Are Blonds/Blondes Dumb? A paper I had hanging around I never did try to publish.

EYE COLOUR AND HEMISPHERICITY                                       Dr Anthony Fallone


The present study derives substantially from part of a Ph.D. thesis of the author submitted to the University of Edinburgh. The author is grateful to Professor Colwyn Trevarthen for supervising the project; thanks are due, too, to the students whose efforts have made this particular study possible.

Summary
Gordon's Cognitive Laterality Battery (1986) was used to test the hypothesis that light-eyed individuals are less cognitively able than the dark-eyed, following the theories of Gary and Glover (1976) and more recent reaction time studies (Fallone, 1993). The hypothesis was supported: in this sample light-eyed University students were significantly ( p<. 02) less able than dark-eyed students on a battery of cognitive tests. If eye colour is a hitherto largely unconsidered but significant variable in cognitive testing this casts some doubt on previous research which has disregarded such a potentially influential confounding characteristic. Introduction Worthy (1974) claimed that light-eyed people are more sensitive to form, while the dark-eyed respond more to colour; Gary and Glover (1975) tested this, finding that in a series of tasks, the dark-eyed made more form errors than colour errors, the light-eyed performing conversely. In tasks which measured ability in paper and pencil work, play, and dexterity, dark-eyed individuals scored more highly than light-eyed individuals, although no difference was found in the areas of number work and gross motor skills. Light-eyed, light-haired, and fair-skinned individuals seem more susceptible to hypoglycaemia (Gary and Glover, 1976). Happy and Collins (1972) are cited by Gary and Glover (1976) as having theorised that there is a link between autism and light eye and hair colour in Caucasians: lack of melanin may leave nerve cells in the ascending reticular activating system less protected, sometimes causing a defect in noradrenergic pathways which could bring about autistic symptoms. They found that there was an over-representation of relatively low pigmented children with autism in their sample, a significant        ( p<.05) difference by comparison with dark children with autism, something noted earlier by Kastein (1966). Gary and Glover conducted a study into the possibility of being able to predict learning disability from eye colour; their experiment had an N of 5,552 males and 4,012 females identified as having learning disabilities. Their conclusions were that the dark-eyed seemed less likely to be learning disabled, regardless of their sex, and that light-eyed individuals were more likely to be learning disabled, especially if they were male. The same study found that the light-eyed are much more likely to suffer diagnosed medical conditions and have various physical anomalies, the ratio being 30 to1. Those individuals low in pigmentation are therefore 30 times more likely than the highly pigmented to suffer heart, endocrine, and central nervous system disorders, and in addition are more likely to be unusually obese, thin, short, tall, or have premature or delayed secondary sexual characteristics. The sample was Caucasian because including non-Caucasians would grossly bias the findings, essentially nullifying the research. As eye colour is such an easily seen physical trait it has an obvious attraction for researchers; pulse rate, pulse pressure, respiration rate, galvanic skin response and body temperature are some of the physiological variants found with eye colour (Kent, 1956a; Markle, 1976), together with magnitude of pupillary dilation (Gambill, Ogle, Kearns, 1967), oculocardiac reflex (Fry, Hall-Parker, 1978) and resistance to the damaging effects of industrial noise (McFadden, Wightman, 1983). Usually, it is the light-eyed individual who is found to suffer by comparison with the dark-eyed, e.g., blue-eyed people have lower tactile thresholds on the cornea, and report more pain from wearing contact lenses, than do people with brown eyes (Millodot, 1975; Tota, La Marca, 1982). Certain pathologies are more associated with light-eyed people: malignant melanoma are more likely to be developed by light-eyed, light-haired individuals (Gellin, Kopf, Garfinkel, 1969), while patients with light coloured irises are more likely to suffer from Idiopathic Dystonic Syndromes, involuntary movements of the sort found in Huntington's and Parkinson's disease (Korein, 1981). Neural Crest Syndrome is characterised by autonomic dysfunction manifested in pupil abnormalities and pain perception and by the absence of sweating; all reported cases are blond, blue-eyed and fair-skinned (Brown, Podosin, 1966). Sufferers from phenylketonuria are frequently light-eyed (Berg, Stern, 1958). Prader-Willi syndrome has been recognised as marked by light coloured hair and eyes (Creel, Bendel, Wiesner, Wirtschafter, Arthur, King, 1986). Bassett and Dabbs (2001) looked at consumption of alcohol in light- and dark-eyed individuals (10,860 males and 1,862 females, all Caucasian), finding that light eyes consumed significantly more alcohol than dark eyes. It appears that light eye colour, among all the other disorders listed above, is associated with learning disability, dyslexia and autism, where left hemisphere abilities are poor; if that is so, those individuals who do not necessarily suffer the most extreme effects of hypo-pigmentation may, nevertheless, demonstrate reduced left hemisphere cognitive functions, relative to the dark-eyed and dark-haired.
