Gender Differences in Dream Recall
-
Upload
sava-bogdan -
Category
Documents
-
view
214 -
download
0
Transcript of Gender Differences in Dream Recall
-
7/28/2019 Gender Differences in Dream Recall
1/8
doi: 10.1111/j.1365-2869.2008.00626.x
Gender differences in dream recall: a meta-analysis
M I C H A E L S C H R E D L and I R I S R E I N H A R DCentral Institute of Mental Health, Mannheim, Germany
Accepted in revised form 23 October 2007; received 20 July 2007
S U M M A R Y Many studies have reported gender differences in dream recall. Data from 175
independent studies have been included in the analyses. Overall, estimated effect sizes in
five age groups of healthy persons differed significantly from zero. Variables like
measurement method and publication year did not affect the gender difference but age
groups showed different effect sizes. The smallest effect size was found for children
(0.097), the largest for adolescents (0.364), whereas the three adult groups ranged from
0.242 to 0.270. The findings suggest that the age-dependent gender differences in dream
recall might be explained by gender-specific dream socialization. Longitudinal studiesin this area, however, are still lacking.
keyword s dream recall, gender differences, meta-analysis
I N T R O D U C T I O N
Dreaming, defined as mental activity during sleep, is presumed
to occur every night in every person. This presumption is based
on the high recall rates following REM awakenings and even
NREM awakenings in the sleep laboratory (Nielsen, 2000).
Despite this universality, dream recall varies considerably
between individuals and within one person from night to night.Over the decades, many factors have been studied which might
explain this variation (Schredl, 2007). Personality dimensions,
like openness to experience or related measures (boundary
thinness and absorption), creativity and the frequency of
nocturnal awakenings showed small but significant correla-
tions with dream recall frequency (DRF; e.g. Schredl et al.,
2003b). Other factors like trait repression, extraversion, sleep
duration and visual memory have yielded mixed results (for an
extensive review, see Schredl and Montasser, 199697).
In several large-scale studies (Borbely, 1984; Giambra et al.,
1996; Pagel et al., 1995; Schredl, 2002b; Schredl and Piel, 2003;
Wynaendts-Francken, 1907) with sample sizes ranging from
265 to 2328 persons, gender was related to DRF. I.e., on
average, women tend to recall their dreams more often than
men, although effect sizes vary from 0.15 (Borbely, 1984) to
0.55 (Nielsen et al., 2000). A recent representative study
(Stepansky et al., 1998) including 1000 participants, however,
reported no gender difference in DRF. Two factors might
explain this discrepancy. First, Giambra et al. (1996) found
that gender differences were more pronounced in the 25- to 55-
year age range, whereas for young persons as well as for elderly
persons DRF does not differ substantially between men and
women. Stepansky et al. (1998) did not analyze the data forthe different age groups they studied.
Secondly, the studies applied different methods for measur-
ing DRF. The most often used measurement approaches are
questionnaire scales, dream diaries and awakenings in the sleep
laboratory. The correlation coefficients between different types
of questionnaire scales, scales with absolute categories and
scales with relative categories (r = 0.647; N = 444; Schredl,
2004a) and different measurement types (diary questionnaire
scale: r = 0.69; N = 338; Cohen, 1979) are quite high. DRF
measured by awakenings in the sleep laboratory is even related
to home dream recall: with high dream recallers (more than
three dreams per week), Goodenough et al. (1959) obtained a
recall rate of 93% for REM awakenings, whereas in only 46%
of the REM awakenings were low dream recallers (less than
once a month) able to report a dream. The retest reliability of
the questionnaire scales is satisfactory if absolute answering
categories have been applied (e.g. several times a week, once a
week, about twice a month, about once a month, less than
once a month; r = 0.85; N = 198; Schredl, 2004b). Scales
with relative categories like never, rarely, sometimes, often,
very often yielded lower retest reliability coefficients (r = 0.59;
N = 106; Bernstein and Belicki, 199596). A two-week dream
Correspondence: Michael Schredl, PhD, Sleep laboratory, Central
Institute of Mental Health, PO Box 12 21 20, 68072 Mannheim,
Germany. Tel.: +49 621 1703 1782; fax: +49 621 1703 1785; e-mail:
Please contact the author for a complete list of references regarding the
studies included in this meta-analysis.
J. Sleep Res. (2008) 17, 125131 Dreaming
2008 European Sleep Research Society 125
-
7/28/2019 Gender Differences in Dream Recall
2/8
diary showed sufficient reliability (internal consistency:
r = 0.818; N = 196; Schredl and Fulda, 2005) and retest
reliability: r = 0.67; N = 106 (Bernstein and Belicki, 1995
96). Stability coefficients for the REM awakening technique
have not been reported in the literature, possibly due to the
expense of this measurement technique (cf. Schredl, 2007).
Other measures for dream recall were applied less often: (1)
eliciting once whether or not a dream was recalled last night
(e.g. Botman and Crovitz, 198990); (2) determining whetheror not the person is able to report a most recent dream (e.g.
