Dreaming, Vol. 9, No. 4, 1999

The Relationship of Dream Content and Changes in Daytime

Mood in Traumatized Vs. Non-Traumatized Children

Raija-Leena Punamäki

Department of Psychology
Applied Division
P.O. Box 4
00014-University of Helsinki
FINLAND
Fax: +358-9-191 2 33 79
Tel: +358-9-191 2 32 53

E-mail: raija.punamaki-gitai@helsinki.fi



Abstract

The study examined how the mood changes from night to morning, and how dysphoric dream contents associate with this change among children who live in traumatic environment and their controls from peaceful area. The sample consisted of 413 Palestinian boys and girls of 6-15 years of age, the mean age being 11.22+2.64. The participants filled in a seven-day dream diary in which they recorded their recalled dreams every morning. First, the results, confirmed that mood change from evening to morning is a general dream function: age and gender are not related to the change. The mood chance was rather associated with what and whom the children dreamt about. Second, the hypothesis of the trauma group showing less change in dysphoric dream content and in the intensity of negative morning mood across a period of time of seven days was not confirmed. On the contrary, the results showed that both dreams incorporating dysphoric themes and negative morning mood decreased only among children living in traumatic conditions. Third, it was hypothesized that there is a stronger association between presleep negative mood and dysphoric dreams, as well as between the dysphoric dreams and negative morning mood among children living in traumatic environment than among children from peaceful area. Contrary to the hypothesis, results for the trauma group revealed a reverse association between evening mood and dream contents: the more afraid, angry and worried children felt in the evening, the more Happy recreation dreams they reported, and the happier evening mood they reported, the more Threatening stranger dreams they had. However, concurring with the hypothesis, a direct association was found between dysphoric dreams and negative morning mood in the trauma group. The more children dreamt about Threatening strangers, the more afraid, angry and worried they felt in the morning. The discussion proposes a model of the correcting or balancing dream function that is characterized by an reverse assimilation of incorporating evening mood into dreams, and by a direct accommodation of dream content into morning mood.

Key words: Dreams, Emotions, Traumatic events




Dream content and change of mood from evening to morning

There is evidence that how one feels before going to sleep and upon awakening the following morning are systematically different from each other: the intensity and variability of feelings decreases from night to morning (Kramer, 1993). A number of researchers have proposed that dreams alter subjective feelings and mood. However, little is known about how these changes in emotional state across the night are related to dream content.

Kramer (1993) puts forward the wake-dream-wake paradigm as the basis for exploring the function of dreaming in affecting mood changes. Prior wakefulness impacts sleep and sleep impacts subsequent wakefulness. Dreaming thus serves both assimilative and accommodative functions. The assimilative function refers to the 'day residuals' incorporating disturbing events and aiming at solving emotional problems. The assimilative dream function can achieve some corrective (Kramer & Roth, 1973), compensatory and mastery (Wright & Koulack, 1987) goals. The accommodative function involves the dream experience having an enduring effect on waking feelings (Wasserman & Ballif, 1984) and a potential impact on the dreamer's future life (Jung, 1916/1960).

The decrease of the intensity of negative mood from evening to morning is considered a general function of dreaming that does not depend on factors like a person's age, gender or living conditions (Kramer, 1993). Research indicates, however, that some circumstances such as depression and traumatic experiences affect the correlation between dream content and mood. Research has shown that there is no decrease in feeling intensity across the night among depressive patients (Cartwright, 1984; 1991). Their dysphoric dreams characterized by masochistic relationships negatively influenced their morning mood. Research on Vietnam veterans shows that about 50% of PTSD patients' dreams were manifestly about their military experiences. Their dreams typically repeated same traumatic scene, and there was neither a decrease in the intensity of negative emotions across the night nor across longer periods of time (Hartmann, 1985; van Kolk, 1987; Van der Kolk, Blitz, Burr, Sherry & Hartmann, 1984).

This study deals with children, and analogously I hypothesize that there will not be a decrease in the intensity of negative mood from evening to morning among children who live in violent conditions and are personally exposed to traumatic experiences. Furthermore, I hypothesize that among children who are exposed to traumatic events, there will be less change in dysphoric dream content during a seven day period than among children unaffected by traumatic events. Subsequently, children living in traumatic conditions show less change in their morning mood during a seven day period than children living in peaceful environment.

In addition to the repetitive horrific contents, characteristic to the dreams of traumatized people are a lack of dreamlike and bizarre quality (Van der Kolk, Blitz, Burr, Sherry, & Hartmann, 1984), anxious, hostile, and persecuting themes, and dominance of the past (Lavie & Kaminer (1991). These kind of dysphoric dreams may be the reason for the lack of attenuation of mood from evening to morning, because they fail both the assimilating and accommodating dream function.

Dreaming is generally responsive to the experiences that precede it during the day, but they are usually juxtapositioning in a symbolic, novel, condensed and metaphoric form (Porte & Hobson, 1987). On the contrary, among trauma victims the occurrence of anxiety, persecution and aggression dreams suggest that fusing waking-life experiences into bizarre, condensed and camouflaging ways into dreams is impaired. Dreaming about the traumatic scene in a realistic form indicates rather a direct mirroring of the waking-life experience. The dreaming process of trauma victims thus forcefully incorporates intrusion of painful memories and overwhelming negative feelings in to dream, but fails in corrective, camouflaging and mastery function. Subsequently, the emotionally provocative experiences are intensively and vividly present also in the dreams (Hartmann, 1991; Kramer, 1993). Accordingly I hypothesize that among children exposed to traumatic events, there will be a stronger association between presleep negative mood and dysphoric dreams than among children who are not affected by traumatic events.

