Abstract
Study Objectives:
To conduct a comprehensive and comparative study of prospectively collected bad dream and nightmare reports using a broad range of dream content variables.
Design:
Correlational and descriptive.
Setting:
Participants' homes.
Participants:
Three hundred thirty-one adult volunteers (55 men, 275 women, 1 not specified; mean age = 32.4 ± 14.8 y).
Interventions:
N/A.
Measurement and Results:
Five hundred seventy-two participants kept a written record of all of their remembered dreams in a log for 2 to 5 consecutive weeks. A total of 9,796 dream reports were collected and the content of 253 nightmares and 431 bad dreams reported by 331 participants was investigated. Physical aggression was the most frequently reported theme in nightmares, whereas interpersonal conflicts predominated in bad dreams. Nightmares were rated by participants as being substantially more emotionally intense than were bad dreams. Thirty-five percent of nightmares and 55% of bad dreams contained primary emotions other than fear. When compared to bad dreams, nightmares were more bizarre and contained substantially more aggressions, failures, and unfortunate endings.
Conclusions:
The results have important implications on how nightmares are conceptualized and defined and support the view that when compared to bad dreams, nightmares represent a somewhat rarer—and more severe—expression of the same basic phenomenon.
Citation:
Robert G; Zadra A. Thematic and content analysis of idiopathic nightmares and bad dreams. SLEEP 2014;37(2):409-417.
Keywords: Bad dreams, dream content, nightmare, parasomnia, sex differences
INTRODUCTION
Considerable progress has been made over the past 25 y on the prevalence, etiology, and psychopathological correlates of nightmares1–3 and theoretical models have been proposed1,4 Empirical findings on nightmares' actual content, however, remain surprisingly scarce. Although findings exist on the content of trauma-related dreams and nightmares5–8 as well as recurrent dreams,9,10 little is known about the contents of idiopathic nightmares, the most common form of disturbed dreaming.11
Many reasons underlie the need for a comprehensive picture of idiopathic nightmares' content. First, as detailed in the following paragraphs, much of what is known about nightmare content is based on clinical anecdotes, questionnaire-based studies, or purely descriptive reports. Second, even though standard diagnostic texts such as the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV12) and International Classification os Sleep Disorders, Second Edition (ICSD-213) define nightmares as resulting in an awakening from sleep, little is known about the dream content that precedes such awakenings. Third, nightmares have been conceptualized as frightening dreams for more than 40 y,14 yet few studies15–17 have actually investigated the range of dysphoric emotions implicated in nightmares. Although these more recent findings are reflected in the revised ICSD-213 definition of nightmares as disturbing mental experiences rather than as frightening dreams, they are based on student populations and the results of the most comprehensive of these studies15 have yet to be replicated. Finally, although women report having substantially more nightmares than do men,3,18 sex effects in nightmare content remain virtually unexplored.
Table 1 presents the key characteristics and findings from studies of idiopathic nightmares' content. According to Table 1, studies vary greatly in the population examined, on how nightmares are defined, and in the instruments used to investigate nightmare content. In more than half of these studies, the definition provided for nightmares is inconsistent with standard diagnostic texts12,13 as the awakening criterion is not specified. These studies thus include negatively toned dreams that do not awaken the sleeper, otherwise known as bad dreams.19,20 Because bad dreams are about four times more frequent than nightmares,19,21 definitions that encompass bad dreams result in the investigation of a more common and broader dimension of disturbed dreaming. Nightmares have also been shown to be more emotionally intense than bad dreams,15 leading to the view that nightmares represent a somewhat rarer—and more severe—expression of the same basic phenomenon. The extent to which these two types of disturbed dreaming are related remains unclear and recent reviews outline the importance of addressing this question empirically.1,3,22
Table 1.
Summary of studies on the content of idiopathic nightmares in adults
Although daily prospective logs are considered the gold standard for the study of nightmares,3 most studies reviewed in Table 1 are based on questionnaire or interview methods. Questionnaires or similar retrospective instruments can yield inaccurate dream reports due to the fragile nature of dreams' long-term recall as well as memory and saliency biases.23 For instance, one comparison24 of dream content obtained from participants' questionnaires and 2-w logs found no relationship between the estimated frequency for the appearance of aggressive, friendly, and sexual elements and their actual frequency in the dream reports. Similarly, people's beliefs about the presence of anxiety in their dreams is often unrelated to the actual affective tone of their everyday dreams as recorded prospectively in sleep logs.25 Furthermore, when asked to think of a nightmare, most people are likely to recall particularly intense, unusual, or otherwise salient nightmares rather than more typical experiences. This may explain why themes of falling and of being chased are among the most frequently reported themes in studies based on questionnaire or interview data while appearing much less frequently in prospective logs (see Table 1).
Finally, although there exist well-established and validated coding systems for the quantification of dream content exist,26,27 they have been rarely used to investigate the contents of disturbed dreaming.28–30 Furthermore, only one study,31 which served to refine the current research study's methodology, used a validated instrument to investigate content differences in prospectively collected bad dream and nightmare narratives.
