Dreaming, Vol. 7 No. 4, 1997

Determinants and Mental Health Effects of Dream Recall

Among Children Living in Traumatic Conditions


Raija-Leena Punamäki1


This study examined whether the repression, mood congruent memory, and salience models would explain the frequency of diary recorded dream recall in two groups, one repeatedly exposed to trauma and the other living in relatively peaceful circumstances. The trauma group included 268 Palestinian children and adolescents living in a politically violent area in Gaza; the comparison group included 144 Palestinian children and adolescents living in a peaceful area in Galilee. In general, the more children were exposed to trauma, the more frequently they recalled their dreams: the trauma group reported more dreams than the comparison group, and, within the trauma group, children who were repeatedly exposed to traumatic events recalled more dreams than those exposed to fewer trauma. Of the three models of dream recall, two were supported. First, salient (i.e., bizarre, vivid, emotional, active, and narratively coherent) dreams were more frequently recalled, and, second, those in which the dream mood (atmosphere and feeling) was congruent with waking mood were more frequently recalled. However, contrary to expectations, repressive coping strategies (e.g., paralysis, denial, numbing, and distraction) were associated with more frequent dream recall. Moreover, although, in general, dream recall was correlated with problems in psychological adjustment, the relationship was symptom specific: frequent dream recall shielded children from somatic and anxiety symptoms but made them more susceptible to depressive symptoms.

KEY WORDS: dream recall; traumatic stress; psychological adjustment; dream salience; mood-congruent memory; repression.




Although laboratory studies indicate that more dreams occur than are spontaneously recalled, explanations for failures in dream recall vary. According to the repression hypothesis, shameful and painful dream material is eliminated from awareness and forgotten. On the other hand, the mood congruent memory model suggests that, when the mood connotations of dream material are inconsistent with waking mood, dreams are less likely to be remembered. Finally, according to the salience hypothesis, people do not recall dreams that lack vividness, emotionality, activity, and narrative coherence. It remains unknown whether these are competing explanations or whether each may contribute to an overall understanding of dream recall. Clarifying these possibilities may be especially important in traumatic conditions if, as is sometimes proposed, remembering dreams contributes to psychological adjustment (Brown & Donderi, 1987; Cartwright & Lloyd, 1994).

The present study explored the relations between traumatic events, dream recall, and psychological adjustment. It examined the role of repression, mood congruence, and salience on the frequency of dream recall, as well as the mental health role of dream recall among children and adolescents living in violent and in peaceful environments.

Trauma, Age, Gender, and Dream Recall

Studies of whether traumatic events increase or decrease the frequency of dream recall have provided contradictory results (for reviews, see Punamäki, 1997; Ross, Ball, Sullivan, & Caroff, 1989). Traumatic experiences have been shown to increase dream recall among American war veterans (Greenberg, Pearlman, & Gampel, 1972; Ross, Ball, Dinges, Kribbs, Morrison, Silver, & Mulvaney, 1994), bereaved older persons (Reynolds et al., 1992; 1993), and divorced women (Cartwright, 1983). Also, Terr (1983, 1991) observed increased dream recall, especially of dreams repeating the trauma scene, among children exposed to kidnapping. In contrast, evidence that trauma victims recall their dreams less than do controls has been obtained in studies of Israeli war veterans (Hefez, Metz, & Lavie, 1987), holocaust survivors (Kaminer & Lavie, 1991), and children living in violent and dangerous conditions (Nasef, 1993; Rofe & Lewin, 1982).

The severity and duration of traumatic events may be among the factors that explain these discrepancies. Repeated exposure to life-threatening events, such as occurs during incarceration in a concentration camp (Kaminer & Lavie, 1991) or exposure to shelling (Rofe & Lewin, 1980; 1982), has been associated with decreases in dream recall, whereas exposure to occasional and less severe trauma, such as major life changes, reportedly increase dream recall (Cartwright, 1986). Perhaps recurrent life-threatening events evoke different dream recall dynamics than occasional stressful events. Repeated exposure to severe trauma may exceed people’s capacity to cope, and, in these circumstances, failures of dream recall may be one aspect of the more general difficulty of adjusting to traumatic experiences (Turner & Gorst-Unsworth, 1990). Less persistent and severe trauma, in contrast, may enhance dream recall and provide opportunities for the kind of self-exploration that facilitates emotional-cognitive processing of stressful events.

To examine this possibility, the present study was designed to include one group of children and adolescents who lived in a persistently life-threatening and violent environment, and another group for whom occasional trauma occurred against a background of relative peace. Furthermore, children in the trauma group varied considerably, both in the frequency and type of trauma to which they were exposed. Although studies of the effects of traumatic events on children’s dream recall are rare, our earlier research suggested that an event that directly threatened a child’s own security (e.g., being imprisoned or threatened) was more traumatic than witnessing violence toward others (Qouta, Punamäki, & El Sarraj, 1997). Accordingly, it was expected that, in the trauma group, being a personal victim of violence would disrupt dream recall, whereas witnessing violence toward one’s family members might increase dream recall.

The study of children and adolescents, as undertaken in the present investigation, introduces another factor that may influence the relations between trauma and dream recall. Because dream structure is isomorphic with cognitive development (Foulkes, 1982; 1985) and a dreamer’s cognitive skills are related to dream recall (Butler & Watson, 1985; Cicogna, Cavallero, & Bosinelli, 1986; Foulkes, 1982; Foulkes, Hollifield, Bradley, Terry, & Sullivan, 1991), the child’s age may be an important determinant of dream recall. Thus, in this study we expected that older children would recall more dreams than would the younger ones, although we also examined whether traumatic living conditions might affect that relationship.

Because of evidence that gender influences dream recall (Foulkes, 1982; 1985; Tonay, 1993), perhaps especially under stressful conditions (Armitage, 1992), the present study was also designed to examine this factor. Girls and women report more dreams than boys and men (Hartmann, 1984; 1989; Rofe & Lewin, 1982), perhaps because women are more motivated to attend to and remember dream content (Tonay, 1993) or perhaps because they are more thin boundaried, i.e., sensitive, vulnerable, and imaginative (Hartmann, Elkin, & Garg, 1991; Cowen & Levin, 1995). Of particular relevance to the present study is evidence that stressful events increase dream recall among women and decrease it among men (Armitage, 1992). Therefore, although we expected girls to recall more dreams than the boys, we also explored whether traumatic living conditions might alter that relationship.

Repression and Dream Recall

The repression of shameful and painful dream material, such as that related to trauma, is often considered an important explanation for dream recall failure. According to psychodynamic theory, only after disturbing aspects of the original dream thoughts are represented in disguised or symbolic form are they allowed into consciousness. The dream work is the process by which these original contents are censored on their way toward the manifest dream (Adam-Silvan & Silvan, 1990; Freud, 1900/1953), although an insufficiently censored dream experience may be subjected to further repression by being forgotten upon awakening.

Empirical investigations of repression and dream recall failure are contradictory and conceptually problematic. Some correlational studies have indirectly confirmed the repression hypothesis by observing an association between infrequent dream recall and personality characteristics linked to repression, such as field dependence (Schonbar, 1965); other studies have failed to find a relation between diminished dream recall and such indicators of repression (Cohen & Wolfe, 1993; Tonay, 1993). Experimental studies of threatening and shameful pre-sleep stimuli also have produced conflicting results. Cartwright, Bernick, Borowitz, and Kling (1969) found that pre-sleep viewing of erotic films increased failures of dream recall, and Goodenough Witkin, Lewis, Koulack, and Cohen (1974) confirmed that laboratory stress increased the frequency of recall failure among field-dependent (i.e., repressive) but not among field-independent subjects. However, other studies have found no effect of stressful and aggressive stimuli on subsequent dream recall (Goodenough, Witkin, Koulack, & Cohen, 1975).

