  cognitive development (measured by psychometric intelligence and verbal  ability tests) depends on parental socioeconomic status (SES), especially on wealth. We reanalyzed their data and show that the quality of education given by parents is crucial (beta-PEB=.58) and not the diffuse aggregate measure of SES (beta-SES=.11). Additionally, we  compare their sample with a similar but larger sample (Hoff, 2003) showing the same pattern of results. Possible causal factors (associated environmental and genetic factors) are discussed.

https://drive.google.com/file/d/0B3c4TxciNeJZSEJPOXgxLTlrNTA/view?usp=sharing

As you will see, the original study had a sample size of 42, of whom 17 were African American. So, a very small sample on which to make any statements about large issues, and furthermore, a sample which conflates social and ethnic effects. Is there any hope for psychology?

Thursday, 8 January 2015

Charlie Hebdo

The purpose of terrorism is to terrify. In that aim the perpetrators of the attack on the staff of Charlie Hebdo in Paris have been successful. They have shown that journalists in Paris cannot be protected against men carrying heavy weapons willing to kill anyone in their way. Leading journalists and cartoonists, familiar to all French readers, have been gunned down despite Police in the street, bodyguards in the building and locks on the doors. Pens were not mightier than machine guns that day, though they may be in the long run. In the grisly trade of political assassination this has been a successful mission.

The status quo of domestic policing in Europe is that most people are peaceful, and that the few who are violent are doing so for their own limited criminal purposes and want to avoid detection and capture. They run away with their stolen goods, and are usually eventually tracked down.  These particular perpetrators of the attack on Charlie Hebdo have already been identified as being two brothers with criminal backgrounds. They have moved on to Posture criminality, in which the aim is to stroll and swagger rather than run, to cause the greatest hurt possible, to be triumphant in murder so as to glorify themselves and their adopted cause. Conventional policing will always have a difficulty when confronted with armed death cults, who kill for the sake of making a larger political and religious point. “Don’t ever make fun of Islam” is their boast, and they are in charge of deciding what constitutes fun.

However, it would, be mistaken to describe these miscreants as “professional”. Professional soldiers follow rules of engagement, and do not kill unarmed civilians or even armed men who are wounded, incapable of defending themselves, and have surrendered. “Dispassionate” or “cold blooded” would be closer to the mark, and even that may be part of a posture, the martyr’s bravado, the swagger of a fool who considers himself fully justified because of a set of arguments which admit no contradiction.

Liberal democracies are a soft touch on these matters, and that is their attraction to the locals and to immigrant arrivals, and also part of their strength. Avoiding over-reaction is a measure of the long-sightedness of a tolerant society. However, there is also a danger in under-reacting. “Business as usual” is a good policy most of the time, but it can also slide into appeasement, withdrawal, and the cowed acquiescence of a fearful public, restricting their opinions and freedoms. What democracies must now work out is how they deal with people who treat open societies as if they were war zones, and who regard their half-examined convictions as being justification enough for slaughter. France and Europe as a whole will have been frightened by this attack. It is not too much to speak of them going through a collective symbolic trauma. However, they should not shirk from recognising that Islam has a prominent militant wing, a death meme which finds resonance in many a self-important young idiot, and with a supportive fringe of unknown size who will condemn the act and then list the supposed causes in the distant sense of “understanding why they did it, wrong as it was”.

After being subjected to criminal and political violence many of my traumatised patients felt that they would never see humanity in the same way ever again. That is understandable, though an over-reaction. Equally, seeing this event as an individual aberration is an error. A corrosive idea is doing the rounds of the impressionable: you are justified in killing those who criticise your religion. Pens also wrote the murderous stories contained in some sermons and religious texts. Pens, and perhaps more, will be required to counter it.

Wednesday, 7 January 2015

Author replies

 

Emanuel Jauk has come back to me about comments I made on his paper, making some clarifications, and pointing out that although I had uploaded one of his papers he had sent me another relevant paper which I hadn’t uploaded, so here is a link to it, with apologies to Emanuel:

https://drive.google.com/file/d/0B3c4TxciNeJZek00aXpudTBWUU0/view?usp=sharing

I had said of his presentation:

"The paper is slightly unusual, in that it seeks to model creative processes using IQ120 as a cut-off, and so runs contrary to the general findings of the Lubinski and Benbow work that there is no cut-off point, and that the brighter you are the more creative you are in real life."

Emanuel replies: “The paper you refer to was published in Intelligence 2013 (i sent it to you; it is a different one than the document you uploaded). We observed an IQ-120-cutoff for the relationship between intelligence and divergent thinking ability (referred to as "creative potential" in this manuscript).

I had said:
"The paper [...] runs contrary to the general findings of the Lubinski and Benbow work [... .] I think the difference lies in the fact that [...] their volunteers were assessed on creativity tests, rather than real life achievements."

Emanuel replies:
“We did actually assess real-life achievements, too, and found no cutoff-point. Instead, we observed a linear increase of creative achievement as intelligence increases. Thus, the findings do in no way contradict those from the Lubinski group; they rather support them. In the second paper (EJP 2014; the one you uploaded) deals with creative achievement in more detail and shows that intelligence has additive and interactive effects on it.

Emanuel concludes:

“I understand that all the different creativity measures and tests might lead to some confusion, but it is important for me to highlight that if we consider the full complexity of the phenomenon, the big picture seems to be fairly consistent, actually.”

 

In summary, there appear to be break points if you look at creativity test performance, but not when you look at actual achievements in the real world.

 

Order restored.

Tuesday, 6 January 2015

Creative memories

Perhaps I am slightly jaundiced about the benefits of memory, because my traumatised patients were trying so hard to forget, but the dream of boosting one’s intelligence never dies, and expanding one’s memory is seen as the royal path to that special power.

So, what can be achieved by training people in the N-back technique? You may have already guessed that my calm and balanced view is that the whole enterprise is a colossal waste of time which would be better spent learning how to play the cello.

Oshin Vartanian lives in Toronto, a city with a high tower with a glass floor where, if one is so inclined, one can dance with women one has never met purely for the vertiginous pleasure of waltzing over the streets which are clearly and terrifyingly visible very far below. My compliments to the unknown lady with whom I danced, and her wise word: “Just dance, and don’t look down”.

Here, as promised, is Oshin’s Powerpoint presentation:

https://drive.google.com/file/d/0B3c4TxciNeJZb2U2MnlSWDlpRlE/view?usp=sharing

Here is one of the papers on which his talk was based, published in the journal Neuroscience in 2013. Briefly, he found that a short regimen of WM training was associated with greater neural efficiency during engagement in a divergent thinking task, but not improved performance as measured by efficiency.

https://drive.google.com/file/d/0B3c4TxciNeJZcy1GcUpackM5VEk/view?usp=sharing

He has an on-going study due to finish this week in which the team are testing the relative effectiveness of a 10-day training regimen on the adaptive dual n-back task (vs. non-adaptive dual n-back task) on divergent thinking—but this time measuring originality, flexibility and appropriateness in addition to fluency. They will also score all responses using the consensual assessment technique for the quality of responses which Paul Silvia described in his presentation (see earlier post).

From my point of view the sample sizes are too small, and I wouldn’t bother to put any of these heads in a scanner, simply because reliable results require sample sizes approaching 500 subjects. However, these slight but significant results for N back keep alive the dream that you, by learning to remember pointless things which happened a few moments ago, might conceivably become more fluid in your intelligence and might even become more creative in your cello improvisations.

However, don’t use the N-back technique to learn to play the cello. Get a good music teacher, preferably one not interested in neuro-psychology.