Our aim was to use a range of content variables to obtain a comprehensive and comparative description of prospectively collected bad dream and nightmare narratives. Based on previous findings (see Table 1) and in accordance with the view that nightmares and bad dreams are expressions of same basic phenomenon that varies in intensity, the following predictions were tested: (1) Physical aggressions and interpersonal conflicts will be the most frequently reported themes in bad dreams and nightmares, but a broader range of themes will characterize bad dreams; (2) Emotions in nightmares will be significantly more intense than in bad dreams; (3) Fear will be the most frequently reported emotion in nightmares and bad dreams but a greater proportion of bad dreams will contain other primary emotions; (4) When compared to bad dreams, nightmares will contain more negative dream content, including aggression, failure, and misfortune. Trigger factors within the dream narratives, the nature of the dreams' ending, and their degree of bizarreness were also investigated, as were sex effects. Given the absence of findings in relation to these variables, no specific predictions were formulated
METHODS
Participants
Participants were nonpaid volunteers recruited through media announcements between 2000 and 2010 from undergraduate classes (n = 295) as well as the general population (n = 277). Prospective participants were told that the study concerned the relation between dreams and measures of personality and that we were interested in both high and low recallers of dreams and in all types of dreams.
Procedures
Participants first completed a series of questionnaires including measures of personality and well-being as part of a separate program of research. They were then required to provide upon awakening a complete written description for each remembered dream in a daily log for 2 to 5 consecutive w. In addition to each recalled dream's narrative, participants had to report the date, the main emotions present (if any), the emotion's intensity on a five-point Likert scale, and to note whether the dream was a lucid dream, a nightmare, a bad dream, or a flying dream, for which participants were provided written definitions. Sleep terrors were also defined to distinguish them from nightmares. The protocol was accepted by the university's Ethics Committee and signed consent obtained from each participant.
Consistent with previous work,15,19 nightmares were defined as very disturbing dreams in which the unpleasant visual imagery and/or emotions cause the person to wake up (i.e., the dream's unpleasant content woke the dreamer while the dream was ongoing). Bad dreams were defined as very disturbing dreams that do not cause a person to awaken (e.g., the dream occurred earlier in the night prior to awakening, the person remembered it only after being awakened by external factors such as an alarm clock, or the person only remembered the dream later during the day). Thus, the definitions provided for nightmares and bad dreams were equivalent except for the waking criterion.
Measures
Thematic Content
Categories for the classification of the thematic content of bad dreams and nightmares were based on the literature, on a modified version of the Typical Dreams Questionnaire,32 and pilot testing.31 Whenever possible, conceptually related categories used in previous studies were grouped to avoid overlap and only categories capturing more than 3% of the narratives' content were retained. Table 2 presents the final 12 thematic categories used to classify nightmare and bad dream narratives. When dream reports contained more than one theme, raters had the option of identifying a secondary theme if its occurrence was not the direct consequence of the main theme (e.g., a character becomes ill only after being physically attacked).
Table 2.
Classification of thematic categories in nightmares and bad dreams
Emotional Content
Consistent with previous work on emotional content of bad dreams and nightmares,15 the main dream emotions explicitly reported by the participants were grouped into seven content categories shown to account for approximately 90% of emotions reported in disturbed dreaming. The categories were: fear (e.g., terrified, horrified, frightened, scared, panicky), anger (e.g., furious, angry, irritated, outraged), sadness (e.g., depressed, lonely, hopeless, heartbroken), confusion (e.g., puzzled, perplexed, bewildered), disgust, guilt (e.g., remorse, regret), and frustration. An eighth category encompassed all positive emotions and a ninth category labeled “other” comprised emotions that did not fit into the preceding categories.
Dream Content
The following variables were used from the Hall and Van de Castle27 system. Friendly and aggressive interactions. This scale measures the frequency of emotionally-toned social interactions. Success and failure. These variables measure the successful handling of some difficulty encountered by a character or an incapacity of the character to achieve a desired goal because of personal limitations and inadequacies. Good fortune and misfortune. Good fortune is scored when something beneficial happens to a character that is completely adventitious whereas misfortunes refer to any mishap, adversity, harm, danger, or threat that happens to a character as a result of circumstances over which they have no control. Two scales were used to investigate bizarreness. The Rationality Scale33 measures the likelihood of occurrence of the dream content and the degree of his adherence to natural laws whereas the Everydayness Scale33 examines the degree to which the dream content approximates that of everyday life.
