Wednesday, 30 September 2015

Does culture cultivate, or do you need a good plough?


The search for culture-free or culture-fair tests has proved endless, because “culture” can be used so broadly as to encompass virtually anything a human does. People live in society, and societies transmit the habits of previous generations. There was a time in the debates after Jensen’s 1969 paper when psychologists believed that they could estimate the cultural loading of a test by inspecting the items. Indeed, my very first published paper attacked Jensen for arguing that the Wechsler subtest of Block Design was relatively culture-free, such that black-white differences on that test were probably genetic, whereas I felt it depended on access to constructional toys.

How does one determine the cultural loading of an intelligence test item? A Dutch team have plunged into these waters (strictly speaking they are below sea level, but no matter) and have rated subtests thus: Cultural load was operationalized as the average proportion of items that were adjusted in each subtest of the WAIS-III when the scale was adapted for use in 13 countries (Georgas et al., 2003).  To my eye that is certainly a language adjustment, though I wonder whether it allows for the different availabilities of artefacts in the home (not that I can think of an easy way to measure that).

Kees-Jan Kan, Jelte M. Wicherts, Conor V. Dolan, and Han L. J. van der Maas. On the Nature and Nurture of Intelligence and Specific Cognitive Abilities: The More Heritable, the More Culture Dependent. Psychological Science 24(12) 2420–2428




On this measure, Vocabulary is the most culture dependent subtest. On first glance that makes sense: the easiest way to learn a language is to be immersed in the particular culture that speaks it. However, that merely covers translating words from one language to another. Even within a culture, even the most ethno-centric citizens do not learn all the available words: intelligence is required for that.

From an early post on Vocabulary: Some people have the simplistic notion that vocabulary must be determined by mere exposure to spoken language. That is necessary, but far from sufficient, as even children work out. They notice patterns, informal rules, and the contexts in which communication takes place.  “The acquisition of meaning is based on the eduction of meaning from the contexts in which the words are encountered”. (So, even if the word “eduction” in the quotation from page 146 of Jensen’s “Bias in mental Testing” is unfamiliar, you will not be surprised to deduce that it means “To assume or work out from given facts; deduce”).The meaning of a word is acquired in some contexts which permits at least some partial inference as to its meaning. By hearing or reading the word in different contexts, through a process of generalization, discrimination and eduction one can guess at the essence of the meaning of the word, so as to use it (experimentally) oneself the next time a similar context presents itself. Words move from being unfamiliar to familiar, from familiar but not really understood to being familiar and partly understood (at which stage the explanations given about the meaning of the word are threadbare and inaccurate), and from there to being explained by use of synonyms (though those can range from partial to full understanding as shown by the power of the explanations and definitions).

The Methods section is explicit about how things were calculated, one step at a time: a model approach to be commended.


WAIS subtest heritabilities

The culturally loaded tests have higher heritabilities.

The authors conclude:

Each subtest’s proportion of variance in IQ shared with general intelligence was a function of cultural load: The more culture loaded, the higher this proportion. In addition, in adult samples, culture-loaded tests tended to have greater heritability coefficients than did culture-reduced tests, and there was a relationship between subtest’s proportion of variance shared with general intelligence and heritability. In child samples, these relationships were in the same direction, but correlations were small and insignificant.

They sound a cautionary note about the data, but their substantive point is:

A correlation between, for instance, g loading and heritability coefficient is in line with the hypothesis that the g factor is the most heritable factor (Jensen, 1998), but a test of the significance of this correlation does not provide the means to test whether the g factor is indeed the most heritable factor1 (Dolan & Hamaker, 2001). The method merely serves to evaluate competing theories of intelligence (Rushton & Jensen, 2009): A significant correlation denotes that a phenomenon exists that is in need of theoretical explanation. Theories that account for the correlation are stronger (with respect to this correlation) than are theories that do not account for it or are silent about it. The same line of reasoning holds for the correlations of cultural load with g loadings and heritability coefficients.

Having given their conclusions, the team then go against normal sequence and start a discussion.

Our result showing that culture-loaded knowledge tests (crystallized tests) are more strongly related to general intelligence than are culture-reduced cognitive processing tests (fluid tests) fits better with the idea that g loadings reflect societal demands (Dickens, 2008) than that they reflect cognitive demands (Jensen, 1987). Furthermore, in adult samples, our finding that the heritability coefficients of culture-loaded tests tend to be larger than those of culture-reduced tests calls for an explanation, given that this result does not follow from the subtest-complexity and investment hypotheses of g theory and fluid-crystallized theory.

After discussing some options they plump for genotype-environment covariance.

Because the acquisition of knowledge depends on cognitive processing, individuals who develop relatively high levels of cognitive-processing abilities tend to achieve relatively high levels of knowledge. High achievers are more likely to end up in cognitively demanding environments that encourage and facilitate the further development of a wide range of knowledge and skills. The contents and organization of these environments largely reflect societal demands. These societal demands thus influence the degree of dynamical interaction among cognitive processes and knowledge and, hence, their intercorrelations. In this way, the societal demands determine IQ-subtest loadings on the general factor of intelligence and, eventually, the degree to which broad-sense heritability coefficients of IQ subtests include the effects of (growing) genotype-environment covariance. In view of theoretical parsimony, we conclude that the assumption of a true causal g can be incorporated but that this is not required.

This paper presents interesting, counter-intuitive findings, which deserve replication on other samples and other psychometric tests. As to their favoured genotype-environment effect, I don’t see how bright people can obtain high levels of knowledge without being bright in the first place. They don’t develop intelligence, they have that ability in varying degree and use it to develop their knowledge to varying degrees. I am still working this out, but I think that ability is prior, and therefore more likely to be causal.

See what you think.

Monday, 28 September 2015

Blood Moon: Recalling Eratosthenes


Although the Observatory at Greenwich still be-straddles the globe as the origin of longitude, and thus of Time itself, London long ago ceased to be a good vantage point for examining the heavens. Anthracite coal conquered the world, but it besmirched London’s skies, and then the Clean Air Act coincided with ever-stronger electric street lighting, so light replaced soot as the celestial pollutant, fading distant stars. Of more moment, British skies are always covered with clouds, so astronomy is well-nigh impossible. And yet, and yet, last night the London sky was sparkling clear, so the whole supermoon lunar eclipse was visible for the seven stages from first to final contact.

At this numinous and transient moment I sleepily tried to explain to myself what was happening. The earth was travelling round the sun, but not at a speed sufficient to account for what I was observing. The earth was rotating on its axis, but rotation does not cause shadow. The moon was apparently fixed in the sky, yet it was slowly falling into a shadow caused by the nearest heavenly body, the Earth, and the atmosphere of that home planet was causing selective filtering of light-waves, taking out the shorter ones, and leaving the red.

If Eratosthenes, the third-ever Chief Librarian of the great library at Alexandria, had been standing beside me at my London window, he would have been the perfect teacher, had he not, as would have been more likely, been using the event to avoid chit-chat and make his own observations and calculations.

Astronomical events were the first and greatest of puzzles faced by our ancestors, the stuff of creation myths, superstitions, rituals and eventually sceptical surmise: the dawn of science. Astronomy required a leap of understanding: that the all too solid earth on which we stand might also be just one hurtling dot among the many visible (and invisible) one in the skies. Tycho Brache, the last of the naked-eye astronomers, was chronicling the regressive paths of the planets, but could not fully agree with Copernicus’ interpretation of those wandering planets in terms of the Earth’s own orbit round the sun.  The Copernican shift of perspective was a leap of Piagetian proportions, in which an observant maturing child eventually understands that what they see, and what a doll placed in an assembled mini-landscape sees, are not one an the same thing. The ego-centric perspective of early childhood is attenuated by a growing appreciation of the perspectives of other minds. It is similar in intellectual status to a growing child noticing that kindergarten children are becoming smaller, but eventually realizing that is only because he has grown bigger, not because new generations have shrunk.