Hartmann (1991) has argued that, in various repressive ways, traumatized people create "thick boundaries" to protect themselves. Consistent with his formulation, Terr (1991) observed that children tended to forget not only their own behavior during the traumatic incident, but also their subsequent trauma-related symptoms. In some cases, the repression of painful memories of panic attacks, intrusive images, and nightmares became a generalized personality style, which, according to Terr, protected children’s feelings of invulnerability. Thus, generalized repressive strategies (i.e., thick boundariedness) may diminish the reported occurrence of anxiety dreams and nightmares following trauma (Kaminer & Lavie, 1991; Rofe & Lewin, 1982).

In an investigation that is particularly relevant to the present effort, Rofe and Lewin (1980; 1982) observed that a repressive personality style was associated with infrequent dream recall among children exposed to the shelling and violence of war. They argued that living in a traumatic environment motivated children to develop repressive response patterns, such as denial, distraction, avoidance, and numbing (Rofe & Lewin, 1982). Children adopting these strategies for psychologically distancing themselves from traumatic events also recalled fewer dreams. Perhaps especially among children who constantly live in dangerous and violent conditions, repressive styles are related to infrequent dream recall.

Thus, the repression hypothesis suggests that the more children use repressive coping strategies, such as denial, distraction, paralysis, and numbing, the more frequently they will forget their dreams. In the present investigation, we examined whether the association between repressive coping strategies and infrequent dream recall may be especially strong among children exposed to recurrent and severe traumatic events.

Mood Congruence and Dream Recall

Some dream researchers have challenged the notion that repression is a sufficient explanation for variations in dream recall. They argue that the rules shaping daytime memory similarly determine dream recall (Antrobus, 1991; Cicogna, Cavallero, & Bosinelli, 1986). For example, the mood congruent memory model suggests that, in general, people are more likely to remember events with affective connotations that correspond to their present mood (Blaney, 1986; Bradley & Mathews, 1988). Similarly, the mood state-dependent memory model maintains that, if people have learned something in a certain mood, they remember it best in a similar mood (Eich, 1989; 1995; Bartlett & Sandtrock, 1979). These models suggest that the difficulty of remembering dreams is due to discrepancies between the dream’s affective connotations (or perhaps the dream mood state) and the dreamer’s mood upon awakening in the morning.

Accordingly, in the present study, it was expected that when the affective connotations of dream content did not correspond with waking mood, dream recall would be diminished.

Salience and Dream Recall

Whereas the repression and mood congruent memory models emphasize the role of dreamed action and feeling in dream recall, Cohen’s (1979) salience hypothesis of dream recall suggests that the style of the dream narrative determines whether it is remembered. Salient dreams are those that are vivid (i.e., provide experiential detail), emotional (i.e., express positive or negative affect), bizarre (i.e., include strange or incredible elements), and involve dreamer activity (Cohen, 1979). Research has confirmed that the dream reports of frequent recallers are more vivid, emotional, and bizarre than those of non-recallers (Cohen & MacNeilage, 1974; Schredl, Kleinferchen, & Gell, 1996).

However, there is no consensus about the impact of traumatic experiences on dream salience, and, therefore, dream recall. Some research indirectly suggests that traumatic events increase dream salience, including emotional intensity (Cartwright, 1984) and the vividness with which aversive themes are incorporated (Lavie & Kaminer, 1991; Van der Kolk, Blitz, Burr, Sherry, & Hartmann, 1984). On the other hand, dreams following trauma often replicate the painful scene and are more realistic than bizarre. For instance, Vietnam veterans typically report repetitive anxiety dreams that exactly replicate actual combat events (Van der Kolk, Blitz, Burr, Sherry, & Hartmann, 1984).

These findings have direct implications for predictions based on the salience model. If children exposed to traumatic events have more vivid and emotional dreams, it would be expected that they would more frequently recall their dreams. On the other hand, if children exposed to trauma have less bizarre dreams, the salience model predicts that they would less frequently recall them. The present study was designed to explore these possibilities, as well as whether traumatic living conditions would affect the relations between dream salience and recall.

Dream Recall and Psychological Adjustment

Clarifying the determinants of dream recall in traumatic conditions is especially important if, as has been proposed, remembering dreams contributes to psychological adjustment. And yet, it remains uncertain whether dream recall has such a constructive effect. Correlational studies have documented both a positive relation (Hill, 1974; Kramer, Schoen, & Kinney, 1983) and the lack of any relation (Tonay, 1993) between dream recall and adjustment. Investigations of the mental health role of dream recall during periods of acute psychological distress have been more promising. For example, the frequent recall of recurrent dreams may be an indication of unresolved trauma (Cartwright & Romanek, 1978; Brown & Donderi, 1986). Brown and Donderi (1986) found that, when a psychological conflict had been resolved, recurrent dreams were reported less frequently than when a psychological conflict was ongoing. Moreover, past-recurrent dreamers not only were better adjusted than those currently experiencing recurrent dreams, but they also reported greater well-being than a control group. Recalling recurrent dreams during a conflict seemed to have a positive long-term mental health effect.

An increase in dream recall during psychological distress and a decrease afterwards may explain why Kaminer and Lavie (1991) found that well-adjusted holocaust survivors reported lower dream recall than controls. Characteristic of successful recovery is that survivors gradually integrate the traumatic experience into their present personality until eventually the disturbing affect associated with trauma-related memories is diminished. In these circumstances, dreams related to traumatic memories no longer provide the "emotional surge" that instigates spontaneous awakening and related dream recall (Kramer, 1993).

Although research to date does not clarify whether dream recall or dreaming per se serves a mental health function (cf. Cartwright, 1991), results to date are at least compatible with the notion that remembering dreams is an immediate response to current psychological distress that serves a long-term adaptive role in processing emotional information, solving problems, and integrating painful experiences (Breger, Hunter, & Lane, 1971; Hartmann, 1995). Accordingly, in the present study, it was expected that frequent dream recall would be associated with poor psychological adjustment, especially among children exposed to recurrent and severe traumatic events.


In summary, this research examined, first, whether dream recall, measured by diary reports, varies as a function of trauma severity, gender, and age. Second, we explored whether the repression, mood congruent memory, and dream salience models would predict the frequency of dream recall. Third, we examined whether frequent dream recall is associated with poor psychological adjustment, especially when children are exposed to recurrent and severe traumatic events.



The participants were 412 Palestinian children and adolescents, of whom 268 lived in a violent and dangerous environment in the Gaza strip and 144 lived in a peaceful area in Galilee, Israel. (The participants from Gaza will be referred to as the trauma group and those from Galilee as the comparison group.) Participants’ ages ranged from six to 15 years, with a mean 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 (F(1,411) = 3.09, ns). The trauma and comparison groups were similar in gender, c 2 = .006, df = 1, ns, n = 410, and in age, F(1,408) = 2.58, ns.

The trauma group lived in violent conditions of military occupation and nationalistic struggle against it. Gaza children typically had experienced military confrontations, curfews, political strikes, and restrictions on travelling, and they often had lost their family members through death or imprisonment (B’Tselem, 1993; Cohen & Golan, 1991). The Galilean group, even if belonging to the Palestinian nation, had not experienced any of this kind of violence since 1976, when "The Day of Land" demonstrations ended violently.

The children and adolescents in the trauma group were selected by using random and systematic sampling procedures (Pedhazur & 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 that were possible. Second, following a random start, every third house in each street was visited. The comparison group was selected in a similar way. The Galilean village was first divided into four areas in order to secure the representation of all social strata. In each area, following a random start, every third house on each street was visited. Both in Gaza and in Galilee, it was agreed that in one household all children who wanted to participate would be given the instructions and research materials.


The researcher, together with a Palestinian psychologist, contacted the Gaza children and adolescents in their homes in September and October of 1993. At the same time, in Galilee, a local psychologist contacted the children and adolescents belonging to the comparison group. There were no refusals to participate in the study among the families in either Gaza or Galilee. This favourable response apparently was due to the familiarity of the local field workers, the cultural code of hospitality, and curiosity about the study topic.

