ABSTRACT
The topic of this paper is whether it is worth applying the propositional approach to text processing and readability to the improvement of longer expository texts.
Two versions of a biology text were constructed: a coherent one would, according to the propositional approach, lead to better learning outcomes, than a non-coherent one. To check for instructional alternatives, a reward for good learning results was either announced or not In a field experiment, 167 students at school read one text under one condition and had to give a written reproduction immediately after reading and again after two weeks.
Reproductions were merely weakly influenced by the texts’ propositional structure; visible effects disappeared after two weeks. In contrast, reward caused strong effects, which even became stronger in time.
Finally, some reasons are presented, which account for the fact that previous results within the propositional approach might not be valid for longer expository texts.