Narrative Structure
To clarify how and when dreams turn into bad dreams or nightmares, each narrative was scored for the following variables.31 The presence of triggering factors, which initiate a negative turn of events, were classified into one of four categories: negative event (e.g., “Looking up at the sky I see a missile coming right down at us”); cognitions (e.g., “I suddenly realize that if I am floating in the air, it's because I am dead”); emotions (e.g., “My sister was there and I was very afraid of her”); or a combination thereof. Whether the trigger occurred during the first, second, or last third of the narrative was also determined by dividing the number of words before the trigger by the narrative's total word count. Fewer than 2% of dream reports contained information explicitly noted by the participant as being out of sequence (e.g., remembering a detail from the dream's beginning after having completed the narrative). In these cases the material was moved to its correct place (as specified by the participant) before segmenting the report according to its word count. Each narrative's ending was scored according to the following outcomes: negative (e.g., dying or being caught); partially positive (e.g., dreamer escapes danger but a partner is injured); positive (e.g., dreamer succeeds in fighting off an aggressor). Finally, four categories were used to determine each nightmare's cause of the awakening as reported by the dreamer: emotional intensity (e.g., “I was so afraid that I woke up”); immediate threat (e.g., “The man rushed toward me with a knife and I woke up”); sudden perceptually striking event (e.g., “The officer rang the alarm, it made a loud noise and I woke up”); intentional awakening (e.g., “I realized that I was dreaming so I woke myself up”).
All content variables were scored independently by two extensively trained raters who refined their scoring on an unrelated sample of 225 bad dreams and nightmares. Approximately 25% of all the total narratives were scored by both judges. Because kappa coefficients are unreliable when applied to variables with infrequent occurrences (as is the case with several dream content categories), interjudge reliability was assessed with Gwet's34 AC1 statistic for interrater reliability. Results showed a good to excellent agreement across all dream content categories with AC1 values ranging from 0.62 to 0.98, and 10 of the 15 variables obtaining values above 0.80.
RESULTS
Participants reported a mean of 17.1 ± 11.0 dreams in their logs for a total of 9,796 dream reports. One hundred sixty-four of the 572 participants (29%: 125 women, 36 men, three sex not specified) were excluded from the study because they did not report any nightmares or bad dreams in their logs. Of the 281 nightmares and 1,016 bad dreams collected from the remaining 408 participants, a total of 22 nightmares (8%) and 73 bad dreams (7%) were excluded because they contained too few words (< 25) to permit a reliable content analysis (eight nightmares, 50 bad dreams), the report's content was too vague or confused (12 nightmares and 23 bad dreams), or the experiences appeared to be sleep terrors (two nightmares). Eighty-five bad dreams were further excluded as the participants explicitly reported that the dream's content resulted in an awakening, which was inconsistent with the operational definition provided. Of the remaining 331 participants, 61 (18.4%; 15 men, 46 women) reported nightmares exclusively, 133 (40.2%; 20 men, 113 women) reported bad dreams exclusively, and 136 (41.1%; 20 men, 116 women) reported both. To prevent an overrepresentation of participants reporting many bad dreams or nightmares, a maximum of two nightmares and two bad dreams were selected per participant. The first nightmare or bad dream reported in the log was automatically included and, if available, a second report was randomly selected from the log.
The final sample was thus comprised of 253 nightmares and 431 bad dreams reported by 331 participants (55 men, 275 women, one not specified; mean age = 32.4 ± 14.8 y). Men and women did not differ significantly in age (male = 34.9 ± 15.6 y; female = 31.9 ± 14.6 y). The mean number of words per narrative was 144.4 ± 114.3. Nightmare reports contained significantly more words (165.8 ± 140.9) than did the bad dreams (131.8 ± 93.2), (t (1, 681) = 3.79, P < 0.001), with a small effect size (d = 0.30). The only significant sex difference for report length was that women's bad dreams contained more words (136.8 ± 95.5) than did the men's (103.2 ± 72.1), t (1, 428) = 2.62, P < 0.01, with a small effect size (0.36).
Thematic Content
Fifty-six percent of the narratives had a single theme and 44% had two themes. Nightmares were significantly more likely than bad dreams to contain two themes (52.2%, versus 39.7%; χ2 = 10.1, P = 0.001) but this difference disappeared when report length was controlled for t (1, 682) = 0.73, P = 0.465). The distribution of thematic categories across the nightmares and bad dreams are presented in Table 3. Themes involving physical aggression and interpersonal conflicts were the most frequent, followed by failure/helplessness, health-related concerns/death, and apprehension/worry. All other themes appeared in fewer than 10% of the narratives. Nightmares were significantly more likely to contain themes of physical aggression, being chased, evil forces, and accidents, whereas themes of interpersonal conflicts were significantly more frequent in bad dreams.
Table 3.
Nightmare and bad dream themes
Emotions
Table 4 presents the mean emotional intensity of nightmares and bad dreams as well as the proportion of different emotions contained by each type of disturbing dream. Nightmares were rated by participants as being significantly more intense than were the bad dreams, with a corresponding large effect size. Fear was the most frequently reported emotion in both types of dreams but appeared in a significantly greater proportion of nightmares. There were no significant differences between nightmares and bad dreams on any of the other categories.
Table 4.
Main emotions and mean emotional intensity in nightmares and bad dreams
Hall and Van de Castle
Because dream report length differed significantly between nightmares and bad dreams, word count was controlled for by dividing the total number of mentions of each variable by the report's number of words and the result multiplied by 100. Mann-Whitney U tests were used to compare nightmare and bad dream narratives because content variables were positively skewed. When compared to nightmares, bad dreams contained (per 100 words) significantly more mentions of friendliness (0.46 ± 0.79 versus 0.26 ± 0.49, P < 0.05) whereas nightmares contained significantly more mentions of aggression (1.38 ± 1.54 versus 1.09 ± 1.24, P < 0.05) and failure (0.09 ± 0.34 versus 03 ± 0.20, P < 0.01). The minimal U value for the significant effects was 45852.5. Nightmares and bad dreams did not differ on measures of misfortune, good fortune, or success (P > 0.05).
