This paper reports on an inquiry into parameters which influence difficulty of a word problem. The study that we present is a part of a larger research project and focuses on a particular case where groups of fourth graders have been given four versions of one word problem which differed in two parameters - the representation of a numerical value and a number of message levels.
Although no substantial difference in success rate has been observed between the versions, we give a brief account of differences in mistakes pupils made and contrast them with mistakes of third graders who have been asked to solve one version of the problem as well.