Sleep Medicine Reviews 15 (2011) 115e121

Contents lists available at ScienceDirect

Sleep Medicine Reviews

journal homepage: www.elsevier.com/locate/smrv

CLINICAL REVIEW

Gender differences in nightmare frequency: A meta-analysis

Michael Schredl*, Iris Reinhard
Central Institute of Mental Health, Mannheim, Germany

a r t i c l e i n f o s u m m a r y

Article history: Many studies have reported gender differences in nightmare frequency. In order to study this difference
Received 18 March 2010 systematically, data from 111 independent studies have been included in the meta-analysis reported
Received in revised form here. Overall, estimated effect sizes regarding the gender difference in nightmare frequency differed
14 June 2010
significantly from zero in three age groups of healthy persons (adolescents, young adults, and middle-
Accepted 15 June 2010
Available online 3 September 2010
aged adults), whereas for children and older persons no substantial gender difference in nightmare
frequency could be demonstrated. There are several candidate variables like dream recall frequency,
depression, childhood trauma, and insomnia which might explain this gender difference because these
Keywords:
Gender differences
variables are related to nightmare frequency and show stable gender differences themselves. Systematic
Nightmare frequency research studying the effect of these variables on the gender difference in nightmare frequency, though,
Meta-analysis is still lacking. In the present study it was found that women tend to report nightmares more often than
men but this gender difference was not found in children and older persons. Starting with adolescence,
the gender difference narrowed with increasing age. In addition, studies with binary coded items showed
a markedly smaller effect size for the gender difference in nightmare frequency compared to the studies
using multiple categories in a rating scale. How nightmares were defined did not affect the gender
difference. In the analyses of all studies and also in the analysis for the children alone the data source
(children vs. parents) turned out to be the most influential variable on the gender difference (reporting,
age). Other results are also presented. Investigating factors explaining the gender difference in nightmare
frequency might be helpful in deepening the understanding regarding nightmare etiology and possibly
gender differences in other mental disorders like depression or posttraumatic stress disorder.
Ó 2010 Elsevier Ltd. All rights reserved.

Introduction question arises as to whether the gender difference in nightmare

frequency is age-dependent since nightmare frequency itself has
Nightmares are defined as disturbing mental experiences that a peak between 6 and 10 yrs of age15 and decreases with age.16
generally occur during rapid eye movement (REM) sleep and often In evaluating the consistencies or inconsistencies in nightmare
result in awakening (international classification of sleep disorders – prevalence research, two methodological issues have to be consid-
2nd version (ICSD-21)). Nightmare complaints are quite frequent in ered: nightmare definition of the particular study and nightmare
the general population: about 5% of the adults in representative measures. Three aspects of defining nightmares have been changing
samples stated they suffer from nightmares.2e5 In their extensive with time: the type of emotions within the nightmare, explicit
review of the nightmare literature, Levin and Nielsen6 concluded that awakening criterion, and the differentiation from night terrors.
there is a robust gender difference in nightmares with women Whereas earlier classification systems focused on fear as the
reporting nightmares more often than men. Their claim is supported predominant affect in nightmares (Diagnostic and Statistical Manual
by large-scale surveys in adolescents7,8 and young adults.9e11 On the of Mental Disorders– 4th version (DSM-IV), American Psychiatric
other hand, several studies with large sample sizes12e14 did not report Association17), several studies18e20 indicated that while fear is
gender differences of nightmare frequency in children and the data indeed present in most nightmares (about 80e85%) other emotions
basis regarding nightmare frequency in the elderly is scarce. So, the like anger, rage, embarrassment, disgust, and sadness are also
important. Wood and Bootzin,21 for example, used the nightmare
definition “a dream that frightens the dreamer” thereby deviating
from the current ICSD-2 definition. Levin10 explicitly asked for
* Corresponding author. Sleep laboratory, Central Institute of Mental Health, PO
a dream that awakens the sleeper whereas Nielsen et al.7 differen-
Box 12 21 20, 68072 Mannheim, Germany. Tel.: þ49 621 1703 1782; fax: þ49 621
1703 1785. tiated between nightmares (with awakening) and bad dreams
E-mail address: Michael.Schredl@zi-mannheim.de (M. Schredl). (without awakening) and Schredl et al. 22 elicited the bad dream

1087-0792/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.smrv.2010.06.002
116 M. Schredl, I. Reinhard / Sleep Medicine Reviews 15 (2011) 115e121