How well do contemporary citizens understand astronomy? I should add, understand it without looking it up and repeating it, only to fall into ignorance again? If the Flynn Effect is real, then it will be far easier for average persons to understand eclipses, night and day, summer and winter. I do not have up to date data on pass rates, but here is an interesting finding from a random survey of British adults in 1992, which is 122 years after the 1870 Education Act and 472 years after the publication of Copernicus’ De revolutionibus orbium coelestium in 1543.

John Durant, Geoffrey Evans and Geoffrey Thomas. (1992) Public understanding of science in Britain: the role of medicine in the popular representation of science. Public Understand. Sci. 1 161-182.



So, 30% of British adults thought that the observed passage of the sun in the sky meant that our star obligingly whizzed round us. 16% imagined it did so once a day, as per visual observation of the same apparent phenomenon. More recent findings gratefully received.

If I concentrate on matters involving events observable to our ancestors, and avoid calendrical calculations, science education websites show that popular misconceptions still include the following: that the phases of the moon are caused by the moon going into and out of the earth’s shadow; that the moon has a side which is in perpetual darkness; that the moon does not rotate; that the phases of the moon are completed in exactly the number of days it takes to completes its orbit of the earth; that the moon is somehow larger on the horizon than when it is high in the sky; that the four seasons are the result of the changing distance from the sun; and that heavier objects fall faster than lighter objects.

I know that we no longer live on the land, and therefore are more distant from nature, and from the peregrinations of the sun and moon. I know that ignorance about astronomy is very largely a matter of education, but it is likely also that some education has been given but was not retained, because egocentric observation is deemed sufficient by many people. I know that I can make errors, and that many people make errors in simple Newtonian physics (imagining that if you drop an object when running it will describe an arc backwards to the ground, not forwards to the ground).  I know all this, but if we were really getting brighter over the last two centuries we would be able to work out much of this for ourselves, as Eratosthenes worked out the circumference of the world when he heard a casual remark about a well to the south of Alexandria where on one particular day of the year the sun shone down to the very bottom of the well.

Just for amusement, here are some Northern hemisphere university graduates explaining the astronomical causes of seasonal variations.

Here are some popular misconceptions tracked down and explained.

So, these were some thoughts whilst watching the super-moon lunar eclipse last night, among the blazing street lights of urban London. If by some magical process Eratosthenes had stood next to me I am sure that I would have best served him only by listening to him attentively.

Sunday, 27 September 2015


On 19 June the blog reached 500,000 and just a moment ago, 102 days later, it achieved 600,000. I am aware that most citizens will continue with their post-Sunday lunch nap, but I judge that my select group of readers will at least raise one eyebrow before returning to their even more serious reading. The total is far higher than I conceived possible when I began the blog two years and 10 months ago, when I felt lucky if I got 20 readers a day. The current daily page-view rate is roughly a thousand.




The snapshot of the past month shows a pronounced ISIR conference peak, and the most popular posts in that period are all about the conference.


The all-time greats remain the familiar old posts, with a few additions in the lower ranks:



Where do readers come from?



Readers are almost 6 times more likely to be in the US than the UK. The peak age for readers is 18-35 range, going down gradually to age 65+. So, the message is getting through to those who have most of their working life ahead of them, and might make decisions based on intelligence results. Readers are interested in science, news and politics. In terms of how they get to the blog, last month 4080 visitors came directly (loyal established readers?)  3912 through social media (loyal plus new readers?)  2822 through “organic search” (searchers for knowledge?) and  2634 through referral (new readers willing to admit they are searching for knowledge?).

There is a tendency for the longer essays to get correspondingly longer reading times, suggesting readers stick with the content. The item on whether Asians were bright but lacking in curiosity made people read for 6 minutes, whilst shorter conference announcements got half that duration of attention, all consistent with visitors being real readers.

Twitter is my hyper-active front-runner for the slower meditative page-turners of my blog. Precis is good for the mind. I have a few more followers (now at 1,262) which is welcome. Of course, I want the right sort of followers: those who contribute to knowledge, even when they just ask questions. I tweet sparingly (on average 3 or 4 a day) and virtually always only about blog posts or published work. I get 107 retweets per 100 tweets, and almost always respond to tweets with further answers.










Request: Could I ask university teachers and researchers to select one of their students, and get them to critique one of my posts, or to take up one of the suggestions for further research? It would be good to have more psychology student readers (not that I know how many of my readers are in that category) and I think that will come if psychology teachers get their students to have a look at the blog.

If you have any suggestions for getting more blog readers please let me know.

The Donate button is down on the bottom right. Just one gentle press does the trick. $35 buys readership of a pay-walled paper (this is just a price guideline) $20 buys a printer ink cartridge, $15 any number of coloured pencils. What more do I need, other than to keep up my enthusiasm, secure in the knowledge my readers are changing the world?

Thursday, 24 September 2015

Types of psychology lecture


It was said of Presidential Addresses at the British Psychological Society Annual Conference that they fell into three types:

Whither Now?

Patients I have Cured


These three encapsulated the philosophical, clinical and physiological traditions in the society.

With that in mind, I have looked back at the very recent International Society for Intelligence Research conference in Albuquerque to see if I can detect a classificatory structure, tripartite or otherwise. Here, without too much poetic licence, is a possible troika of themes:

Technical: use of statistics and modelling techniques, understanding the limitations and characteristics of particular intelligence measures, arguing about the hierarchical structure of intelligence

Correlational: Real-life associations with IQ, and examples of the predictive power of intelligence.

Genetic: the genetic underpinnings of intelligence and related behaviours.

The technical theme is very specialised. It makes crucial points about intelligence measures and how results can be modelled and analysed. Tracking down whether tests show “measurement invariance” is essential if you want dependable findings. Understanding all this is crucial for researchers. Speaking personally, I find some of the discussions about the structure of group factors less interesting.

The correlational theme is enormous in scope, and accounts for the bulk of published results. Intelligence runs through psychology like carbon through biology. All of human life is there. Intelligence is the most replicated result in psychology, and with the largest sample sizes, sometimes in the millions. There is much still to learn, and the results keep coming in.

The genetic theme is transformational. This is the leading edge of intelligence research. The tempo seems to be about one major publication every 2 or 3 months. Sample sizes are usually above 100,000 and sometimes 300,000. These papers usually find links between the genome and human behaviour which are statistically significant but moderate in effect size, and very probably caused by very many genes of small effect, and which also have effects on other things. I get forewarning of a few of these publications, and will comment further when papers in the review pipeline get published. Tracking down the genetics of intelligence is happening now, with impacts which most people don’t yet appreciate.

The classificatory scheme is a mere sketch, and very open to counter-claims. It might be better to follow the path outlined by Borges in classifying animals in “Celestial Emporium of Benevolent Knowledge” : Those that belong to the Emperor; Embalmed ones; Those that are trained; Suckling pigs; Mermaids; Fabulous ones; Stray dogs; Those that are included in this classification; Those that tremble as if they were mad; Innumerable ones; Those drawn with a very fine camel hair brush; Those that have just broken the flower vase;  Those that resemble flies from a distance.

Perhaps all lectures should be judged by the criterion “Have they just broken the flower vase?”


Sunday, 20 September 2015

#ISIR15 ends, celestial carriages await


So, thus ends a stellar conference.

One of the delights of a conference is to sit next to like-minded and knowledgeable confederates who feed me comments, evaluations and questions which need to be asked. The audience should have the last word. So, here is a selection of what such persons said to me about the talks.

James Lee’s talk went very well. At the beginning I thought it was an over-sell, but boy, the flow of his argument is terrifically clear.