Hypotheses
(1) University students with light eyes will score significantly less well over all cognitive tests than students with dark eyes;
(2) University students with light eyes will manifest a significant right-hemisphere cognitive profile (CLQ).
Method 
Gordon's Cognitive Laterality Battery
Harold Gordon (1986) devised an eight test Cognitive Laterality Battery (CLB) providing a cognitive profile, normed against a large sample. 75% of the adult norms were from University students, with more postgraduates than undergraduates. 25% were non-student subjects, increasing the variability of the sample. Estimates based on comparing the population means to maximum scores made Serial Sounds and Orientation-3d as the most difficult of the sub-tests, with the verbal fluency and Form Completion tests next. Nevertheless, the sub-tests of the CLB were chosen from long-used neuropsychological or psychometric instruments. These were known to be adversely affected by brain damage in one or the other cerebral hemispheres. For example, performance on the Form Completion sub-test was well known to be poor after right hemisphere insults. The assumptions on which the CLB was based are that activity in the neurosystem related to Appositionality is more distributed to or active in the right hemisphere, with the Propositional neurosystem more distributed to or active in the left, with precise anatomical locations remaining unfixed.
The Sub-tests 
a) Serial Sounds. Subjects listen to sounds played from an audio tape such as a dog barking, horse neighing, bugle blowing, and doorbell ringing in sequences which gradually increase in number of sounds; the task is to remember their order and write it down correctly.
b) Serial Numbers. Numbers are spoken on audiotape in sequences that lengthen from three to a maximum of nine and then decrease to 6; the task is to write down the correct sequences of numbers; it resembles the Wechsler digit span test.
c) Word Production-Letters. Subjects are asked to write down as many words as possible in one minute. Scoring is a simple total of all words that are not proper names or different tenses from three trials starting with three different letters.
d) Word Production-Categories. A category is stated (e.g., flowers) and once more it is the number of words in that category produced within one minute that is scored. Two categories are given, the score being the total for both categories.
e) Localisation. 24 Small crosses are displayed in a rectangular frame on a projection screen for three seconds; subjects are required to mark equivalent answer blank rectangles where they thought they saw the crosses. Scoring is the total millimetre error deviation from the correct positions.
f) Orientation-3d. 24 three-dimensional S-shaped constructions of 10 stacked cubes are presented on a screen in sets of three; the task is to mentally reorientate the shapes in order to find out which two are the same. The third is a mirror image which can never be the same as the other two. (Shepherd and Metzler, 1971)
g) Form Completion (Gestalt closure). Partially erased white on blue silhouette drawings of common objects are projected on a screen; the subjects must use their imaginations to decide what is represented, writing this down in a word or two (Thurstone and Jeffrey, 1966).
h) Touching Blocks. Drawings of stacks of 7-10 stacked rectangular blocks are projected on a screen for 45 seconds each; for each stimulus slide five of the rectangular blocks are numbered; the blocks are stacked in such a way that two to eight blocks may be touching any numbered block. The task is to count up the touching blocks above, below, and on each side of the numbered blocks (MacQuarrie, 1953).