frequency only. Most studies, however, do not give a specific night- people, e.g., with high neuroticism scores, overestimate the distress
mare definition so it is left to the participants as to how they define of nightmares irrespective of their frequency has not been sup-
a nightmare. Problems with the estimation as to whether a night- ported26; neuroticism was correlated with nightmare frequency
mare has been the reason for waking up or not have been studied by but not with the intensity or amount of distress attributed to single
Blagrove and Haywood.23 Their results indicate that the persons’ nightmares. Whether the gender difference in nightmare distress
estimates are quite accurate. Another difference across studies is the parallels the gender difference in nightmare frequency has not
differentiation between nightmare frequency (“How often did you been studied.
experience nightmares over the last few months?”9); and nightmare The present meta-analysis was carried out in order to determine
suffering (“Do you suffer from frequent nightmares?”24). The concept whether there is a substantial gender difference in nightmare
of nightmare distress was elaborated by Bilicki25 including aspects frequency if all published studies are aggregated numerically.
like seeking help due to the nightmares, nightmares affecting well- Second, the effect of age on the gender difference in nightmare
being, etc. Her findings that nightmare distress is more closely frequency was studied. Lastly, whether different measurement
associated with personality traits like neuroticism than nightmare methods might explain differences in the findings was tested.
frequency itself were not confirmed by the study of Schredl et al.26
The main critique was that Belicki’s scale is confounded with Method
nightmare frequency (using five-point frequency scales for eliciting
nightmare distress). If distress of single nightmares was measured Literature search
independently from nightmare frequency nightmare frequency was
more strongly correlated to neuroticism than this distress variable.26 A literature review of the PubMed and PsycInf databases was
For a conceptual viewpoint, nightmare distress might function as carried out on September 12, 2008. For PubMed, the Medical
variable mediating the relationship between nightmare frequency Subject Heading “Dreams” was used because the more specific term
and psychopathology.27 It seems plausible that the variance of “Nightmares” is not available. The number of hits was 4931.
nightmare definitions applied in the different studies (or lack of The search term “nightmares” yielded 575 hits in the PsycInf
definition) has an effect on nightmare prevalence or frequency and database. Reference lists of the papers were also checked for
thus might also affect the gender difference in nightmare frequency. appropriate studies. Furthermore, four unpublished studies of the
Lastly, very few studies21 include an explicit description of the authors were included.
difference between nightmare and night terrors with night terrors Overall, over 6000 publications were inspected and included if
being an non-rapid eye movement (NREM) parasomnia with nightmare frequency was reported for males and females sepa-
massive panic but with difficulties in remembering vivid dream rately and the data were sufficient for computing effect sizes.
imagery.1 In a recent large-scaled study, Schredl et al.,28 using self Additional information was obtained from Zadra and Donderi,29
ratings and parental questionnaires, found that a possible confusion Nielsen et al., 11 Roberts and Lennings,36 Robert and Zadra,37 and
of these two parasomnias did not strongly affect the prevalence rates Simard et al.13 The data of our own studies were reanalyzed to
of nightmares (bias less than 10% of the prevalence rate). obtain effect sizes for gender differences in nightmare frequency if
In addition to the varying nightmare definitions, the prevalence they were not presented in the original publication.
studies used different methods of assessing nightmare frequency: In order to maximize overall sample size of the meta-analyses,
binary items (“Do you suffer from nightmares?”), rating scales three different sets of studies were analyzed. Overall, 111 inde-
(“How often do you remember nightmares?”) with different pendent samples reporting gender differences in dream recall have
formats (ordinal, interval, open-end), and dream diaries (for an been identified. Of these studies, 98 included healthy persons and
overview see6). It is also plausible that these different measure- 13 included different patient groups. The patient samples included
ment techniques might have an effect on nightmare prevalence different groups: patients with mental disorders (N ¼ 6 studies:
measures. For diary-based studies,29 much higher prevalence rates psychiatric inpatients with various diagnoses,38 outpatients with
were found in comparison to retrospective measures. It is still not various diagnoses,39 patients with neurotic illness not further
clear whether this can be interpreted as an underestimation bias specified,40 patients with panic disorder,41 adult patients with
concerning the retrospective measure or an increase of nightmare Attention deficit/Hyperactivity Disorder,42 and children with
frequency by focusing on nightmares when keeping a dream diary Attention deficit/Hyperactivity Disorder43), patients with sleep-
or a nightmare log. The increase of dream recall by keeping a diary related breathing disorders (N ¼ 3), patients with sleep disorders
is well documented,30 especially in low dream recallers,31 a finding not further specified (N ¼ 2), asthma patients (N ¼ 1), and insomnia
that might explain the increase in nightmare frequency being larger patients (N ¼ 1).
than the increase of dream recall by keeping a nightmare log Several studies included data for different age groups, so that
(because nightmare frequency is lower at the beginning of the a separate analysis regarding age group could be performed with
study period compared to overall dream recall). A methodological a population of 118 studies (only including non-patient samples) by
issue related to eliciting nightmare frequency in children is the including the effect sizes of each age group. Again, in the overall
question as to whether the children were asked (self-report) or the analysis with independent samples, only the effect size of the total
parents provided the data (parental data). Several studies indicate population was included.
that prevalence rates are considerably lower if parental data were
obtained in comparison to the children’s self-reports.28,32e34 The Coding
interpretation of these findings is still not clear. For example, it
seems possible that children do not tell their parents all their The following variables were extracted from the studies:
nightmares and thus the parental data are an underestimation of Publication year, total sample size, number of males, number of
the child’s nightmare frequency.28 Systematic research concerning females, mean age of the total sample, measurement method
how these methodological issues affect the gender difference in (questionnaire scale, dream diary), and group (healthy persons,
nightmare frequency is completely lacking. patients). For the questionnaire scales only, it was determined
Several studies25,26,35 focused on nightmare distress, a construct whether the scale included absolute categories like “several times
which can be differentiated from nightmare frequency even though a week”, “once a week”, etc., or relative categories like “never”,
the intercorrelations are high r ¼ 0.42.35 The hypothesis that some “rarely”, “sometimes” and “often”. In addition, it was coded
 M. Schredl, I. Reinhard / Sleep Medicine Reviews 15 (2011) 115e121 117