I heard a lot of audience reaction after Steve Pinker's talk.  Comments like 'inspired, entertained, definitely going to try harder to write more clearly'.  There was quite a podium-rush after his talk - felt for his safety! Really good to have the hall ringing with laughter at the end of a long day.

PhD student Sephira Ryman gave a standout talk.  She asked: since men and women have similar mean intelligence, yet women have smaller brain sizes, are there other features that differ?  She found that gray matter volume is important for men, but white matter network connectivity was more important for women. Evidence from her sample of 244 persons that men and women may arrive at their intelligence by slightly different means.

Paul Sackett and Nathan Kuncel utterly destroyed the idea that SAT tests do not predict college performance. Their "ginormous" dataset comprised over a million students. A droll and data rich talk, they left myths about the non-utility of standardised tests lying like road-kill on the highway of evidence.

PhD student Helen Davis gave a fascinating talk that contrasted the spatial abilities and mobility patterns of two traditionally-living (forager-horticulturalist) peoples: the Maya and the Tsimane.  The lifestyle and ecology differ between the societies and this is reflected in their spatial abilities and movement patterns. The typically found sex difference in spatial ability (men outperform women) was only found in the Maya where men travel greater distances to find food for their children.

Alice Dreger. Simply barnstorming. Brilliant, and packed with rich content.  We must keep in contact with her.

Tim Bates' talks are consistently a highlight of any meeting he speaks at.  His careful replications, showing null results of famous memes that tear through the classroom like flu, are a pleasure to hear.

IN CONCLUSION – See you in St Petersburg, 15-17 July 2016

Russian and UK school kids



Elaine White1,2 , Margherita Malanchini1,2 , Dina Zueva2 , Olga Bogdanova2 , Yulia Kovas1,2

1 Goldsmiths, University of London, UK,

2 Tomsk State University, Russia.

Research suggests that within any country, almost the whole spectrum of individual variation in academic achievement is observed in any school or classroom, with only a small portion of within-population variance attributable to differences across teachers, classes and school (e.g. Asbury et al., 2008).

It may be that shared effects of class/teacher are weaker or stronger as a function of such factors as teacher training, curricula, educational norms, and cultural stereotypes (e.g. Kovas et al., 2013). As longitudinal research into teacher/classroom effects are limited to date and neglect the contribution of non-cognitive factors, this study investigates teacher/classroom effects on academic achievement, across several points of the academic year in two countries.

This longitudinal study follows 622 11-12 year old Russian and UK secondary school students at several waves across one academic year. As students have subject-specific teachers for the first time in their education, comparisons can be made between their classrooms for two subjects, maths and geography. The students from 3 urban schools completed a range of tests and self-report questionnaires during their maths lesson. Data were collected to assess cognitive and non-cognitive factors in relation to academic progress. The students’ school achievement data were also obtained.

We explore differences: across the two countries; within and between classes; across the two school subjects; and motivational factors. Preliminary results (from the first 3 waves) suggest stability of the measures, maths ability and maths self-efficacy, over time. A reciprocal relationship was shown between maths ability and maths self-efficacy across time 1 and time 2. This suggests that higher performance increases self-efficacy and higher self-efficacy increases performance. This reciprocal relationship remains when controlling for IQ and the relationship strengthens between ability at time 1 and self-efficacy at time 2. A negative relationship, which appears between ability at time 2 and self-efficacy at time 3, is likely to be the result of performance feedback.

This research investigates potential differences between Russian and UK education systems comparing classroom environments of mathematics in contrast to geography. Although taught and utilised differently, both academic subjects contain similar attributes. Both Russian and UK secondary school students have specific subject teachers for the first time in their education. UK students have the same teacher for all subjects during primary school and changes yearly, whereas Russian students have the same teacher throughout the four years of their primary education. The study therefore provides an ideal comparison of cognitive and non-cognitive factors across subject and classroom environments. Identifying factors moderating classroom effects is important for educational policy and provision.

Remember the 7 tribes of intellect


Take a dozen eggs. Better still, take several dozen eggs and compare them to another several dozen eggs. Eggs are eggs, and an omelette make.

However, from the individual differences perspective, humans differ. Brighter kids learn faster, about 5 times faster than their slower classmates. Take a whole school district and you will find a few children who learn 7 times faster, hence

Here’s a deal: we will improve our experimental designs if they will measure, even very briefly, the ability and personality of their experimental subjects.Even a simple brief vocabulary test, plus a digit span test or speeded coding task would provide useful information, and if parents could be persuaded to do the same we would have a handle on a major source of unexamined variance in experimental designs.

As Sara says: There is a world outside of experimental designs


Sara A. Hart Florida State University,

There has been a growing body of work, which suggests that the individual traits that a child brings into an intervention project have an interactive effect on literacy learning. Even within intervention studies shown to be impactful at the mean level, there are individual differences in how children responded to the intervention.

I contend that there are numerous (typically unmeasured) sources of these individual differences, and for this talk I will present data examining the role of both crystallized and fluid intelligence in predicting individual differences in response-to-intervention, with data pooled across multiple projects allowing for generalization beyond any given intervention protocol. Integrative Data Analysis (IDA; Curran & Hussong, 2009) was used to create a pooled source of Project KIDS raw data of 545 kindergarten and first grade children (age M = 5.6yrs) who had previously participated in one of three literacy-based randomized control trial interventions in the treatment group.

IDA allows for raw data from each project to be combined and heterogeneity, such as age and project, controlled for. Reading was measured as pre- and post-intervention scores on the Woodcock Johnson Tests of Achievement Letter-Word Identification (LWID) subtest, crystallized intelligence was measured using a pre-test mean raw score across the KBIT-2 Verbal Knowledge and Riddles subtests, and fluid intelligence was measured using a pre-test raw score from the KBIT-2 Matrices subtest.

As a first step of IDA, a moderated nonlinear factor analysis was used to create scale scores which are project invariant for the constructs of interest. I then used Proc Mixed to calculate covariance adjusted scores to model change from pre-test to post-test for LWID, operationalizing “response-to-intervention”. Quantile regression was then used to model both crystallized and fluid intelligence predicting response to-intervention.

The models indicated that both crystallized and fluid intelligence were statistically significant predictors across the distribution of response-to-intervention, although for both, the effect was statistically greater for the students who made the greatest gains due to the intervention.

These results indicate that brighter children do even better in an intervention that is impactful for most students. Although certainly not surprising for the audience of ISIR, child traits such as intelligence are not often included in determining response-to-intervention in education studies, and I argue that it is important moderator that should be considered. Beyond these specific findings, I will discuss how we will use these pooled data to exploring many other sources of moderation of response-to-intervention, including other cognitive traits, behavioral traits, the environment and family history. This work will expand the understanding of how and why some children are more successful when receiving gold standard educational interventions.

Older fathers still have bright children


This is an interesting paper, but I note it refers to European populations. It may not hold true of societies in which many children are the product of older men accumulating many younger wives.



Ruben C. Arslan 1 , Kai P. Willführ 2 , Emma M. Frans 3 , Mikko Myrskyla 4 , Catarina Almqvist 3 & Lars Penke

1 Georg August University Göttingen, Germany,

2 MPI for Demographic Research, Rostock, Germany.

3 Karolinska Institut, Stockholm, Sweden.

4 MPI for Demographic Research, Rostock, Germany.

Ruben Arslan


Paternal age at offspring conception seems to be the main driver of single nucleotide de novo mutations (Kong et al.., 2012). Different theories posit that intelligence is linked to mutation load as a fitness indicator or simply owing to its genetic complexity. Based on evolutionary genetic theory we predicted negative paternal age effects on offspring fitness and intelligence in the normal range. To investigate effects on fitness, we used church records from three pre-industrial Western populations and governmental data from 20th century Sweden. We used a sibling control design and accounted for confounds including maternal age, birth order and parental loss. Main analyses had an aggregate N > 1.3 million.

