Effects of Sentence Length, Content, and Transformational Form in a Memory Overflow Task

Savin and Perchonock (1965) used a memory-overflow paradigm to investigate sentence memory. For a series of trials, they required subjects to remember a sentence and as many words as possible from a subsequent eight-word string. They found that word recall decreased as the sentences’ transformational complexity increased. Subsequently, Matthews (1968) revealed that the original study had confounded transformational complexity with sentence length, and that the latter was found to produce the memory effects. In the present study, the variables transformational form, sentence length, and number of content words in the sentence (also previously confounded) were manipulated, and the effects of these factors were separated with regression analyses. It was found that the best model, using criteria of prediction and simplicity, was based on content words alone.