To better understand how nightmare and bad dream content differs from everyday dreams, the proportion of the narratives containing at least one mention of each content variable was computed and compared to the Hall and Van de Castle normative data,27 which have been replicated in several studies.26,35 Comparisons were made as a function of sex and type of disturbing dream (nightmares, bad dreams, and both combined). Results are presented in Table 5. For both males and females, the proportion of nightmares and of bad dreams containing one or more misfortune was almost twice that of everyday dreams, whereas the proportion containing at least one aggression was about 1.5 times greater. Men's bad dreams and women's nightmares were also significantly less likely to contain friendliness than were normative equivalents. Nightmares and bad dreams contained fewer failures than the established norms, but differences were not always significant due to the variable's infrequent occurrence. Although the aggression/friendliness percent could not be statistically compared to the norms, they appeared equivalent to normative values across sexes.
Table 5.
Hall and Van de Castle categories: comparisons to male/female norms
Narrative Structure
As shown in Table 6, negative events were identified as triggers in about 75% of all dream narratives and appeared in the first third of the dream approximately 60% of the time. Nightmares and bad dreams did not differ significantly on trigger variables (P > 0.05). A significantly higher proportion of nightmares than bad dreams contained an unfortunate ending whereas partially and totally positive endings characterized a greater proportion of bad dreams. Nightmares were significantly more bizarre (i.e., less rational and more unlike everyday life) than were bad dreams. Finally, participants spontaneously noted the cause of their awakening in 36.5% of the nightmare reports. The most frequently invoked explanations included an immediate threat (42.4%), the intensity of the emotions experienced (25%), a salient element (18.5%), and an intentional awakening from the nightmare (14.1%).
Table 6.
Other dream content variables
Sex Differences
Sex effects were first separately explored within nightmares and bad dreams but because observed patterns were very similar to those found for all disturbing dreams combined, they are presented for the overall dream dataset. When compared to women, men were significantly more likely to report themes involving disaster/calamity (9.4% versus 4.7%; χ2 = 3.93, P < 0.05), as well as insects (9.4% versus 4.5%; χ2 = 4.33, P < 0.05). Men's dreams contained significantly less friendliness per 100 words (0.21 ± 0.50 versus 42 ± 0.73, U = 25309, P = 0.001). The occurrence of negative events was significantly more likely to be identified as triggers in men's dreams than in women's (90.6% versus 73.8%; χ2 = 13.98, P < 0.001) but appeared less frequently during the middle portion of the narrative (22.6% versus 32.1%; χ2 = 3.78, P < 0.05). No other signifi-cant sex differences were found (e.g., emotional intensity, main emotion, ending, bizarreness).
Within-Subject Comparisons
Because 136 participants reported at least one nightmare as well as one or more bad dreams in their logs, additional within-subject analyses were conducted using one nightmare, one bad dream, and one randomly selected control (non-nightmare and non-bad dream) dream report from each of these participants. Hall and Van de Castle content categories were scored per 100 words as well as in accordance with the “at least one per dream” procedure. Because the two methods yielded virtually identical results and given that several content variables were absent from most dream reports, only the proportion of dreams with at least one occurrence of each content category is presented here. As can be seen in Table 7, participants' nightmares and bad dreams contained significantly more words than did their everyday dreams and both types of disturbing dreams were also significantly more emotionally intense, more bizarre, and more likely to contain misfortunes than were the control dreams. In addition, participants' nightmares were significantly more emotionally intense, more bizarre, and more likely to contain aggressions than were their bad dreams.
Table 7.
Comparison of nightmares, bad dreams, and everyday dreams in participants having reported at least one of each in their daily logs
The top three thematic categories found in our total sample of nightmares and bad dreams (i.e., physical aggression, interpersonal conflicts, and failure/helplessness) characterized 81.4% of the 136 participants' nightmares, 71.3% of their bad dreams, and only 38.2% of their control dreams. Moreover, themes involving accidents, insects, and environmental anomalies occurred in fewer than 2% of these participants' everyday dreams whereas themes of being chased and of evil forces were entirely absent from the control dataset.
DISCUSSION
To summarize the main findings of the current study: (1) The most frequently reported theme in nightmares was physical aggression whereas interpersonal conflicts predominated in bad dreams; (2) Nightmares were rated by participants as being significantly more emotionally intense than were bad dreams; (3) Although fear was the most frequently reported main emotion in nightmares and bad dreams, almost half of all disturbing dreams contained primary emotions other than fear; (4) When compared to bad dreams, nightmares were more bizarre and contained significantly more aggressions, failures, and unfortunate endings; (5) Several content differences differentiated men and women's disturbing dreams; (6) Many of the key findings on nightmares and bad dreams were further supported by within-subject analyses. These findings are discussed in turn.