whether the effect size of the study is based on dichotomous In Table 1, the estimated effect sizes for the five age groups
questions (for example, having nightmares once a week or more (healthy persons only) are depicted. The total number of studies is
often3) or several categories included in a rating scale (e.g.,9). Most 118 because for several studies two or more effect sizes specifically
of the studies did not give a clear definition of what nightmares are computed for the age groups have been included. The results
(disturbing REM dreams usually resulting in awakening). In order (independent analyses for each age group) indicate that substantial
to study this methodological issue, studies were coded according to gender differences were present in three of the five age groups,
their definitions of nightmares: disturbing dreams with while the effect sizes for children and elder adults were not
a wakening, bad dreams (without awakening), and no definition at significantly different from zero.
all. Several studies did not differentiate between the frequency of The estimated effect size for the distress scale is depicted in
nightmares and the distress experienced by nightmares. In the Table 2 and is comparable with the figure of nightmare frequency in
study of Bixler et al.,2 for example, the participants were asked if the young adult group. Five of the seven studies included students
they had a problem with frightening dreams. According to this with the study’s mean ages ranging from 20 to 25. The mean ages of
classification, the studies were coded as to whether they used the two other studies were 10 and 14 yrs. For the patient groups
simple nightmare frequency measures or a frequency measure (weighted mean age: 51.5 yrs), estimated effect size was signifi-
implying some evaluation of the nightmare distress (but not cantly different from zero and quite similar to the figure for the
a formal measure of nightmare distress; those studies were middle-aged adults.
analyzed separately, see below). A mixed model was computed to test the common effects of age
Especially in the children sample, parents were asked about the group, nightmare definitions, measurement method, and publica-
nightmare frequency of their children, so whether the effect size tion year (see Table 3). The strongest effect was found for the data
was based on self reports or on parental estimates was coded. source variable. The effect size was d ¼ 0.020 for parental data,
According to the mean ages of the samples, studies were grou- and 0.163 for self-report data (controlled for all other variables). To
ped into five age groups: children (10 yrs), adolescents further investigate the effect of this variable on the effect size in the
(10 < x < 18 yrs), young adults (18  x < 30 yrs), middle-aged adults children group, a separate analysis with the data source variable
(30  x < 60 yrs), older adults (60 yrs). and the sample’s mean ages (see Table 4) was carried out. The data
Seven studies were also included because they measured source variable was again significant (p ¼ 0.0035), the sample’s
nightmare distress using mostly the Belicki25 scale. mean age showed a statistical trend (p ¼ 0.0584) with a negative
coefficient. i.e., the effect size of the gender difference in nightmare
Meta-analytical procedure frequency decreased with age. Testing the studies separately, the
mean effect size of the studies with parental data did not differ
Study-level effect sizes. The effect sizes of studies applying significantly from zero whereas the mean effect size based on self-
questionnaire scales and dream diaries were computed or recon- report data was significantly larger than zero (see Table 5). The
structed by the author using the formula given by Rustenbach44 for weighted mean age of the studies with parental data was 7.3 yrs,
Hedges’s g (see also45). For gender differences based on differences whereas the weighted mean age of the self-report studies
in percentages, Rustenbach44 provided a formula for computing the amounted to 9.3 yrs.
effect sizes of this difference divided by the pooled variance, so this The second measurement variable with a significant effect on
value is comparable to Hedges’s g. the gender difference in nightmare frequency was the distinction
Aggregate effect sizes. First, the Q-statistics of Cochran (e.g.,46) between binary measures and rating scales. The studies with binary
were computed to test for homogeneity of the different sets of measures (Yes/No) showed a significantly smaller effect size than
effect sizes. Since most of the tests yielded significant results and the studies using more sophisticated rating scales. Other
thus indicated a substantial inter-study variance, the random- measurement variables like questionnaire vs. diary, absolute vs.
effects approach was selected to test whether the effect sizes differ relative categories, nightmare definition, scale definition
from zero. A mixed model was computed to determine common (frequency vs. frequency and distress mixed), and publication year
effects of different factors (fixed effects) and check for inter-study did not affect the gender difference in nightmare frequency.
variation (random factors). For these models, the Satterthwaite The age group variable yielded a marginally significant effect.
approximation was applied to compute degrees of freedom. The Post hoc analysis revealed that young adults differed significantly
analyses were carried out with the SAS 9.2 statistical analysis from the middle-aged adults (F ¼ 7.55, df ¼ 1, 67.2, p ¼ 0.0077).
software (SAS Institute Inc., Cary, NC, USA). All other contrasts among the age groups were not significant.

Results Discussion

The meta-analysis is based on 111 independent samples, Overall, the meta-analysis indicates that there is a small but
including 98 samples consisting of healthy persons and 13 patient substantial gender difference in nightmare frequency; i.e., women
samples 6 psychiatric patient samples, 6 samples with sleep- tend to report nightmares more often than men. This gender
disordered patients, and one sample with somatic patients). difference, however, was not found in children and older persons.

Table 1
Effect sizes for different age groups (healthy persons).

Age group Studies Sample size Females/males Estimated effect size Confidence interval (95%) Q
Children (10 yrs) 29 42119 20829/21290 0.031 0.008e0.070 74.84***
Adolescents (10 < x < 18 yrs) 20 35333 19269/16064 0.219 0.155e0.282 106.87***
Young adults (18  x < 30 yrs) 42 40162 24221/15941 0.263 0.210e0.317 139.89***
Middle-aged adults (30  x < 60 yrs) 19 61174 33852/27322 0.147 0.099e0.194 63.11***
Older adults (60 yrs) 8 8351 5095/3256 0.095 0.066e0.257 22.92**

Q: Q-statistics: **p  0.01; ***p  0.001.

118 M. Schredl, I. Reinhard / Sleep Medicine Reviews 15 (2011) 115e121

Table 2
Effect sizes for nightmare distress measures and the patient groups.

Method Studies Sample size Females/males Estimated effect size Confidence interval (95%) Q
Nightmare distress (normal samples) 7 1956 1210/746 0.318 0.092e0.544 22.06**
Patients 13 10477 3895/6582 0.178 0.090e0.267 32.19**

Q: Q-statistics: **p  0.01.