Winget et al., 1972); and (3) determining whether or not the
person remembers specific dream themes (e.g. Schredl and Piel,
2003).
The present meta-analysis was carried out to determine
whether there is a substantial gender difference in dream recall
if the studies are aggregated numerically. Secondly, whether
different measurement methods might explain differences in
the findings was tested. Lastly, the effect of age on the gender
difference in dream recall was studied.
M E T H O D
Literature search
A literature review of the PubMed, PsycInf and Psyndex
databases with the search terms dream recall was carried out
on 18 April 2006. The number of hits was 641. The broad
search term was applied because searching sex differences and
dream recall yielded only 36 studies supporting the impres-
sion that the key words of the papers rarely included the
finding regarding gender difference in dream recall. In addi-
tion, all the studies cited in the literature reviews of Gooden-
ough (1978), Belicki (1986), Schredl and Montasser (199697)
and Schredl (1999, 2007) were included. Reference lists of thepapers were also checked for appropriate studies. Further-
more, non-published studies of the author (N = 7 samples
with healthy persons, N = 8 patient samples) were included.
Overall, over 1000 publications were inspected and included
if dream recall was reported for men and women separately
and the data were sufficient for computing effect sizes.
Additional information was obtained from Colace (1998,
2006), Colace and Violani (1993), Colace et al. (1997), Strauch
and Meier (2004), Levin et al. (2003), Blagrove et al. (2003)
and Hartmann (1989). In addition, Art Funkhouser provided
an unpublished data set (N = 1019). The data of own studies
were reanalyzed if gender differences were not presented in the
original publication.
To maximize overall sample size of the meta-analyses, three
different sets of studies were analyzed. Overall, 175 indepen-
dent samples reporting gender differences in dream recall have
been identified. Of these studies, 147 studies included healthy
persons and 28 studies included different patient groups. The
patient samples were categorized into four groups: psychiatric
patients [N = 12 studies; different disorders not specified
(N = 3), depression (N = 3), psychotherapy patients
(N = 3), attention deficithyperactivity disorder (N = 2),
panic disorder (N = 1)], somatic patients [N = 5 studies;
asthma (N = 2), brain lesions (N = 1), patients who under-
went EEG diagnosis in a somatic clinic (N = 1), patients with
myasthenia gravis (N = 1)] and sleep-disordered patients
[N = 9 studies; sleep apnea syndrome (N = 3), insomnia
(N = 2), restless legs syndrome (N = 1), Pavor noctur-
nussleepwalking (N = 1), hypersomnia (N = 1), narcolepsy
(N = 1)]. The two studies including blind persons were
subsumed under patients because the sleepwake patternmight differ from seeing persons. The rationale of this
categorization is derived from findings indicating that depres-
sive mood is related to reduced dream recall (cf. Schredl, 2007),
whereas sleep disorders, especially insomnia and narcolepsy,
are associated with heightened DRF (Schredl, 1998; Schredl
et al., 1998a). This general shifts in levels of dream recall might
affect the gender difference in this variable. For somatic
patients and blind persons, differences in DRF in comparison
with healthy controls have not been studied systematically (cf.
Schredl, 2007).
For analyzing each measurement method separately, the
overall number of studies including healthy controls increased
to 169 because several studies applied more than one
measurement method. I.e., for each measurement method,
the analyses are still based on independent effect sizes.
If several measures were applied in one study, the measure
with the highest expenditure was selected: first REM awak-
enings, then the dream diary and lastly the questionnaire scale.
Dream diaries included checklists with eliciting only dream
recall and diaries with explicit recording of dream content if
recalled. If more than one questionnaire scale was included,
the most reliable method was chosen (absolute categories or
longer measurement interval). This selection procedure was
chosen to maximize sample sizes for the high expenditure
methods.Similarly, several studies included data for different age
groups, so that a separate analysis regarding age group could
be performed with a sample of 163 studies (healthy controls)
by including the effect sizes of each age group. Again, in the
overall analysis with independent samples only the effect size
of the total sample was included.
Coding
The following variables were extracted from the studies:
publication year, total sample size, number of men, number
of women, mean age of the total sample, measurement method
(questionnaire scale, dream diary, REM awakenings, able to
report most recent dream, dream recall of the previous night,
able to report general dream themes) and group (healthy
persons, psychiatric patients, sleep-disordered patients, so-
matic patients and blind persons). For the questionnaire scales
only, it was determined whether the scale included absolute
categories such as several times a week, once a week, etc. or
relative categories like never, rarely, sometimes and often.
In addition, it was coded whether the study was conducted by
the author to estimate whether including a large number of
126 M. Schredl and I. Reinhard
2008 European Sleep Research Society, J. Sleep Res., 17, 125131
-
7/28/2019 Gender Differences in Dream Recall
3/8
published and unpublished studies by the author might have
affected the overall results.