Although the frequent occurrence of themes involving physical aggression, interpersonal conflict, and failure/helplessness was expected, other content categories described in previously reviewed studies, including themes of falling, being paralyzed, and suffocation were rarely if ever reported in the current study. For instance, five of the studies reviewed in Table 1 reported elevated frequencies for the theme of falling, with percentages ranging between 21-73%. By contrast, themes of falling occurred so rarely in our sample (1.5% of all nightmares and bad dreams) that it was dropped as separate category and reclassified under accidents. Furthermore, themes involving feelings of paralysis or suffocation were entirely absent from our dataset. Two factors may account for these differences. The first concerns the prospective log-based nature of our data versus retrospective reports (e.g., questionnaires) that characterize past studies. Although themes of falling or being paralyzed appear infrequently in dream logs, their high saliency makes them particularly memorable and thus more likely to be recalled in interviews or questionnaires long after their occur-rence. Second, themes of falling, being paralyzed, or suffocation may well represent other commonly experienced parasomnias such as hypnic jerks, isolated sleep paralysis, or sleep terrors. When faced with broadly defined questionnaire items, people may be more likely to report these types of sleep experiences as if they were nightmares. Providing participants with definitions for bad dreams, nightmares as well as sleep terrors and requiring actual dream narratives to be reported in their logs may have reduced the inclusion of other sleep phenomena.
Although nightmares and bad dreams shared the same top three thematic categories (i.e., physical aggression, interpersonal conflict, and failure/helplessness), the proportion of nightmares containing physical aggression was more than twice that of bad dreams, and interpersonal conflicts were 1.5 times more frequent in bad dreams than in nightmares. Being chased and the presence of an evil force were the next most frequently reported themes for nightmares, whereas themes of health-related concerns and apprehension/worry were next for bad dreams. These findings suggest that nightmare content is more strongly related to themes involving a direct threat to physical integrity than are bad dreams which, in turn, are more likely to present a broader range of thematic contents. The data also lend support to the DSM-IV's12 clinical description of nightmares as involving threats to survival, security, or self-esteem. In fact, the current findings show that physical threats are more likely to characterize nightmares whereas psychological threats, including threats to self-esteem, are more prominent in bad dreams. Also noteworthy is the fact that these top three thematic categories occurred about twice as often in participants' nightmares and bad dreams than in their everyday control dreams and that other common themes including insects, being chased, and the presence of evil forces was almost entirely absent from the control dreams.
Our data on the intensity and content of emotions in nightmares and bad dreams support the idea that nightmares are significantly more emotionally intense than bad dreams and that a significantly greater proportion of bad dreams than nightmares (55% versus 35%) contain emotions other than fear, including anger, sadness, and frustration. These findings are consistent with another prospectively based study,15 although participants in that study were not required to provide narrative descriptions of their disturbed dreaming. These data show that the awakening criterion can be used as an indirect measure of disturbing dreams' intensity and support the direction taken by the American Academy of Sleep Medicine13,35 in defining nightmares as disturbing mental experiences rather than frightening dreams. Moreover, by restricting disturbing dreams to frightening dreams, studies may be losing out on approximately 30% of nightmares and 50% of bad dreams experienced by the general adult population.
Apart from a pilot study on women's nightmares,30 this was the first study to have investigated the content of nightmares and bad dreams using the Hall and Van de Castle system,27,36 the best validated and most widely used instrument for quantifying dream content. A comparison of our data to the Hall and Van de Castle norms reveals that men's and women's disturbing dreams contain significantly more aggressions and misfortunes as well as fewer friendly interactions than do everyday dreams. These results are consistent with those reported by Taub et al.28 However, the finding that failures occurred significantly more frequently in the norms than in our disturbing dreams was unexpected. Because failures in dreams result from a character's “personal limitations and inadequacies”27 this finding suggests that bad dreams and nightmares are less likely to reflect issues of personal competence than do everyday dreams.
Our exploratory work on negative triggers suggests that events external to the dreamer are typically responsible for turning everyday dreams into disturbing ones and that the first portion of bad dream and nightmare reports can be relatively uneventful. Not surprisingly, nightmares were significantly more likely than bad dreams to end on a negative note (78% versus 62%). However, it should be noted that 22% of all nightmares and 38% of bad dreams contained either a partially positive or entirely positive outcome (e.g., taking control over a situation, being finally saved or rescued). Although positively toned recurrent dreams have been described,36 this is the first time that positive content elements have been investigated in nightmare and bad dream narratives. The study of these kinds of content variations within nightmares and their possible differential relation to nightmare distress or well-being19,37,38 warrants further investigation.
Nightmares were significantly more bizarre than bad dreams, a difference that remained significant even after controlling for dream report length. When compared to previous findings39 based on the same scales, these results suggest that nightmares and bad dreams are more bizarre than are everyday dreams. Moreover, the results from the within subjects analyses directly support the idea that nightmares are significantly more bizarre than bad dreams which, in turn, are significantly more bizarre than everyday dreams.
