H567/02 Section A - Grant et al. (1998): context-dependent memory

Context as a retrieval cue

Environmental details present during learning may become associated with the learned information. Reinstating those details later can provide retrieval cues, helping the learner access the information.

Earlier research often used unrelated word lists and found clearer context effects for recall than recognition. Grant et al. asked whether the effect extended to meaningful material and ordinary academic tests. Short-answer questions require the learner to produce an answer; multiple-choice questions offer possible answers for recognition. Recognition alternatives might themselves provide strong cues, reducing dependence on environmental context.

The practical concern was that students often study with background sound but sit tests in quiet rooms. If matching matters, a student can adjust the study environment to the expected testing environment. That is a reasoned application to a common situation, rather than a claim that all sound necessarily harms memory.

Four context conditions

The experiment independently varied sound during study and sound during testing. Each factor had silent and noisy levels, producing a $2\times2$ independent measures factorial design. Each participant experienced only one of the four context combinations, though everyone completed both test formats.

Eight members of a psychology laboratory class acted as experimenters. Each recruited five acquaintances: 40 participants, 17 women and 23 men, aged 17–56, with mean age 23.4 years. Each experimenter assigned participants to all four conditions and an additional assigned condition, randomly allocating their participants to those slots. This balanced experimenter involvement across conditions. One silent-study/silent-test participant's results were excluded because both scores were more than 2.5 standard deviations below the combined mean; 39 were analysed.

The material was a two-page, three-column article on psychoimmunology, the relationship between psychological processes and immune function. Copies of one master cafeteria recording supplied background conversation, chair movement and dish sounds. Experimenters used their own cassette players and headphones; sound was moderately loud. The dependent variables were scores on ten short-answer questions and sixteen four-option multiple-choice questions.

Reading and testing procedure

The procedure controlled several features that might otherwise explain a context effect. Instructions described a voluntary class project and informed participants that comprehension would be tested using both formats. They read the article once as for a class assignment, and could underline or highlight it. Experimenters recorded reading time rather than imposing a fixed time.

Everyone wore headphones during study and testing. Silent participants heard nothing; noisy participants heard the cafeteria recording and were told to ignore it. Keeping headphones present in every condition reduced the chance that wearing headphones, rather than the sound, explained differences.

A break of about two minutes preceded testing to reduce reliance on immediate short-term retention. The short-answer test always came first, so its answers could not be copied from recognition of the later multiple-choice alternatives. Questions followed the article's order and concerned information explicitly stated in it. Participants were debriefed at the end; the procedure lasted approximately 30 minutes.

Using the same article, recording and questions improves comparability and makes the task replicable. Nevertheless, different personal equipment and individually managed sound volume could introduce variation between experimenters. Fixed test order addresses answer contamination, but also means test format is confounded with being tested first or second.