A Propositional score is derived from the four Verbal-sequential tests, and an Appositional one from the Visual-spatial sub-tests; the Propositional score is then taken away from the Appositional, leaving a "Right-hemisphere profile", the extent to which the individual appears to be more able visual-spatially. The profile so derived is called a "CLQ", or Cognitive Laterality Quotient. Finally, the scores overall are added together and divided by the number of sub-tests in order to give a measure of cognitive ability, the "CPQ", or Cognitive Performance Quotient. This is similar to a score in intelligence, although the CLB does not measure everything measured in IQ tests. Many of Gordon's sub-tests are derived from IQ tests; however, he is very clear that he intends them to be used in a neuropsychological context, deriving their reason for use from clinical studies of brain lesions and pathologies.
Nonetheless, some of Cattell's primary factors are tested by Gordon's CLB (all descriptions of the Cattell primary factors taken from Kline, 1991):
"W"=Word fluency, the rapid production of words, without semantic connection but conforming to an initial letter requirement (found by Cattell in 1933, and used ever since in IQ tests); "V"=understanding words or ideas; it is likely that this plays a great part in successful performance on the "Word Production-Categories" sub-test, with some aid from "W", and from Fl, or "ideational fluency". "V" is supposed to be the most reliable indicator of "crystallised intelligence".
"Ms"=span memory, the short-term recall of digits or letters, as used in the Wechsler IQ tests. It is uncertain if the recall of series of sounds taps the same memory.
"S"=spatial factor; the ability to visualise two- or three-dimensional objects when their orientation is changed. The 3d Spatial Rotation-Orientation sub-test clearly fits this factor.
"Cs"=speed of closure factor; this apparently taps the ability to quickly complete a gestalt when parts of the stimulus are missing. Interestingly, speed of visual closure correlates 0.610 with word fluency.
It may then be possible to predict that Form Completion, the Gestalt sub-test, will load quite strongly on any verbal-sequential Factor in a Factor Analysis of these research results. Kline (1991) states that verbal ability is highly loaded on intelligence tests, especially those used for selection to Universities and schools. He says that people in the arts are usually high on this factor, although it should not be confused with intelligence as such, because many scientists are relatively weak in verbal ability. Second order ability factors have been derived from the intercorrelations of the first order factors: "gf"=fluid intelligence, which includes memory span, flexibility of closure and intellectual speed, tested in the CLB by the two sequential memory sub-tests and the Gestalt Form Completion sub-test, together with the overall fast timing of the whole battery.
"gc"=crystallised intelligence, tested by the verbal fluency tests.
"Pv"=Visualisation, tested by the Orientation, Touching Blocks, and Localisation sub-tests. "gr"=Retrieval capacity or general fluency, ideational and associational fluency which may well play a strong part in scoring well on the Word Production-Categories sub-test.
"gf"=Cognitive speed factor, something which, again, the overall fast timing of the CLB tests.
These five factors, says Kline, embrace most of the variance in human ability. Clearly, the family resemblance, although not identical, is close between an IQ test and the CLB. A transformation procedure is possible to derive IQ equivalents from the CLB Z-scores, with 0 equalling 100 on the IQ scores, and +3 (3 Standard Deviations for an IQ test) equalling approximately an IQ of 148, -3 an IQ of approximately 60 or 70, although this can only be a very approximate measure. The literature certainly seems to suggest that at least the CPQ could be predictive of IQ. Eye Colour Eye colour in the form of intensity of iris pigmentation was noted in the student sample, with the demarcating midway point between the classification of 'Light' and 'Dark' that of hazel eyes, the lighter and darker varieties of these falling on either side of the category line; green was put into the darker category and light grey into the lighter category. Actual colour tended to be ignored, with very dark blue eyes classified in the "Dark' category, while pale brown eyes fell into the 'Light' category (It had been decided that a fault with some previous work was that the eye colour classification used bore little relationship to this intensity of pigmentation).
