Abstract Index Conference Index

ABSTRACT

Accuracy of Dream Recall: I. Data from Real Dreams

Authors: William H. Moorcroft, Brandis McFarland, Bill Larkin, Bryan Castelda, Erin Cenefelt, Steven Ondrashek, Natalie Stage, Laura Koenig, Devon Whitehead, & Angela Sweeny

Presenter
William (Bill) H. Moorcroft, Ph.D., Luther College, Decorah, Iowa
Professor of Psychology and Director of the Luther College Sleep and Dreaming Laboratory
 email: moorcrwi@luther.edu

4. Summary of Presentation.

Fourteen subjects each recalled a dream following a REMS awakening then again the next morning, a week later, and a month later. Although the gist of the recalled dream never varied, considerable variation was seen in detailed elements when the four recalls were compared casting reservations about dream recall accuracy.

5. A) Learning Objectives.
1. to evaluate how much of the actual dream content is recalled following a REMS awakening
2. to appreciate how much recall of a dream can change over time
3. to appreciate the individual differences in dream recall accuracy

B) evaluation questions
1. How good is the recall of a dream following REMS awakening?
2. How much does the recall of a dream change with the passage of time?
3. How great are the individual differences in dream recall?

8. Abstract.

Both doing dream work and the understanding of the process of dreaming necessarily depend on recalls of dream experiences. Yet seeming little attention has been paid to how accurately dreams are recalled following REMS awakenings or at later times. Like recall of stories (Bartlett, 1930) and waking experiences (Loftus, 1979) it is possible that the recalled memory of dreams may be imperfect and subject to change with the passage of time.
Recall of a single dream from each of 17 subjects (14 usable) were tape recorded on four occasions: following a REMS awakening, the next morning, one week later, and one month later. Each transcribed recall was divided into distinct "storyboard" elements. (Another person did likewise for half of the recalls yielding 91.0 % agreement.) Then pooling all the elements from all recalls but eliminating redundancies produced a composite dream. (Although some elements in the actual dream may not be in the composite and elements in the composite may not have been a part of the actual dream we believe this composite is the closest likeness to the actual dream that can be obtained.)
To reduce the problem of differences in recall report lengths, percentages of similarities and changes were used for all comparisons. Each comparison was tested for significance beginning with a one-way (time of recall) repeated measures ANOVA followed either by Tukey's HSD post hoc test or, if the ANOVA was not significant or the appropriate test, a single sample T-test collapsed across the time of recall dimension.
The gist never changed in any series of recalls but the average (mean) recall contained only 49.9 ± 11.2 (standard deviation)% of the composite elements (F(3,39)=2.611; t=-16.205* Vs. 100% recall). The REMS awakening recalls were somewhat better with a mean of 55.6± 15.3% of the composite elements (t=10.494* Vs. 100% recall) but a mean of 56.5± 24.1% of the elements reported following the REMS awakening were absent in subsequent recalls (F(2,26)=1.532; t=8.438* Vs. 0% absent). Additionally, the mean composite elements not reported in the REMS awakening recall but added in subsequent recalls was 22.2± 9.7% (F(2,26)=0.270; t=8.304* Vs. 0% added). (However, these are not pure measures of how many identical elements were present or absent in the four recall samples throughout the entire month. In many cases something forgotten in any one recall may have been remembered in another at the same time that something else was forgotten and vice versa.) Finally a mean of 53.7± 36.5% of the elements added during the morning or one week recall were retained in a subsequent recall (F(1,10)=0.189; t=-4.009* Vs. 100% retention and t=4.644* Vs. 0% retention).
Beyond the gist of the dream, these results argue against accurate recall, even those following REMS awakening. Additionally, these results show that there are also significant numbers of omissions and additions to recalls of the same dream done at different times. Finally, individual differences in accuracy of dream recall appear to be great.
*p<0.05
References:
Bartlett, F. C. (1932). Remembering: A study in experimental and social psychology. London: Cambridge University Press.
Loftus, Elizabeth F. (1979). Eyewitness testimony. Cambridge: Harvard University Press.