Even though men's and women's narratives showed similar content profiles, some significant sex differences were noted. Nightmares in men were more likely than those in women to contain themes of disaster/calamity, a finding consistent with Schredl's40 retrospective study showing a higher frequency of war/terror themes in men's nightmares. In addition, although themes of interpersonal conflicts occurred in about 34% of men and women's bad dreams, they were twice as frequent in women's nightmares than in men's (23% versus 11%). Dreams of interpersonal conflicts may thus elicit a more intense emotional response in women, leading to a greater proportion of such dreams ending in a nightmare awakening.
Although the current study did not aim to test predictions from theories on nightmares or dream function, our findings may have implications for some of these models. Many of the thematic categories found to characterize a majority of nightmares are globally consistent with the Threat Simulation Theory4,41 of dreams, which stipulates that the function of dreaming is to simulate threatening events with the intent of improving the dreamer's capability to recognize and avoid threats in real life. However, it remains unclear to what extent nightmares contain realistic threat perceptions and efficient or successful avoidance responses. Our finding that 55% of bad dreams and 35% of nightmares contain primary emotions (as reported by the dreamer) other than fear suggests that the focus on fear memories put forth in recently proposed neuro-cognitive models1,3 of disturbed dreaming may need to be broadened. Finally, a large body of evidence26,42–44 supports the continuity hypothesis of dreaming, which postulates that dream content reflects waking states and concerns. While the occurrence of idiopathic nightmares has been associated to waking life stressors,45–47 the extent to which the content of bad dreams and nightmares reflect the nature of these stressors remains unknown. Although emotionally charged dreams may depict such stressors metaphorically, the relation between specific waking concerns and the central themes highlighted in the current study (e.g., interpersonal conflicts, health-related concerns) warrants further study.
Although the current work represents the largest and most in-depth investigation of nightmare and bad dream content to date, it has several limitations. First, our subject pool, although large, was biased toward individuals who were interested in dreams. Second, our sample of bad dreams and nightmares was overrepresented by women. Third, we do not know how many of our participants had a history of trauma and, consequently, what proportion of our disturbing dreams may have been trauma-related.
Finally, based on the 9,796 dream reports collected from the study's 572 participants, nightmares were found to make up 2.9% of all prospectively collected dream narratives whereas bad dreams were about 3.5 times more prevalent, accounting for 10.8% of the dream reports. Consistent with previous observations,19 most participants reporting nightmares in their daily logs also reported bad dreams whereas many individuals only experience bad dreams. Our results reveal that bad dreams are significantly more emotionally intense, more bizarre, and more likely to contain misfortunes than everyday control dreams. They also converge in showing that nightmares are in many ways more intense than bad dreams while nevertheless sharing several content characteristics. These findings indicate that nightmares and bad dreams share phenomenological similarities and provide strong support for the view that nightmares represent a rarer and more severe expression of the same basic phenomenon.
DISCLOSURE STATEMENT
This was not an industry supported study. The authors have indicated no financial conflicts of interest. Research supported by the Social Sciences and Humanities Research Council of Canada (grant # 410-2007-1957 to Dr. Zadra).
REFERENCES
- 1.Nielsen T, Levin R. Nightmares: A new neurocognitive model. Sleep Med Rev. 2007;11:295–310. doi: 10.1016/j.smrv.2007.03.004. [DOI] [PubMed] [Google Scholar]
- 2.Wittmann L, Schredl M, Kramer M. Dreaming in Posttraumatic Stress Disorder: A critical review of phenomenology, psychophysiology and treatment. Psychother Psychosom. 2006;76:25–39. doi: 10.1159/000096362. [DOI] [PubMed] [Google Scholar]
- 3.Levin R, Nielsen TA. Disturbed dreaming, posttraumatic stress disorder, and affect distress: A review and neurocognitive model. Psychol Bull. 2007;133:482–528. doi: 10.1037/0033-2909.133.3.482. [DOI] [PubMed] [Google Scholar]
