Abstract:

In the present study, we examined retrieval practice using text material, asking whether the effect might differ if the texts are narrative versus expository in nature. Examining potential differences between these two types of text material may be important, not only theoretically, but also from an educational standpoint, since both narrative (e.g., literary fiction) and expository (e.g., scientific material) texts are frequently used in the classroom.

Method and Participants:
Participants were undergraduate students at a mid-sized university in the southern United States. A total of 120 students participated in the experiment for course credit and were randomly assigned to one of four conditions.

Materials and Design:
The researcher selected two text passages from Weaver and Bryant (1995). Based on Weaver and Bryant’s assertion that expository passages are more difficult to comprehend, we chose an expository text (Euripides) from the easy difficulty category of the Flesch (1951) readability scale, whereas we chose a narrative text passage (The Owl) from the standard difficult category. Half of the participants were assigned to read The Owl and half were assigned to read Euripides. Similarly, half of each text type was assigned to either the repeated study or testing group. Intermediate and final tests were free recall.

Procedure:
All participants were given 10 minutes for the initial reading of their assigned passages. After the first reading, participants in the repeated study group received three additional 5-minute study periods. Between study periods, participants worked on mazes for 2 minutes. Participants in the test condition also worked mazes for 2 minutes and then were given 5 minutes to recall as much as possible from the paragraph they had just studied. Participants were scheduled for a follow-up session three days later and dismissed. In the follow-up session all participants were given 10 minutes to recall as much of the previously studied passage as they could.

Results and discussion:
The final free-recall test was scored in terms of idea units (e.g., Wolfe, 2005;). “The Owl” contained 51 idea units and “Euripides” contained 44 idea units. Participants received 1 point for each idea unit correctly recalled. Scoring was based on gist recall, which was defined as the recall of either the exact wording or close synonyms of the idea units (Wolfe & Mienko, 2007). The tests were scored by two raters with inter-rater reliability of r =.96. A 2 (condition: study vs. test) X 2 (text: narrative vs. expository) ANOVA revealed a main effect of text type, F(1,116) = 10.80, p = .002, ήp2 = .184, indicating that overall, recall of the narrative text was greater than that of the expository text.
There was no effect of condition indicating no overall effect of testing across the two text types. However, there was a text type by condition interaction F(1, 116) = 6.33, p = .015, ήp2 = .116), revealing that for the expository text, testing produced greater recall than restudying whereas for the narrative text, restudying lead to borderline significantly greater recall than testing.
These results indicate that although narrative passages tend to be easier to remember than expository tests, the two texts differ markedly in the effects of retrieval practice on later recall. Although retrieval practice resulted in a testing effect for the expository passage, it had a detrimental effect on later recall of the narrative passage.

Keywords:

Retrieval practice, narrative text, expository text, memory, learning.