To investigate effects on intelligence, we compared siblings in the German Socio-Economic Panel (N = 1479). Furthermore we were the first to directly adjust for measured parental intelligence, the most obvious confound, in data from the Minnesota Twin Family Study (N = 1898 twin pairs). We found clear support for mutational paternal age effects on offspring survival, mating and reproductive success. Weaker effects were found in 20th century Sweden, possibly indicating a diminished strength of purifying selection. However, we found no mutational paternal age effect on offspring intelligence, which was corroborated further by a Swedish study of half a million men (D’Onofrio et al.., 2014).

Although paternal age effects seem to be an appropriate way to characterize the effect of de novo mutations on fitness, no effect was found on intelligence in the normal range. Genomic research supports this result. The inferred genetic architecture of intelligence does not seem to make it fragile and vulnerable to increases in paternal age-driven mutation or to decreases in purifying selection.

Creativity and fluid intelligence


Finally, Brazilians take the stage, in the form of Paulistano Ricardo Primi, who gives a creative take on creativity, looking at the difficult-to-measure genre of figural drawing. Brazil needs to figure more in international intelligence research, particularly on the large matter of genetic differences, since Brazil’s history is very different from that of the US, and the contrast can provide a test of cultural explanations for black/white intelligence differences. That bigger project will have to wait, but here is what they have done on their drawing task, also using Bootstrap to test model fit.



Ricardo Primi 1 , Nelson Hauck-Filho 1 , Tatiana de Cássia Nakano 2

1 Universidade São Francisco,

2 PUC-Campinas.


This study examines the association of fluid intelligence and creativity. In divergent thinking tests it is common to observe that later responses tend to be more creative than earlier ones – this is called serial order effect. Recent view of the role of executive function on divergent production predicts that high fluid intelligence subjects will have creative responses already in the beginning of divergent thinking tasks. This indicates a central role of executive functions –inhibiting common less creative responses and management interference on idea production.

Most studies observing these relationships are done in verbal tasks. This research tests if this relationship can be found on divergent productions of figural drawings. Participants in the present study were 585 children and high school students with ages from 7 to 17 (mean = 11.11 years, SD = 2.02; 52.5% female). All participants provided demographic information on a self-report questionnaire, and undertook a cognitive assessment battery (verbal, abstract, logic and numeric reasoning) supplemented by a creativity task, whose data we analyzed in the present study.

This creativity task consisted of 10 stimuli, which participants were required to complete using paper and pencil. Independent raters subsequently coded each resulting drawing in a scale from 1 to 5 to reflect the extent to which it approached a set of criteria defining creative responses. Data analysis was conducted using Mplus 7.11. Factor growth mixture modelling were performed in order to detect groups of potentially differing patterns of performance (ratings) from the first to the last stimulus of the task.

Bayesian Information Criterion (BIC) and the Bootstrap Likelihood Ratio Test (BLRT) suggested that a three-class solution was a better fit to the data (entropy = .77) when compared to alternative 1-, 2-, and 4-class solutions. Latent classes revealed a large group (83.36%) of individuals with initially modest scores and descending performances along the 10 stimuli, as well as two small groups of individuals with high initial scores—one (12.52%) with a descending performance, and the other (4.12%) with a stable high performance across the whole task.

Last two groups have significantly higher scores in Gf. This study shows that executive processes of top down voluntary control are important components for production of creative responses. This demonstrates a higher role of intelligence on creative idea production. It shows a high role of fluid intelligence in idea production.

Sex differences in chattering and counting


People come in two types: those that chatter and those that count.

Women are more likely to chatter, men to count. Women incline to non-STEM subjects, men to STEM; as a consequence they go on to have different jobs and careers.

For both sexes, chatterers turn to humanities, counters to STEM.

This may be evidence of a sex difference.

Tom and his team’s excellent work makes it plain that there are some natural inclinations and ability profiles (tilts), not only between sexes, but within sexes and between those with different cognitive abilities. I don’t see this as something which needs remediation and some special campaign (and doubt that Tom thinks so). By all means offer maths classes and writing classes to those that want them. Indeed, offer them to those that don’t want them, so that the Two Cultures don’t divide permanently.



Thomas R. Coyle, Miranda C. Richmond and Anissa C. Snyder, University of Texas at San Antonio,

Although g is the best predictor of life outcomes (academic achievement), non-g factors may also predict such outcomes. One non-g factor with predictive validity is ability tilt, the difference in math and verbal scores on tests (SAT). There are two types of tilt: math tilt (math>verbal) and verbal tilt (verbal>math). Whereas prior studies have examined tilt in the profoundly gifted (1 in 10,000 in ability), the current study examines sex differences in tilt in nongifted subjects. Nongifted subjects show fatter ability profiles (less tilt), which may lower the predictive validity of tilt.

Unlike prior studies of nongifted subjects, the current study is the first to examine sex differences in tilt for outcomes after college (occupations, but also college majors and specific abilities).

Subjects (866 males and 1084 females) were drawn from the National Longitudinal Survey of Youth, a representative sample in the U.S. Tilt was based on math minus verbal scores on the SAT, ACT, and PSAT (in high school). College majors were STEM majors (science, technology, engineering, math) and humanities majors (English, history, languages). Occupations (around age 30) were STEM jobs (chemist, biologist, engineer) and verbally-loaded jobs (media, law, counseling). Specifc abilities (math, verbal) were based on the Armed Services Vocational Aptitude Battery (ASVAB).

Sex differences in levels (or frequency) of tilt were examined with ANOVAs (chi-squares), and tilt relations with specific abilities were examined with structural equation modelling. Significant effects are reported at p<.05. Tilt was unrelated to g (based on the ASVAB) for males and females (r < .10), confirming its non-g status.

Math tilt (math>verbal) on all tests (SAT, ACT, PSAT) was more common for males, whereas verbal tilt (verbal>math) was more common for females. In addition, STEM majors and jobs were more common for males, whereas humanities majors and verbally-loaded jobs were more common for females. For both sexes, STEM majors and jobs were associated with math tilt, and humanities majors and verbal jobs were associated with verbal tilt (r ~ .35). Also for both sexes, math tilt predicted math ability (on the ASVAB), and verbal tilt predicted verbal ability (betas ~ .30), confirming the construct validity of tilt.

The results were confirmed for all tests (SAT, ACT, PSAT). Tis study is the first to examine sex differences in tilt for nongifted subjects. Tilt was unrelated to g, confirming its non-g status, but still differentiated males and females. Males tended to show math tilt, which predicted STEM outcomes, whereas females tended to show verbal tilt, which predicted verbal outcomes. The absence of sex differences in tilt relations (with jobs and majors) suggests that males and females with math tilt prefer STEM whereas those with verbal tilt prefer humanities.

An important question for public policy is why fewer females show math tilt (which may reduce STEM participation). Future research will examine tilt at earlier ages (elementary school) and also examine other abilities (spatial ability) that may contribute to sex differences in tilt and STEM.

Multiple objects, unitary ability?


Tracking multiple objects is one of the essential tasks at any conference, though the objects would prefer to be called distinguished researchers, which is what they are. Researchers, speakers, confederates, people who need to be persuaded about something, all swirl around the Barcelona conference room and the associated coffee spaces, with minimal supervision. I suppose, in the interests of keeping track on them, I could ask them to wear identifying hats. The task of tracking multiple targets must have been important in our hunter gatherer past, and so might be a candidate task for a dedicated module somewhere in the brain. I am a generalist rather than a modules man, but I make an exception for faces. (I never forget a face, but in your case I'll be glad to make an exception). I assume that the ability to track multiple objects is an advantage if one is shadowing people through city streets.

Jeremy Wilmer and colleagues have had a look at this, and believe they have identified that multiple object tracking is driven by a relatively specific cognitive mechanism that is not reducible to general ability.


