Undergraduates meet with difficulties when they have to solve comparison problems in which key words in the problem prime inappropriate operations (e.g., the word "less" priming subtraction when addition is required to solve the problem). This is a comprehension error. In this paper, I examine one factor suspected of making the comprehension of some problems more difficult than others: the computational requirement of the problem. In two experiments, the size of the operand involved in the computation and the number of mathematical operations required to solve a problem were manipulated. Both factors affected the number of comprehension errors leading to the conclusion that increasing the computational requirement of a problem drove some of the solvers to use a superficial mode of understanding the text.