The Experiment 
Participants 
Mixed-sex groups of Edinburgh University students (N=109) aged between 17 to 50 years were tested with the CLB. The CLB takes between 70 and 80 minutes to administer. The order of testing is: 1) Serial Sounds; 2) Localisation; 3) Serial Numbers; 4) Orientation; 5) Word Production-Letters; 6) Word Production-Categories; 7) Form Completion; 8) Touching Blocks.
Results 
Eye Colour 

There was a general superiority in cognitive scoring for the total sample by those in the dark eyed category over those in the light eyed.

An analysis of variance was carried out; there were significant differences overall (F1,108=6.327 p<. 02). To analyse these in greater detail, t-tests were conducted. Significant differences between eye colours were found in the Serial Numbers (df=109; t-value=2.639; p<.005) and Word Production-Categories (df=109; t-value=3.299; p<.001) sub-tests. The Propositional (df=109; t-value=2.760; p<.005) and Appositional sub-totals (df=109; t-value=1.741; p< .05) were both significant, all showing the dark-eyes superiority. Of the supposed right hemisphere sub-tests only Touching Blocks was significant, like the others showing the light-eyed to be less able (df=109; t-value=2.116; p<.02). The major effect was in the Propositional area, which indicated that light-eyed individuals were generally less able on verbal-sequential tasks than dark-eyed individuals. There were no sub-tests where the light-eyed outscored the dark-eyed. Eye Colour-High/Low CPQ Groups Scores The middle order scores on the overall performance measure, the Cognitive Performance Quotient (CPQ), were removed, leaving a ‘top 30’ and a ‘bottom 30’. A Chi-Square test showed that there was an association between eye colour and position in the High/Low 30 of CPQ scores (N=56, df=1, Phi=.372, Chi=6.319, p<.02). 75% of the dark-eyed were to be found in the High 30, while only 37.5% of the light-eyed scored that well. The light-eyed made up 76.9% of the Low 30. There seems to have been a significant clustering of light eyes in the Low and of the dark eyes in the High 30. A One-factor analysis of variance was carried out on the z-score data for these two groups, showing a significant difference between eye colours (F1,54=10.145, p<.005); by comparison with the overall analysis of variance it is clear that dropping out the central corpus of middle range CPQ scores intensified the eye colour difference.

The Propositional and Appositional sub-scores remained significantly different ( p<.05), while most interestingly the Cognitive Laterality Quotient, the measure of 'right hemisphericity', became significantly different between eye colours (32 Light=.353, 24 Dark= -.063, df=54, t-value=1.884,  p<.05), apparently showing that the Low 30 light-eyed students were significantly more right hemisphere dominant than the High 30 dark-eyed students. Discussion 76.9% of the Low 30 for CPQ were light-eyed (as compared with 59.504% of the whole sample), a result which may reflect the speeded element in the CLB, if Gary and Glover's (1975) findings are considered. Both the Propositional ( p<.02) and Appositional ( p<.05) sub-tests registered a significant difference; perhaps the speed difference between eye colours would always be a handicap for the light-eyed for a speeded paper and pencil test (light-eyed favouring a slower, more planned and thoughtful approach to activities, according to Worthy). Within only the top and bottom 30 scores similar significant overall differences appeared between the two eye colour groups, with the dark-eyed significantly more left hemispheric in their dominance. What was of greatest interest was that the light-eyed were significantly more right hemispheric at the extremes of the range of CPQ scores than the dark-eyed. It must be emphasised that the cognitive ability differences found here between light- and dark-eyed individuals are relative only: both groups scored well in relation to Gordon's norms, but the dark-eyed usually scored more highly than the light-eyed; in a sample from the normal population much greater differences would be found as a simple function of an increase in the sample variance, with many more scoring well below the norms. At both levels of analysis the hypothesis was supported; it seems that in this sample light-eyed University students were significantly less able than dark-eyed students on a battery of cognitive tests. If eye colour is a hitherto largely unconsidered but significant variable in cognitive testing (Fallone and Baluch, 1993) this casts some doubt on previous cognitive research which has disregarded such a potentially influential confounding characteristic.