Jeremy B Wilmer 1 , Ken Nakayama 2 , Laura Germine 3

1 Wellesley College,

2 Harvard University.

3 Massachusetts General Hospital.


Jeremy Wilmer

Though recent decades have seen massive improvements in our understanding of cognitive and neural mechanisms of behavior, the task of dissecting human ability into its basic cognitive and neural components has remained elusive (Deary, 2000, 2010).

Here, we examine a potential new tool for such dissection: multiple object tracking (MOT), the elementary cognitive task (ECT) of keeping track of several moving objects amongst moving distractor objects. MOT’s neural substrates have been well characterized (Jovicich et al., 2001) and the bulk of reliable individual differences in MOT can be explained by a single posterior event-related potential (ERP) component (Drew & Vogel, 2008). Yet of the 145 papers on MOT to date, only one other has examined individual diferences (Oksama & Hyona, 2004).

Our aim was to characterize the cognitive correlates of MOT. To this end, we developed a brief (8-10 min), reliable (Cronbach’s alpha = .88), web-based measure of MOT performance. We posted this test and several others on our website, which has succeeded in sustainedly attracting a large number of motivated volunteer participants (~1.5 million since 2008) by providing engaging tests that give performance feedback.

We tested a total of 19,724 participants (10,014 male) of varying ages (5th, 50th, and 95th percentiles 14, 25, and 58 years, respectively), educations (e.g. 34% of 25+ year olds had no more than a high school diploma), and ethnicities (46% non-European). MOT dissociated strongly from reliable (Cronbach’s alpha > .80), well-validated tests of sustained visual attention (r(1930)=-.03) and vocabulary (r(2810)=-.09), evidence that individual differences in MOT performance are driven by a relatively specific cognitive mechanism that is not reducible to a single general ability. MOT associated strongly with reliable (Cronbach’s alpha > 0.80), well-validated tests of spatial working memory (r(10841)=.44) and rapid spatial attentional switching (r(533)=.41), evidence that MOT captures a mechanism important for visuospatial performance.

Finally, MOT robustly predicted SAT-math performance (r(2467)=.29), but less so SAT-verbal performance (r(2467)=.09), evidence for a relatively specific link of MOT performance to math potential. Here, we report a study of MOT that is nearly two orders of magnitude larger than the entire prior literature on MOT.

Our results demonstrate that individual differences in MOT can be reliability and validly measured. We develop, norm, and validate a brief, reliable, new web-based test of MOT. By characterizing several key cognitive correlates of (and dissociations from) MOT, we place MOT firmly in a broader framework of human abilities. Given advances in our understanding of the neural correlates of MOT, we suggest that MOT could aid future efforts to dissect human ability into its core neuro-cognitive components.

#ISIR15 Last day


Many talking points from yesterday, and I hope to get more of the presentations to send on to you later. However, the morning begins with a look at a very severe problem. Given that intelligence is so important, why is the public debate three or four decades behind the times?

Alice Dreger kicks off with Understanding Science Journalists and Why They Misunderstand You. Alice is an historian of science, science writer, and patient rights activist, and will draw several cautionary tales from her latest book, Galileo’s Middle Finger: Heretics, Activists, and the Search for Justice in Science.

For many scientists whose work touches on identity politics, dealing with science journalists can be a fraught endeavor. Work with them, and you risk being misrepresented. Decline to work with them . . . and you risk being misrepresented. Tis lecture will draw on the speaker’s 20 years’ experience working with and within science journalism to explore how the field has changed in the last two decades and how scientists can protect themselves today in the media.

She has a very good title, and she helps us understand our predicament: we have so many new and interesting results, but the concepts “intelligence” and particularly “IQ” are toxic. As Pinker tweeted:  Irony: Replicability crisis in psych DOESN'T apply to IQ: huge n's, replicable results. But people hate the message.

As a researcher or explainer I have personal experience of having to wade through so many misconceptions, and sometimes politely veiled disdain, that it is hard to keep absolutely calm and give the results.

But I am locked into my version of the conference bubble. Has there been any reporting on the conference, any at all? If there has been, can you send me some links?

Saturday, 19 September 2015

Two feasts, The Curse of Knowledge and Good Night


Tonight the traditional banquet will take place, involving some calories, some delicate imbibing of local wines, and the production of more decibels than a stadium of orators. By a further tradition, no specific mention is made of subjects discussed, nor of the persons showing up very late the following day. Anyone can oversleep.

Before that banquet, another feast. Steven Pinker, who has written so well about psychological matters, will be talking about the craft of writing. He identifies one of the strangest states into which any thinking person can fall, particularly when they consort with other thinkers: the delusion that, having understood something, anyone else can also understand it, and probably already knows it, or will very soon .

By way of tantalus, he says:

As a user, promoter, and occasional practitioner of intelligence research, I discuss a little-appreciated aspect of scientific practice that has significant implications for the future of the field: writing. Most academics are terrible communicators. Why do the world’s most cerebral people find it so hard to convey their ideas? And how can we learn to do better?

The classic style manuals are based on the personal intuitions of journalists and writing instructors and tend to mix some helpful hints with some harmful folklore. I suggest that the modern sciences of mind and language can provide sounder and more systematic guidance to writers today. Thoughtful writers should begin with a clear idealization of the simulated scenario in which they are communicating with their readers. They must overcome The Curse of Knowledge – the inability to imagine what it’s like not to know what they do know. They should understand how syntax works and thus how best to deploy the grammatical resources of the English language. And they should know how to think about the rules of correct usage, distinguishing the ones that enhance clarity and grace from the shibboleths and superstitions.

White matter efficiency predicts women’s IQ, but not men’s

Here is a curious and interesting finding, best understood without mentioning either Mars or Venus. A prediction based on white matter efficiency works for the women subjects in this study, but not for the men. Do women really rely more on efficient brain operation, while men just organise their crania like a bachelor’s bedroom?

FRONTO-PARIETAL GRAY MATTER AND WHITE MATTER EFFICIENCY DIFFERENTIALLY PREDICT INTELLIGENCE IN MALES AND FEMALES* Sephira G Ryman1 , Ronald Yeo1 , Katie Witkiewitz1 , Martijn van den Heuvel2 , Marcel de Reus2 , Andrei Vakhtin1 , Ranee Flores3 , Christopher Wertz3 , Christine Meadows3 , Rex E. Jung3

1 University of New Mexico,

2 University Medical Center Utrecht.

3 University of New Mexico HSC.

Intelligence is associated with communication efficiency of a widespread neural network as well as variation in gray matter volume, particularly within the fronto-parietal regions of the human brain. Recent reports of gender differences in the relationships between brain measures and intelligence highlight the need to differentiate how the brains of males and females may rely on different neural structures when completing measures of intelligence.

The current study utilizes a network approach in conjunction with structural equation modelling to examine potential gender differences in the relationship between white matter efficiency, fronto-parietal gray matter volume, and general cognitive ability.

Two hundred and forty-four (21.77+/−3.29 years; 125 males) subjects participated in the current study. Individual brain networks were modelled on the basis of the set of reconstructed fiber tracts. Network white matter efficiency was calculated as the average inverse shortest path length across the whole network. Two confirmatory factor analyses (CFA) were conducted to identify the factor structure and measurement invariance of a general cognitive factor (GCA) and a fronto-parietal gray matter volume factor (FPG). Lastly, we proceeded to fit a structural equation model relating the GCA latent factor to both the FPG and white matter efficiency. A nested model comparison and examination of the interactions were conducted to determine if the regression coefficients differed in each gender.

