Does shape constancy relate to drawing ability? Two failures to replicate

People vary in their ability to make accurate representational drawings. Cohen and Jones (2008) have suggested that individuals who draw poorly have problems in the perception of objects, so that the extent of shape constancy (phenomenal regression) correlates with drawing ability, there being a "robust negative relation between perceptual errors resulting from shape constancy and drawing accuracy". In this study we describe two attempts to
replicate that finding, in Study 1 in 30 non-art students, and in Study 2 in 106 art students. Study 1 found a correlation that was statistically significant, but in the opposite direction to that reported by Cohen and Jones, and in Study 2 the correlation was very close to zero. Combining these results meta-analytically with those of Cohen and Jones finds a nonsignificant correlation. Taken overall, these findings throw doubt on the Cohen and Jones'
hypothesis that those with less phenomenal regression have better representational drawing ability.