The field work proceeded in two stages. First, the children and adolescents were asked to participate in the study, and the questionnaires and dream diaries, with instructions, were given to them. Second, after seven days, the field workers returned and collected the completed dream diaries and questionnaires.

The instructions were written in a tightly-scripted and standardized statement in order to guarantee consistency from home to home. The dream diary and questionnaires were explained carefully, page by page, to the child or children, and standardized examples were given to demonstrate how to report dreams and respond to the questions. The children’s understanding of the explanation was checked by asking them to provide examples of how to fill in the questionnaires. The mothers of preliterate children filled in the diaries as dictated by the children themselves. This latter procedure was used with subjects who were less than eight years old (n=35).

Nearly all of the dream diaries were returned; only one child in Gaza and two in Galilee failed to do so. Of the total sample, four dream diaries were deleted due to missing information.


Traumatic Events Checklist

The Traumatic Events Checklist consisted of 15 items describing events that children and adolescents typically experienced during the Intifada, the national struggle for independence (Abu Hein, Qouta, Thabet, & El Sarraj, 1993; Summerfield, 1993). It asked about events such as night raids, the detention of family members, beatings, and being wounded. The participants were asked to check any event that they had experienced during the Intifada. The composite score from the checklist (range: 0 to 15) was more suitable for assessing the prevalence of traumatic experiences in the trauma group than in the comparison group. However, two items, death of a family member and being wounded (through violence or an accident), were relevant for both groups, and they were used as indicators of traumatic events in analyses of the entire sample.

In the trauma group, subscales reflecting different kinds of traumatic experience were created to assess (a) violence directed toward the children themselves, including being wounded, being imprisoned, and receiving threats to their lives, and (b) violence directed toward their families, including the death or imprisonment of a family member and the destruction of property.

All counted events occurred during the preceding six years of the Intifada, although we were unable to assess how recently each event had occurred. Earlier field work and pilot studies had indicated that, in a situation involving chronic violence, children and sometimes their parents found it difficult to estimate the timing of events.

Dream Diary: Dream Recall

A semi-structured diary was developed to record dreaming during a seven-day period. For each night, a separate sheet was available. On each morning, the participant was asked to record the dream or dreams recalled from the previous night. Each sheet began with, "Last night I dreamt that...", followed by 20 blank lines. Participants were also instructed to indicate whether they felt they had been dreaming but could not remember the content.

For each night, each sheet was scored as either no dream recall (scored 1) or dream recall (scored 2). There were four children who reported two or more dreams on at least one morning. Their multiple dream reports for each night were treated as a single narrative. Dream recall scores were summed across the seven nights so that the dream recall variable ranged between 7 (no dreams reported during the week) and 14 (a dream reported every morning). Participants reporting no dreams during the week were, of course, excluded from analyses that involve dream content variables.

Dream Reports: General

Dream reports were analyzed using scales adapted from those developed by Foulkes (1985) and Gottschalk, Winget, and Gleser (1969). Two research assistants, blind to group membership and unfamiliar with the research hypotheses, were trained to score dream content. All dream variables were coded using a qualitatively scored WP-program (Sulkunen & Kekäläinen, 1992) that made it possible to code recurrences of the different content categories. Since the reported dream could receive more than one score for each category, we tallied the frequency with which a content category was scored for each dream. The reliability of these scoring procedures was measured by assessing inter-rater agreement for the dreams of 70 randomly selected subjects (a total of 196 dreams). Kappa values ranged from 0.79 to 0.94, which can be considered good agreement (Fleiss, 1981). Details of the scoring procedures are available from the author.

Dream Reports: Salience

Salience of dream content was assessed using ratings of (a) bizarreness, i.e., unfamiliar, strange, or incredible scenes and actors; (b) vividness, i.e., visual, descriptive, and experiential detail; (c) activity, i.e., whether the dreamer was observing, participating in, or dominating dream events, (d) positive emotionality, i.e., a pleasant or joyful dream atmosphere with happy endings, (e) negative emotionality, i.e., life-threats, persecution, and frightening characters, and (f) narrative coherence, i.e., meaningfully related scenes without shifting actors and events.

The total number of dreams reported by a participant correlated, of course, with the overall frequency of dream content scores. Therefore, the dream content scores for all of a participant’s dreams were summed and divided by the number of dreams reported by that participant. The skewedness of each distribution was tested, and when excessive, summary scores were standardized by using arcsine transformations (Cohen & Cohen, 1983).

Dream Reports: Mood Congruence

Several mood congruence parameters were constructed to assess the level of correspondence between the affective connotations of the dream (dream atmosphere and dream feelings) and morning mood during the seven day study period. Dream atmosphere scores were (a) pleasant and joyful and (b) unpleasant and unfortunate. Dream feelings scores were (a) fear and worry, (b) anger and rage, (c) sadness and loss, and (d) happiness and joy. The seven-day dream diary also included structured ratings of morning mood, adapted from Russel (1980). Children were asked to evaluate their mood in the morning when waking up. They were given a list of seven feelings (calm, happy, sad, afraid, angry, worried, and excited) and asked to estimate: "When waking up I was feeling _______" (1) not at all, (2) a little, (3) very much.

A mood-congruence index was created by calculating correlations between dream atmosphere/feeling scores and morning mood ratings across the seven days. Correlations were calculated for each of the following paired scores and ratings: (1) pleasant and joyful dream atmosphere with calm morning mood, (2) happy and joyful dream feelings with happy morning mood, (3) dream feelings of sadness and loss with sad morning mood, (4) fear and worry dream feelings with afraid and worried morning moods, (5) anger and rage dream feelings with angry morning mood, and (6) unpleasant and unfortunate dream atmosphere with excited morning mood. These correlations were transformed into z-scores prior to statistical analyses.

Repressive Coping Strategies

Repressive coping strategies were assessed using an unfinished sentences task that involved seven stressful events selected from a total of 25 that had been adapted from Rotter and Rafferty (1950). To complete this semi-projective test, children were instructed to continue the incomplete sentences with the first thoughts that come to mind. All the unfinished sentences referred to scenes of military violence: Somebody’s house is destroyed by soldiers . . . ; I see a soldier . . . ; Somebody is taken into the prison...; People are under curfew...; I hear that there are demonstrations . . .; I hear about the shooting of my people . . .

The children’s sentence completions were classified in terms of 24 behavioral, cognitive, and emotional categories of coping (Pearlin & Schooler, 1978; Punamäki & Suleiman, 1990; Punamäki & Puhakka, 1997). Categories for the behavioral level of coping ranged from paralysis to activity, for the cognitive level of coping from denial and defence to problem solving, and for the emotional level of coping from numbing to courageous coping. From these primary categories, those reflecting repressive tendencies were selected and summed across the seven stressful situations. These included: paralysis (e.g., I am like a dead man; I am staring at him), denial of the occurrence and importance of stress and related feelings (e.g., I did not see anybody being wounded; I feel angry but put on a brave face and appear indifferent), distraction (e.g., I isolate myself and do not think about it; I do not dare to look and close my eyes), and numbing (e.g., I do not feel anything; I feel strange).

Reliability of all of the primary coping scores was assessed on a sample of 1024 responses to six unfinished sentences. Over-all inter-rater agreement was .82, suggesting strong agreement between the two judges (Fleiss, 1981). As an indication of the construct validity of this repressive coping measure, our earlier research showed that repression increased with age, which concurs with developmental theories (Punamäki & Puhakka, 1997).

The repressive coping style scores correlated positively (.16-.33), and, therefore, were submitted to a principal components analysis. A single-factor solution was adopted (accounting for 39% of the variance), and the factor scores from this solution were used to indicate repressive coping.