Results of the GCA CFA indicate adequate model fit [χ² (5) = 6.82, p = 0.23, RMSEA = 0.04 (90% CI [0.00-0.10]), CFI = .98], with the GCA factor exhibiting metric invariance and partial scalar invariance across gender. Results of the FPG CFA indicate adequate model fit [χ²(2) = 0.36, p = 0.84, RMSEA < 0.01 (90% CI [.00-.07]), CFI = 1.00], exhibiting metric, scalar, and residual invariance across gender. A nested model comparison of the model predicting GCA before and after constraining the regression coefficients across gender resulted in a significant χ² difference between the models (χ²(2) = 8.33, p = 0.01. There was a significant relationship between FPG and CGA in males and females. In contrast, white matter efficiency significantly predicted GCA in females, but not males.

The current study aimed to identify the relationship between intelligence, fronto-parietal gray matter volume, and white matter connectivity. Results of the structural equation model highlight that the latent factor of fronto-parietal gray matter volume predicts GCA, with greater fronto-parietal gray matter corresponding to greater GCA scores. Of interest, while the relationship between fronto-parietal gray matter and GCA is consistent across males and females (with larger effect sizes in males), white matter efficiency demonstrated differential effects across gender.

The current study provides further evidence for this trade of in brain structure seen in males and females, suggesting that women likely rely more on efficient brain organization when performing measures of intelligence.

Boost your IQ (hard matrices problems only)

Here is an intriguing finding, from a team I last saw in a very agreeable restaurant in Graz. They have found a way of boosting performance on difficult items of a matrix problem task, by using in-task transcranial stimulation. They are tentative, but speculate that the externally imposed effect might mimic some cortical control process, which could be very productive.


Image result for aljoscha neubauer


Aljoscha C. Neubauer1 , Martin Wammerl1 , Mathias Benedek1 , Emanuel Jauk1 , Norbert Jausovec2

1 University of Graz,

2 University of Maribor.

Recently, the issue was raised whether fuid intelligence can be increased through (mostly working memory) training. Despite some initial supportive evidence (Jaeggi, et al., 2008) more recent studies did not find support for training effects (e.g. Colom et al., 2013; Owen et al., 2010). Moreover, the method is critically discussed for methodical problems (Shipstead, et al., 2012). It seems questionable whether intelligence can be enhanced through cognitive training in a consistent and long-lasting manner (Haier, 2014).

We tested whether fluid intelligence (gf) could be enhanced by directly influencing brain activity through transcranial alternating current stimulation (tACS) applied to the parietal cortex and at the same time measuring concomitant changes in brain activity by fMRI. In a double-blind, sham-controlled experiment 20 participants performed two fluid intelligence tasks (matrices task, and paper folding & cutting, PFC) after either transcranial alternating current stimulation (tACS) at theta frequency or sham stimulation was applied.

The stimulation site was the lef parietal cortex (P3), because of its key role for intelligence and intelligence-related functions (e.g., working memory capacity, verbal memory, spatial working memory). Sham and verum conditions were realized within-participants (using parallel test versions of both tests) with the two sessions separated by 28 days. While working on the two tasks stimulation-induced brain activity changes were recorded using fMRI.

Results indicated task- and difficulty-specific stimulation effects: When solving difficult items of matrices test verum tACS significantly increased gf performance, as compared to sham. No difference emerged for easy items.

In the second gf task (PFC), tACS had no effect. For Raven matrices test whole-brain analyses showed that left parietal brain stimulation was accompanied by less right sided activation in areas of the frontal lobe, fusiform gyrus and occipital lobe, as well as left sided middle occipital gyrus. Additional ROI analyses revealed a tendency for less activation in the left inferior parietal lobule.

We conclude that left parietal theta tACS could enhance performance in a reasoning task but only for difficult items; and not for a gf task requiring mental rotation. We presume that theta frequency resembles a general cognitive control process, which might be of importance for gf performance. Neurophysiologically, the tACS-induced reductions of brain activation primarily concerned brain areas that partly overlapped with the task-negative network (e.g., precuneus). This seems in line with the neural efficiency hypothesis, that higher gf is related to lower activation, mainly in task-negative brain regions (Basten et al., 2013; Dunst et al., 2014; Neubauer & Fink, 2009).

But so far, the current findings can only be interpreted as transient increases in gf test performance rather than in gf per se. Though the enhancement could be intelligence-specific, also other processes might account for tACS induced increases in gf performance (cf. Haier, 2014).




David H Schroeder

Johnson O’Connor Research Foundation,

Research Department staff


Although there have been many studies of the heritability of general intelligence (g), much less attention has been directed at specific cognitive abilities. In this study, data for specific abilities from monozygotic (MZ) and dizygotic (DZ) twins and also siblings were examined.

Samples: All of the participants in this study were clients of the Johnson O’Connor Research Foundation’s vocational-guidance program. For twins, there were 94 MZ and 261 DZ pairs and for siblings, 1,641 pairs. There were roughly equal numbers of males and females, and age ranged from 14 to 47 although most of the participants were high-school- or college-age. Effects of age were partialled from all scores.

Participants were generally upper-middle-class and college-oriented.

Measures: Participants took 15 tests of specific cognitive abilities, which yield 4 group factors and a g factor. They also took 11 tests of non-cognitive abilities including 3 tests of music-related abilities (e.g., Pitch Discrimination). Heritability estimates for the factors ranged from .85 for the Spatial Ability and g factors to .66 for Speed of Reasoning and Memory. For the specific-cognitive-ability tests, many of the values lined up with their factors—e.g., .94 and .62 for the spatial tests and .62 and .52 for the memory tests.

Interestingly, Rhythm Memory, one of the musical-ability tests, had a relatively high value, .80. The highest values for the shared-environment component were for the vocabulary test and 2 of the musical-ability tests.

The results with only same-sex DZ twins were generally similar to the results with the full DZ group. The large sample of siblings revealed a moderating effect for parents’ educational level, with higher familiarities for participants with higher levels of parental education. These results indicate a substantial genetic influence for specific cognitive abilities and group factors in addition to g.

The moderating effect for parents’ education, which is broadly consistent with previous studies of family SES, adds an important element to our understanding of genetic factors and suggests further avenues for research.

Scholar in 86 SNPs?


It is good to see James Lee on the speakers list, because he told me long ago that samples of at least 100,000 would be required to link any genes with behaviour, and that he might be under-estimating the sample size required. Now he finds some interesting links with a sample of 305,000 which look very promising.



1 University of Minnesota Twin Cities,

2 Social Science Genetic Association Consortium.

Genome-wide association studies (GWAS) have revealed much about the biological pathways responsible for phenotypic variation in many anthropometric traits and diseases. Such studies also have the potential to shed light on the developmental and mechanistic bases of behavioral traits.

Toward this end we have undertaken a GWAS of educational attainment (EA), an outcome that shows phenotypic and genetic correlations with cognitive performance, personality traits, and other psychological phenotypes. We performed a GWAS meta-analysis of ~305,000 individuals, applying a variety of methods to address quality control and potential confounding. We estimated the genetic correlations of several different traits with EA, in essence by determining whether single-nucleotide polymorphisms (SNPs) showing large statistical signals in a GWAS meta-analysis of one trait also tend to show such signals in a meta-analysis of another. We used a variety of bio-informatic tools to shed light on the biological mechanisms giving rise to variation in EA and the mediating traits afecting this outcome. We identified 86 independent SNPs associated with EA (p < 5E-8). The robust agreement between estimates of effect sizes derived from population samples and family cohorts suggests that very little of this signal is due to confounding.

We found that both cognitive performance (0.82) and intracranial volume (0.39) show substantial genetic correlations with EA. Many of the biological pathways significantly enriched by our signals are active in early development, molding features of the brain such as its overall size, morphology, and architecture of axon-dendrite connections. Genes where de novo mutations can cause intellectual disability or autism spectrum disorder also significantly enrich the loci where our GWAS signals are found. We nominate a number of biological pathways and individual genes of likely importance in the etiology of EA and mediating phenotypes such as cognitive performance.