The failure of decrease of emotional intensity across the night among trauma victims may serve an example of their difficulty to translate painful events into new symbolized and masked forms. As the dysfunctional dreams decline to correct traumatic experiences, they, as a consequence, continue to exert their effects on negative mood in the morning (Cartwright, 1991; Piccone, Jacobs, Kramer, & Roth, 1977; van der Kolk, 1987).

In the same vein, certain types of nightmares fail their accommodating function in improving mood across the night or breaking the dream content and morning mood association (Belicki, 1992). This relates with the assumption that exposure to trauma makes the boundary between waking and dreaming realities especially transparent. It is revealed by research showing an increased dream recall (Punamäki, 1997a; Ross, Ball, Dinges, Kribbs, Morrison, Silver, & Mulvaney, 1994) and nightmare distress (van der Kolk, 1987) among trauma victims. The dysphoric dream atmosphere and emotions subsequently tend to dominate the morning mood and general well-being. Accordingly, I hypothesize that there will be a stronger association between dysphoric dreams and negative morning mood among children exposed to traumatic events than among children unaffected by trauma.

Even if a number of researchers argue that dreams alter subjective feelings and mood, little is known about what kind of dreams are dysfunctional, predicting changes towards negative, or beneficial, predicting positive mood changes across the night. Research suggest that the dream actors and repertoire of activities in dreams are crucial in affecting the changes in the dreamer's mood (Cartwright, 1984; 1992; Kramer, 1993; Kramer & Roth, 1980). The actors that appear in dreams are important determinants of mood change apparently because they represent significant current psychological concerns of the dreamer. For instance, Cartwright (1984) showed that the absence of an actor who is central to the current problem, such as a spouse in divorce, seems especially to suggest unsuccessful dreaming. Kramer (1993) found that the repertoire of dream actors engaging in various activities was crucial in decreasing an unhappy mood across the night. Analogically, depressive people typically show unhappy mood, and their dreams have been found to be are narrow, involve only one actor, usually the self-character, who repetitiously is failing, deprived and attacked (Cartwright, 1992). A research showed that when depression lightened, hostility and anxiety dreams decreased and dreams about intimate relationships increased (Kramer, Trinder, Whithman, & Baldridge, 1969).

Focusing on a limited aspect of dreams, such as characters, may be misleading. It is important to learn how a more comprehensive dream structures account for the subjective feelings that result from sleep. Belicki (1992), for instance, urged more systematic consideration on different types of nightmare content relating to types of waking emotional concerns (.e.g. anxiety dreams with fearfulness, grief dreams with depressive mood). Accordingly, in this research I attempt to undertake a more comprehensive approach to examine the dream characteristics that predict either the attenuation or intensification of negative affects from evening to morning. Analysing the dimensionality of dreams reveals, in addition to dream actors, other dream characteristics such as emotional and cognitive states, participation and success of outcome, as well as dream bizarreness and vividness.

To conclude, the present study examines how the mood changes from night to morning, and how dysphoric dream contents associate with this change among children who live in traumatic environment and their controls from peaceful area. First, it is hypothesized that attenuation of negative mood from evening to morning occurs only among children from peaceful area, whereas there is not decrease in intensity of negative mood during the night among children living in traumatic environment. Second, I hypothesize that there is less change in dysphoric dream content and less change in morning mood during a seven day period in the trauma group as compared with the control group. Third, it is hypothesized that the assimilating dream function is stronger in the trauma than in the control group, i.e., there will be a stronger association between presleep negative mood and dysphoric dreams among children living in traumatic environment than among children from peaceful area. Forth, it is hypothesized that in a traumatic environment, dysphoric dreams are likely to counteract the usual positive mood-attenuating effects of dreaming. Therefore, there will be a stronger association between dysphoric dreams and negative mood in the morning among children living in traumatic environment than among children from peaceful area.

Research shows, interestingly enough, that while general variables of gender and age are related to dream contents (Foulkes, 1990; Cartwright, 1992), they do not affect the mood change across the night (Kramer, 1993). Therefore, I propose that gender and age are associated with dream content but not with the changes in the mood from evening to morning.

Method

Subjects

The subjects were 412 Palestinian children and adolescents, of whom 268 lived in conditions of political violence in the Gaza strip and 144 in a peaceful area in Galilee, Israel. The subjects from Gaza are referred to as the trauma group and those from Galilee as the control group. Their ages ranged from six to 15 years, with a mean age of 11.22+2.64. The proportion of boys (55%) was slightly larger than that of girls (45%), but there were no gender differences in age (X^2=3.09, df=2, p=ns., n=410). The trauma and control groups were similar in gender, X(1)=.0006, df=1, p=ns., n=412, and in age, F(408,1)=2.58, p=ns., n=410.

The children and adolescents in the trauma group were selected using random and systematic sampling procedures (Pedhazur & Schmelkin, 1991). First, the places of residence were chosen according to the population distribution in the Gaza Strip. One town was chosen and two refugee camps were allotted from eight possible ones. The division guaranteed also representation of social strata of worker- and middle-class families. Second, following a random start, every third house in a street was visited. The control group was from one Galilean village. First, the village was divided into four areas in order to secure the representation of worker and middle-class families. In each area, following a random start, every third house on each street was visited. Both samples included children from various social-economic status, but were not matched it. The economic situation is extremely bad in the Gaza strip, and therefore the trauma group objectively suffered from more economic hardships than the control group. Concerning the education, children in both groups were primary and secondary school pupils. However, due to the military violence the Gaza children's education was frequently interrupted by curfews, confrontations, bombings and strikes.

Procedure

The researcher, together with a Palestinian psychologist, contacted the Gaza children and adolescents in their homes in September and October 1993. In Galilee, a local psychologist simultaneously contacted the control group children. The field work proceeded in two stages. First, the children and adolescents were asked to participate in the study and the dream diaries and instructions were given to them. The parents' (usually the mother's) consent was first asked. Second, after seven days, the field workers returned and collected the filled-in dream diaries.