According to the mean ages of the samples, studies were
grouped into five age groups: children (10 years), Adolescents
(10 < x < 18 years), young adults (18 x < 30 years), mid-
dle-aged adults (30 x < 60 years) and elder adults
(60 years). In a few studies, mean values of age were missing,
but it was reported that the participants were students. As
almost all mean ages of student samples ranged between 18and 25 years, these studies were categorized under the young
adults group.
Meta-analytical Procedure
Study-level effect sizes
The effect sizes of studies applying questionnaire scales and
dream diaries were computed or reconstructed by the author
using the formula given by Rustenbach (2003) for Hedges g
(see also Rosenthal, 1994). The other measurement methods
yielded differences in percentages (e.g. REM awakening
technique, dream recall of the last night and recall of a most
recent dream). Rustenbach (2003) provided a formula for
computing effect sizes of this difference divided by the pooled
variance; so, this value is comparable to Hedges g.
Aggregate effect sizes
First, the Q-statistics of Cochran (e.g. Hardy and Thompson,
1998) were computed to test for homogeneity of the different
sets of effect sizes. As most of the tests yielded significant
results and thus indicated a substantial interstudy variance, the
random effects approach was selected to test whether the effect
sizes differed from zero. A mixed model was computed to
determine common effects of different factors (fixed effects)
and check for interstudy variation (random factor). For these
models, the Satterthwaite approximation was applied to
compute degrees of freedom. Furthermore, a bivariate
approach developed by Van Houwelingen et al. (2002) was
applied for comparing effect sizes stemming from studies with
questionnaire scales and dream diaries to account for possiblecorrelation in the 16 studies which included both measures.
The analyses were carried out using sas 9.1.3 (SAS Institute
Inc., Cary, NC, USA).
R E S U L T S
The meta-analysis is based on 175 independent samples,
including 147 samples consisting of healthy persons and 28
patient samples (12 psychiatric patients, nine sleep-disordered
patients, five somatic patients and two samples with blind
persons).
In Table 1, the estimated effect sizes for the five age groups
(healthy persons only) are depicted. The total number of
studies is 163 because for several studies two or more effect
sizes specifically computed for the age groups have been
included. The results (independent analyses for each age
group) indicate that substantial gender difference was present
in all five age groups, although the effect size for children is the
smallest.
The estimated effect sizes for the six measurement methods
are depicted in Table 2. Studies were only included if healthy
persons were investigated. The number of studies (169) exceeds
147 because quite a lot of studies used more than one
Table 1 Effect sizes for different age groups (healthy persons)
Age group Studies Sample size Women men
Estimated
effect size
Confidence interval
(95%) Q
Children (10 years) 15 4834 2332 2502 0.097 0.0350.159 6.02
Adolescents (10 < x < 18 years) 18 5725 3133 2592 0.364 0.2720.456 33.91**
Young adults (18 x < 30 years) 81 15 781 9238 6543 0.242 0.1940.289 121.32**
Middle-aged adults (30 x < 60 years) 36 16 250 9139 7111 0.270 0.2060.334 86.03***
Older adults (60 years) 13 1795 1028 767 0.243 0.1250.362 22.80*
Q-statistics: *P 0.05; **P 0.01; ***P 0.001.
Table 2 Effect sizes for different measurement methods (healthy persons)
Method Studies Sample size Women men
Estimated
effect size
Confidence interval
(95%) Q
Questionnaire scales 93 32 995 18 691 14 304 0.246 0.1860.306 294.96***
Dream diary 46 4548 2947 1601 0.309 0.2110.406 81.49***
REM awakenings 11 535 269 266 0.164 )0.0370.365 5.63
Most recent dream (yes no) 9 2115 1278 837 0.237 0.0760.398 16.12*
Recalled dream previous night 5 991 560 431 0.123 )0.0590.305 7.53
Dream themes (yes no) 5 6346 3367 2979 0.223 0.1530.293 4.74
Q-statistics: *P 0.05; ***P 0.001.
Gender and dream recall 127
2008 European Sleep Research Society, J. Sleep Res., 17, 125131
-
7/28/2019 Gender Differences in Dream Recall
4/8
measurement technique. Gender differences measured with
questionnaire scales, dream diaries, a question concerning the
most recent dream or dream themes were significantly different
from zero. Although the effect sizes for REM awakenings and
the question about dream recall last night were positive, the
confidence intervals include zero and, thus, were not signifi-
cantly different from zero.
For the patient groups, estimated effect sizes are presented in
Table 3. The sleep-disordered patients also showed a sub-stantial gender difference similar to that of healthy persons.
The confidence intervals of the other two groups were large
and, thus, the tests were non-significant, even though the effect
sizes were positive. For the two studies including blind
persons, a random model analysis could not be performed.
The study effect sizes are g = 0.439 and g = 0.122.
A mixed model was computed to test the common effects of
age group, measurement method, publication year and
whether the study was carried out by the author (Table 4).