Our results provide insight into the developmental mechanisms giving rise to individual differences in these traits and a starting point for an analysis of evolutionary change in the human lineage.

Genius? Avoid rare functional genes


For some time now I have been waiting for the results of the genetic analysis of genius, even while knowing that intelligence is heritable, but created by many, many genes of small effect. So, needle, haystack, and the needles aren’t just for sewing, but do other things as well.

However, on the good side, I believe that Plomin will nail the genes for intelligence before they nail him. This paper brings us a step closer to understanding the genetic underpinnings of exceptionally high intelligence, suggesting it happens when some intelligence damaging genes are not present. The team have looked at only the subset of DNA that encodes proteins,  because mutations in those sequences are much more likely to have severe consequences.

Image result for robert plomin king's college london

A GENOME-WIDE ANALYSIS OF PUTATIVE FUNCTIONAL AND EXONIC VARIATION ASSOCIATED WITH EXTREMELY HIGH INTELLIGENCE Robert Plomin1 , Sarah L. Spain2 , Inti Pedroso2 , Neli Kadeva2 , Mike B. Miller3 , William G. Iacono3 , Matt McGue3 , Evie Stergiakouli4 , George Davey Smith4 , Martha Putallaz5 , David Lubinski6 , Emma L. Meaburn7 , Michael A. Simpson2

1 Institute of Psychiatry, Psychology & Neuroscience; King’s College London,

2 Division of Genetics and Molecular Medicine, King’s College London.

3 Department of Psychology University of Minnesota.

4 MRC Integrative Epidemiology Unit University of Bristol.

5 Duke University Talent Identifcation Program Duke University.

6 Department of Psychology and Human Development Vanderbilt University. 7 Department of Psychological Sciences Birkbeck, University of London.

Although individual differences in intelligence are highly heritable, molecular genetic analyses to date have had limited success in identifying specific loci responsible for its heritability. The present study is the first to investigate exome variation in individuals of extremely high intelligence. Under the quantitative genetic model, sampling from the high extreme of the distribution should provide increased power to detect associations.

We performed a case-control association analysis with 1,409 individuals drawn from the top 0.0003 (IQ > 170) of the population distribution of intelligence and 3,253 unselected population-based controls. Our analysis focused on putative functional exonic variants assayed on the Illumina Human Exome BeadChip. We did not observe any individual protein-altering variants that are reproducibly associated with extremely high intelligence and within the entire distribution of intelligence. Moreover, no significant associations were found for multiple rare alleles within individual genes.

However, analyses using genome-wide similarity between unrelated individuals (Genome-wide Complex Trait Analysis) indicate that the genotyped functional protein-altering variation yields a heritability estimate of 17.4% (SE 0.017) based on a liability model. In addition, investigation of nominally significant associations revealed fewer rare alleles associated with extremely high intelligence than would be expected under the null hypothesis.

A common theme emerging from genetic studies of intelligence, similar to all complex traits and common disorders, is its highly polygenic nature with its heritability explained by many variants of small effect. While the unique extreme sampling design used in the current study provides improved power to detect associations in certain situations it has also provided challenges for direct replication. Nevertheless, the evidence for the contribution of protein-altering variants to the heritability of intelligence and the evidence that rare functional alleles are detrimental to intelligence provides a framework for defining the role of individual rare alleles.

g for Latinos

When genomic analysis becomes cheap and reliable we will be able to give our position in population space with exquisite accuracy, but for the time being we will have to use the occasional classification of genetic and cultural groups can by their language. Yes, languages often parallel genetics (why not) but Hispanic is a broad church, though 63% are Mexican.


Miranda C Richmond and Thomas R Coyle

University of Texas at San Antonio,

This study examines relations among g, non-g factors (unrelated or weakly related to g), and academic performance. The non-g factors are grit and conscientiousness. Conscientiousness has been shown to predict academic performance beyond g, and grit has been shown to predict outcomes beyond conscientiousness. Ethnic differences in g and non-g factors have been examined primarily for Blacks and Whites.

The current study examines ethnic differences in g and non-g factors for Hispanic and Whites. In addition, the study compares the relative influence of conscientiousness and grit (beyond g) in predicting college GPA. Subjects (215 female, 80 male, mean age = 18.6 (±1.1) years) were drawn from the subject pool at the University of Texas at San Antonio. g was estimated using SAT subtest scores (math, verbal, writing), Wonderlic scores, and Raven’s Advanced Progressive Matrices scores. Conscientiousness was measured with the Big Five Inventory, and grit was measured with the Grit short form (Grit-S). College GPAs were based on academic records.

Ethnic differences in the predictive validity of g and non-g factors for GPA were assessed using multigroup (182 Hispanic, 83 White) structural equation modelling (SEM). Grit and conscientiousness were weakly related to g (r = -.13, -.06 respectively), confirming their non-g status. Multigroup analyses indicated that the direct effect of g on GPA (controlling non-g factors) was stronger for Whites than Hispanics (Whites: β = .446 (.079), p < .001; Hispanics: β = .398 (.045), p <.001).

Although the relation between g and conscientiousness was not significant for either group, the relation between g and grit was negative for Hispanics (β = -.086 (.026), p < .01) and positive for Whites (β = .144 (.038), p < .001). In addition, conscientiousness reliably predicted GPA (controlling grit) for both groups (Whites: β = .255 (.094), p < .01; Hispanics: β = .294 (.053), p < .001). Indirect effects were non-significant for both groups.

This study examined relations among g, non-g factors (grit and conscientiousness), and GPA for Hispanics and Whites. g predicted GPA (controlling non-g factors) better for Whites than Hispanics, suggesting that g is differentially predictive for Hispanics and Whites. This pattern is consistent with prior research showing that g predicts academic achievement better for higher ability groups (Coyle, Snyder et al., 2011). In addition, conscientiousness predicted GPA for both groups (controlling grit), but grit did not predict GPA for either group (controlling conscientiousness), suggesting that conscientiousness is a better predictor of academic achievement. These results have implications for the identification of non-g traits that might enhance student success.

Kids walking about

As far as I can recall, in my childhood kids walked about a lot, and fairly long distances from about 5 or 6 onwards. When we got a bit older we rode bicycles, and that was a big deal. Most of all, we wanted to drive cars, but weren’t allowed to, so we built crude versions of them. Real cars we drove on beaches as soon as we could reach the pedals, and then legally on the highway aged 16.

Can anything be learned by studying the walking about habits of primitive tribes, the sort of people who do not have iPhones?

Helen Davis has been doing this, with preliminary results showing that kids in such communities walk about a great deal more than in our vehicle prone wealthier world. Even in such Elysian tribes women are more worried about the great wide open, and less able to deal with it, though the differences seem rather small to me.

Helen has also done an experiment with chickens, but I don’t think she will be reporting on that at the conference.

Helen Elizabeth Davis portrait



Helen Elizabeth Davis, Karen Kramer, Elizabeth Cashdan

University of Utah,

Across a wide range of societies, males range farther than females, are more confident in their abilities to navigate and do better at many spatial activities. While these gender differences are well-documented in the west, much less is known about spatial exploration and cognition among people, especially children, in traditional societies. To address this empirical gap, the proposed research focuses on three main questions. 1) What is the age and gender patterning of spatial abilities among Maya and Tsimane adults and adolescents? 2) How do differences in harm avoidance perceptions shape boys’ and girls’ spatial behavior and cognitive reasoning? 3) How do these patterns vary across children whose families are experiencing novel socioeconomic changes in subsistence and education?

Using data collected in a remote rural community in the Yucatan peninsula of Mexico (N=130) and among the Tsimane forager-horticulturalists of central Bolivia (N= ~120) this project assesses the developmental processes that underlie spatial learning in a non-western context among adults and children (during the juvenile period (8-18)).