Psychological Symptoms Scale

The 34-item Psychological Symptoms Scale asked participants to indicate whether they suffered from each listed symptom often (scored 1), sometimes (scored 2), or never (scored 3). The scale includes cognitive, somatic, anxiety, depressive, and aggressive symptoms (Macksound, Aber, Dyregrov, & Raundalen, 1990; Eth & Pynoos, 1985). To examine the dimensionality of the scale, the scores were subjected to principal components analysis with an orthogonal rotation. A five-factor solution (explaining 39% of the variance) was adopted. Subscales based on these factors, as well as a total psychological symptoms variable, were used in the analyses. The reliability (Cronbach’s alpha) of the total psychological symptoms score was .88. The interpretation of each factor follows:

I. Depressive Symptoms: sad, lonely, and desperate feelings; physical signs of depression (e.g., I feel terribly sad and hopeless; I have lost my appetite). This scale included eight items, its reliability was .76, and the factor accounted for 18.4% of the total variance.

II. Somatic and Anxiety Symptoms: emotional and somatic indicators of anxiety; the intrusion of trauma-related fears (e.g., I am feeling sick and worried; I have pains all over my body; I am constantly thinking of bad and frightening things). This scale included nine items, its reliability was .74, and the factor accounted for 6.7% of the total variance.

III. Sleeping Difficulties: nightmares and bad dreams; fears disturbing sleep during the night (e.g., I often wake up in the middle of the night; I am afraid of darkness). This scale included six items, its reliability was .64, and it accounted for 5.5% of the total variance.

IV. Cognitive Problems: level of school performance, reading, and writing; difficulty concentrating and remembering (e.g., I get bad marks at school; I have difficulty concentrating). This scale included six items, its reliability was only .58, and it accounted for 4.6% of the total variance.

V. Aggressive Symptoms: difficulty controlling hostility; destructive behavior; quarrelling and fighting with adults and peers (e.g., I feel like hitting somebody; I have fights with other kids). This scale included five items, its reliability was only .57, and it accounted for 4.0% of the total variance.


Descriptive Statistics

Table 1 presents, for the trauma group in Gaza and the comparison group in Galilee, the proportion of boys and girls, the percentage of children experiencing traumatic events, the percentage of children reporting no dream recall, the percentage of children reporting recall every night, and the means and standard deviations for dream recall, repressive tendencies, psychological symptoms, dream salience, and mood congruence.

The comparison group from Galilee lived in a peaceful environment and was, as expected, not as extensively exposed to traumatic experiences as the group from Gaza. The death of a close person and being wounded were the only equivalent traumatic events in the two groups. Results showed that 12% of the trauma group and 6% of the comparison group reported losing a family member through death. The groups did not, however, differ in how frequently they were wounded (through violence or accident).

Age and gender were related to the frequency of exposure to traumatic experiences in the trauma (Gaza) group. Boys reported such exposure more frequently than girls, F(1,264) = 12.42, p = .001, and older children and adolescents reported such exposure more often than did younger ones, F(2,264) = 5.35, p = .005. An age by gender interaction indicated that traumatic events increased with age only among boys, F(2,264) = 5.13, p =.007.

In the entire sample, 10.2% (n = 42) reported no dreams at all and 12.8% (n = 53) reported dreams on each of the seven nights of the study period. The groups differed in levels of dream recall, the trauma group remembering their dreams more frequently than did the comparison group. In the trauma group, 7.8% and in the comparison group 14.6% of the children did not report any dreams during the seven nights, while 19% in the trauma and only 1.4% in the comparison group reported dreams every night.

The trauma group coped more frequently by using the repressive styles of paralysis, denial, and numbing than did the comparison group. On the other hand, the trauma and comparison groups differed little from each other in the overall prevalence of psychological symptoms, although the trauma group reported more sleeping difficulties than did the comparison group.

The dreams of children in the trauma group included more negative emotionality and less narrative coherence than the dreams of the comparison group. Finally, few differences between the groups were found in the mood congruence measures. Only the correlation between unpleasant dreams and excited morning mood was stronger in the trauma than in the comparison group.

Trauma, Gender, Age, and Dream Recall

A 2 (trauma vs. comparison group) by 3 (6-8, 9-12, vs. 13-15 year-olds) by 2 (gender) between-subjects ANOVA was carried out on the frequency of dream recall. Results showed that traumatic environment, F(1,409) = 36.97, p = .0001, and gender, F(1,409) = 9.44, p = .002, had main effects on dream recall. The children in the trauma group remembered more dreams than the comparison group, and girls remembered more dreams than boys. The interaction between group, age, and gender, F(2,409) = 3.27, p = .039, revealed that older girls recalled their dreams more frequently in the trauma group especially.

select for 
Table 1. 

Descriptive Statistics in the Trauma and Comparison Groups

To examine these same relationships in the trauma group alone, a 3 (low [1-5 events], medium [6-10], vs. high [11-15] trauma levels) by 3 (6-8, 9-12, vs. 13-15 years-olds) by 2 (gender) between-subjects ANOVA was carried out on the frequency of dream recall. Results confirmed that children reporting high levels of trauma exposure recalled their dreams more frequently than did those with low levels of exposure F(2,264) = 4.15, p = .017, and that girls recalled dreams more frequently than boys, F(1,264) = 8.66, p = .004. The gender by traumatic events interaction indicated further that the level of traumatic events was differently associated with dream recall among boys and girls, F(2,264) = 4.31, p = .014. As indicated in Figure 1, among the girls increased dream recall was especially evident when they were exposed to high levels of traumatic events. Among the boys a medium level of trauma exposure already increased their dream recall, but no further increase was associated with high levels of trauma exposure.

select for
Fig 2. 

Relationships between (a) child-targeted and
 (b) family-targeted traumatic events 
on the dream recall among boys and girls in the trauma (Gaza) group.

To examine whether being a personal victim of violence or witnessing violence among others differentially influenced dream recall, two separate ANOVAs on dream recall frequency were performed for the trauma group only, one using child-targeted and the other family-targeted traumatic event scores as independent variables. Results indicated that, contrary to expectations, both family-targeted events, F(2,264) = 2.78, p = .064, and child-targeted events, F(2,264) = 2.53, p = .082, marginally increased dream recall. Further analysis showed, however, that the gender of the child was decisive in determining the association between type of traumatic event and dream recall. As indicated in Figure 2a, child-targeted violence increased dream recall only among boys, F(2,147) = 3.15, p = .041. As shown in Figure 2b, family-targeted violence increased dream recall only among girls, F(2,118) = 4.64, p = .012.

select for
Fig 1.

Effect of the interaction between gender and traumatic events 
on dream recall in the trauma (Gaza) group.


Repression, Trauma, and Dream Recall

Tests of the repression, mood congruence, and salience hypotheses were first performed for the entire sample (combining trauma and comparison groups), using the sum of experiences with death and with wounding as an index of trauma exposure. Then the models were tested again for the trauma group only, using the level of personal exposure (low [1-5 events], medium [6-10], vs. high [11-15] trauma levels) as an index of trauma.

select for 
Fig 3. 

Effect of the interaction between repression and traumatic events on dream recall among the Gaza boys.

To examine whether repressive coping strategies are associated with infrequent dream recall in general, and especially among children exposed to high levels of trauma, a multiple regression analysis was performed using (a) the repressive coping score, (b) traumatic events involving death and wounding, (c) the interaction (cross product) between repressive coping and traumatic events, (d) age, and (e) gender as independent variables, and dream recall frequency as the dependent variable.

The regression model for the entire sample explained 8% of the variation in dream recall, F(5,404) = 7.40, p = .0001. Contrary to expectations, the more children used repressive coping strategies of paralysis, denial, numbing, and distraction, the more frequently they remembered their dreams (beta =.23, t = 4.81, p = .00001). Girls more frequently recalled their dreams than did the boys (beta = .13, t = 2.70, p = .007), but experience with death and wounding did not predict dream recall (beta = .05, t = 1.07 p =.ns). The traumatic events by repression interaction also was non-significant (beta = .03, t =.54, p =.ns), indicating, contrary to the hypothesis, that the association between repressive coping and dream recall did not differ according to the level of exposure to death and wounding.