References 
Bassett, J. F. and Dabbs, J. M. Jr. (2001). Eye color predicts alcohol use in two archival samples, Personality and Individual Differences, Vol 31 (4), pp. 535-539
Berg, J. M., Stern, J. (1958) Iris color in Phenylketonuria. Annals of Human Genetics, 22, 370-372. Brown, J. W., Podosin, R. (1966) A syndrome of the neural crest. Archives of Neurology, 15, 294-301. Creel, D. J., Bendel, C. M., Wiesner, G. L., Wirtschafter, J. D., Arthur, D. C., King, R. A. (1986) Abnormalities of the central visual pathways in Prader-Willi Syndrome associated with hypopigmentation. The New England Journal of Medicine, Vol. 314, No. 25, 1606-1609.
Fallone, A. R., Baluch, B. (1993) Eye colour: an unconsidered variable in cognitive research? Perceptual and Motor Skills , 77, 11123-1127
Fry, E. N. S., Hall-Parker, J. B. (1978) Eye hue and the oculocardiac reflex, British Journal of Opthalmology, 62, 116-117.
Galaburda, A. (16th September, 1989) Paper presented at the Rodin Dyslexia Conference, University of North Wales, Bangor, Wales.
Gambill, H. D., Ogle, K. N., Kearns, T. P. (1967) Mydriatic effect of 4 drugs determined with pupillograph. Archives of Opthalmology, 77, 740-746.
Gary, A.L., Glover, J. (1975) Melanin as a predictor in the acquisition of developmental skills. The Journal of Psychology, 90, 185-190.
Gellin, G. A., Kopf, A. W., Garfinkel, L. (1969) Malignant melanoma: a controlled study of possible associated factors. Archives of Dermatology, Vol. 99, 43-48.
Gordon, H. W. (1986) The Cognitive Laterality Battery: tests of specialised cognitive function. International Journal of Neuroscience, 29, 223-244.
Greiner, A. C., Nicolson, G. A. (1965) Schizophrenia-melanosis. Lancet, ii, 1165.
Happy, R. Collins, J. K. (1972) Melanin in the ascending reticular activating system and its possible relationship to autism. Medical Journal of Australia, 2, 1484-1486.
Kastein, S., Trace, B. (1966) The birth of language: the case history of a non-verbal child, Springfield, Il.: Charles C. Thomas.
Kent, I. (1956a) Human iris pigmentation: 1. a concept of individual reactivity with implications in health and disease. Canadian Psychiatry Association Journal,1, 99-104.
Kline, P., Intelligence: The Psychometric View, London: Routledge, 1991.
Korein, J. (1981) Iris pigmentation in Idiopathic Dystonic Syndromes including Torticollis. Annals of Neurology, Vol. 10, No. 1, 53-55.
MacQuarrie, T. W. (1953) MacQuarrie Test of Mechanical Ability. Monterey, CA.: California Test Bureau.
Markle, A. (1976) Eye colour and responsiveness to arousing stimuli. Perceptual and Motor Skills, 43, 127-133.
McFadden, D., Wightman, F. (1983) Audition: some relations between normal and pathological hearing, Annual Review of Psychology, 34, 95-128.
Millodot, M. (1975) Do blue-eyed people have more sensitive corneas than brown-eyed people? Nature, 255, 151-152.
Shepherd, R. N., and Metzler, J. (1971) Mental rotation of three-dimensional objects. Science, 171, 701-703.
Thurstone, L. L., and Jeffrey, T. E. Closure Speed. Chicago: Industrial Relations Center.
Tota, G., La Marca, F. (1982) Correlations between corneal sensitivity and iris colour. Atti della Fondazione Giorgio Ronchi, Vol.37 {1}, 59-69.
Worthy, M. (1974) Eye color, sex and race. Anderson, S. C.: Droke House/Hallux