Path analysis and multi-level modelling help to evaluate data collected through interviews, tests of cognitive performance and spatial ability between boys and girls. Preliminary results among the Maya suggest that among adults, females report significantly higher levels of spatial anxiety (mean = 2.62, sd = .474) than males (mean = 2.45, sd = .486; t= −2.158, df=152, p = .033). Males were also more proficient at spatial ability tasks (t = −3.601, df=139, p <.001). However, early comparative results among the Tsimane suggest that environmental and social variation may play a significant role in the degree of difference found between sexes in these two populations.

This project presents a unique opportunity to learn whether age and gender differences in spatial cognition and anxiety, as documented in the west, are a generalizable feature of human ontogeny or are conditioned by specific aspects of growing up in a protected environment (Blackwell et al. 2011; Barkley and Gabriel 2007).

Analyses of Maya and Tsimane activity data suggest that, compared to people living in a market-based society, boys and girls are more independent and explore a broader physical environment at younger ages. However, recently, activities and behaviors have shifted as groups transition into the market economy, which offers a unique look at the cultural and behavioral impacts that shift will have on traditional child development within and across age groups.

Hick’s Law, reaction times and fluid g

I would be proud to have submitted any of the papers presented at ISIR15. Here is one I was initially tempted to skip over, because I know you are busy people, running the world and inventing stuff. However, on closer inspection it reminds me of what I imagined in would do in psychological research, but never got round to, what with one thing and another.

I should talk about this at more length another time, when I do not have a press of papers to tell you about, but if psychology is to progress it must tackle some fundamental problems. In a reaction time experiment, what is the effect of response uncertainty, such as choosing between 2 or 4 options?

Hick (1952) On the rate of gain of information. Quarterly Journal of Experimental Psychology Volume 4, Issue 1, 1952

The principal finding is that the rate of gain of information is, on the average, constant with respect to time, within the duration of one perceptual-motor act, and has a value of the order of five “bits” per second.

I suggest you forget about contingent negative variation for the time being.


1 University of Utah,

A large body of research has examined correlations between reaction time (RT) and intellectual ability, though the neural correlates of that relationship and its potentially moderating task effects are comparatively understudied. Markers of anticipatory neural activity such as the EEG Contingent Negative Variation (CNV) represent compelling possible mediators of RT-ability relations, and provide a means to study neural processes that may drive task-specific effects.

The goal of this study was to assess the effect of increasing numbers of response alternatives on CNV amplitudes preceding RT, and to evaluate relations therein to fluid intelligence (Gf). Forty-three, right-handed, undergraduate psychology majors were administered the WAIS-IV and performed a version of the Hick paradigm during EEG recording. The Perceptual Reasoning Index served as the measure of Gf. The Hick paradigm consisted of 96 trials each of simple, 2-choice, and 4-choice RT conditions, which were randomized within blocks. Each trial involved a cue indicating the condition, followed by a warning stimulus, a variable inter-stimulus interval (range 1-2.4 sec.), and a reaction stimulus.

Participants responded with their right hand. EEG data were epoched time-locked to the onset of the warning stimulus. CNVs were quantifed as the average amplitude from 875-1000 msec. post-stimulus at the central electrode Cz, and neighboring left (C3) and right-central electrodes (C4).

RT increased while CNVs generally decreased as a function of increasing response alternatives. Linear and quadratic trends were significant for both variables, indicating a smaller condition effect between the choice RT conditions. CNVs were larger overall at Cz relative to C3 and C4. Tis efect interacted with condition where activation decreased from simple to choice RT at Cz and C3, with no condition effect at C4. Gf significantly predicted overall, 2, and 4-choice RT (r ≥ -.36, p < .02), and the latter effects were significantly stronger than that for simple RT (z ≥ -2.0, p < .05).

Last, in the 4-choice condition, CNVs showed differential effects such that only C4 activation predicted RT (r = .41, p =.006), whereas only Cz showed a trend toward predicting Gf (r = -.29, p = .06). RT and anticipatory neural activation showed similar effects of increasing response alternatives. Detrimental trends of increasing alternatives on both RT and CNVs were negatively accelerated, indicating a larger difference between simple and choice RT than between different numbers of choice alternatives. The study also replicated findings of stronger RT-ability relations for choice than simple RT (e.g., Deary, Der, & Ford, 2001), but suggested complex effects on CNVs.

Taken together, the results suggest that identifying the mechanisms involved in simple vs. choice RT performance may help shed light on the neural substrates of RT-ability relations. Follow-up analyses of response-locked trials will clarify the time course and topography of activation most central to variation in RT and Gf.

Genetics of IQ and lifespan


At first, when an association was found between IQ and lifespan, many plausible reasons came to mind. Brighter people, it was argued, were the first to heed health warnings, adopt healthy life styles, read the medicine instructions carefully, and avoid putting their tender parts in meat grinders. All this is true. Brighter people stop smoking sooner than others, don’t get into so many accidents, and so on.

However, part of the IQ/lifespan links when you control for simple reaction times in middle to late age, suggesting there are underlying systems in common. If you are blessed with a good system, and have “system integrity” (a hypothetical construct dreamed up by Ian Deary when he looked at the first sets of data) then you also get the benefit of a healthy body and a longer life. Mens sana in corpore sano indeed.

Image result for rosalind arden kcl

Now Arden and team have found a genetic link to the overlap between intelligence and lifespan. Fingers crossed for all my (highly intelligent) readers.


Rosalind Arden1 , Michelle Luciano2 , Ian J Deary2 , Chandra A Reynolds3 , Nancy L Pedersen4 , Brenda L Plassman5 , Matt McGue6,7 , Kaare Christensen8 , Peter M Visscher9

1 CPNSS London School of Economics,

2 University of Edinburgh,

3 University of California, Riverside,

4 Department of Medical Epidemiology and Biostatistics Karolinska Institutet Stockholm, Sweden.

5 Department of Psychiatry and Behavioral Sciences Duke University Medical Center, Durham, NC.

6 Department of Psychology, University of Minnesota.

7 The Danish Aging Research Center, University of Southern Denmark.

8 The Danish Twin Registry and Te Danish Aging Research Center University of Southern Denmark, Odense, Denmark.

9 Queensland Brain Institute The University of Queensland Diamantina Institute, Translational Research Institute (TRI) Woolloongabba, QLD 4102 Australia.

Several studies in the new field of cognitive epidemiology have shown that higher intelligence predicts longer lifespan. This positive correlation might arise from socio-economic status influencing both intelligence and health; intelligence leading to better health behaviors, and/or some shared genetic factors influencing both intelligence and health. Distinguishing among these hypotheses is crucial for medicine and public health, but can only be accomplished by studying a genetically-informative sample.

We analyzed data from three genetically-informative samples containing information on intelligence and mortality: 1) 377 pairs of male veterans from the NAS-NRC US World War II Twin Registry, 2) 246 pairs of twins from the Swedish Twin Registry and 3) 784 pairs of twins from the Danish Twin Registry. The age at which intelligence was measured differed between the samples.

We used three methods of genetic analysis to examine the relationship between intelligence and life-span: we calculated the proportion of the more intelligent twins who out-lived their co-twin; we regressed within-twin-pair life-span differences on within-twin pair intelligence differences; and we used the resulting regression coefficients to model the additive genetic covariance.

The combined (and all three individual samples) showed a small positive phenotypic correlation between intelligence and lifespan. In the combined sample observed r=.12 (95% confidence interval .06 to .18). The additive genetic covariance model supported a genetic relationship between intelligence and lifespan. In the combined sample the genetic contribution to the covariance was 95%; in the US study, 84%; in the Swedish study, 86%, and in the Danish study, 85%. The finding of common genetic effects between lifespan and intelligence has important implications for public health and those interested in the genetics of intelligence, lifespan or inequalities in health outcomes including lifespan.