Lesson 11 of 11 · 20 min

Context-dependent memory and Grant et al. (1998)

Analyse Grant et al.'s (1998) 2 × 2 experiment on studying and testing in silence or noise, interpret its interaction and conclusion, and analyse data from a modified version of it.

What you will be able to do

  • Explain context-dependent memory with retrieval cues.
  • Describe the aim, 2 × 2 design, variables, procedure and controls of Grant et al. (1998).
  • Interpret the results as an interaction between study and test conditions, and state the conclusion.
  • Analyse and interpret data from a modified Grant et al. (1998) experiment, including means, a t-test result and the limitations of the evidence.

You walk into the kitchen and cannot remember what you came for. You walk back to your desk, and it comes to you. The thought was stored all along; what was missing was the cue, and the cue was the place where the thought began.

Context-dependent memory is the finding that retrieval is better when the external context (the place, the background sound, the smell) at retrieval matches the context at encoding. The context is encoded along with the information, so being back in it provides retrieval cues. It is a kind of retrieval failure in reverse: when the cues are missing, information that is stored is harder to reach.

History

Earlier evidence

Godden and Baddeley (1975) had divers learn lists of words either on a beach or 6 metres underwater, then recall them on land or underwater. Words were recalled better when the learning and recall environments matched. But word lists are not what students learn. Grant et al. (1998) asked whether the effect holds for meaningful material, of the kind studied for exams.

Key idea

Grant et al. (1998)

Aim. To investigate whether context-dependent memory occurs for meaningful material: whether matching the noise conditions at study and test improves performance on tests of a studied article.

Participants. 40 people aged about 17 to 56, recruited by the student researchers from people they knew (a convenience sample), 10 in each condition.

Procedure. Each participant read an article on psychoimmunology, wearing headphones. In the silent condition nothing was played through them; in the noisy condition they heard a recording of background café noise at a moderate volume. They then answered a short-answer test and a multiple-choice test on the article, either in the same condition or in the other one.

Design. A 2 × 2 independent groups design: two independent variables, each with two levels, giving four conditions. Each participant was allocated to one condition.

The 2 × 2 design of Grant et al. (1998). Shaded conditions (a match between study and test) are where context-dependent memory predicts the best scores.
Study conditionTested in silenceTested with noise
Studied in silenceSilent–silent (match)Silent–noisy (mismatch)
Studied with noiseNoisy–silent (mismatch)Noisy–noisy (match)

Variables.

  • IV 1: the study condition (silent or noisy while reading).
  • IV 2: the test condition (silent or noisy while answering).
  • DV: memory for the article, operationalised as the score on the short-answer test and the score on the multiple-choice test.
  • Controlled variables: the same article and tests for everyone; headphones worn by everyone in every condition (so wearing them could not explain a difference); the same recording at the same volume; tested individually.

The short-answer test is a measure of recall; the multiple-choice test is a measure of recognition.

Why did participants in the silent conditions of Grant et al. (1998) wear headphones even though nothing was played through them?

Choose one answer

The results. On both tests, performance was better when the study and test conditions matched (silent–silent and noisy–noisy) than when they did not. Studying in silence or in noise did not itself change scores: what mattered was whether the conditions matched.

To see what that pattern looks like in numbers, here is a class replication of the full 2 × 2 design. Its numbers are illustrative class data, not Grant et al.'s, but they show the same pattern.

Mean scores in a class replication of Grant et al. (1998), 10 students per condition (illustrative class data). The short-answer test had 10 questions and the multiple-choice test 16.
Condition (study–test)Short-answer test (out of 10)Multiple-choice test (out of 16)
Silent–silent (match)6.614.0
Silent–noisy (mismatch)5.212.4
Noisy–silent (mismatch)4.912.6
Noisy–noisy (match)6.313.8

Short-answer scores in the class replication (illustrative class data). The lines cross: scores were higher whenever study and test conditions matched. This crossing is an interaction.

Reading the results. Look at what is not there first. Studying in silence was not clearly better overall (6.6+5.22=5.9\frac{6.6 + 5.2}{2} = 5.9 against 4.9+6.32=5.6\frac{4.9 + 6.3}{2} = 5.6), and being tested in silence was no better overall (5.75 either way). Neither IV on its own made much difference (no main effect). What mattered was the combination: whichever condition people studied in, they scored higher when tested in the same condition. When the effect of one IV depends on the level of the other, the two IVs interact, and on a graph the lines cross.

