Background and aim

1. Link to the theory: Williams et al. (1992) supports the activation synthesis theory of dreaming, in which the cerebral cortex builds a dream out of random activity generated during REM sleep.
2. Aim: to show that dreams are a product of the different brain activity that occurs during sleep.
3. Hypothesis: the bizarre content of dreams would be different from the bizarre content of waking fantasies, because the two are products of different types of brain activity.
4. Bizarreness: how strange, illogical or disjointed a report is, including remoteness of time and place and breaks in the plot.
5. A waking fantasy is a daydream, something imagined while awake.

Method

1. Design: a natural experiment that also used self-report.
2. Variables: the independent variable was whether the report was of a dream or of a waking fantasy, and the dependent variable was how bizarre it was, measured as a bizarreness density score.
3. Sample: twelve biopsychology students from an American university, aged 23 to 45, ten of whom were female.
4. Materials: a written journal kept by each student for one term, plus the two scoring scales the judges used.
5. Procedure: the students recorded any dreams they could recall and any fantasies they experienced while awake.
6. The researchers then selected 60 dreams and 60 fantasies, choosing the reports that were more detailed and had visual content.
7. Each report was scored on two scales, one for the locus of the bizarreness and one for its type, taken sentence by sentence.
8. A bizarreness density score was worked out by dividing the number of bizarre items in a report by the number of sentence units in it.
9. Three separate judges did the scoring, and their agreement was checked for inter-rater reliability.
10. The judges did not know whether they were reading a dream or a fantasy, and each time they were asked afterwards which they thought it had been.

Results

1. The three judges showed good inter-rater reliability, so the scoring was consistent between them.
2. There was a significant difference between the mean density scores for dreams and for fantasies, with the mean for dreams about three times higher.
3. The largest difference of all was on plot discontinuity.
4. Seven of the twelve participants had dreams scoring significantly higher for bizarreness than their own fantasies.
5. The judges were 88.7% accurate at telling a dream report from a fantasy report.

Conclusions

1. The bizarreness of dreams follows from brain activity during REM sleep, which is why dream reports carry more of it, including remoteness of time and place.
2. The 88.7% accuracy suggests dreaming and fantasising involve different types of cognitive activity.
3. How it relates to the theory: dream reports were far stranger than waking fantasies, which fits content that begins as random neural activity.

How the brain differs when dreaming and when awake

1. Awake, the brain works on a steady stream of information from the senses and can check what it thinks against what is in front of it.
2. In REM sleep the body is paralysed and sensory information never reaches the areas that interpret it, so the brain works on activity it generates itself.
3. That activity also reaches the limbic system, which deals with emotion, so a dream report carries strong feeling as well as strange content.

Criticisms

1. Objective and blind scoring is a strength, because three judges who did not know which type of report they held used an operationalised density score with checked inter-rater reliability.
2. Ecological validity is a further strength, because recording at home across a whole term is much closer to normal sleeping than a night wired up in a laboratory.
3. Testability is a strength too, because the theory produced a prediction that could be tested and could have failed.
4. The sample is a weakness, because twelve people is too small a group to generalise from, and ten of the twelve were female, so the study may be gender biased.
5. Self-report relies on memory, insight and the ability to write an experience down, and dreams were recorded much later than fantasies, so dream detail is more likely to have been forgotten and rebuilt.
6. Construct validity is a further weakness, because rich qualitative data was turned into numerical scores, so much of the detail of the experience was lost.
7. Control of the independent variable was limited, because there was no guarantee the reported dreams happened in REM sleep, and fantasies recorded while drowsy may already have been dream-like.

Turning this into marks

1. When a strength and a weakness are both asked for, make them genuinely separate, so the small sample and the reliance on self-report count as two while two versions of the sample problem count as one.
2. Name waking fantasies as the comparison, because writing that dreams were compared with being awake does not secure the knowledge mark.
3. The mark most often lost is quoting the bizarreness scores without saying that greater bizarreness is what random activation during REM sleep predicts.