Solely, the age group variable yielded a significant effect. A
post hoc analysis revealed that children differed significantly
from the three pooled adult groups (F = 4.61, d.f. = 1, 146,
P = 0.0335) and adolescents (F = 15.90, d.f. = 1, 71.8,
P = 0.0002). In addition, adolescents differed from the three
pooled adult groups (F = 4.74, d.f. = 1, 147, P = 0.0310).
The other variables (publication year, measurement method
and whether the study was carried out by the author) did not
affect the gender difference in dream recall.
A bivariate mixed model including 16 studies using both
measures (diaries and questionnaires) did not reveal significant
differences between the effect sizes obtained by questionnaire
scales and the effect sizes of dream diary studies (N = 123
studies; estimated difference in effect sizes: 0.063, CI )0.046 to
0.173, t = 1.14, P = 0.2562). A subsequent analysis using 80
independent studies using two types of scale format [absolute
answering categories (N = 55) and relative answering catego-
ries (N = 25)] also yielded a non-significant finding [estimated
difference in effect sizes: 0.074 (CI )0.080 to 0.227), t = 0.98,
P = 0.3333].
Lastly, the difference between healthy persons (N = 147studies) and the patient groups (N = 28) was not significant
(estimated difference in effect sizes: )0.058, CI )0.160 to 0.043,
t = )1.13, P = 0.2597). However, the comparison of the 12
psychiatric patient groups with the healthy persons yielded a
significant finding (estimated difference in effect sizes: )0.175;
CI )0.338 to )0.011, t = )2.11, P = 0.0361).
D I S C U S S I O N
Overall, the meta-analysis indicates that there is a small but
substantial gender difference in dream recall; i.e. women tend
to recall their dreams more often than men.
The measurement method did not affect the gender differ-
ence (mixed model analysis for healthy persons) and, thus, it is
unlikely that different algorithms for determining study effect
sizes due to different measurement levels (nominal, ordinal and
interval) might have biased the results (Table 2). However, for
two measurement methods (REM awakenings and recalling
dreams from the previous night) the effect sizes were positive
but not significantly different from zero. Several factors might
explain this result. First, the sample sizes are smaller than
Table 3 Effect sizes for different patient groups
Method Studies Sample size Women menEstimatedeffect size
Confidence interval(95%) Q
Psychiatric patients 12 888 539 349 0.088 )0.0650.242 8.46
Somatic patients 5 283 143 140 0.252 )0.2320.736 9.63*
Sleep disordered patients 9 1853 729 1124 0.242 0.1050.379 11.03
Q-statistics: *P 0.05. Due to the small sample size, a random effects model could not be computed for the two studies with blind persons.
Table 4 Mixed model for testing effects of covariates (healthy persons)
Factor F-value
Degrees of
freedom P-value
Age group:children (15), adolescents (14), young adults (76),
middle-aged adults (34), elder adults (8)
4.17 (4, 92.4) 0.0038
Measurement method:
questionnaire (77), dream diary (45),
REM awakenings (11), most recent
dream last night dream themes (14)
0.41 (3, 135) 0.7442
Study conducted by the author:
yes (45), no (102)
0.51 (1, 95.7) 0.4753
Publication year 1.88 (1, 66.6) 0.1753
N = 147 independent study effect sizes; Ntotal = 44 460, Nwomen = 24 916, Nmen = 19 544.
128 M. Schredl and I. Reinhard
2008 European Sleep Research Society, J. Sleep Res., 17, 125131
-
7/28/2019 Gender Differences in Dream Recall
5/8
those for the other methods which might be explained by the
considerable expense of sleep laboratory research. Secondly,
the reliability, e.g. retest reliability, is not known for these two
methods, whereas questionnaire scales showed high retest
reliability indices (Schredl, 2004b) and dream diaries have high
interitem consistency (Schredl and Fulda, 2005). A third factor
which might play a role in explaining the smaller gender
difference in the percentage of recall after REM awakenings is
addressed by Strauch and Meier (2004). To obtain as manydream reports as possible, the researchers often select high
dream recallers for participating in sleep laboratory studies.
However, one should keep in mind that the effect sizes did not
differ significantly from the overall effect size and an analysis
including more participants might have yielded a significant
result, even if the effect sizes are smaller than those of the other
methods.
A separate analysis was carried out for the difference
between questionnaire scales and dream diaries because a
number of larger studies included both measures. In contrast
to Beaulieu-Prevost and Zadra (2007) who, in their meta-
analysis, found a significant effect of measurement method on
the correlation coefficients between DRF and several person-
ality measures, the present findings did not support the idea
that methodological biases, e.g. memory biases for the
retrospective measures, affect the findings. Despite the weak-
nesses each measurement method possesses (cf. Schredl, 2007),
the present meta-analysis indicates that all methods yielded
comparable effect sizes for the gender difference in dream
recall.
