Biases in the relationship between dream threats and level of anxiety upon awakening

Abstract

Objectives: Controlling report length in dream content analysis comprises a significant methodological problem. Individual differences occur in report length which can influence category coding and rating scales. Differences are also found in dream content by sex and age. The aim of this study is to determine the bias of certain variables in dream content analysis when using rating scales, coding systems and questionnaires. As such, an evaluation was performed of the bias of these variables on the relationship between anxiety upon awakening, social threats (ST) and terrifying threats (TT) established in a previous study. Methods: The sample consisted of 215 dreams collected in dreamers' homes (63 belonged to men and 152 to women). The dreamer's level of anxiety upon awakening was assessed with the CEAD. The level of social and terrifying threats in the content of the dreams was also assessed. Other variables entered into the analysis were sex, age, dream length, number of hours before answering the questionnaire, number of hours' sleep and the frequency with which the dreamer suffers nightmares. Results: Use of the Mann Whitney U found significant differences by sex in the dreamer's nightmare frequency (z=-2.53 p=.011), in terrifying threats in the dream (z=-2.03 p= .042) and by dream time (z=-2.51 p=.012). The Spearman Rho correlation coefficient indicated a positive relationship between anxiety upon awakening and nightmare frequency (Rho=.26 p<.001). Social and terrifying threats were also positively correlated with word count and the number of dream characters (Rho=.37 p<.001, Rho=.17 p=.010). Both anxiety upon awakening and social and terrifying threats were negatively correlated with the age of the dreamer (RhoCEAD-AGE=-.20 p=.006, RhoST-AGE=-.30 p<.001, RhoTT-AGE=-.37 p<.001). Possible biases due to sex, age, word count and the number of characters were statistically controlled by means of partial correlation. Through the use of partial correlations, the significance between anxiety upon awakening, social threats and terrifying threats in the dream was observed to be maintained (rCEAD-TS=.17 p=.025, rCEAD-TT=.19 p=.011). Conclusion: The sex, age of the dreamer, the report word count and the number of dream characters must be controlled in research into dream content. In addition, after eliminating these biases, a significant relationship was confirmed between threats which appear in the dream and the dreamer's level of anxiety upon awakening.