The narrowing of the gender difference with increasing age high nightmare frequencies. Other variables which affect the
might be attributed to the decline of nightmare frequency with nightmare frequency per se like using a questionnaire scale vs.
age16,47 since for other possible risk factors like depression48 or a dream diary or using different definitions (bad dreams vs.
insomnia49 the gap regarding prevalence did not change with nightmare with awaking) did not have an effect on the gender
age. In the youngest age group, the gender difference was still difference in nightmare frequency. This underlines the robust-
significant when aggregating the studies applying self-report ness of the gender difference.
measures. The sample’s mean ages ranged from 7.5 yrs to 10 yrs The result that nightmare definition did not affect the gender
(see Table 4). For still younger children, only studies with difference supports the notion of Levin and Nielsen6 that there is
parental estimates were available and showed no gender differ- continuum from bad dreams to nightmares and etiological factors
ence. The negative correlation of the gender difference with age are comparable. The data source (children vs. parents) is the most
makes it plausible that methodological issues might play an influential variable on the gender difference (reporting, age) in the
important role (see Discussion below) because one would expect analyses of all studies but also in the analysis for the children alone
a steady increase of the gender difference within this age range (see Tables 4 and 5). Muris et al.33 reported that differences
or even a zero correlation, i.e., the gender difference might be between children’s self-reports and parental estimates were quite
present even in the youngest children. This question could not small for younger children (4e6 yrs) but considerably larger for
yet be answered due to the validity issues that arise with using older children (7e12 yrs). This supports the hypothesis of Schredl
parents as the data source. Interestingly, the gender difference in et al.28 that reporting the dreams to their parents might explain this
nightmare frequency seems to occur prior to the gender differ- discrepancy because older children do not seek direct help for
ence in dream recall frequency.50 This might be of interest in nightmares from their parents whereas younger children often go
explaining, for example, the gender difference in adolescent to the bedroom of their parents in the night. This would also
depression, nightmares e at least frequent nightmares e might explain the different effect sizes between the self-report studies
be precursors or risk factors for developing a mental disorder like and the studies including parental estimates because girls talk
depression or anxiety disorder. about their bad dreams/nightmares more often.51 Given this
Two methodological variables had a significant effect on methodological problem, a conclusive statement about the gender
gender difference. First, studies with binary coded items showed
a markedly smaller effect size for the gender difference in
nightmare frequency compared to the studies using multiple Table 4
categories in a rating scale. A plausible explanation of this Effect sizes of the gender difference in nightmare frequency in children.(10 yrs)
difference is provided by two studies7,51 showing the largest Authors Data Mean Girls Boys Total Effect
gender difference for the “sometimes” category and not for source age size
frequent nightmares (d ¼ 0.26 vs. d ¼ 0.01, 13 yrs-olds,7). The Beltramini and Hertzig63 P 3 53 54 107 0.075
studies using binary coded items aimed mainly at persons with Shepherd et al.64 P 5 209 251 460 0.066
Schredl et al. P 5.5 674 711 1385 0.063
(unpublished data)
Smedje et al.65 P 5.8 893 951 1844 0.114
Table 3
Foster and Anderson66 P 6 142 154 296 0.059
Mixed model for testing effects of covariates (normal samples).
Döpfner et al., 1996 in P 6 1620 1644 3264 0.025
Factor F-value Degrees of p-value Kraenz et al.12
freedom Haffner et al.14 P 6 2044 2319 4363 0.020
Kraenz et al.12 P 6 3135 3329 6464 0.003
Age group 4.50 (4, 64.7) 0.0511
Achenbach and Edelbrock67 P 6.5 300 300 600 0.051
Children (27), adolescents (14),
Fisher et al.68 P 7.4 112 149 261 0.000
young adults (37), middle-aged
O’Connor et al.69 P 7.6 113 103 216 0.083
adults (17), elder adults (3)
MacFarlane et al.70 P 8 114 122 236 0.126
Measurement method 0.23 (1, 98) 0.6351
Paavonen et al.71 P 8.5 2867 2964 5831 0.084
Questionnaire (94), dream diary (4)
Massetani et al.72 P 8.7 338 305 643 0.083
Answering categories 2.28 (1, 98) 0.1354
Schredl et al.73 P 8.8 152 148 300 0.238
Absolute categories (60),
Liu et al.74 P 8.87 1006 998 2004 0.100
relative categories (38)
Vela-Bueno et al.75 P 9 232 255 487 0.121
Measurement type 5.11 (1, 85.8) 0.0263
Kimmins76 P 9 2000 2000 4000 0.079
Binary measures (39), rating scales (59)
Lapouse and Monk77 P 9 237 245 482 0.159
Nightmare definition 0.31 (2, 98) 0.7309
Agargün et al.78 P 9.1 499 472 971 0.040
Awaking criterion (10), bad dreams
Cluydts and De Roeck79 P 9.5 655 548 1203 0.036
(7), no explicit definition (81)
Scale definition 0.00 (1, 56) 0.9645 Muris et al.80 S 7.5 85 91 176 0.094
Nightmare frequency (90), Muris et al.33 S 7.9 98 92 190 0.058
distress and frequency (8) Schredl et al.28 S 9.5 2346 2161 4507 0.089
Data source 14.21 (1, 64.9) 0.0004 Eitner et al.81 S 9.6 546 568 1114 0.217
Self report (76), parental estimates (22) Keßels82 S 9.7 50 53 103 0.483
Publication year 0.01 (1, 85.3) 0.9076 Schredl and Sartorius83 S 9.9 51 49 100 0.383
Abdel-Khalek84 S 10 239 240 479 0.027
(N ¼ 98 independent study effect sizes; Ntotal ¼ 187185, Nfemales ¼ 103305,
Schredl and Pallmer51 S 10 19 14 33 0.197
Nmales ¼ 83880).
 M. Schredl, I. Reinhard / Sleep Medicine Reviews 15 (2011) 115e121 119

Table 5
Effect sizes for nightmare frequency in the children group (10 yrs).