Similarly, publication year and the over-representation of
studies conducted by the first author [due to availability of
gender-specific information and seven unpublished studies
(healthy persons)] did not affect the effect sizes, demonstrating
again the robustness of the gender difference in dream recall.The meta-analysis revealed a significant effect of age on the
gender differences in dream recall, while controlling for other
possibly confounding effects (publication year, measurement
method, only studies with healthy persons were included). The
similar effect sizes in the adult groups contradict the findings of
Giambra et al. (1996) who reported a marked gender differ-
ence between 25 and 55 years of age but not for young adults
and older adults (N = 2328) but are in line with the
insignificant age gender interaction (analysis depicted in
Schredl and Piel, 2005) in four large representative surveys
reported by Schredl and Piel (2003). The significant effects
found in all age groups are also not in accordance with the
findings of Stepansky et al. (1998). Why they did not find a
substantial gender difference in their representative sample
cannot explained by the present meta-analytic results. Schredl
et al. (1996a) reported stability for DRF over adulthood by
eliciting current DRF and retrospectively asking for DRF
experienced in the participants early twenties. However,
longitudinal studies are completely lacking; so, the stability
of DRF has to be supported by future studies.
The significant smaller gender difference in dream recall in
children compared with adults has not yet been reported in the
literature, mainly because the sizes of children samples have
been quite small and, therefore, do not allow generalizations.
One might speculate, that there may be a gender-specific
dream socialization with a peak in adolescence, e.g. it might
be possible that girls are encouraged more often by their
caregivers or peers to talk about their dreams which, in turn,
increases DRF by focusing on dreams (cf. Schredl, 2007).
Schredl and Sartorius (2006) were able to demonstrate a small
but significant correlation between DRF in children and theirmothers, indicating that the dream topic might be more
present in families with adults who recall their dreams more
often. That seems to support the idea of a dream socialization
process, but whether this is gender specific has to be demon-
strated. It is not very likely that genetic effects might explain
this association because twin studies (Cohen, 1973a; Gedda
and Brenci, 1979) did not find any significant differences in
concordance rates between monozygotic and dizygotic twin
pairs. But then again longitudinal studies starting in early
childhood with larger samples to support the idea of gender-
specific dream socialization are lacking.
Patients with sleep disorders and somatic disorders showed
practically the same effect sizes regarding the gender difference
in dream recall as healthy persons. This indicated that sleep
variables might not be able to explain the gender difference in
dream recall (see discussion below). Psychiatric patients,
however, did not show a significant gender difference and
their estimated effect size was significantly lower than the effect
sizes of healthy persons. One might speculate that reduced
dream recall, which is often related to depressive mood and is a
major symptom in many psychiatric disorders (e.g. Schredl,
1995; Schredl and Engelhardt, 2001), might cover up the
gender difference in dream recall normally found in healthy
persons.
Given the substantial and robust finding of gender differ-ences in dream recall, the question of what factors might
explain this difference arises. Possible factors have to meet two
criteria: first, they should show a stable gender difference and,
secondly, they should correlate substantially with DRF. With
regard to the first criteria, there are a variety of factors which
might be related to dream recall and show stable gender
differences like verbal intelligence (Halpern, 2000; Hyde and
Linn, 1988), verbal memory (Chipman and Kimura, 1998),
frequency of talking about emotional matters (Kring and
Gordon, 1998), recall of emotional experiences (Seidlitz and
Diener, 1998), sleep quality (Schredl et al., 1998b), prevalence
of insomnia (Zhang and Wing, 2006) and interest in dreams
(Schredl, 2000). Most of these studies showed small-to-
moderate effect sizes indicating that women do better in verbal
intelligence and verbal memory talks, talk more often about
emotional matters and showed higher recall of emotional
experiences. Regarding sleep, women tend to wake up more
often at night, and sleep quality is reduced. The empirical data
linking these variables to DRF, however, yielded mixed
findings. Tonay (1993), for example, reported no significant
relationship between verbal intelligence and dream recall.
Similarly, Armitage et al. (1987), Schredl et al. (1996a) and
Gender and dream recall 129
2008 European Sleep Research Society, J. Sleep Res., 17, 125131
-
7/28/2019 Gender Differences in Dream Recall
6/8
Blagrove and Akehurst (2000) did not find an effect of verbal
memory capacity on DRF. The findings that heightened DRF
is associated with a better recall of autobiographical episodes
from childhood (Robbins and Tanck, 1978; Schredl et al.,
1996b) might indicate that the capacity of recalling emotional
experiences might be associated with DRF. But specific
experiments eliciting recall of emotional relevant stimuli like
those carried out by Seidlitz and Diener (1998) have not
included a DRF measure. Similarly, no empirical datum isavailable for the relationship between the frequency of talking
about emotional matters and DRF. On the other hand, several
studies (cf. Schredl et al., 1997, 1998a, 2003b) indicate that low
sleep quality, frequency of nocturnal awakenings and the
diagnosis of insomnia are related to heightened dream recall.
