Five component problem-solving skills (general strategy, planning, logical thinking, algebraic variables, and debugging) were identified as common elements of both computer programming and mathematical problem-solving. Based on the similarities of these general skills in specific contexts, a theory was generated that the skills would transfer and that experience in computer programming would cause an improvement in mathematical problem-solving achievement.

A path model was constructed to illustrate this hypothesized causal relationship between computer programming and mathematical problem-solving achievement. In order to control for other relevant variables, the model also included mathematics experience, access to a home computer, ability, socioeconomic status, and gender. The model was tested with a sample of 800 high school students in seven southwest Virginia high schools.

Results indicated that ability had the largest causal effect on mathematical problem-solving achievement. Three variables had a moderate effect: computer programming experience, mathematics experience, and gender. The other two variables in the model (access to a home computer and socioeconomic status) were only very slightly related to mathematical problem-solving achievement.

The conclusion of the study was that there was evidence to support the theory of transfer of skills from computer programming experience to mathematical problem-solving. Once ability and gender were controlled, computer programming experience and mathematics experience both had causal effects on mathematical problem-solving achievement. This suggests that to maximize mathematical problem-solving scores, a curriculum should include both mathematics and computer programming experiences.