The Cohen and Bennett (1997) is a very interesting article that sheds some light about visual accuracy and some of the skills being used when we try to accurately draw an object that we perceive. To operationally define this type of task can be difficult, but Cohen and Bennett decided to define it as an accurate drawing as one that can be recognized as a particular object at a particular time in a particular space, rendered with little addition of visual detail that cannot be seen in the object represented or with little deletion of visual stimuli. Additionally, an artist must perceive the object accurately, decide which areas to represent and how to represent them, they must have the motor coordination to translate this onto paper, and finally they must be able to objectively assess accuracy or inaccuracies. Common error lies in misperception of the object, incorrect representational decisions, lack of motor coordination, and misperception of one's drawing.
In their first and perhaps most interesting experiment, they designed a within-subject design with three separate drawing conditions. There was a tracing condition, a distance condition, and a traditional condition. The subjects traced in the tracing condition. In the distance condition, the subjects could see the original picture from a distance through a glass table and drew their picture on a transparent sheet. In the traditional condition, subjects saw the pictures taped to the wall from a distance. The two photos that were traced were a picture of a woman's face and a picture of a generator. The results showed the tracing condition as more accurate than the distance condition, and the distance condition as more accurate than the traditional condition. More interestingly, the traditional drawings of the face were much less accurate the drawings of the generator. Since the two tracing conditions eliminated representational decisions and motor coordination, the drawing inaccuracies of the faces can be attributed to the inability of the artists to perceive them. The generator drawings contained right angles that were more easily represented. This raises an important question. Why can't most people draw what they see? Or better yet, why can't people draw faces?