Results for the trauma group alone showed that the model explained 13% of the variation in dream recall, F(5,259) = 7.90, p = .0001. Again, contrary to expectations, the more children coped by using repressive strategies, the more frequently they remembered their dreams (beta = .25, t = 4.21, p = .00001). The model also confirmed that girls recalled their dreams more frequently than did boys (beta = .18, t = 2.98, p = .003), and, unlike in the analysis of the entire sample, high levels of trauma exposure were associated with elevated dream recall (beta = .17, t = 2.80, p = .005). The traumatic events by repression interaction effect was non-significant (beta = .09, t = 1.58, p = .11), indicating again, contrary to the hypothesis, that the association between repressive coping and dream recall did not differ between children exposed to high and low levels of traumatic events in the Gaza group.

As gender seemed important in shaping children’s ways of responding to traumatic events, we tested the preceding regression models on dream recall separately for boys and girls in the trauma group (see Table 2). A significant interaction effect between traumatic events and repressive coping was found among boys. As shown in Figure 3, repression is associated with strikingly elevated dream recall among boys who live in a violent and dangerous environment and who are exposed to high levels of traumatic events. The regression model for girls showed that for them the association between repression and frequent dream recall was not dependent on the level of exposure to traumatic events.

select for 
Table 2. 

Moderating Effects of Repression on Traumatic Events and Dream Recall Association 
Among Boys and Girls in the Trauma (Gaza) Group.


Mood Congruent Memory and Dream Recall

To examine whether the correspondence between dream mood (atmosphere and feelings) and morning mood would predict high dream recall, a regression analysis was performed using (a) the mood-congruent memory parameters, (b) the level of exposure to death and wounding, (c) age, and (d) gender as independent variables, and dream recall frequency as the dependent variable.

As indicated in Table 3, and in accordance with the hypothesis, high positive correlations between dream mood and morning mood significantly predicted dream recall frequency. The model explained 16% of the variation in dream recall, F(9,365) = 7.63, p = .0001. The significant beta values revealed that correspondence between unpleasant dreams and excited morning mood and correspondence between fearful and worrying dreams and fearful morning mood predicted dream recall.

The same regression model applied to the trauma group alone explained 23% of the variation in dream recall, F(9,240) = 8.02, p = .0001. It further confirmed the mood congruence hypothesis, indicating that correspondence between unpleasant dreams and excited morning mood (beta = .33, t = 5.13, p = .0001; r = .53) and correspondence between sad dreams and sad morning mood (Beta =.11, t =1.90, p =.037; r = .03) predicted frequent dream recall in the trauma group.

select for 
Table 3.
Regression Model for Mood-congruent Memory Parameters 
on Dream Recall in the Entire Sample


Salience and Dream Recall

To test the salience hypothesis with the entire sample, a multiple regression analysis was performed using dream recall as the dependent variable and the salience scores for (a) bizarreness, (b) vividness, (c) positive emotionality, (d) negative emotionality, (e) narrative coherence, and (f) dreamer activity, as well as (g) traumatic exposure to death and wounding, (h) age, and (i) gender as independent variables. Also, (j) the interaction between traumatic events and emotionality (negative), and (k) the interaction between traumatic events and bizarreness were added as independent variables (see Table 4).

The regression model for dream salience explained 52% of the variation in dream recall for the whole sample, F(11,305) = 29.68, p = .0001. In accord with the hypothesis, results showed that the more bizarre, vivid, emotionally positive, active, and narratively coherent children’s dreams were, the more frequently they recalled them.

The non-significant interaction terms between traumatic events and emotional intensity and between traumatic events and bizarreness indicate, contrary to expectations, that the associations between emotional intensity, bizarreness, and dream recall frequency did not vary according to the level of exposure to death and wounding.

Results for the trauma group alone showed that the same model explained 47% of the variation in dream recall, F(11,250) = 21.06, p = .00001. The results further confirmed the salience hypothesis by revealing that children remembered their dreams more frequently if they were emotionally negative (beta = .40, t = 8.34, p = .00001), active (beta = .26, t = 4.87, p = .001) bizarre (beta = .25, t = 4.91, p = .001), and vivid (beta = .16, t = 3.33, p = .001). Again the traumatic events by bizarreness and traumatic events by negative emotionality interaction effects were non-significant (beta = .06, t = 1.26, p = .27; and beta = .01, t = 0.24, p = .81, respectively), indicating, contrary to the hypothesis, that the association between salient dream characteristics and dream recall did not differ with exposure to traumatic events in the Gaza group.

select for 
Table 4. 

Regression Model for Salient Dream Characteristics
 on Dream Recall in the Entire Sample.


Traumatic Events, Dream Recall, and Psychological Adjustment

To test the hypothesis that frequent dream recall is associated with poor psychological adjustment, especially among traumatized children, a series of multiple regression analyses were performed in which (a) the frequency of dream recall, (b) exposure to death and wounding, (c) the interaction (cross product) between dream recall and trauma exposure, (d) age, and (e) gender were the independent variables, and depression, somatic and anxiety symptoms, sleeping difficulties, cognitive problems, aggressive symptoms, and the total sum of symptoms were successively used as the dependent variables.

Table 5 presents the significant regression models in this series of analyses for the entire sample. As expected, results showed that the more frequently children remembered their dreams, the more they suffered from psychological symptoms, especially from somatic and anxiety symptoms. The regression models also revealed that older children reported more depressive, somatic and anxiety symptoms, but fewer sleeping difficulties than younger children. Exposure to death and wounding was positively associated only with somatic and anxiety symptoms.

select for 
Table 5. 

Significant Multiple Regression Analysis on 
Psychological Symptoms in the Entire Sample.


The significant trauma exposure by dream recall interactions indicated, again in accord with the hypothesis, that the association between exposure to trauma and psychological symptoms depended on the frequency of dream recall. The interaction effects between traumatic experiences and dream recall on depressive symptoms and on somatic and anxiety symptoms are presented graphically in Figure 4a and Figure 4b. They show that experiences of death and wounding were more markedly associated with depressive symptoms among children who frequently recalled their dreams. On the other hand, exposure to these traumatic events was more markedly associated with somatic and anxiety symptoms among children who infrequently remembered their dreams. In other words, infrequent dream recall protected children from developing depressive symptoms when they were exposed to experiences of death and wounding, but made them more susceptible to somatic and anxiety symptoms.

In the trauma group alone, the regression model for somatic and anxiety symptoms explained 13% of the variation, F(5,248) = 7.48, p = .0001). In accord with the hypothesis, somatic and anxiety symptoms were more severe among children who frequently recalled their dreams (beta = .12, t = 2.01, p = .045). A non-significant trauma exposure by dream recall interaction indicated that, contrary to the hypothesis, the association between exposure to traumatic events and total psychological symptoms did not differ according to the frequency of dream recall. However, the more children in the trauma group were personally exposed to traumatic events, the more they suffered from somatic and anxiety symptoms (beta = .26, t = 4.15, p = .0001).

Traumatic events were more important than dream recall in predicting the total of psychological symptoms (F(5,248) = 2.37, p = .039; the model explained 8%) and depressive symptoms (F(5,251) = 3.00, p = .011; the model explained 5%) in the Gaza group. The more traumatic events children had, the more they suffered from all psychological (beta = .25, t = 3.76, p = .0002) and depressive (beta = .16, t = 2.40, p = .016) symptoms.

select for 
Fig 4. 

Effect of the interactions between dream recall and trauma events on 
(a) depressive and 
(b) somatic and anxiety symptoms 
in the entire sample.



The present results provide evidence that exposure to recurrent and severe trauma is associated with elevated dream recall. Children living in a persistently violent and dangerous environment in Gaza remembered their dreams more frequently than those living in a peaceful area in Galilee. Also, within the Gaza group, recurrent personal exposure to experiences such as loss of a family member, destruction of one’s house, and being detained was associated with an increased dream recall.