- 4.Revonsuo A. The reinterpretation of dreams: an evolutionary hypothesis of the function of dreaming. Behav Brain Sci. 2000;23:877–901. doi: 10.1017/s0140525x00004015. [DOI] [PubMed] [Google Scholar]
- 5.Duval M, Zadra A. Frequency and content of dreams associated with trauma. Sleep Med Clinics. 2010;5:249–60. [Google Scholar]
- 6.Esposito K, Benitez A, Barza L, Mellman T. Evaluation of dream content in combat- related PTSD. J Trauma Stress. 1999;12:681–7. doi: 10.1023/A:1024725319777. [DOI] [PubMed] [Google Scholar]
- 7.Wood JM, Bootzin RR, Rosenhan D, Nolen-Hoeksema S, Jourden F. Effects of the 1989 San Francisco earthquake on frequency and content of nightmares. J Abnorm Psychol. 1992;101:219–24. doi: 10.1037//0021-843x.101.2.219. [DOI] [PubMed] [Google Scholar]
- 8.Phelps AJ, Forbes D, Hopwood M, Creamer M. Trauma-related dreams of Australian veterans with PTSD: content, affect and phenomenology. Aust. N Z J Psychiatry. 2011;45:853–60. doi: 10.3109/00048674.2011.599314. [DOI] [PubMed] [Google Scholar]
- 9.Zadra A, Desjardins S, Marcotte E. Evolutionary function of dreams: A test of the threat simulation theory in recurrent dreams. Conscious Cogn. 2006;15:450–63. doi: 10.1016/j.concog.2005.02.002. [DOI] [PubMed] [Google Scholar]
- 10.Robbins PR, Tanck RH. A comparison of recurrent dreams reported from childhood and recent recurrent dreams. Imag Cog Person. 1991;11:259–62. [Google Scholar]
- 11.Nielsen T, Zadra A. Idiopathic nightmares and dream disturbances associated with sleep- wake transitions. In: Kryger M, Roth T, Dement W, editors. Principles and practice of sleep medicine. 5th ed. Philadelphia: Saunders; 2011. pp. 1106–15. [Google Scholar]
- 12.American Psychiatric Association. Diagnostic and statistical manual of mental disorders: DSM-IV-TR. 4th ed. Washington, DC: American Psychiatric Association; 2000. [Google Scholar]
- 13.American Academy of Sleep Medicine. International classification of sleep disorders: diagnostic and coding manual. 2nd ed. Westchester, IL: American Academy of Sleep Medicine; 2005. [Google Scholar]
- 14.Hersen M. Nightmare behavior: A review. Psych Bull. 1972;78:37–48. doi: 10.1037/h0032960. [DOI] [PubMed] [Google Scholar]
- 15.Zadra A, Pilon M, Donderi DC. Variety and intensity of emotions in nightmares and bad dreams. J Nerv Ment Dis. 2006;194:249–54. doi: 10.1097/01.nmd.0000207359.46223.dc. [DOI] [PubMed] [Google Scholar]
- 16.Dunn KK, Barrett D. Characteristics of nightmare subjects and their nightmares. Psychiatr J Univ Ott. 1988;13:91–3. [PubMed] [Google Scholar]
- 17.Belicki K, Cuddy MA. Nightmares: Facts, fictions and future directions. In: Gackenbach J, Sheikh AA, editors. Dream images: A call to mental arms. New York: Baywood; 1991. pp. 99–113. [Google Scholar]
- 18.Schredl M, Reinhard I. Gender differences in nightmare frequency: a meta-analysis. Sleep Med Rev. 2011;15:115–21. doi: 10.1016/j.smrv.2010.06.002. [DOI] [PubMed] [Google Scholar]
- 19.Zadra A, Donderi D. Nightmares and bad dreams: Their prevalence and relationship to well-being. J Abnorm Psychol. 2000;109:273–81. [PubMed] [Google Scholar]
- 20.Blagrove M, Haywood S. Evaluating the awakening criterion in the definition of nightmares: How certain are people in judging whether a nightmare woke them up? J Sleep Res. 2006;15:117–24. doi: 10.1111/j.1365-2869.2006.00507.x. [DOI] [PubMed] [Google Scholar]
- 21.Robert G, Zadra A. Measuring nightmare and bad dream frequency: impact of retrospective and prospective instruments. J Sleep Res. 2008;17:132–9. doi: 10.1111/j.1365-2869.2008.00649.x. [DOI] [PubMed] [Google Scholar]
- 22.Spoormaker VI, Schredl M, van den Bout J. Nightmares: from anxiety symptom to sleep disorder. Sleep Med Rev. 2006;10:19–31. doi: 10.1016/j.smrv.2005.06.001. [DOI] [PubMed] [Google Scholar]
- 23.Schredl M. Questionnaires and diaries as research instruments in dream research: Methodological issues. Dreaming. 2002;12:17–26. [Google Scholar]
- 24.Bernstein DM, Belicki K. On the psychometric properties of retrospective dream content questionnaires. Imag Cogn Person. 1995;15:351–64. [Google Scholar]
- 25.Beaulieu-Prevost D, Zadra A. How dream recall frequency shapes people's beliefs about the content of their dreams. N Am J Psychol. 2005;7:253–64. [Google Scholar]
- 26.Domhoff GW. Finding meaning in dreams: A quantitative approach. New York: Plenum; 1996. [Google Scholar]
- 27.Hall CS, Van de Castle RL. The content analysis of dreams. New York: Appleton-Century-Crofts; 1966. [Google Scholar]