The meta-analysis of Beaulieu-Prevost and Zadra (2007)
showed that interest in dreams is correlated with DRF.
To date, only two studies (Schredl, 2000, 200203) investi-
gated the direct effect of possible explanatory factors on the
gender difference in dream recall by applying statistical
methods like regression analyses and partial correlations.
Schredl (2000) was not able to find a significant reduction in
the gender difference in dream recall by introducing sleep
quality, frequency of nocturnal awakening (questionnaire
measure) into the regression analysis. When the engagement
in dreams variable, which showed a marked gender difference
(d = 0.71; N = 722), was additionally included, the gender
difference was no longer significant. Although the findings
suggest that the gender difference in interest in dreams
explains the gender difference in DRF, it is difficult to
determine a causal chain. On the one hand, it seems plausible
that only persons who recall their dreams develop an interest
in them; and, on the other hand, it has been demonstrated that
persons who are interested in their dreams and focus on
dreams can increase their dream recall markedly (cf. Schredl,2007). Including what has been written above, the two
variables, dream recall and engagement or interest in dreams,
might be affected by socialization which, in turn, might be
gender specific. Therefore, interest in dreams should also be
measured in longitudinal studies to determine possible causal
links.
Exploratory analyses of other factors explaining the gender
difference in dream recall, such as emotional intensity of
dreams, emotional tone of dreams, time of getting up in the
morning and nightmare frequency, failed to produce signif-
icant results (Schredl, 200203). The score of a rating scale
measuring subjective meaning of dreams, however, did have an
effect and reduced the gender difference in DRF if it was
statistically controlled (Schredl, 200203), again pointing to
the relationship between interest in dreams and dream recall.
In addition to these two studies (Schredl, 2000, 200203), it
would be interesting to study the effect of the above-listed
factors such as verbal intelligence, verbal memory, recall of
emotional experience and frequency of talking about emo-
tional matters on the gender difference in dream recall by
applying suitable statistical methods such as those reported
above.
Another factor which might explain the gender difference in
dream recall might be sex role orientation, i.e. the hypothesis
would be that feminine persons tend to recall their dreams
more often than masculine persons and, thus, this dimension
might explain why women differ from men with regard to
dream recall. Although several studies (Cohen, 1973b; Porach,
1971; Waterman et al., 1988; Wood et al., 1988) investigated
the effect of sex role orientation on dream content, no research
has been published on the relationship between the masculinefeminine dimension and dream recall. I.e. future studies should
test whether there is a substantial correlation between sex role
orientation and dream recall and if sex role orientation
explains the reported gender difference in dream recall.
To summarize, the meta-analysis including a considerable
number of studies was able to demonstrate a small but
substantial gender difference in dream recall which is indepen-
dent of methodological factors, e.g. different measurement
methods. The smaller gender difference in childhood might
indicate that there is a gender-specific dream socialization.
However, longitudinal studies in this field are lacking and are
necessary to support the conclusion derived from the meta-
analytic findings.
R E F E R E N C E S
Armitage, R., Stelmack, R. M. and McCarry, P. Individual differences
in event-related potentials and recognition memory for words with
manipulation of instructional set. Psychophysiology, 1987, 24: 576
577.
Beaulieu-Prevost, D. and Zadra, A. Absorption, psychological bound-
aries and attitude towards dreams as correlates of dream recall: two
decades of research seen through a meta-analysis. J. Sleep Res.,
2007, 16: 5159.
Belicki, K. Recalling dreams: an examination of daily variation and
individual differences. In: J. Gackenbach (Ed.) Sleep and Dreams: A
Sourcebook. Garland, New York, 1986: 187206.Bernstein, D. M. and Belicki, K. On the psychometric properties of
retrospective dream content questionnaires. Imagin. Cogn. Pers.,
199596, 15: 351364.
Blagrove, M. and Akehurst, L. Personality and dream recall frequency:
further negative findings. Dreaming, 2000, 10: 139148.
Blagrove, M. T., Button, L. and Bradshaw, C. Dream recall frequency
is associated with openness to experience but not sleep length. Sleep
Suppl., 2003, 26: A91A92.
Borbely, A. Schlafgewohnheiten, Schlafqualita t und Schlafmittelkon-
sum der Schweizer Bevo lkerung: Ergebnisse einer Repra sentativum-
frage. Schweiz Arzteztg, 1984, 65: 16061613.
Botman, H. I. and Crovitz, H. F. Dream reports and autobiographical
memory. Imagin. Cogn. Pers., 198990, 9: 213224.
Chipman, K. and Kimura, D. An investigation of sex differences on
incidental memory for verbal and pictorial material. Learn. Individ.
Differ., 1998, 10: 259272.
Cohen, D. B. A comparison of genetic and social contributions
to dream recall frequency. J. Abnorm. Psychol., 1973a, 82: 368371.
Cohen, D. B. Sex role orientation and dream recall. J. Abnorm.