Method Studies Sample size Females/males Estimated effect size Confidence interval (95%) Q
Parental estimates 21 35417 17395/18022 0.006 0.032e0.044 42.79***
Self-report 8 6702 3434/3268 0.121 0.037e0.204 0.15

Q: Q-statistics: ***p  0.001.

difference in nightmare frequency in children younger than 7 yrs gender difference was evident but not for the parental estimates.
cannot yet be made. Future studies should elicit nightmare Schredl et al.28 argued that the discrepancy between these two
frequency not only via parental questionnaires but also include measures might be explained by the fact that girls report their
self-report measures. Using a diary kept by the parents who ask dreams more often than boys51 e this has also been found in
their children each morning about their dreams and nightmares adults.52 The decrease of the gender difference with old age provides
might be a valuable option for measuring nightmare frequency an argument against a simple self-report bias regarding nightmare
without report and/or recall biases. frequency. In addition, self-report biases might be minimized by
The gender difference regarding the distress experience is using dream diaries and the present analyses showed that there was
comparable with the gender difference in nightmare frequency for no effect on the gender difference in nightmare frequency. Whether
this age group (young adults). This finding indicates that a bias in self-report biases really affect nightmare frequency in general or
overestimating the distressing effect of nightmares is very unlikely. only the gender difference in nightmare frequency are questions not
The effect size of the patient group is also comparable to the effect yet addressed in a systematic way.
size in the normal group of the same age range, indicating that The gender differences in the prevalence of childhood sexual
differences in sleep physiology are unlikely able to explain the abuse,53 depression,48 posttraumatic stress disorder,54 and
gender difference in nightmare frequency. i.e., many studies indi- insomnia49 are well documented. Childhood sexual abuse is
cate that women suffer from insomnia more often than men49 and, related to heightened nightmare frequency in adult wom-
thus, higher nightmare frequency in women might be due to their en.55e57 Patients with depressive disorders experience more
lower sleep quality. Sleep quality, however, is markedly reduced in negatively-toned dreams,58,59 and insomnia is related to
patients with mental disorders (male and female patients alike) heightened nightmare frequency.60 In posttraumatic stress
and, thus, the comparable effect size regarding the gender differ- disorder, the occurrence of nightmares is included in the diag-
ence in nightmare frequency between healthy controls and patients nostic criteria and 50e70% of patients with PTSD report night-
suggests that the lower sleep quality in women does not explain the mares.61 Studies investigating these possible confounders on the
higher nightmare frequency. gender difference in nightmare frequency have not yet been
Given the substantial and robust finding of gender differences in carried out. The hypothesis that psychosocial risk factors for
nightmare frequency, the question arises as to what factors might developing a depression like low instrumentality, ruminative
explain this difference. Possible factors have to meet two criteria; style of coping or emotion-focused coping is related to the
first, they should correlate substantially with nightmare frequency occurrence of nightmares have never been tested empirically.
and, second, they should show a stable gender difference. Levin and The fifth hypothesis put forward by Levin and Nielsen6 e
Nielsen6 cited five possible factors that might explain the gender possible gender differences in emotional brain processes e is of
differences in nightmare frequency: (1) self-report biases, i.e., special interest for dream and nightmare research because
women are more likely to report distressing experiences; (2) risk imaging studies have shown that the limbic system is very
factors like childhood sexual abuse, depression, and insomnia; (3) active during REM sleep.62 Although the proposed factors are
processes that explain the higher depression prevalence in women; plausible, systematic research studying the effect of these vari-
(4) coping styles; and (5) biological differences in emotional brain ables on the gender difference in nightmare frequency still
processes. needs to be performed.
The most obvious variable which might affect the gender To summarize, the meta-analysis including a considerable
difference in nightmare frequency is the frequency of how often number of studies was able to demonstrate a small but substantial
dreams are recalled overall, i.e., the hypothesis is that women gender difference in nightmare frequency in adolescents and
report more nightmares because they recall their dreams more young and middle-aged adults but not in children or older
often. A robust gender difference in dream recall frequency has persons. Whereas several methodological factors like nightmare
been demonstrated by the meta-analysis of Schredl and Reinhard50 definition or type of scale did not have an effect on the gender
and dream recall is closely related to nightmare frequency difference, the results from studies using parental estimates
(r ¼ 0.335, N ¼ 444,9). However, Nielsen et al.7 reported that e differed from those using self-reports of nightmare frequency in
when dividing the total sample into infrequent and frequent dream children. There are several candidate variables which might
recallers e nightmare frequency was still more prevalent in girls explain this gender difference like dream recall frequency,
compared to boys in both samples. Dividing the sample into two depression, childhood trauma, and insomnia because these vari-
groups is a crude approach, it would have been more appropriate to ables are related to nightmare frequency and themselves show
use a regression technique in order to compare the effect of the stable gender differences. Systematic research studying the effect
biological sex variable and dream recall frequency on the gender of variables that might explain the gender difference in nightmare
difference in nightmare frequency. So, the effect of the gender frequency like depression, childhood trauma, dream recall
difference in dream recall frequency on the gender difference in frequency, posttraumatic stress disorder is still lacking. Investi-
nightmare frequency warrants further investigation. gating factors explaining the gender difference in nightmare
A possible effect of the first factor listed by Levin and Nielsen,6 frequency might be helpful in deepening the understanding
self-report biases, might be supported by the significant effect of regarding nightmare etiology and possibly gender differences in
the data source (parental vs. children’s self-reports) on the gender other mental disorders like depression or posttraumatic stress
difference in nightmare frequency: for the self-report measures, the disorder.
120 M. Schredl, I. Reinhard / Sleep Medicine Reviews 15 (2011) 115e121

