Effect of Test Items and Degree of Similarity upon Interference in Learning from Prose Materials

Abstract

An experiment was performed to assess the effect of retroactive interference upon learning from prose materials. A test-no test variable was incorporated into a 2 × 2 design along with two levels of organizational similarity. Immediate and delayed tests were used to measure the retention of information stated directly in the passage and that which was based upon inference. Including the test after the original learning passage significantly affected retention. Within the no-test condition, subjects who read an interpolated passage which was highly similar to the original passage had significantly fewer correct responses when tested immediately on information that was stated directly in the original passage.