Test design

Historical precedents

All of the components of the CDAT have been used in research prior its development (Clark & Zimmerman, 1997). Clark (1989) mentions Thorndike’s 1913 Scale For the Merit of Drawings By Pupils 8 to Fifteen Years Old. Clark and Wilson (1991) list a number of art tests discussed in gifted educational psychology literature of the 1980s including:

• Meier Art Tests (Meier, N. C., 1929, 1942, 1963)
• Knauber Art Ability Test (Knauber, A. 1932)
• The Horn Aptitude Inventory (Horn, C. C., 1935, 1953)

These and a number of other historical precedents (Clark, Zimmerman, & Zurmuhlen, 1987) guided Clark to develop a test based on a few simple drawing tasks. None of the previous tests, however, were being used at the time for identifying artistically talented students or investigating issues of artistic talent. Judging these tests’ scoring methods and exemplars to be dated, Clark and Wilson (1991) did not consider them valid measures of high ability in art.

Focus on drawing

Clark cited Koppitz’s (1968) evaluation of pencil drawing as the most agreeable task for working with children in psychological study:

There is one that is more meaningful, more interesting, and more enjoyable than all others, and this technique is drawing, just drawing with pencil and paper. I know the value of drawing at first hand, having used it myself both as a child and as an adult to help me through periods of crises and inner turmoil. (in Clark, 1989, p. 98)

Clark’s (1989) justification of drawing as the focus of the art tests rested on the following points, many supported by various other researchers across disciplines:

• Drawing ability was purported to be a normally distributed characteristic. Clark reported that he tested and confirmed this assertion.
• Drawing ability was considered fundamental to art skill and interest and “recognized as basic to expression in all art forms” (p. 99).
• Drawing was regarded as an appropriate task to administer in the context of scientific testing.
• Drawing was characterized as satisfying a child’s need for expression of thoughts in graphic language.
• Drawing ability was seen as related to general intelligence among younger children as well as to specific giftedness in art among adolescents.
• Drawing performance was considered evidence of skill and “knowledge of the problems of art and artists. ” (Wilson, Hurwitz, & Wilson, 1987, in Clark, 1989, p. 99).
• Drawing tasks that may favor representation are well suited to the developmental level of most subjects in upper elementary school and older, who “are at a developmental level where they are consciously and deliberately developing emerging abilities to depict things realistically” (p. 102).

The four drawing tasks

Four carefully selected drawing tasks serve as a control variable in the test:

Work samples require completion of the same task, using the same amount of time, with the same materials and instructions by all students in order to compare student performance. Completion of the same task by all students provides a more legitimate basis for analyzing and comparing children’s art development than an examination of different products rendered in a variety of media. (Clark & Zimmerman, 2001, p. 109)

In order to solicit responses with measurable degrees of difference in drawing ability, Clark (1989) selected tasks from among those previously used in art tests:

Task	Justification	Clark’s (unpublished manuscript) cited sources
1. Draw an interesting house as if you were looking at it from across the street.	Earliest implementation in Thorndike’s 1913 Scale For the Merit of Drawings.	Thorndike (1913), Kline & Carey (1922), Lewerenz (1927, and Lark-Horovitz (1942)
2. Draw a person who is running very fast.	Appeared on the 1981 National Assessment of Educational Progress (NAEP).	Kline & Carey (1922), Lewerenz (1927), Lark-Horovitz (1942), and NAEP (1977)
3. Draw a picture of you and your friends playing in a school yard	“Many researchers have included some type of crowd drawing” (p. 100).	Eisner (1967), NAEP (1977)
4. Make a fantasy drawing from your imagination.	“There are literally dozens of researchers who have used a similar task” (p. 100).	Lewerenz (1927), Lark-Horovitz (1942)

Criteria for scoring

Clark (1989) stated that his aim for the scoring criteria on the CDAT was to achieve greater levels of objectivity than the earlier art tests. He designed the CDAT with 12 scoring criteria in four categories based on common constructs of art learning of the 1980s: “(1) sensory properties (line, shape, texture, value), (2) formal properties (rhythm, balance, unity, composition), (3) expressive properties (mood, originality), and (4) technical properties (technique, correctness of solution)” (p. 100).

These four concepts of art learning may seem familiar to contemporary art educators. The first seven seem to have endured in the field as elements and principles, though they are not specifically named in the National Core Arts Standards compiled in 2014 by the State Education Agency Directors of Arts Education (SEADAE).

When collecting quantitative data to be used for identifying high-ability visual art students, each of these 12 criteria are to be scored by raters using an unlabeled Likert-type scale from 1 to 5. (The one label provided, in fact, is for a score of 5: “5 on each item is used to give credit only for ‘unique, innovative, unusual’ responses” (Clark, 1993, p. 75). A 13th criterion for the appropriateness of a title is applied to the fourth task, the fantasy drawing. Judges provide a total of 49 item scores from 1 to 5, with 245 possible points.

Standardization

Clark and Zimmerman (2001) reported a normal distribution of scores obtained from 946 subjects in grade 3 in rural settings of Indiana, New Mexico, and North Carolina, with a standard deviation of 29.96 and a mean of 73.2.

Interpreting validity

Clark’s (1989) report of a 1984 study to validate the CDAT scores focused on their correlation with ratings scales by teachers of artistically talented students in a specialized art program. Examples of behaviors rated highly by teachers include creating an adequate self portrait and demonstrating effort, commitment, and positive attitude in the studio. Clark reported that he found a predictive correlation between high scores on the CDAT and rating scales of performance in diverse art classes, including photography, ceramics, sculpture, printmaking, computer graphics, and even theater and dance. The reported average correlation of r=.36 (n=60) between test item scores and teacher ratings can be interpreted as 13% of the high test scores being accounted for among students highly rated by their teachers.

A very high level of correlation reported between the scores on the four drawing tasks (average r=.70) suggests that similar characteristics are measured in the four tasks. Clark (1993) described the tasks, however, as demanding “very different abilities and skills” (p. 74) as detailed in the chart below. He saw the correlation as support for the effectiveness of the CDAT for research. The correlation of scores across tasks may be interpreted as evidence of the CDAT triangulating an underlying drawing ability.

Task	Clark’s (1993) description of skills called upon
1. Draw an interesting house as if you were looking at it from across the street.	“Drawing a house requires skills such as depicting perspective, textures, differential and meaningful shapes and sizes, and recognizable d etails” (p. 74).
2. Draw a person who is running very fast.	“Drawing a person running requires portrayal of a human figure in action as well as body proportions and recognizable detail ” (p. 74).
3. Draw a picture of you and your friends playing in a school yard.	“Drawing a group of persons on a playground requires portraying and composing receding space and groupings of figures in that space” (p. 74).
4. Make a fantasy drawing from your imagination.	The fantasy drawing task allows subjects to use their imaginations to portray whatever they wish, including things they know how to draw well” *pp. 74-75).