We also explored whether certain types of trauma would decrease dream recall and other types increase it. We assumed that being a personal victim of violence constituted a more serious trauma than witnessing violence toward others, and expected that personal exposure to violence would decrease dream recall, whereas witnessing violence might increase it. Results indicated instead that, although both child- and family-targeted trauma marginally increased dream recall, these effects were gender specific. Among boys, violence targeted toward the child was predictive of frequent dream recall, while, among girls, witnessing violence targeted toward others was associated with frequent dream recall. These findings confirm research with adults that suggest gender differences in processing stressful experiences (Armitage, 1992), although clearly the nature of the stressful experience must be considered more carefully.

Gender-specific dreaming habits and the different political roles of Palestinian boys and girls may explain the present results. Girls have been found to dream frequently about familiar people and human relationships (Foulkes, 1985), to be interested in interior and home issues, and to worry about and care for other people (Lytton & Romney, 1991; Maccoby, 1988). Consequently, girls may experience violence and humiliation directed toward their family members as especially threatening to their security. For them, family-targeted violence was especially impactful, as reflected in increased dream recall. In contrast, the boys participated more actively in the Intifada fighting than did the girls (Kostelny & Garbarino, 1994; Kuttab, 1988; Qouta, Punamäki, & El Sarraj, 1995), and they were themselves detained and wounded more often. Such personal ordeals were apparently more disturbing for the boys, perhaps prompting their incorporation into highly memorable dreams.

There is little available research on the gender-specific effects of traumatic events on dreams and dream recall. Cartwright’s (1983; 1984; 1986) work on divorce and dreaming focused on women, and research on war veterans’ dreaming habits has focused on men (Ross, Ball, Sullivan, & Caroff, 1989; Ross, Ball, Dinges, Kribbs, Morrison, Silver, & Mulvaney, 1994). The systematic comparison of how different types of trauma affect dream recall among men and women is required in future research.

The general result that girls reported more dreams than boys is consistent with previous studies, but, examined more closely, the present data indicate that the pattern of the relationship between traumatic events and dream recall also differed according to gender. Among boys, increased dream recall was evident at moderate levels of exposure to traumatic events, while, among girls, only at high levels of exposure to traumatic events was there an increase in dream recall. Thus, boys’ dream recall seems to be more responsive to moderate levels of trauma exposure, without further increases at higher trauma levels. In contrast, girls’ dream recall seems especially responsive to high levels of trauma exposure. Perhaps, in extremely traumatic conditions, girls tend to remember and boys to forget their dreams.

The hypothesis that older children would recall more dreams than younger children was not generally substantiated in this study. Only among girls in the trauma group, older children recalled more dreams than did the younger children. The rule that cognitive maturation allows children increasingly to remember and narrate their dreams (Foulkes, 1985) was here gender- and environment-specific.

Explanations for Dream Recall

The present results confirmed the salience and mood congruent memory hypotheses of dream recall, indicating that both dream style or structure and dreaming-waking relationships are important in explaining the remembering of dreams. However, the repression hypothesis was not confirmed.


The dreams of frequent recallers involved, as the salience hypothesis suggests, vivid experiential detail, narratively coherent themes, intense emotions, dreamer activities, and more bizarre content than the dreams of infrequent recallers. One may cautiously suggest that, both among adults (Cipolli, Bolzani, Cornoldi, De Beni, & Fagioli, 1993) and children, similar principles explain dream recall and waking memory.

However, contrary to our expectations, the associations between bizarreness and dream recall and between negative emotionality and dream recall were unaffected by levels of exposure to traumatic events. We observed that, in the Gaza group, the more children were exposed to trauma, the more negatively emotional (r = .18, p = .017, N=222) and the less bizarre (r = -.14, p = .051, N=222) were the dreams that they reported. A similar pattern has been documented among adults: trauma victims often report dreams in which they, in some sense, realistically relive an emotionally painful experience (Kaminer & Lavie, 1991; Van der Kolk et al., 1984). However, in the present study, regardless of exposure to trauma, emotionality and bizarreness were both associated with increased dream recall. This introduces an interesting tension between the qualities of trauma related dreams that facilitate recall (e.g., their emotional intensity) and those that disrupt it (e.g., their lack of bizarreness) It is evident that we need to refine our characterization of the association between trauma-related dream contents and dream recall.

Mood Congruence

There was closer correspondence between dream mood (atmosphere and feelings) and waking mood among frequent dream recallers than among infrequent recallers. Moreover, congruence between negative dream content and negative morning mood was more powerful in predicting dream recall than congruence between positive dream content and positive morning mood. Kramer (1993) similarly found a closer association between negative dream characteristics and negative morning mood than between positive dream characteristics and positive morning mood. Perhaps negative dreams have a more powerful effect on morning mood and, consequently, they more strongly prompt mood congruent dream recall.

The congruence between dream mood and morning mood may be especially robust among people exposed to traumatic experiences. That possibility is consistent with the observation that the sleep of traumatized people is lighter and more "transparent" (Dagan, Lavie, & Bleich, 1991) and that they show general alertness. Terr (1981, 1983) observed that traumatized children intensely feared that further traumas would happen, which also intensified other apprehensions, such as fear of strangers, darkness, and being alone. Further research is needed to examine whether the congruence between the mental processes during sleep and those that follow awakening is greater among trauma victims and whether such congruence might explain their more frequent dream recall. Some preliminary evidence for this proposal comes from the present study: unpleasant dreams correlated more strongly with negative morning mood in the trauma than the comparison group (see Table 1).


Relationships between repressive tendencies and dream recall contradicted the repression hypothesis. It was expected that children who frequently use repression when coping with stress would frequently forget their dreams, because both tendencies serve to attenuate painful feelings and memories (Hartmann, 1991; Rofe & Lewin, 1982). Results indicated, however, that the more children used repressive coping strategies, such as paralysis, denial, distraction, and numbing, the more frequently they recalled their dreams.

These results suggest a kind of compensation dynamic between waking stress responses and dream recall (Koulack, 1993). Children who felt paralyzed in their actions, denied danger, and were affectively numb during the daytime tended to re-experience painful events and the associated emotions in "a nighttime safe place," as Hartmann (1995) puts it. The elevated dream recall among children with repressive tendencies suggests such a compensatory dynamic. Their memorable dreams may involve trauma-related feelings and images that they could not afford to fully experience during the day.

This idea was indirectly confirmed by results showing that the association between repression and frequent dream recall was especially strong among boys who had personally suffered through traumatic events and who were living in the dangerous and violent environment in Gaza. These boys were the "heroic youngsters" participating in the Intifada, the national struggle for independence, but they had also lost their family members and homes and witnessed destruction, humiliation, and violence. Their experiences were thus good candidates for repression. Their attempt to develop repressive strategies, thick boundaries, and a heroic attitude may have helped to attenuate painful experiences during the day, but different dynamics affected their dreams. The overwhelming pain that they repressed during the day could be ventilated only in the memorable dreams that emerged at night.

Research confirms that children refuse to recall bad experiences and to express feelings in dangerous and threatening conditions (Rofe & Lewin, 1982; Terr, 1983). They tend to delay and ration painful and frightening emotions. Processing them in dreams may be one way of rationing these overwhelmingly distressing experiences. Our results confirmed that the trauma group from Gaza coped with stress by using more denial, numbing, and paralysis than did children from the peaceful area of Galilee (Table 1). Our results challenge, however, the notion that the same repressive tendency is consistently present during waking and sleeping.

The present exploration of these issues can be faulted for relying on only one dimension of children’s coping strategies as an indication of repression. The measurement procedures used here involve defences that can be understood as less severe and more flexible than repression proper. For instance, paralysis and numbing may involve conscious, resourceful, and even creative ways of mastering extreme stress. Children may elastically employ them according to environmental demands as a way to neutralize intense affect, including affect related to dreams. Such neutralization can occur without repression per se, and thus without sacrificing dream recall. Repression per se, in contrast to the defences studied here, is considered a primitive defence, characterized by failure to remember painful events (Andrews, Singh, & Bond, 1993), including dreams.