The instructions for reporting dreams were written as a tightly-scripted and standardized statement in order to guarantee replication in each house. The field workers explained the dream diary and the questionnaires to the child or group of children page by page, and gave standardized examples of how to report dreams and respond to the questions. They checked whether the children had understood the idea by asking them to provide some examples of how to fill in the questionnaire. Young children dictated their dreams to their mothers, who then wrote them to the diaries. This involved 10 six and 15 seven year-olds.

The rate of return of the dream diaries was almost total: one child in Gaza and two in Galilee failed to fill in and return them. Of the total sample, four dream diaries were discounted due to missing information.

A pilot study examining children's ability to recall and document their dreams was conducted in a summer camp. The researcher visited the camp for seven days and gave ten girls and seven boys a notebook with similar instructions as in the proper study. She reviewed the dream reports every morning, in some cases transcribed them and interviewed children about their dreaming habits. Even if the reporting conditions in the summer camp and home differed, the pilot study provided some norms for children's dream recall and reporting styles.

Measures

Dream contents. A semi-structured dream and sleep diary was developed for the purpose of the study. The recording period was seven days, and for each night a sheet was prepared. The children were asked to record every morning the dream or dreams that they had had the previous night. They were given a sheet with 20 lines to be filled in: "Last night I dreamt that...".

The dreams were scored by using the systems of Foulkes (1982;1985), and Gottschalk, Winget and Gleser (1969). The former is based on cognitive theory and the dream characteristic scores include the scenes, actors, bizarreness, vividness, atmosphere, feelings and outcomes of the dream. The latter follows the psychoanalytical approach and involve scores of anxiety and overt and covert aggression.

Two research assistants were trained to categorize the dreams. They were unaware of the trauma-control group division in the sample, and of the specific hypotheses. A qualitative scoring program developed by Sulkunen and Kekäläinen (1992) was chosen as a categorizing tool. It enables the occurrence of all different aspects of a dream to be registered statistically. The reported dream can thus receive more than one dream characteristic score, and the values of dream content scales refer to the total number of scores counted. All the dreams from all dreamers served as a basis of scoring the dream characteristic variables. The value for each child was thereafter accounted by summing up the occurrence of each category across all dreams that he or she had reported.

The reliability was measured according to interrater agreement on the scores of 196 dreams of 70 randomly selected subjects. The reliabilities are presented after each category. They showed good agreement - Kappa values greater than .80, suggesting strong between-judges consensus (Fleiss, 1981).

The contents of dream characteristic scores are the following: The actors in the dreams were categorized as (a) Family members: father, mother, siblings, (b) Known people: neighbours, relatives, teachers, (c) Male strangers, (d) Female strangers, (e) Peers, (f) Soldiers, (g) Own fighters, (h) Animals. The setting of the dream was scored as (a) the dreamer's home, (b) school and (c) outdoors, such as the beach or the playground. Atmosphere was scored as (a) unpleasant, (b) neutral (c) both pleasant and unpleasant, and (d) pleasant (reliability 90.7%). The emotional state of the dreamer or other participants included the categories (a) Fearful and worrying (reliability 100%), (b) Angry and raging (reliability 89.3%), (c) Sad and experiencing feelings of loss (reliability 92.1%), and (d) Joyful and happy (88.7%). The cognitive state involved thoughts and perceptions of (a) the dreamer him/herself, (b) other people, and (c) both the dreamer and others. Manifest activity involved (a) verbal and (b) motor activities. The quality of Human relationships involved (a) Attack and fright (reliability 91.1%), (b) Avoidance (reliability 86.1%) and (c) Approach and affiliation (reliability 85.1%). Bizarreness refers to the level of strangeness, fantasy and credibility. The separate measurements of character and setting distortion were the following: (a) all mundane, (b) both bizarre and mundane characters and settings, and (c) all bizarre and distorted (reliability 93.7). Vividness refers to (a) high and (b) low levels of vision of colours and perspectives, the sense of texture and touch, the hearing of emotional tones and the appraisal of qualities (reliability 93.1%). Participation indicates whether the dreamer's role was that of (a) an observer, (b) a non-dominant participant, or (c) a dominating participant (reliability 94%). The outcome of the dream was scored as (a) happy and favourable (reliability 93.7%) or (b) Unhappy and unfavourable (reliability 88.3%). Anxiety and persecution dreams involved death, injury, destruction, separation, guilt, shame and threat, and being paralysed by terror and being haunted. Interrater agreement was 99.1%. Hostility and aggression dreams incorporated overt and covert aggression. Overt aggression is revealed in dreams in which the dreamer is the subject committing acts involving killing, threatening, causing suffering and expressing hostility. Covert aggression is revealed when somebody else is the subject committing aggressive and hostile acts, and the dreamer is an observer. The interrater reliability for combined aggression score was 92.5%.

To examine the dimensionality of dream content, all the dream category characteristic scores (accounted across seven nights for each child) were subjected to principal component analysis using varimax rotation. A five factors solution was adopted using the criterion of eigenvalues being greater than one and the factors accounting for 46.2% of the total variance. The solution was also conceptually sound, yielding both dysphoric and beneficial dream contents. Table 1 represents the component structure with dream dimensions, corresponding dream characteristic scores, eigenvalues, percentages of explained variation, and the Cronbach alpha reliabilities for each subscale.

Select for Table # 1

Composite dream variables are the factor sum scores of items contributing to each of the five factors. The dysphoric dreams include Threatening stranger dreams and Bizarre animal dreams, whereas Family dreams, Happy recreation dreams, and Peer group and school dreams reveals beneficial contents.