Average the two matching conditions and the two mismatching ones for the short-answer test:

match: 6.6+6.32=6.45mismatch: 5.2+4.92=5.05\text{match: } \frac{6.6 + 6.3}{2} = 6.45 \qquad \text{mismatch: } \frac{5.2 + 4.9}{2} = 5.05

Matching conditions scored 1.4 marks (out of 10) higher. The multiple-choice test shows the same pattern.

Key idea

The conclusion

Context-dependent memory occurs for meaningful material, on both recall (short-answer) and recognition (multiple-choice) tests: performance is better when the study and test environments match. For students, the implication the researchers drew is to study in conditions like those of the exam: most exams are taken in silence, so study in quiet.

Interactive

Look at all four conditions

Choose a study condition and a test condition; the matching cells score higher.

Using the multiple-choice means in the table, calculate the difference between the mean of the matching conditions and the mean of the mismatching conditions.

in marks

Note

Limitations of the evidence

  • Sample. 40 participants over four conditions is 10 per condition: small, and a convenience sample of the researchers' acquaintances, aged about 17 to 56. The results may not generalise to school students.
  • Experimenter effects. The student researchers knew the hypothesis and tested people they knew, so their expectations could have influenced the participants.
  • Ecological validity. One reading of one article, tested in the same session, is not weeks of study and an exam days later. Only one kind of context (one recording of background noise) was tested.
  • Independent groups. Different people in each condition means individual differences in reading ability could add to the differences between conditions.

A modified experiment. In your own investigation (and in the data test) you will meet modified versions of Grant et al., with a different context: music or silence, one room or another, a scent present or absent, chewing gum or not. A common modification reduces the 2 × 2 design to two independent groups: one group studies and is tested in the same condition, the other in different conditions, and the two means are compared with an unpaired t-test.

Interactive

Analyse a modified experiment

Illustrative class data. The test result is hidden: work out whether the difference is significant, then show the test to check.

Worked example

Interpret data from a modified Grant et al. (1998) experiment

  • Complex unfamiliar
  • Technology active

A student modified Grant et al. (1998). Twenty Year 12 students read a one-page article. Ten, chosen at random, read it and took a 20-question test with the same instrumental music playing (same context); the other ten read it with music and took the test in silence (different context). Test scores:

Same context14121611151712141315
Different context1211151014139131212

Same context: standard deviation 1.91. Different context: standard deviation 1.79. An unpaired t-test gave t(18)=2.17t(18) = 2.17, p=0.043p = 0.043.

(a) State a null and an alternative hypothesis. (b) Calculate the mean score of each group. (c) Draw a conclusion. (d) Suggest one way the uncertainty in the conclusion could be reduced.

5 steps

Try it yourself first, then check the working one step at a time.

Method

  1. Hypotheses name the IV and the DV and the population.

  2. Means from totals; compare p with 0.05.

  3. Qualify the conclusion: sample size, p near the threshold, one kind of context.

Watch out for

  • Saying the experiment proves the hypothesis: it supports it.

  • Writing a directional alternative hypothesis when the test was two-tailed.

Interactive

Control the threats in a modified experiment

Switch on controls and see which threats each one removes.

Exam tip

Data questions on Grant et al.

  • Name both IVs and say each has two levels; the DV is the test score (say which test).
  • Describe the pattern as matching versus mismatching, quoting means, not "noise made people worse".
  • An interaction shows as crossing (or converging) lines; parallel lines would mean no interaction.
  • For a conclusion, use the p-value: p < 0.05 supports the hypothesis that matching conditions improve scores; then give a limitation (sample size, convenience sampling, one kind of context).

Summary

  • Context-dependent memory: retrieval is better when the context at retrieval matches the context at encoding, because the context acts as a cue.

  • Grant et al. (1998): 40 people read an article in silence or with café noise and were tested in silence or noise: a 2 × 2 independent groups design.

  • IVs: study condition and test condition. DV: short-answer (recall) and multiple-choice (recognition) scores. Headphones worn by all.

  • On both tests, scores were higher when conditions matched; studying in silence or noise did not itself change scores: an interaction.

  • Conclusion: context-dependent memory occurs for meaningful material; study in conditions like the exam.

  • Limitations: small convenience sample, experimenters who knew the hypothesis, one article and one kind of context.

Finished the lesson?

Try the 2 quick checks above and the lesson completes itself, or mark it complete now.

Lesson 11 of 110% read

Save your progress with a free account