To conclude, children who frequently recalled their dreams were those (a) who were exposed to traumatic events, (b) who coped with stress by using repressive strategies of paralysis, denial, numbing, and distraction, (c) whose negative mood in the morning was congruent with unpleasant dream atmosphere and feelings, and (c) whose dreams involved dreamer activity and were vivid, emotional, narratively coherent, and bizarre. These results support an integrated explanation for remembering dreams, appealing both to aspects of the recall process (mood congruent memory) and to characteristics of dream content and structure (salience).

Dream Recall and Adjustment

The hypothesis that frequent dream recall is associated with poor psychological adjustment was substantiated in this study. The more frequently children recalled dreams, the more they suffered from psychological symptoms, especially depressive or somatic and anxiety symptoms. (Sleeping difficulties, on the other hand, were explained by gender and age, older children and girls suffering them more frequently.)

We expected that frequent dream recall would be associated with poor adjustment especially among children suffering through severe traumatic events. The results were, however, symptom specific, indicating both a moderating and an intensifying effect of dream recall. Frequent dream recall intensified the impact of death and wounding on depressive symptoms, whereas they moderated the effects of these trauma on anxiety and somatic symptoms. One possibility is that frequent dream recall marks the expression of feelings associated with experiences of loss. Such dream-induced ventilation may, therefore, elevate depressive responses but nonetheless curb somatic and anxiety symptoms, which are more ambiguous and problematic responses to trauma.

The results for the Gaza children confirm earlier research showing that elevated dream recall accompanied ongoing and acute trauma (Brown & Donderi, 1987; Cartwright & Lloyd, 1994). Frequent dream recall was associated with poor mental health and apparently did not moderate the impact of personal exposure on any of the symptoms. However, the present cross-sectional study does not allow any conclusion about whether frequent dreaming would serve a positive long-term mental health function in the course of recovery from these traumas. Research has shown that living one’s childhood in the context of war, national struggle, and political violence moulds intrapsychic processes, such as dreaming (Masalha, 1994; Nashef, 1993; Rofe & Lewin, 1982). Yet, more detailed analysis of the interaction between the nature of trauma and cognitive-emotional processes affecting children’s long-term adaptation is essential.

The present research design also requires that causal interpretations be cautiously entertained. For instance, rather than the notion that dream recall compensates for the effects of repressive tendencies, one may as plausibly suggest that the remembering of distressing dreams motivates children to develop repressive strategies in order to cope with the effects of a traumatic environment. Moreover, this research depends upon children’s diary reports of their dreams. The results, therefore, may not be readily compared with research based on polysomnographic recordings with adults. Nonetheless, except for studies by Rofe and Lewin (1982) and by Nasef (1993), there are few investigations of the dreams of children living in traumatic conditions. Moreover, although tentatively offered the compensatory dynamics suggested to explain associations between repressive coping and frequent dream recall, and the observed symptom-specific effects of dream recall on depression and on somatic and anxiety symptoms warrant further study. They may describe important aspects of psychological adjustment among children who suffer from trauma or constantly live in violent and dangerous circumstances.



Abu Hein, F., Qouta, S., Thabet, A., & El Sarraj, E. (1993). Trauma and mental health of children in Gaza. British Medical Journal, 306, 1129.

Adam-Silvan, A., & Silvan, M. (1990). ‘A dream is the fulfilment of a wish’: Traumatic dream, repetition compulsion, and pleasure principle. International Journal of Psycho-Analysis, 71, 513-522.

Andrews, G., Singh, M., & Bond, M. (1993). The defence style questionnaire. Journal of Nervous and Mental Disease, 181, 246-256.

American Psychiatric Association (1987). Diagnostic and Statistical Manual of Mental Disorders (3rd edn, revised; DSM-III-R). Washington, DC: APA.

Antrobus, J. (1991). Dreaming: Cognitive processes during cortical activation and high afferent thresholds. Psychological Review, 98, 96-121.

Armitage, R. (1992) Gender differences and the effect of stress on dream recall: A 30-day diary report. Dreaming, 2, 137-141.

Bartlett, J. C., & Santrock, J. W. (1979). Affect-dependent episodic memory in young children. Child Development, 50, 513-518.

Blaney, P. H. (1986). Affect and memory: a review. Psychological Bulletin, 99, 229-246.

Bradley, B. P., & Mathews, A. (1988). Memory bias in recovered clinical depressives. Cognition and Emotion, 2, 235-245.

Breger, L., Hunter, I., & Lane, R. W. (1971). The effect of stress on dreams. Psychological Issues. VII (3), Monograph 27.

Brown, R. J., & Donderi, D.(1986). Dream content and self-reported well-being among recurrent dreamers, past-recurrent dreamers, and nonrecurrent dreamers. Journal of Personality and Social Psychology, 50, 612-623.

B’Tselem (1993). Human rights violations in the occupied territories 1992/1993. Jerusalem: B’Tselem.

Buttler, S. F., & Watson, R. (1985). Individual differences in memory for dreams: The role of cognitive skills. Perception and Motor Skills, 61, 823-828.

Cartwright, R. (1983). Rapid eye movement sleep characteristics during and after mood-disturbing events. Archives General Psychiatry, 40, 197-201.

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

Cartwright, R. (1986). Affect and dream work from an information processing point of view. Journal of Mind & Behavior, 7, 411-427.

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., Bernick, N., Borowitz, O., & Kling, A. (1969). Effects of an erotic movie on the sleep and dream of young men. Archieves of General Psychiatry, 20, 263-271.

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

Cartwright, R. D., & Lloyd, S.R. (1994). Early REM sleep: A compensatory change in depression? Psychiatry Research, 51, 245-253.

Cartwright, R. D., & Romanek, I. (1978). Repetitive dreams of normal subjects. Sleep Research, 7, 174.

Cicogna, P., Cavallero, C., & Bosinelli, M. (1986). Differential access to memory traces in the production of mental experience. International Journal Psychological Physiology, 4, 209-216.

Cohen, D.B. (1979). Sleep and dreaming: Origins, nature and functions. New York: Pergamon.

Cohen, J., & Cohen, P. (1983). Applied Multiple Regression/ Correlation Analysis for the Behavioral Sciences. Hillsdale, NJ.: Lawrence Erlbaum Associates, Publishers.

Cohen, S., & Golan, D. (1991). The interrogation of Palestinians during the intifada: Ill-treatment, "moderate physical pressure" or torture? Jerusalem: B’Tselem.

Cohen, D.B., & MacNeilage, P.F. (1974). A test of the salience hypothesis of dream recall. Journal of Consulting and Clinical Psychology, 42, 699-703

Cohen, D.B., & Wolfe, G. (1973). Dream recall and repression: Evidence for an alternative hypothesis. Journal of Consulting and Clinical Psychology, 41, 349-353.

Cowen, D., & Levin, R. (1995). The use of the Hartmann boundary questionnaire with an adolescent population. Dreaming, 5, 105-114.

Dagan, Y., Lavie, P., & Bleich, A. (1991). Elevated awakening thresholds in sleep stage 3-4 in war-related post-traumatic stress disorder. Biological Psychiatry, 30, 618-622.

Eich, E. (1989). Theoretical issue in state dependent memory. In H.L. Roediger, III & F.I.M. Craik (Eds.), Varieties of memory and consciousness. Essays in honour of Endel Tulving (pp. 331-354). Hillsdale: Lawrence Erlbaum.

Eich, E. (1995). Mood as a mediator of place dependent memory. Journal of Experimental Psychology: General, 124, 293-308.

Eth, S., & Pynoos, R. (1985). Post-traumatic stress disorder in childhood. Washington, D.C.: American Psychiatric Association.

Fleiss, J.L. (1981) Statistical methods for rates and proportions. 2nd ed. New York: Wiley & Sons.

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., Hollifield, M., Bradley, L., Terry, R., & Sullivan, B. (1991). Waking self-understanding, REM-dream self representation, and cognitive ability variables at ages 5-8. Dreaming, 1, 41-51.

Freud, S. (1900). The interpretation of dreams. The Standard edition, 4 and 5. London: Hogarth Press, 1953.