I Family dreams incorporated family members as actors, home settings, both pleasant and unpleasant atmosphere, approaching and affiliating human relationships, and thoughts and perceptions of both the dreamer and others.

II Threatening stranger dreams involved strangers as actors, an unpleasant atmosphere, the dreamer as an observer, low vividness, attacking human relationships, unfavourable outcomes, and anxiety, persecution and hostile and aggressive themes.

III Happy recreation dreams typically involved familiar actors, beach and playground settings, a pleasant atmosphere, non-dominating participation, happy feelings, thoughts and perceptions of the dreamer him/herself, and favourable outcomes.

IV Peer group and school dreams incorporated peer groups in school settings, and both bizarre and mundane characters and scenes.

V Bizarre animal dreams involved bizarre and distorted characters and scenes, animals as dream actors, fearful feelings and avoidant relationships.

Evening mood. The subjects were instructed to rate their mood in the evening just before going to bed. They were given a list of seven feelings (calm, happy, sad, afraid, angry, worried and excited) and asked to estimate: "Before going to bed I am feeling like ..." (1) not at all, (2) a little, (3) very much.

Separate composite variables were calculated for each evening mood by summing up their values across the seven evenings. Test-retest reliabilities (Split-half) for the seven evening feeling items were: Calm .83; Happy .83; Sad .94; Afraid .94; Angry .91; Worried .94; Excited .95.

Morning mood. The subjects were asked to evaluate their mood in the morning when waking up. The list presented seven feelings (calm, happy, sad, afraid, angry, worried and excited) and children were asked to estimate: "When waking up I was feeling like ..." (1) not at all, (2) a little, (3) very much. Separate composite variables were calculated for each morning mood by summing up their values across the seven mornings. Test-retest reliabilities (Split-half) for the seven morning feeling items were: Calm .85; Happy .83; Sad .94; Afraid .93; Angry .94; Worried .94; Excited .94.

Mood change from evening to morning was calculated by subtracting separately each evening mood score from the corresponding morning mood score. (e.g. morning calmness minus evening calmness). These scores were then aggregated for seven nights to indicate the evening to morning mood change scores for calmness, happiness, sadness, fear, angry, worry and excitement. The number of dreams recalled -variable is the sum of the children's dream reports, revealing whether (1) no dream was reported or (2) one or more dreams were reported. It ranged between 7 and 14. The zero-recall subjects are naturally excluded from the analyses that involve dream content.

Results

The hypotheses suggest that the relationship between mood and dream content is different among children living in traumatic and peaceful environments. Therefore, most of the analyses are conducted separately in the trauma and control groups. The groups differed in the frequency of dream recall and contents of their dreams. The trauma group recalled dreams more frequently than the control group, F(1,412) = 18.81, p=.0001. Furthermore, the trauma group had more Threatening stranger dreams, F(1,369) = 51.86, p=.0001, and Family dreams F(1,369) = 9.38, p=.002, while the control group had more Peer group and school dreams F(1,369) = 5.85, p=.01.

Mood change from evening to morning

and across the seven mornings

It was hypothesized that the decrease of negative mood from evening to morning occurs only in the control group, while in the trauma group there is no decrease in intensity of negative mood. Paired T-tests were performed for each samples of evening and morning moods of sadness, fear, angry and worry, indicating negative moods, and calmness, happiness and excitement, indicating positive mood.

The hypothesis was not substantiated, and results showed that there was no systematic decrease in mood intensity from evening to morning in either group. Only the intensity of calm feelings decreased from evening to morning in the trauma group, t(269)=-5.39, p<.0001.

In order to examine whether negative mood varied across a longer period of time, that of the seven mornings, a MANOVA repeated measures analysis of variance was carried out separately in the trauma and control groups. The scores of sadness, fear, anger and worry, and calmness, happiness and excitement for seven mornings were used as within-subject variables, and age and gender as between-subject variables. Contrary to the hypothesis significant decreased of negative mood across the seven mornings was found only in the trauma group, whereas no decrease was found in the control group. Figure 1 illustrates the negative mood across seven nights in the trauma group.

Select for Figure 1

In the trauma group, children reported significantly less sadness, F(6,1560) = 3.49, p<.002, fear, F(6,1560) = 3.63, p<.001, and marginally less anger, F(6,1560) = 1.84, p<.08 in the mornings during the seven days' period. The within-subject effects of negative morning mood were nonsignificant in the control group. The within-subject effects were nonsignificant for the positive morning moods of calmness and happiness in both groups. However, excitement decreased across the seven nights in the trauma group, F(6,1560) = 4.37, p<.0001. As assumed, no significant age or gender between-subject effects were found.

Variation of dysphoric dream content across the Seven Nights

We hypothesized that there would be less variation in dysphoric dream content across the seven nights in the trauma group than in the control group. To examine this, the scores for anxiety and persecution, and hostility and aggression dreams(2)

across seven nights were submitted as within-subject variables in a MANCOVA repeated measures analysis separately in the trauma and control groups. The number of recalled dreams was a covariate. Gender and age (6-8, 9-11, 12-15) were used as between-subject variables.

Figure 2 presents the means of the anxiety and hostility (overt) dreams across the seven nights among the trauma and control groups. Contrary to the hypothesis, the results show that the intensity of dysphoric dream contents significantly varied, i.e. decreased, during the seven nights' period only for the children in the trauma group. The within-subject effects were highly significant on anxiety and persecution (F(6,1374=6.16, p<.0001), overt aggression and hostility (F(6,1560)=6.77, p<.0001), and on covert aggression and hostility (F(6,1560=6.17, p<.0001) dreams in the trauma group. There were no significant age or gender between-subject effects.