*7. Nielsen TA, Leberge L, Paquet J, Tremblay RE, Vitaro F, Montplaisir J. Devel-
opment of disturbing dreams during adolescence and their relation to anxiety
Practice points symptoms. Sleep 2000;23:727e36.
8. Choquet M, Tesson F, Stevenot A, Prevost E, Antheaume M. Les adolescents et
1. Women tend to report nightmares more often than men, leur sommeil: approche epidemiologique. [The adolescents and their sleep: an
especially during adolescence and young adulthood. epidemiological approach]. Neuropsychiatrie de l’Enfance et de l’Adolescence
1988;36:399e410.
2. Methodological issues like nightmare definition,
*9. Schredl M. Effects of state and trait factors on nightmare frequency. European
measurement type (questionnaire vs. diary), and Archives of Psychiatry and Clinical Neuroscience 2003;253:241e7.
publication year did not affect the gender difference in 10. Levin R. Sleep and dream characteristics of frequent nightmare subjects in
nightmare frequency. a universitary population. Dreaming 1994;4:127e37.
3. Nightmare distress was also higher in women 11. Nielsen TA, Stenstrom P, Levin R. Nightmare frequency as a function of age,
compared to men. gender, and September 11, 2001: findings from an internet questionnaire.
Dreaming 2006;16:145e58.
4. Nightmare frequencies of children obtained by asking 12. Kraenz S, Fricke L, Wiater A, Mitschke A, Breuer U, Lehmkuhl G. Häufigkeit und
the parents about their children’s dreams and night- Belastungsfaktoren bei Schlafstörungen im Einschulalter. Praxis der Kinder-
mares might be underestimated. psychologie und Kinderpsychiatrie 2004;53:3e18.
5. Retrospective measures (questionnaire scales) might *13. Simard V, Nielsen TA, Tremblay RE, Boivin M, Montplaisir JY. Longitudinal
yield lower figures for nightmare frequency compared study of bad dreams in preschool-aged children: prevalence, demographic
correlates, risk and protective factors. Sleep 2008;31:62e70. S1eS4.
to dream diaries or nightmare logs. 14. Haffner J, Esther C, Münch H, Parzer P, Raue B, Steen R, et al. Verhaltensauf-
fälligkeiten im Einschulungsalter aus elterlicher Perspektive e Ergebnisse zu
Prävalenz und Risikofaktoren in einer epidemiologischen Studie. Praxis der
Kinderpsychologie und Kinderpsychiatrie 2002;51:675e96.
15. Sandoval D, Krakow B, Schrader R, Tandberg G. Adult nightmare sufferers: can
they be identified and treated in childhood? Sleep Research 1997;26:256.
Research agenda 16. Salvio M-A, Wood JM, Schwartz J, Eichling PS. Nightmare prevalence in the
healthy elderly. Psychology and Aging 1992;7:324e5.
17. American Psychiatric Association. Diagnostisches und Statistisches Manual
1. Sophisticated studies are needed to obtain valid
Psychischer Störungen (DSM IV). Göttingen: Beltz; 1996.
measures of nightmare frequency in childhood. As 18. Dunn KK, Barrett D. Characteristics of nightmare subjects and their night-
retrospective questionnaires completed by the parents mares. Psychiatric Journal of the University of Ottawa 1988;13:91e3.
might not be valid, longitudinal studies using logs are 19. Rose MW, Perlis ML, Kaszniak AW. Self-reported dream emotion: nightmares
recommended. and vivid dreams. Sleep Research 1992;21:132.
*20. Zadra A, Pilon M, Donderi DC. Variety and intensity of emotions in nightmares
2. Systematic research studying the effect of variables that
and bad dreams. Journal of Nervous and Mental Disease 2006;194:249e54.
might explain the gender difference in nightmare *21. Wood JM, Bootzin RR. The prevalence of nightmares and their independence
frequency like depression, childhood trauma, dream from anxiety. Journal of Abnormal Psychology 1990;99:64e8.
recall frequency, posttraumatic stress disorder is needed. 22. Schredl M, Pallmer R, Montasser A. Anxiety dreams in school-aged children.
3. Investigating factors explaining the gender difference in Dreaming 1996;6:265e70.
nightmare frequency might be helpful in deepening the *23. Blagrove M, Haywood S. Evaluating the awakening criterion in the definition
of nightmares: how certain are people in judging whether a nightmare woke
understanding of possible gender differences in other them up? Journal of Sleep Research 2006;15:117e24.
mental disorders like depression or posttraumatic 24. Cirignotta F, Zucconi M, Mondini S, Lenzi PL, Lugaresi E. Enuresis, sleepwalking