Psychol., 1973b, 82: 246252.
Cohen, D. B. Remembering and forgetting dreams. In: J. F. Kihlstrom
and F. J. Evans (Eds) Functional Disorders of Memory. Wiley, New
York, 1979: 239274.
Colace, C. Sulla valutazione delle credibilita dei sogni raccontati dai
bambini: uno studio preliminare. Psichiatria dell infanzia e dell
adolescenza, 1998, 65: 518.
130 M. Schredl and I. Reinhard
2008 European Sleep Research Society, J. Sleep Res., 17, 125131
-
7/28/2019 Gender Differences in Dream Recall
7/8
Colace, C. Childrens dreaming: a study based on questionnaire
completed by parents. Sleep Hypn., 2006, 8: 2033.
Colace, C. and Violani, C. La bizzarria del sogno infantile come
correlato della capacita di provare sensi di colpa. Psichiatria dell
infanzia e dell adolescenza, 1993, 60: 367376.
Colace, C., Tuci, B. and Ferendeles, R. Bizarreness in early childrens
dreams collected in the home setting: preliminary data. Sleep Res.,
1997, 26: 241.
Gedda, L. and Brenci, G. Sleep and dream characteristics in twins.
Acta Genet. Med. Gemellol., 1979, 28: 237239.
Giambra, L. M., Jung, R. E. and Grodsky, A. Age changes in dreamrecall in adulthood. Dreaming, 1996, 6: 1731.
Goodenough, D. R. Dream recall: history and current status of the
field. In: M. M. Arthur Arkin, S. S. John Antrobus and S. J. Ellman
(Eds) The Mind in Sleep: Psychology and Psychophysiology.
Erlbaum, Hillsdale, NJ, 1978: 113140.
Goodenough, D. R., Shapiro, A., Holden, M. and Steinschriber, L. A
comparison of dreamers and nondreamers. J. Abnorm. Soc.
Psychol., 1959, 59: 295302.
Halpern, D. F. Sex Differences in Cognitive Abilities, 3rd edn.
Lawrence Erlbaum, Mahwah, NJ, 2000.
Hardy, R. J. and Thompson, S. G. Detecting and describing
heterogeneity in meta-analysis. Stat. Med., 1998, 17: 841856.
Hartmann, E. Boundaries of dreams, boundaries of dreamers: Thin
and thick boundaries as a new personality measure. Psychiatr. J.
Univ. Ott., 1989, 14: 557560.Hyde, J. S. and Linn, M. C. Gender differences in verbal ability: a meta
analysis. Psychol. Bull., 1988, 104: 5369.
Kring, A. M. and Gordon, A. H. Sex difference in emotion: expression,
experience, and physiology. J. Pers. Soc. Psychol., 1998, 74: 686
703.
Levin, R., Fireman, G. and Rackley, C. Personality and dream recall
frequency: still further negative findings. Dreaming, 2003, 13: 155
162.
Nielsen, T. A. A review of mentation in REM and NREM sleep:
covert REM sleep as a possible reconcilation of two opposing
models. Behav. Brain Sci., 2000, 23: 851866.
Nielsen, T. A., Leberge, L., Paquet, J., Tremblay, R. E., Vitaro, F. and
Montplaisir, J. Development of disturbing dreams during adoles-
cence and their relation to anxiety symptoms. Sleep, 2000, 23: 727
736.
Pagel, J. F., Vann, B. H. and Altomare, C. A. Reported association of
stress and dreaming: community background levels and changes
with disaster (hurricane Iniki). Dreaming, 1995, 5: 4355.
Porach, L. B. The relationship of masculine and feminine identifica-
tions to dream scores and to menstrual cycle reactions. Diss. Abstr.
Int., 1971, 31: 4558-A.
Robbins, P. R. and Tanck, R. H. Early memories and dream recall.
J. Clin. Psychol., 1978, 34: 729731.
Rosenthal, R. Parametric measures of effect size. In: H. Cooper and L.
V. Hedges (Eds) The Handbook of Research Synthesis. Russell Sage
Foundation, New York, 1994: 231244.
Rustenbach, S. J. Metaanalyse Eine anwendungsorientierte Einfuh-
rung. Hans Huber, Bern, 2003.
Schredl, M. Traumerinnerung bei depressiven Patienten. Psychother.
Psychosom. Med. Psychol., 1995, 45: 414417.
Schredl, M. Traume und Schlafstorungen: Empirische Studie zur
Traumerinnerungshaufigkeit und zum Trauminhalt von schlafgestorten
PatientInnen. Tectum, Marburg, 1998.
Schredl, M. Dream recall. research, clinical implications and future
directions. Sleep Hypn., 1999, 1: 99108, A1A4.
Schredl, M. Gender differences in dream recall. J. Ment. Imagery,
2000, 24: 169176.
Schredl, M. Questionnaires and diaries as research instruments in
dream research: methodological issues. Dreaming, 2002a, 12: 1726.