Select for Figure 2

The association between pre-sleep mood and dream content

To examine whether there is a stronger association between presleep negative mood and dysphoric dreams among children living in traumatic environment than among children from peaceful area, multiple standard regression analyses were carried out separately in the trauma and control groups. The dependent variables were the factor scores of Threatening stranger dreams and Bizarre animal dreams, indicating dysphoric dreams, and Family dreams, Happy recreation dreams, and Peer group and school dreams, indicating beneficial dreams. The independent variables were the sum scores of the calm, happy, sad, fearful, angry, worried and excited mood reported across seven evenings and age, and gender. The regression models for dream content are presented for the trauma group and the control group in Table 2.

Select for  Table 2

The hypothesis was confirmed by showing that in the trauma group, evening mood was more strongly associated with dream contents than in the control group. Regression models were significant for dysphoric Threatening stranger dreams, but also for Happy recreation dreams, Family dreams and Peer group and school dreams in the trauma group, whereas, in the control group, evening mood predicted only Peer group and school dreams. Contrary to the hypothesis, however, negative evening mood was not associated with dysphoric but rather with beneficial dream contents, whereas positive evening mood, in turn, was associated with dysphoric dreams. Results for the trauma group showed that the more afraid, angry and worried children felt in the evening, the more Happy recreation dreams they reported. In the same vein, fearful evening mood was associated a low level Threatening stranger dreams. Results further revealed that the happier the trauma group children felt in the evening, the more Threatening stranger dreams and the less Family and Happy recreation dreams they had, and, finally, the more excited they were, the less Happy recreation dreams they reported. The beta-values of angry mood show, however, an exception of the association between negative evening mood and beneficial dream contents: the angrier children felt in the evening, the more they had both Threatening stranger and Happy recreation dreams.

The results indicated a compensatory or indirect association between evening mood and dream content in the trauma group. The predictors of Peer group and school dreams differed, to some extent, in the trauma and control groups. In traumatic conditions, high excitement and low fear in the evening predicted 11% of dreams incorporating peers and school events. In safe conditions, both negative and positive evening mood predicted about a third (32%) of the variation of Peers group and school dreams.

Table 2 shows that age and gender contributed, to some extent, to the dream content models. In the trauma group, girls reported more Family dreams than boys, and in both groups, older children had more Peer group and school dreams than younger ones. 

Association between dream content and morning mood

Separate standard regression models were used to test the hypothesis that dysfunctional dream content is more strongly associated with morning mood in the trauma group than in the control group. The morning mood scores of calm, happy, sad, fearful, angry, worried and excited feelings were the dependent variables, and the independent variables were the factor scores of Threatening stranger dreams and Bizarre animal dreams, indicating dysphoric dreams, and Family dreams, Happy recreation dreams, and Peer group and school dreams, indicating beneficial dreams, as well as the child's age and gender. Table 3 presents the regression models for morning mood in the trauma and the control groups.

Select Table 3

The hypothesis was confirmed by significant regression models showing that dream contents, indeed, explained the variation of morning mood, except sadness, in the trauma group. Whereas, in the control group dream contents, except Peer and school dreams, were not significant predictors of morning mood, even if the models for morning calmness and happiness were significant.

The results further confirm the hypothesis by showing that the more the children in the trauma group dreamt about Threatening strangers and Bizarre animals, the more fearful they felt in the morning. Furthermore, the more they had Threatening stranger dreams, the more angry and worried, and less happy they felt upon waking up. In the same vein, Happy recreation dreams were associated with happy and excited morning mood. However, contrary to the hypothesis, dysphoric dreams predicted also morning calmness and beneficial dreams the feelings of anger and worry in the morning mood in the trauma group. In the control group, the Peer group and school dreams were associated with happy morning mood.

The gender was associated only with the calm mood, girls feeling more calm in the morning than boys. In the trauma group, older children reported more intensive feelings, both negative and positive, in the mornings than younger children. Also, in the control group the older children reported more morning happiness than younger.

Dream Contents Explaining the Mood Change

Multiple Standard Regression analyses were performed to examine whether dream contents explained the changes in mood between evening and morning. The dependent variables were the scores indicating change in calm, happy, sad, fearful, angry, worried and excited mood across seven nights. The independent variables were the factor scores for Threatening stranger dreams and Bizarre animal dreams, Family dreams, Happy recreation dreams, and Peer group and school dreams, and age and gender.

Select for Table 4 

Table 4 presents the regression models for the evening to morning mood change scores in the trauma and control groups. Results reveal that the regression models were significant only for the trauma group, while in the control group, the dream content was not associated with the change of mood from evening to morning. The results confirmed the hypothesis that dysphoric dream content explains the failure of attenuation of negative feelings across the night among children living in traumatic conditions. The more their dreams incorporated threatening strangers, the more intensive fear and worry, and less calmness they felt in the morning as compared to evening. In the same vein, the more Happy recreation dreams children reported, the more their happiness and excitement increased, and the more fear, anger and worry decreased from evening to morning. Incorporation of Peer group and school events into dreams was associated with decreased excitement.

However, contrary to the hypothesis, Bizarre animal dreams and Family dreams predicted both negative and positive morning mood. Dysphoric dreaming about animals was associated with intensification of fear, but decrease of sadness from evening to morning. In the same vein, beneficial dreaming about family members was associated with both increased calmness and increased fear from evening to morning.

The results concur with the assumption that mood change from evening to morning is a general dream function by showing that age and gender were not related to the change. The mood change was, in turn, strongly associated with what and whom the children dreamt about. As an exception, however, there was more intensification of fear across the night among older than younger children.

Discussion

This study focuses on the relationship between mood change and dream content among children exposure to traumatic events as compared with a control group. It was hypothesized that living in traumatic environment in childhood affects a general dream function of decreasing the intensity of negative mood across the night. It was further assumed that the failure of attenuation of mood intensity might be due to changes in assimilative and accommodative dream functions among traumatized children.

