How did they investigate these questions?
They conducted two experiments. In the first experiment, participants were asked to study one of two text passages (biological anthropology versus geology) and then instructed participants to study in one of four conditions:
1. Provided-Flashcards: Participants were given premade flashcards from the textbooks to study with. These flashcards were heavily detail-level ones.
2. Self-Generated-Flashcards: Participants were asked to create their own flashcards.
3. Conceptual Self-Generated-Flashcards: Participants were instructed to create their own flashcards, but were provided instructions to use a conceptual approach.
4. Free-Study: Participants were instructed to study the text passage as they wished.
In the second experiment, participants were provided with either detail-level or conceptual-level flashcards to study one of the two text passages. In both experiments, participants’ structure building ability was measured. Performance was measured on a multiple-choice and a short-answer test covering both detail and conceptual questions.
What did they find?
Findings from the first experiment:
a) Students were more likely to generate detail-level flashcards if not instructed to do otherwise. In addition, the provided flashcards from textbooks were also heavily biased towards detail-level content. So, without any further instruction and if left to their own devices student will be mainly exposed to or generating flashcards that contain definitions or simple facts. Prompting students to generate conceptual-level flashcards was successful to engage them to create those flashcards, but the authors also mention that they only asked students to generate a few conceptual-level flashcards – making the difference between the different study groups less extreme.
b) No difference was found between the four study groups in relation to their test performance on either test. The authors explain that this could be due to the abovementioned little difference between the three flashcard conditions which all were more biased towards detail-level content. In addition, they acknowledge that even participants in the free study condition may have been using strategies similar to the other study groups – making that group more similar to the others.
c) In this experiment, no benefit was found of students generating their own flashcards. However, a significant difference was found in regard to study time: Participants in both self-generated-flashcards conditions studied for longer (approx. 24 minutes) than participants in the provided-flashcards and free study conditions (approx. 15 minutes). Thus, generating flashcards took them longer to do without an effect on performance. However, it is important to keep the just mentioned limitations in mind before jumping to conclusions. Furthermore, despite the fact that no differences were found between the four study groups, the authors found an overall positive relationship between the number of conceptual-level flashcards generated and test performance: More conceptual-level flashcards was associated with better performance on both tests.
In the second experiment all participants were provided with flashcards (either conceptual-level or detail-level ones). While there was no difference between the two conditions on the multiple-choice test, they found an advantage in short-answer performance for participants in the conceptual-level flashcard condition. Interestingly, the benefit of conceptual-level flashcards was moderated by structure building ability: Particularly participants with low structure building ability benefitted from conceptual-level flashcards whereas for participants with high structure building abilities it did not matter whether conceptual versus detail-level flashcards were used.