Greenberg, R., Pearlman, C., & Gampel, D. (1972). War neuroses and the adaptive function of REM sleep. British Journal of Medical Psychology, 45, 27-33.

Goodenough, D. R., Witking, H. A., Koulack, D., & Cohen, H. (1975). The effects of stress films on dream affect and on respiration and eye-movement activity during rapid-eye-movement sleep. Psychophysiology, 12, 313-320.

Goodenough, D. R., Witkin, H. A., Lewis, H. B, Koulack, D., & Cohen, H. (1974). Repression, interference of field dependence as factors in dream forgetting. Journal of Abnormal Psychology, 83, 32-44.

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.

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

Hartmann, E. (1989). Boundaries of dreams, boundaries of dreamers: Thin and thick boundaries as a new personality measure. Psychiatric Journal of the University of Ottawa, 14, 557-560.

Hartmann, E., Elkin, R., & Garg, M. (1991). Personality and dreaming: The dreams of people with very thick or very thin boundaries. Dreaming, 4, 311-324.

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

Hefez, A., Metz, L., & Lavie, P. (1987). Long-term effects of extreme situational stress on sleep and dreaming. American J Psychiatry, 144, 344-347.

Hill, A.B. (1974). Personality correlates of dream recall. Journal of Consulting and Clinical Psychology, 42, 766-773.

Kaminer, H., & Lavie, P. (1991). Sleep and dreaming in Holocaust survivors: Dramatic decrease in dream recall in well-adjusted survivors. Journal of Nervous and Mental Disease, 179, 664-669

Kostelny, K., & Garbarino, J. (1994). Coping with the consequences of living in danger: the case of Palestinian children and youth. International Journal of Behavioral Development, 17, 595-611.

Koulack, D. (1993). Dreams and adaption to contemporary stress. In A. Moffit, M. Kramer, & R. Hoffman, (Eds.), The function of dreaming (pp. 321-340). Albany: SUNY Press.

Koulack, D. & Goodenough, D., R. (1976). Dream recall and dream recall failure: An arousal-retrieval model. Psychological Bulletin, 83, 975-984.

Kramer, M. (1993). The selective mood regulatory function of dreaming: An update and revision. In Moffit, A., Kramer, M., and Hoffman, R. (Eds.), The functions of dreaming (pp. 139-195). Albany: SUNY Press.

Kramer, M. Schoen, L.S., & Kinney, L. (1984). Psychological and behavioral features of disturbed dreamers. Psychiatric Journal of the University of Ottawa, 9, 102-106.

Kuttab, D. (1988). A profile of the stonethrowers. Journal of Palestinian Studies, 17, 14-23.

Lavie, P. Hefez, A., Halperin, G., & Enoch, D. (1979). Long-term effects of traumatic war-related events on sleep. American Journal of Psychiatry, 136, 175-178.

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

Lytton, H., & Romney, D. M. (1991). Parents’ differential socialization of boys and girls: A meta-analysis. Psychological Bulletin, 109, 267-296.

Maccoby, E. E. (1988). Gender as a social category. Developmental Psychology, 24, 755-765.

Macksound, M., Aber, L., Dyregrov, A., & Raundalen, M. (1990). Post-traumatic stress reaction checklist for children. Columbia University: Center for the Study of Human Rights, Project on Children and War.

Masalha, S. (1993). The effect of prewar conditions on the psychological reactions of Palestinian children to the Gulf war. In L. A. Leavitt & N. A. Fox (Eds.), The psychological effects of war and violence on children (pp. 131-142). Hillsdale, New Jersey: Lawrence Earlbaum Associations

Nashef, Y. (1992). The psychological impact of the Intifada on Palestinian children living in refugee camps in the West Bank, as reflected in their dreams, drawings and behavior. Frankfurt an Main: Peter Lang.

Pearlin, L., & Schooler, C. (1978). The structure of coping. Journal of Health and Social Behavior, 19, 2-21.

Pedhazur, E.J., & Pedhazur-Schmelkin, L. (1991). Measurement, design, and analysis: An integrated approach. Hillsdale, New Jersey: Lawrence Erlbaum.

Punamäki, R.L. (1997). The effect of traumatic experiences on dreaming and sleeping. A review. Manuscript submitted for publication.

Punamäki, R. L., & Puhakka, T. (In press). Determinants and effectiveness of children’s coping with traumatic experiences. International Journal of Behavioral Development (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.

Qouta, S., Punamäki, R. L., & El Sarraj, E. (1995). Relations between traumatic experiences, activity and cognitive and emotional responses among Palestinian children. International Journal of Psychology, 30, 289-304.

Qouta, S., Punamäki, R. L., & El Sarraj, E. (1997). House demolition and mental health; Victims and witnesses. Journal of Social Distress and Homelessness, 6, 202-210.

Reynolds, C. F., Hoch, C. C., Buysse, D. J., Houck, P. R., Schlernitzauer, M., Pasternak, R. E., Frank, E., Mazumdar, S., & Kupfer, D. J. (1992). EEG sleep in spousal bereavement and bereavement-related depression of late life. Biological Psychiatry, 31, 69-82.

Reynolds, C.F., Hoch, C.C., Buysse, D.J., Houck, P.R., Schlernitzauer, M., Pasternak, R.E., Frank, E., Mazumdar, S., & Kupfer, D.J. (1993). Sleep after spousal bereavement: a study of recovery from stress. Biological Psychiatry, 34, 791-797.

Rofe, Y., & Lewin, I. (1980). Daydreaming in a war environment. Journal of Mental Imaginary, 4, 59-75.

Rofe, Y., & Lewin, I. (1982). The effect of war environment on dreams and dream habits. In N. A. Milgram (Ed.), Stress and Anxiety, vol. 8 (pp. 67-79). Washington: Hemisphere.

Ross, R. J., Ball, W. A., Sullivan, K. A., & Caroff, S. N. (1989). Sleep disturbance as the hallmark of posttraumatic stress disorder. American Journal of Psychiatry, 146, 797-707.

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.

Rotter, J. B., & Rafferty. J. E. (1950). Manual: The Rotter incomplete sentence blank. New York: Psychological corporation.

Russel, J.A. (1980). A circumplex model of affect. Journal of Personality and Social Psychology, 36, 1152-1168.

Schlosberg, A., & Benjamin, M. (1978). Sleep patterns in three acute combat fatigue cases. Journal of Clinical Psychiatry, 39, 546-549.

Schonbar, R. A. (1965). Differential dream recall frequency as a component of "life style". Journal of Consulting Psychology, 29, 468-474.

Schredl, M., Kleinferchner, P., & Gell, T. (1996). Dreaming and personality: Thick vs. thin boundaries. Dreaming, 6, 219-223.

Sulkunen, P., & Kekäläinen, O. (1992). WP-index: A programme of analysis of qualitative data. Helsinki: Gaudeamus. (In Finnish).

Summerfield, D. (1993). Health and human rights in Gaza. British Medical Journal, 306, 1416.

Terr, L. C. (1983). Chowchilla revisited: The effects of psychic trauma four years after a school-bus kidnapping. American Journal of Psychiatry, 140, 1543-1550.

Terr, L. C. (1991). Childhood traumas: An outline and overview. American Journal of Psychiatry, 148, 10-19.

Tonay, V. K. (1993). Personality correlates of dream recall: Who remembers? Dreaming, 3, 1-8.

Turner, S. W., & Gorst-Unsworth, C. (1990). Psychological sequelae of torture: A descriptive model. British Journal of Psychiatry, 157, 475-480.

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.



1Correspondence should be directed to Raija-Leena Punamäki, Department of Psychology, Applied Division, P.O. Box 4 (Fabiankatu 28), 00014 University of Helsinki, Finland.

errata: In the original article, figure 1 and figure 2 are switched. We have corrected this online. 

For further discussion of this article, please see the Special Discussion Page

  Copyright ©2002 Association for the Study of Dreams. All Rights Reserved.