As hypothesized, the results failed to show the general dream function of the intensity and variability of feelings decreasing from night to morning among children living in traumatic conditions. However, there was neither decrease of intensity of negative mood among children living in peaceful conditions, which was against the hypothesis.

The assimilative dream function was hypothesized to be stronger among children living in traumatic environment than among children from peaceful area. The hypothesis is based on the idea that stressful and traumatic experiences tend intensively and vividly to occur in dreams, but they often fail juxtapositioning the intrusive and painful images into metaphoric and symbolic forms. Concurring with the hypothesis, evening mood was more strongly associated with dream contents in the trauma than in the control group. However, results disqualified the assumption that negative evening mood would intensively associate with dysphoric dream content, and positive with beneficial dreams.

On the contrary, some corrective and compensatory dynamics between evening mood and dream content were found in the trauma group. Children who felt fearful, angry and worried in the evening, typically reported dreams incorporating Happy recreation scenes. Whereas, the happier evening mood children reported, the more Threatening stranger and the less Happy recreation dreams they had.

The occurrence of a corrective assimilative dream function concurs with research suggesting that dreaming may serve some restorative and balancing purpose in traumatic conditions (Hartmann, 1994; Jung, 1916/1960; Wright & Koulack, 1987). Compensatory dreams provide the dreamer with consolation and relief from a painful reality by incorporating thoughts, emotions and outcomes that are missing from real life (Kramer & Roth, 1973; Stewart & Koulack, 1993; Wright & Koulack, 1987). For children living in extremely dangerous environment, like the Palestinians of this study, it might be especially imperative to correct and change their image of waking life events characterized by threat and violence in their dreams. We have observed that children in war conditions attempt to employ all available mental resources, coping strategy and psychological defence to protect their psychological integrity and to accomplish heroic survival (Punamäki & Suleiman, 1990; Punamaki, 1997b). Dreaming about events incorporating familiar actors, pleasant atmosphere and happy feelings, and avoiding threatening dreams incorporating enemy soldiers and persecution scenes may serve this protection function among children who feel afraid while going to bed.

The finding that children's happy and fearless evening mood was associated with threatening dreams may reveal a mastery function of dreaming. Apparently, children who felt happy 'could afford' to dream about frightening and troublesome issues. Their mastery dreams typically incorporate threatening scenes that entail re-experiencing the trauma and associated affect in a safe place of dreaming (Hartmann, 1991). The results thus suggest that a positive and secure evening mood may be a precondition for mastery dream dynamics. As Kramer's (1991) Mood Regulatory Theory suggests, dreaming serves a selective, affective and regulatory function by correcting the level and range of the person's mood. It is noteworthy, however, that feeling angry was also associated with Threatening stranger dreams. The results indicate that a combination of evening emotions, rather than single positive or negative feelings may predict the dreams.

While the assimilative function showed an indirect, correcting and balancing incorporation of feelings into dreams, the accommodative function was characterized by dreams having a direct effect on morning mood among children living in traumatic environment. The results confirmed the hypothesis that the association between dysphoric dreams and postsleep negative mood is especially strong among traumatized children. The more Threatening stranger dreams children had, the more afraid, angry and worried, and the less happy they felt in the morning.

The results for the trauma group further evidenced that dysfunctional dreams failed in attenuation of negative moods from evening to morning. The more children's dreams incorporated Threatening strangers, the less their feelings of fear, anger and worry decreased across the night. These dreams thus served as an example of failing their accommodating function in improving mood across the night and in breaking the dream content and morning mood association (Belicki, 1992). The association between dysphoric dreams and increase of negative mood from evening to morning was rather explicit, whereas beneficial dreams were associated with both positive and negative change of feelings. The more Happy recreation dreams children reported, the more their happiness and excitement increased and fear decreased from evening to morning, but also they expressed more intensive anger and worry in the morning than in the evening.

The assimilative and accommodative links between dream content and mood are schematized in Figure 3. The correcting and balancing dream function is characterized by an indirect assimilation of incorporating evening mood into dreams, and by a direct accommodation of dream content into morning mood.

Select for Figure 3

The model by Kuiken (1986) may explain why expression of feelings would be associated with corrective rather than direct assimilative dream function. Expressing and becoming conscious about feelings in the evening forms additional mnemonic linkages that facilitates juxtapositioning narrative elements from affective memory sources (Kuiken, Rindslisbacher, & Nielsen, 1991). As a consequence of such a broad array of memories, there is no complete match between evening mood and dream content. In contrast, dreams following emotionally intensive pre-sleep events (such as trauma) without the expression of feeling would match the pre-sleep event in terms of feelings, actors, scenes, and atmosphere. Accordingly, our results suggest that intensity of feelings in the evening may have prompted the activation of a broad range of affective memories which dreaming then fused into distinct narratives. The consequent manifest dreams did not match evening mood because the broadened array of memories provoked either a compensatory or mastery dream phenomenon.

We may assume that, for children, dreaming about peers and school events represent significant current psycho-social concerns. The nature of the assimilative dream function for these dreams differed among children living in traumatic and safe environments. The trauma group children who felt excited and fearless in the evening, dreamt about peers groups and school events. In the control group, the intensity rather than the nature of evening mood predicted the occurrence of peer and school dreams. We have to remember that Palestinian children in Gaza participated in their national liberation task, and subsequently the struggle between feelings of courage and cowardice was fundamental in their social and peer relations. Furthermore, confrontations between children and enemy soldiers frequently took place in the school yards. Taking this political, ideological and security context into account may make it easier to understand why dreaming about peers and school was predicted by feeling daring and excited. In Galilee, where children went to sleep without environmental threat and danger, an intensive repertoire of all various emotions predicted dreaming about peers, i.e. relevant concerns in a child's daily life. Further research is needed, however, to reveal how personal issues, such as social acceptance and friendship, and environmental violence interact in child development and dreams fusing current concerns.