stress disorder. and nightmares: an epidemiological survey in the Republic of San Marino. In:
Guilleminault C, Lugaresi E, editors. Sleepewake disorders: natural history,
epidemiology and long-term evolution. New York: Raven Press; 1983. p. 237e41.
25. Belicki K. Nightmare frequency versus nightmare distress: relation to psycho-
Disclosure statement
pathology and cognitive style. Journal of Abnormal Psychology 1992;101:592e7.
26. Schredl M, Landgraf C, Zeiler O. Nightmare frequency, nightmare distress and
This was not an industry-supported study. Dr. Schredl has neuroticism. North American Journal of Psychology 2003;5:345e50.
received research support from INC Research for a phase III insomnia 27. Levin R, Lantz E, Fireman G, Spendlove S. The relationship between disturbed
dreaming and somatic distress: a prospective investigation. Journal of Nervous
study. Iris Reinhard has indicated no financial conflicts of interest. and Mental Disease 2009;197:606e12.
*28. Schredl M, Fricke-Oerkermann L, Mitschke A, Wiater A, Lehmkuhl G. Factors
Appendix. Supplementary data affecting nightmares in children: parents’ vs. children’s ratings. Eur Child
Adolesc Psychiatry 2009;18:20e5.
*29. Zadra A, Donderi DC. Nightmares and bad dreams: their prevalence and
Supplementary data associated with this article can be found, in relationship to well-being. Journal of Abnormal Psychology 2000;109:273e81.
the online version, at doi:10.1016/j.smrv.2010.06.002. 30. Schredl M. Dream recall: models and empirical data. In: Barrett D,
McNamara P, editors. The new science of dreaming. Content, recall, and
personality correlates, vol. 2. Westport: Praeger; 2007. p. 79e114.
References 31. Schredl M. Questionnaire and diaries as research instruments in dream
research: methodological issues. Dreaming 2002;12:17e26.
1. American Academy of Sleep Medicine. The international classification of sleep 32. Lapouse R, Monk MA. An epidmiologic study of behavior characteristics in
disorders (ICSD-2). Westchester: AASM; 2005. children. American Journal of Public Health 1958;48:1134e44.
2. Bixler EO, Kales A, Soldatos CR, Kales JD, Healey S. Prevalence of sleep disor- 33. Muris P, Merckelbach H, Gadet B, Moulaert V. Fears, worries, and scary dreams
ders in the Los Angeles metropolitan area. American Journal of Psychiatry in 4- to 12-year-old children: their content, developmental pattern, and
1979;136:1257e62. origins. Journal of Clinical Child Psychology 2000;29:43e52.
3. Janson C, Gislason T, De Backer W, Plaschke P, Björnsson E, Hetta J, et al. 34. Owens JA, Spirito A, McGuinn M, Nobile C. Sleep habits and sleep disturbance
Prevalence of sleep disturbances among young adults in three European in elementary school-aged children. Developmental and Behavioral Pediatrics
countries. Sleep 1995;18:589e97. 2000;21:27e36.
4. Ohayon MM, Morselli PL, Guilleminault C. Prevalence of nightmares and their 35. Levin R, Fireman G. Nightmare prevalence, nightmare distress, and self-
relationship to psychopathology and daytime functioning in insomnia reported psychological disturbance. Sleep 2002;25:205e12.
subjects. Sleep 1997;20:340e8. 36. Roberts J, Lennings CJ. Personality, psychopathology and nightmares in young
5. Stepansky R, Holzinger B, Schmeiser-Rieder A, Saletu B, Kunze M, Zeitlhofer J. people. Personality and Individual Differences 2006;41:733e44.
Austrian dream behavior: results of a representative population survey. *37. Robert G, Zadra A. Measuring nightmare and bad dream frequency: impact of ret-
Dreaming 1998;8:23e30. rosepective and prospective instruments. Journal of Sleep Research 2008;17:132e9.
*6. Levin R, Nielsen TA. Disturbed dreaming, posttraumatic stress disorder, and 38. Hersen M. Personality characteristics of nightmare sufferers. Journal of
affect distress: a reviewe and neurocognitive model. Psychological Bulletin Nervous and Mental Disease 1971;153:27e31.
2007;133:482e528. 39. Bernert RA, Joiner TE, Cukrowicz KC, Schmidt NB, Krakow B. Suicidality and
sleep disorders. Sleep 2005;28:1135e41.
40. Tapia F, Werboff J, Winokur G. Recall of some phenomena of sleep:
a comparative study of dreams, somnambulism and enuresis in a control and
* The most important references are denoted by an asterisk. neurotic population. Journal of Nervous and Mental Disease 1958;127:119e23.
 M. Schredl, I. Reinhard / Sleep Medicine Reviews 15 (2011) 115e121 121

