In what must now seem ancient times, namely the late 1960s, it was still usual to begin psychometric testing by asking the child to do the Goodenough Draw a Man Test. What was good about it was the instruction: “Draw me a picture of a boy. Do the best that you can. Make sure that you draw all of him.” (Girls were asked to draw girls). Kids understood, and took the proffered pencil and paper and got started. Mother and child were relieved to find that there were to be no injections (working in a hospital setting, we had to make that point several times). I would then give a few explanations to the mother, and she would leave in relief, without any distress on the part of the vast majority of the children. It was a settling down task, but had a good, understandable marking system which listed the adequacy of the depiction in conceptual terms, and a total score which tracked child development and provided an intelligence estimate. It was a productive use of time.
Now the South London King’s College gang have found that this task is a good predictor of later intelligence and has a genetic component. Yes, the old tests are the best, and there is nothing wrong, and much right, with the twin control method.
Rosalind Arden (an intelligence researcher, not a Shakespearean character) and colleagues have looked at over 14000 participants in the Twins Early Development Study ,comparing their drawing prowess at 4 years of age with later achievements.
Rosalind Arden, Maciej Trzaskowski, Victoria Garfield and Robert Plomin. Genes Influence Young Children's Human Figure Drawings and Their Association with Intelligence a Decade Later. Psychological Science published online 20 August 2014. DOI: 10.1177/0956797614540686
Drawing is no laughing matter: the cave drawings of Chauvet and Lascaux 40,000 years ago are an early instance of when we humans may have started getting thoughtful, able to depict ourselves in a conceptual world, and consequently becoming a little odd, confused and self-conscious. In a word: Arty. I am convinced that Art Criticism started immediately afterwards, with Art History not far behind. Why such a fuss about putting pictures in chronological order? Anyway, humans became self-conscious.
In Florence Goodenough’s test, a drawing receives 1 point for the presence and correct quantity of each of the following bodily features: head, eyes, nose, mouth, ears, hair, body, arms, legs, hands, and feet. The inter-rater reliability for this test is .93 (Naglieri & Maxwell, 1981). The internal consistency of the test (Cronbach’s α), calculated empirically from our sample, was .79. Published test-retest (1-month) stability coefficients for this test, combined with another test of perceptual performance, were .78 in 3-year-olds and .84 in 5-year-olds (McCarthy, 1972, p. 34).
At ages 4 and 14, the children were administered verbal and nonverbal tests of cognitive ability (Oliver & Plomin, 2007; Spinath, Ronald, Harlaar, Price, &Plomin, 2003). At age 4, these tests comprised a sentence-construction test derived from the MacArthur Communicative Development Inventory adapted for the United Kingdom (Fenson, Pethick, & Cox, 1994) and nonverbal tests: an odd-one-out test, a design drawing test, a puzzle test, and 12 items testing conceptual knowledge (taken from the hour-long Parent Report of Children’s Abilities; Price, Eley, Dale, Stevenson, & Saudino, 2000). At age 14, the children were administered (over the Web) 30 items from Raven’s Standard Progressive Matrices (Raven, Court, & Raven, 1996) and a 27-item vocabulary test from the Wechsler Intelligence Scale for Children (Kaplan, Fein, Kramer, Delis, & Morris, 1999).
Surprisingly, the drawing scores correlated almost as much with g at age 14 as did g at age 4, which was measured from a larger test battery.
Table 1. Zero-Order Phenotypic Correlations Among the Key Measures (N = 7,752 pairs)
Variable Drawing at 4 g at age 4
g at age 4 .33 [.32, .35]
g at age 14 .20 [.17, .22] .24 [.21, .27]
Table 2. Heritability of the Key Measures (Estimated From Univariate Models)
Drawing at age 4 (n = 14,874) .29 [.22, .35]
g at age 4 (n = 14,461) .29 [.27, .32]
g at age 14 (n = 4,695) .50 [.38, .61]
Note: The values in square brackets are 95% confidence intervals.
We found that drawings done by MZ twins were significantly more similar than were drawings done by DZ twins. Finding that a behavior is heritable is no longer news; yet if the data had shown that any siblings’ drawing scores were alike, irrespective of zygosity, we would not have been surprised because it seems so plausible that young same-age siblings would emulate each other’s drawings or be guided by parents (irrespective of zygosity). For that reason, we were intrigued to find that scores for a single drawing were as heritable as was g estimated from several different indicators (verbal and nonverbal tests).
In this large sample, a single picture of a 4-year-old child, drawn in around 5 min, had a significant positive phenotypic association with g measured a decade later, and this correlation was as high as the correlation between g at age 4 and g at age 14. This phenotypic association was caused partly by a genetic correlation between drawing at age 4 and g at age 14.
Our data show that the capacity to realize on paper the salient features of a person, in a schema, is an intelligent behavior at age 4. Performance of this drawing task relies on various cognitive, motoric, perceptual, attentional, and motivational capacities. Our estimated positive phenotypic correlation between drawing and contemporaneous intelligence is consistent with estimates from 40 small studies in which the correlations (rs) ranged from .24 to .83 (Scott, 1981; see Willcock, Imuta, & Hayne, 2011). The correlation we observed is also consistent with a large phenotypic study of 7-year-olds that found, perhaps surprisingly, that figure-drawing scores correlated with arithmetic performance (r = .33, n = 14,522) to about the same extent as they correlated with pattern copying (r = .37, n = 14,545; Shepherd, 2012, p. 21).
There is some evidence in the archaeological record that figurative art is more recent than geometric patterning (Pike et al., 2012). If this is correct, then figurative art may track, to some extent, increasing cognitive ability in the human species. Drawing is an ancient human capacity; 32,000 years before the children in our study sat down to draw, unknown people made surviving drawings of great skill and beauty. These images (see Bradshaw Foundation, 2011, for photographs) are among the oldest examples of a human behavior that continues in the same form today. This long history endows the drawing test with ecological validity and relevance to an extent that is unusual in psychometrics.
Drawing marks called finger flutings, made by dragging fingers across wet clay or on soft cave walls, are the oldest known direct evidence of children’s behavior, aside from footprints. Archaeologists have dated these marks to the Upper Paleolithic and ascribed them to young children on the basis of detailed measurements of the groove widths (Sharpe & Van Gelder, 2006). The longevity of children’s drawing behavior indicates that drawing is a natural part of the human species-typical repertoire. Given that drawing enhances the fine-motor skills that children use in writing (Saida & Miyashita, 1979), it may have contributed to the development of pictograms, and eventually writing.
For all we know, Stone Age children were asked to draw on the walls as a settling down procedure before being given a comprehensive test battery of hunting, tracking, spear throwing and long distance communication skills.
In summary, a simple bit of drawing at age 4, scored in a simple manner, provides a small but significant prediction of intelligence at age 14. Truly, intelligence is like carbon in biology: you find it everywhere.