The dreams of trauma victims are often described as repetitious, thematically narrow and paralysing. These dream characteristics apparently fail to improve mood across the night, which, subsequently, may explain poor psychological adjustment among trauma victims (Cartwright & Lamberg, 1993; Punamäki, 1997b). Accordingly, I hypothesized that there would be less variation in dysphoric dream content and negative morning mood across a period of time in the trauma group than in the control group. Contrary to the hypothesis, the results showed, however, that dreams incorporating anxious, persecuting, hostile and aggressive themes systematically decreased (i.e. varied) across the seven nights' period in the trauma group. Similarly, negative morning feelings decreased during this period in the trauma groups, but not in the control group.

Narrating and reflecting one's dreams and feelings, as well as the dream contents per se may explain the decrease in dysphoric dreams and negative affect during the seven days. There is evidence that dream reflection, such as talking about dreams, increases access to feelings (Nielsen, Kuiken, & McGregor, 1989). The expression of feeling, in turn, evokes a broad range of memories to be fused into dreams, which speaks against the narrow repetition of dream content.

A considerable amount of trauma therapy is based on repeating and narrating painful memories, and through that the traumatic events are integrated into a patient's personality and current life. Revitalizing traumatic experiences and ventilating painful and frightening emotions is believed to facilitate mastering over trauma. Mastery dreams such as the dysphoric elements in Threatening stranger and Bizarre animal dreams, seem to play a similar role. Researchers have, indeed, suggested that dreams serve similar functions as psychotherapy in general (Hartmann, 1995), or desensitization therapy in particular (Perlis & Nielsen, 1993). Like therapy, dreaming allows the safe expression of repressed ideas, decontextualizes and symbolizes painful experiences, and provides transient satisfaction. Integration and healing occur in a hyperassociative state of mind through dreams connecting present trouble and past successful coping, and the replication of emotions (Hartmann, 1995).

It would, however, be too sweeping to claim that simple dream reporting makes children feel less afraid, sad, angry or worried in the mornings. Similarly, the idea that dream reporting alone could diminish the occurrence of anxiety, persecution, aggression and hostility dreams seems improbable. These kinds of dream have been found to be very persistent among traumatized people (Lavie & Kaminer, 1991).

It is worthwhile noting that the decrease in dysphoric dreams and negative morning mood across a period of time did not apply to children living in the peaceful area. Both the Gazan and Galilean Arab culture highly values dream reporting, and children are accustomed to think about and narrate their dreams (Nashef, 1993; Punamäki & Joustie, 1997). The reporting of dreams for this research was thus not a novel activity, except that the children were doing it systematically and were conscious that somebody was reading about their dreams.

Our fieldwork took place at the beginning of a militarily relatively calm period in Gaza, at the time of the peace agreement between Israel and the Palestinians. For the first time in the six years of the Intifada, the nightly curfew was lifted, there were fewer night raids and no houses were destroyed as a punishment for political and military activity. Yet, killings and sporadic fighting still took place and political prisoners were not released. Unfortunately we are not able to confirm whether the significant decrease in dysphoric dreaming and negative morning mood was due to the favourable military changes.

The results confirmed the earlier observation that age and gender affect the dream content, but do not influence the change of mood from morning to evening. Girls typically had more family dreams than boys, and older children dreamt more about peers and school events than younger ones. Earlier research showed that girls dreamt more about family members and home settings than boys (Brenneis & Roll, 1976; Punamäki & Joustie, 1997). My results confirmed, as Kramer (1993) suggested, that age and gender did not affect the mood change, but that the attenuation of negative moods from evening to morning was associated with about what and whom children were dreaming. The change in calm, happy, fearful, or worried mood from evening to morning varied according to whether children's dreams incorporated dysphoric or beneficial contents.

Even if the findings on assimilation and accommodation concur with those of Kramer and his team, we are naturally not able to compare the results based on children's subjective reporting with the laboratory research among adults. Kramer and his team sampled dreams from 10-20 subjects during 15 or more nights in a laboratory setting. We collected dream diaries from 412 children for seven consecutive nights. Our study setting did not allow the use of the comprehensive dream factor scores in the analyses assessing the night-by-night changes. The analyses therefore involved specific measures of dysphoric dreams incorporating themes of anxiety and aggression. Our results can only state on a general level that the affective state (mood) is related to dream content, and vice versa. How a change in mood relates to changes in dreaming over one night or longer requires a different research setting.



References

Brenneis, C. B., & Roll, S. (1976). Dream patterns in Anglo and Chicano young adults. Psychiatry, 39, 280-291.

Cartwright, R. d. (1984).Broken dreams: a study of the effects of divorce and depression on dream content. Psychiatry, 47, 251- 259.

Cartwright, R. D. (1991). Dreams that work: The relations of dream incorporation to adaption to stressful events. Psychological point of view. Dreaming, 1, 3-9.

Cartwright, R. D. (1992). "Masochism" in dreaming and its relationship to depression. Dreaming, 2, 79-84.

Cartwright, R. D., & Lamberg, L. (1993). Crisis dreaming: using your dreams to solve your problems. New York: Harber/Collins.

Belicki, K. (1992). Nightmare frequency versus nightmare distress: Relations to psychopathology and cognitive style. Journal of Abnormal Psychology. 101, 592-597.

Foulkes, D. (1982). Children's dreams. Longitudinal studies. New York: John Wiley and Sons.

Foulkes, D. (1985). Dreaming: A cognitive-psychological analysis. Hillsdale, N.J.: Lawrence Erlbaum.

Foulkes, D. (1990). Dreaming and consciousness. European Journal of Cognitive Psychology, 2, 39-55.