Schredl, M. Messung der Tranmerinnerung: Skala und Daten ges
under Personen. Somnologie, 2002b, 6: 3438.
Schredl, M. Factors influencing the gender difference in dream recall
frequency. Imagin. Cogn. Pers., 200203, 22: 3339.
Schredl, M. Traumerinnerung: Modelle und empirische Untersuchungen.
Tectum, Marburg, 2004a.
Schredl, M. Reliability and stability of a dream recall frequency scale.
Percept. Mot. Skills, 2004b, 98: 14221426.
Schredl, M. Dream recall: models and empirical data. In: D. Barrett
and P. McNamara (Eds) The New Science of Dreaming Volume 2:
Content, Recall, and Personality Correlates. Praeger, Westport, 2007:
79114.
Schredl, M. and Engelhardt, H. Dreaming and psychopathology:dream recall and dream content of psychiatric inpatients. Sleep
Hypn., 2001, 3: 4454.
Schredl, M. and Fulda, S. Reliability and stability of dream recall
frequency. Dreaming, 2005, 15: 240244.
Schredl, M. and Montasser, A. Dream recall: state or trait variable?
Part I: model, theories, methodology and trait factors and Part II:
state factors, investigations and final conclusions. Imagin. Cogn.
Pers., 199697, 16: 181210, 239261.
Schredl, M. and Piel, E. Gender differences in dream recall frequency:
data from four representative German samples. Pers. Individ.
Differ., 2003, 35: 11851189.
Schredl, M. and Piel, E. Gender differences in dreaming: are they
stable over time? Pers. Individ. Differ., 2005, 39: 309316.
Schredl, M. and Sartorius, H. Frequency of dream recall by children
and their mothers. Percept. Mot. Skills, 2006, 103: 657658.Schredl, M., Morlock, M. and Bozzer, A. Kindheitserinnerungen und
Tra ume Erwachsener. Zeitschrift fur Psychosomatische Medizin und
Psychoanalyse, 1996b, 42: 2533.
Schredl, M., Schro der, A. and Lo w, H. Traumerleben von a lteren
Menschen - Teil 2: Empirische Studie und Diskussion. Zeitschrift fur
Gerontopsychologie und -psychiatrie, 1996a, 9: 4353.
Schredl, M., Bozzer, A. and Morlock, M. Traumerinnerung und
Schlafsto rungen. Psychother. Psychosom. Med. Psychol., 1997, 47:
108116.
Schredl, M., Scha fer, G., Weber, B. and Heuser, I. Dreaming and
insomnia: dream recall and dream content of patients with insom-
nia. J. Sleep Res., 1998a, 7: 191198.
Schredl, M., Schenck, W., Go rtelmeyer, R. and Heuser, I. Ein-
flufaktoren auf die Schlafqualita t bei Gesunden. Somnologie,
1998b, 2: 99103.
Schredl,M., Wittmann, L.,Ciric,P. andGo tz,S. Factorsof home dream
recall: a structural equation model. J. Sleep Res., 2003b, 12: 133141.
Seidlitz, L. and Diener, E. Sex differences in the recall of affective
experiences. J. Pers. Soc. Psychol., 1998, 74: 262271.
Stepansky, R., Holzinger, B., Schmeiser-Rieder, A., Saletu, B., Kunze,
M. and Zeitlhofer, J. Austrian dream behavior: results of a
representative population survey. Dreaming, 1998, 8: 2330.
Strauch, I. and Meier, B. Dem Traum auf der Spur: Zugang zur
modernen Traumforschung (2 Auflage). Hans Huber, Bern, 2004.
Tonay, V. K. Personality correlates of dream recall: who remembers?
Dreaming, 1993, 3: 18.
Van Houwelingen, H. C., Arends, L. R. and Stijnen, T. Advanced
methods in meta-analysis: multivariate approach and meta-regres-
sion. Stat. Med., 2002, 21: 589624.
Waterman, D., de Jong, M. A. and Magdeliyns, R. Gender, Sex Role
Orientation and Dream Content (Sleep86). Gustav Fischer, Stutt-
gart, 1988.
Winget, C., Kramer, M. and Whitman, R. M. Dreams and demog-
raphy. Can. Psychiatr. Assoc. J., 1972, 17: 203208.
Wood, J. M., Sebba, D. and Griswold, R. J. Stereotyped masculine
interests as related to the sex of dream characters. Sleep Res., 1988,
17: 113.
Wynaendts-Francken, C. J. Tra ume bei Ma nnern und Frauen.
Neurologisches Centralblatt, 1907, 26: 941.
Zhang, B. and Wing, Y.-K. Sex differences in insomnia: a meta-
analysis. Sleep, 2006, 29: 8593.
Gender and dream recall 131
2008 European Sleep Research Society, J. Sleep Res., 17, 125131
-
7/28/2019 Gender Differences in Dream Recall
8/8