41. Schredl M, Kronenberg G, Nonell P, Heuser I. Dream recall, nightmare 64. Shepherd M, Oppenheim B, Mitchell S. Auffälliges Verhalten bei Kindern. Göt-
frequency, and nocturnal panic attacks in patients with panic disorder: their tingen: Vandenhoeck & Ruprecht; 1973.
relationship to nocturnal panic attacks. Journal of Nervous and Mental Disease 65. Smedje H, Broman JE, Hetta J. Parents’ reports of disturbed sleep in 5e7-year-
2001;189:559e62. old Swedish children. Acta Paediatrica 1999;88:858e65.
42. Schredl M, Alm B, Sobanski E. Sleep quality in adult patients with attention 66. Foster JC, Anderson JE. Unpleasent dreams in childhood. Child Development
deficit hyperactivity disorder (ADHD). European Archives of Psychiatry and 1936;7:77e84.
Clinical Neuroscience 2007;257:164e8. 67. Achenbach TM, Edelbrock CS. Behavioral problems and competencies
43. Schredl M, Sartorius H. Dream recall and dream content in children with reported by parents of normal and disturbed children aged four through
attention deficit/hyperactivity disorder. Child Psychiatry and Human Develop- sixteen. Monographs of the Society for Research in Child Development
ment 2010;41:230e8. 1981;46:1e60.
44. Rustenbach SJ. Metaanalyse e Eine anwendungsorientierte Einführung. Bern: 68. Fisher BE, Pauley C, McGuire K. Children’s sleep behavior scale: normative
Hans Huber; 2003. data on 870 children in grades 1 to 6. Percept Mot Skills 1989;68:227e36.
45. Rosenthal R. Parametric measures of effect size. In: Cooper H, Hedges LV, 69. O’Connor M, Foch T, Sherry T, Plonin R. A twin study of specific behavioral
editors. The handbook of research synthesis. New York: Russell Sage Founda- problems of socialization as viewed by parents. Journal of Abnormal Child
tion; 1994. p. 231e44. Psychology 1980;8:189e99.
46. Hardy RJ, Thompson SG. Detecting and describing heterogeneity in meta- 70. MacFarlane JW, Allen L, Honzik MP. A developmental study of the behavior
analysis. Statistics in Medicine 1998;17:841e56. problems of normal children between twenty-one months and fourteen years.
47. Franke H. Über das physiologische und pathologische Schlaf- und Wachver- Berkeley: University of California Press; 1962.
halten von Betagten. Zeitschrift für Gerontologie 1979;12:187e94. 71. Paavonen EJ, Aronen ET, Moilanen I, Piha J, Räsänen E, Tamminen T, et al. Sleep
48. Seedat S, Scott KM, Angermeyer MC, Berglund P, Bromet EJ, Brugha TS, et al. problems of school-aged children: a complementary view. Acta Paediatrica
Cross-national associations between gender and mental disorders in the 2000;89:223e8.
World Health Organization World Mental Health Surveys. Archives of General 72. Massetani R, Lazzaro AT, Piccini P, Bianchi F, Muratorio A. I disturbi del sonno
Psychiatry 2009 Jul;66(7):785e95. in eta evolutiva e Indagine trasversale au 643 soggetti sani (Sleep diorders in
49. Zhang B, Wing YK. Sex differences in insomnia: a meta-analysis. Sleep 2006 growing children e cross-sectional study of 643 healthy subjects). Rivista de
Jan 1;29(1):85e93. Neuroligia 1985;55:149e60.
50. Schredl M, Reinhard I. Gender differences in dream recall: a meta-analysis. 73. Schredl M, Blomeyer D, Görlinger M. Nightmares in children: influencing
Journal of Sleep Research 2008;17:125e31. factors. Somnologie 2000;4:145e9.
51. Schredl M, Pallmer R. Geschlechtsunterschiede in Angstträumen von Schü- 74. Liu X, Sun Z, Uchiyama M, Shibni K, Kim K, Okawa M. Prevalence and corre-
lerInnen. Praxis der Kinderpsychologie und Kinderpsychiatrie 1998;47:463e76. lates of sleep problems in Chinese school children. Sleep 2000;23:1053e62.
52. Schredl M. Sharing dreams: sex and other sociodemographic variables. Percept 75. Vela-Bueno A, Bixler EO, Dobladez-Blanco B, Rubio ME, Mattison RE, Kales A.
Mot Skills 2009;109:235e8. Prevalence of night terrors and nightmares in elementary school children:
53. Walker JL, Carey PD, Mohr N, Stein DJ, Seedat S. Gender differences in the a pilot study. Research Communications in Psychology, Psychiatry and Behavior
prevalence of childhood sexual abuse and in the development of pediatric 1985;10:177e88.
PTSD. Archives of Women’s Mental Health 2004;7(2):111e21. 76. Kimmins CW. Children’s dreams (1931). In: Lee SGM, Mayer AR, editors.
54. Nemeroff CB, Bremner JD, Foa EB, Mayberg HS, North CS, Stein MB. Post- Dreams and dreaming. Middlesex, England: Penguin; 1973. p. 83e109.
traumatic stress disorder: a state-of-the-science review. Journal of Psychiatric 77. Lapouse R, Monk MA. Behavior deviations in a representative sample of
Research 2006 Feb;40(1):1e21. children: variation by sex, age, race, social class and family size. American
55. Cuddy MA, Belicki K. Nightmare frequency and related sleep disturbance as Journal of Orthopsychiatry 1964;34:436e46.
indicators of a history of sexual abuse. Dreaming 1992;2:15e22. 78. Agargun MY, Savas A, Sener S, Bilici M, Ozer OA, Selvi Y, et al. The prevalence
56. Draijer N. Die Rolle von sexuellem Mißbrauch und körperlicher Mißhandlung in of parasomnias in preadolescent school-aged children: a Turkish sample. Sleep
der Ätiologie psychischer Störungen bei Frauen. System Familie 1990;3:59e73. 2004;27:701e5.
57. Garfield PL. Nightmares in the sexually abused female teenager.. Psychiatric 79. Cluydts R, De Roeck J. Sleep complaints in school-aged children, in young healthy
Journal of the University of Ottawa 1987;12:93e97. volonteers and in the sleep disorders center. Stuttgart: Gustav Fischer Verlag;
58. Schredl M, Engelhardt H. Dreaming and psychopathology: dream recall and 1988.
dream content of psychiatric inpatients. Sleep and Hypnosis 2001;3:44e54. 80. Muris P, Merckelbach H, Ollendick TH, King NJ, Bogie N. Children’s nighttime
59. Schredl M, Riemann D, Berger M. The effect of trimipramine on dream recall fears: parent-child ratings of frequency, content, origins, coping behaviors and
and dream emotions in depressive outpatients. Psychiatry Research severity. Behaviour Research and Therapy 2001;39:13e28.
2009;167:279e86. 81. Eitner S, Urschitz MS, Guenther A, Urschitz-Duprat PM, Bohnhorst B,
60. Schredl M. Nightmare frequency in patients with primary insomnia. Interna- Schlaud M, et al. Sleep problems and daytime somnolence in a German
tional Journal of Dream Research 2009;2:85e8. population-based sample of snoring school-aged children. Journal of Sleep
61. Wittmann L, Schredl M, Kramer M. The role of dreaming in posttraumatic Research 2007;16:96e101.
stress disorder. Psychotherapy and Psychosomatics 2007;76:25e39. 82. Keßels T. Angstträume bei neun- bis elfjährigen Kindern. Praxis der Kinder-
62. Maquet P, Peters J-M, Aerts J, Delfiore G, Deguelde C, Luxen A, et al. Functional psychologie und Kinderpsychiatrie 2004;53:19e38.
neuroanatomy of human rapid-eye-movement sleep and dreaming. Nature 83. Schredl M, Sartorius H. Frequency of dream recall by children and their
1996;383:163e6. mothers. Percept Mot Skills 2006;103:657e8.
63. Beltramini AU, Hertzig ME. Sleep and bedtime behavior in pre-school-aged 84. Abdel-Khalek AM. Nightmares: prevalence, age and gender differences among
children. Pediatrics 1983;71:153e8. Kuwaiti children and adolescents. Sleep and Hypnosis 2006;8:33e40.