Gottschalk, L.A., Winget, C.N., & Gleser, G.G. (1969). Manual of instructions for using the Gottschalk-Gleser content analysis scales: Anxiety, hostility, and social alienation -personal disorganization. Berkeley & Los Angeles: University of California Press.

Hall, C.S., & van de Castle, R.L. (1966). The content analysis of dreams. New York: Twayne.

Hartmann, E. (1984). The nightmare. The psychology and biology of terrifying dreams. New York: Basic Books, Inc., Publishers.

Hartmann, E. (1991). Dreams that work or dreams that poison? What does dreaming do? An editorial essay. Dreaming, 1, 23-25.

Hartmann, E, (1995). Making connections in a safe place: Is dreaming psychotherapy? Dreaming, 5, 213-228.

Jung, C. G. (1916/1960). General aspects of dream psychology. In The structure and dynamics of the psyche. Collective works. Vol. 8. New York: Pantheon. pp. 237-280.

Kramer, M. (1993). The selective mood regulatory function of dreaming: an update and revision. In A. Moffitt, M. Kramer, & R. Hoffmann (eds.) The functions of dreaming. Albany: State University of New York Press.

Kramer, M., & Roth, T. (1973). The mood-regulating functions of sleep. In W.P. Koella & P. Levin (Eds). Sleep: Physiology, Biochemistry, Psychology, Pharmacology, Clinical implications, First European Congress on Sleep Research, Basel, 1972, pp. 563-571. Basel: S. Krager.

Kramer, M., & Roth, T. (1980). The relationship of dream content of night-morning mood change. In L. Popoviciu, B. Asgian, & G. Badin (Eds.) Sleep. Forth European Congress on Sleep Research, Tigre-Migres, 1978, pp. 621-624. Basel: S. Krager.

Kramer, M., Trinder, J., Whithman, R., & Baldridge, B.J. (1969). The incidence of masochistic dreams in night collected dreams of depressed subjects. Psychophysiology, 16, 250.

Kuiken, D., Rindlisbacher, P., & Nielsen, T. (1990-1991). Feeling expressions and the incorporation of presleep events into dreams. Imagination, Cognition and Personality, 10, 157-166.

Kuiken, D. (1986). Dreams and self-knowledge. In J. Gackenback (Ed.) Sleep and dreams: A source book, pp. 225-250. New York: Garland.

Lavie, P., & Kaminer, H. (1991). Dreams that poison sleep: Dreaming in Holocaust. Dreaming, 1, 11-21.

McCarley, R. W., & Hobson, J. A. (1989). The form of dreams and the biology of sleep. In B. B. Wolman (Ed.) Handbook of dreams. Research, theories and applications, pp. 76-130. New York: Van Nostrand Reinhold Co.

Nielsen, T. A., Deslauriers, D., & Baylor, G. W. (1991). Emotions in dream and waking event reports. Dreaming, 1, 287-300.

Nielsen, T. A., Kuiken, D. L., & McGregor, D. L. (1989). Effects of dream reflection on waking affect: awareness of feelings, Rorschach movement, and facial EMG. Sleep, 12, 277-286.

Perlis, M. L., & Nielsen, T. A. (1993). Mood regulation, dreaming and nightmares: Evaluation of desensitization function for REM sleep. Dreaming, 3, 243-256

Piccone, P., Jacobs, G., Kramer, M., & Roth, T. (1977). The relationship between daily activities, emotions and dream content. Sleep Research, 6, 133.

Porte, H. S., & Hobson, J. A. (1987). Bizarreness in REM and NREM sleep reports. Sleep Research, 16, 81.

Punamäki, R. L. (1997a). Explanations for and role of dream recall: Traumatic events, mood-congruent memory, dream salience and repression, and the association between dream recall and psychological adjustment. (Submitted to publication).

Punamäki, R. L. (1997b). The role of dreams in protecting psychological well-being in traumatic conditions. International Journal of Developmental Behavior (in press.).

Punamäki, R. L., & Joustie, M. (1997). The role of culture, violent environment, age and gender in affecting dream content and structure. Journal of Cultural Psychology, (in press.)

Punamäki, R. L., & Suleiman, R. (1990). Predictors and effectiveness of coping with political violence among Palestinian children. British Journal of Social Psychology, 29,67-77.

Ross, R. J., Ball, W. A., Dinges, D. F., Kribbs, N. B., Morrison, A. R., Silver, S. M., & Mulvaney, F. D. (1994). Rapid eye movement sleep disturbance in posttraumatic stress disorder. Biological Psychiatry, 35, 195-202.

Taub, J., Kramer, M., Arand, D., & Jacobs, G. (1978). Nightmare dreams and nightmare confabulations. Comprehensive Psychiatry, 19, 285-291.

Van der Kolk, B. (1987). Psychological Trauma. Washington, D.C: American Psychiatric Press.

Van der Kolk, B., Blitz, R., Burr, W., Sherry, S., & Hartmann, E. (1984). Nightmares and trauma: A comparison of nightmares after combat with lifelong nightmares in veterans. American Journal of Psychiatry, 141, 187-190.

Wasserman, I., & Ballif, B.L. (1984). Perceived interactions between the dream and waking divisions of consciousness. Journal of Imagination, Cognition, and Personality, 4, 3-13.

Wright, J., & Koulack, D. (1987). Dreams and contemporary stress. A disruption-avoidance-adaption model. Sleep, 10, 172-172.

1.

1..

2.

0     The analysis uses dream characteristic scores that were measured separately across each seven night. It was not possible to report changes in the dream factor scores, because factoring involved all the dream scores summed up across the seven nights. Factoring of the dream scores for separate nights was technically impossible, because of the large number of naturally missing dream scores. Subsequently, the scores of anxiety, persecution, hostility and aggression dreams are used as indicators of dysphoric dream contents.