- 28.Taub JM, Kramer M, Arand D, Jacobs GA. Nightmare dreams and nightmare confabulations. Compr Psychiatry. 1978;19:285–91. doi: 10.1016/0010-440x(78)90049-4. [DOI] [PubMed] [Google Scholar]
- 29.Desjardins S, Zadra A. Threat simulation in nightmares. Sleep. 2004;27:A61. Abstract Supplement. [Google Scholar]
- 30.Zadra A, Duval M, Begin E, Pilon M. Content analysis of nightmares. Sleep. 2004;27:A64. Abstract Supplement. [Google Scholar]
- 31.Zadra A, Robert G. Why do some dreams become nightmares? Int J Dream Res. 2010;3:S11–2. [Google Scholar]
- 32.Nielsen TA, Zadra AL, Simard V, et al. The typical dreams of Canadian university students. Dreaming. 2003;13:211–35. [Google Scholar]
- 33.Cann DR, Donderi DC. Jungian personality typology and the recall of everyday and archetypal dreams. J Pers Soc Psychol. 1986;50:1021–30. doi: 10.1037//0022-3514.50.5.1021. [DOI] [PubMed] [Google Scholar]
- 34.Gwet KL. Computing inter-rater reliability and its variance in the presence of high agreement. Br J Math Stat Psychol. 2008;61:29–48. doi: 10.1348/000711006X126600. [DOI] [PubMed] [Google Scholar]
- 35.Tonay V. California women and their dreams: A historical and sub-cultural comparison of dream content. Imag Cogn Pers. 1990/1991;10:83–97. [Google Scholar]
- 36.Zadra L. Recurrent dreams: Their relation to life events. In: Barrett D, editor. Trauma and dreams. Cambridge, MA: Harvard University Press; 1996. pp. 231–47. [Google Scholar]
- 37.Belicki K. Nightmare frequency versus nightmare distress: Relations to psychopathology and cognitive style. J Abnorm Psychol. 1992;101:592–7. doi: 10.1037//0021-843x.101.3.592. [DOI] [PubMed] [Google Scholar]
- 38.Blagrove M, Farmer L, Williams E. The relationship of nightmare frequency and nightmare distress to well-being. J Sleep Res. 2004;13:129–36. doi: 10.1111/j.1365-2869.2004.00394.x. [DOI] [PubMed] [Google Scholar]
- 39.Lusignan F-A, Zadra A, Dubuc M-J, Daoust A-M, Mottard J-P, Godbout R. Dream content in chronically-treated persons with schizophrenia. Schizophr Res. 2009;112:164–73. doi: 10.1016/j.schres.2009.03.032. [DOI] [PubMed] [Google Scholar]
- 40.Schredl M. Nightmare frequency and nightmare topics in a representative German sample. Eur Arch Psychiatry Clin Neurosci. 2010;260:565–70. doi: 10.1007/s00406-010-0112-3. [DOI] [PubMed] [Google Scholar]
- 41.Valli K, Revonsuo A, Valli K, Revonsuo A. The threat simulation theory in light of recent empirical evidence: a review. Am J Psychol. 2009;122:17–38. [PubMed] [Google Scholar]
- 42.Domhoff B. Dreams are embodied simulations that dramatize conceptions and concerns: The continuity hypothesis in empirical, theoretical, and historical context. Int J Dream Res. 2011;4:50–62. [Google Scholar]
- 43.Pesant N, Zadra A. Dream content and psychological well-being: A longitudinal study of the continuity hypothesis. J Clin Psychol. 2006;62:111–21. doi: 10.1002/jclp.20212. [DOI] [PubMed] [Google Scholar]
- 44.Schredl M. Characteristics and contents of dreams. Int Rev Neurobiol. 2010;92:135–54. doi: 10.1016/S0074-7742(10)92007-2. [DOI] [PubMed] [Google Scholar]
- 45.Soffer-Dudek N, Shahar G. Daily stress interacts with trait dissociation to predict sleep- related experiences in young adults. J Abnorm Psychol. 2011;120:719–29. doi: 10.1037/a0022941. [DOI] [PubMed] [Google Scholar]
- 46.Blagrove M, Fisher S. Trait-state interactions in the etiology of nightmares. Dreaming. 2009;19:65–74. [Google Scholar]
- 47.Schredl M. Effects of state and trait factors on nightmare frequency. Eur Arch Psychiatry Clin Neurosci. 2003;253:241–7. doi: 10.1007/s00406-003-0438-1. [DOI] [PubMed] [Google Scholar]
- 48.Cason H. The nightmare dream. Psychol Monogr. 1935;46:1–51. [Google Scholar]
- 49.Feldman MJ, Hyman E. Content analysis of nightmare reports. Psychophysiol. 1968;5:221. [Google Scholar]
- 50.Cellucci AJ, Lawrence P. The efficacy of systematic desensitization in reducing nightmares. J Behav Ther Exp Psychiatry. 1978;9:109–14. [Google Scholar]
- 51.Kales JD, Kales A, Soldatos CR, Caldwell AB, Charney DS, Martin ED. Night terrors: Clinical characteristics and personality patterns. Arch Gen Psychiatry. 1980;37:1413–7. doi: 10.1001/archpsyc.1980.01780250099012. [DOI] [PubMed] [Google Scholar]
- 52.Cernovsky ZZ. Content of waking life events and of nightmare themes. Percept Mot Skills. 1984;58:899–902. doi: 10.2466/pms.1984.58.3.899. [DOI] [PubMed] [Google Scholar]
- 53.Hearne KM. A questionnaire and personality study of nightmare sufferers. J Ment Imagery. 1991;15:55–64. [Google Scholar]