Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
This article was downloaded by: [UNAM Ciudad Universitaria] On: 24 December 2014, At: 22:43 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK Journal of Personality Assessment Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/hjpa20 Reliability and Validity of the Social Cognition and Object Relations Scale (SCORS) in the Assessment of Dream Narratives Erin M. Eudell-Simmons , Michelle B. Stein , Jared A. DeFife & Mark J. Hilsenroth Published online: 10 Jun 2010. To cite this article: Erin M. Eudell-Simmons , Michelle B. Stein , Jared A. DeFife & Mark J. Hilsenroth (2005) Reliability and Validity of the Social Cognition and Object Relations Scale (SCORS) in the Assessment of Dream Narratives, Journal of Personality Assessment, 85:3, 325-333, DOI: 10.1207/s15327752jpa8503_09 To link to this article: http://dx.doi.org/10.1207/s15327752jpa8503_09 PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions EUDELL-SIMMONS, SCORSSTEIN, AND DREAMS DEFIFE, HILSENROTH JOURNAL OF PERSONALITY ASSESSMENT, 85(3), 325–333 Copyright © 2005, Lawrence Erlbaum Associates, Inc. Reliability and Validity of the Social Cognition and Object Relations Scale (SCORS) in the Assessment of Dream Narratives Erin M. Eudell-Simmons Downloaded by [UNAM Ciudad Universitaria] at 22:43 24 December 2014 National Naval Medical Center Bethesda, Maryland Michelle B. Stein, Jared A. DeFife, and Mark J. Hilsenroth Derner Institute of Advanced Psychological Studies Adelphi University In this study, we aimed to investigate the application of an externally rated measure of interpersonal representations (Social Cognition and Object Relations Scale [SCORS]; Westen, 1995) to dream narratives. A total of 80 student participants enrolled at a private university in the New York City metropolitan area completed a Dream Log and affect adjective checklist (Wellman, 2002) based on a recalled dream at both 1 month and 3 months following the September 11, 2001, terrorist attacks. Using the dreams provided in this study we examined the interrater reliability of SCORS ratings, the relationship of SCORS variables to an independently rated measure of dream distortion, and the relationship of SCORS variables to participants’ own ratings of dream affect. Results indicated that dreams were reliably rated using the SCORS, 3 cognitive SCORS variables were significantly related to dream distortion, and affective SCORS variables were meaningfully related to participants’ own ratings of affect in their dreams. Findings from this study provide support for the application of SCORS ratings to dream narratives. We discuss implications for further research and clinical application. Dreams are often remembered when they are seen by the dreamer as disturbing or intrusive (e.g., nightmares) or when they reflect an experience significant to the dreamer (Hartmann, 1998; Hill, 1996). As narratives, dream reports are subject to the same mechanisms of memory and selective reporting that occur when patients tell stories about their waking life. According to Mayman (1968), dreams should be considered significant narratives, as they reflect a person’s experience. Dream narratives should, therefore, be amenable to analyses that aim to discover aspects of the individual’s personality that can then be used to facilitate insight, understanding, and change. Eudell-Simmons & Hilsenroth (2005) completed an empirical review on the conceptual uses of dreams in psychotherapy and suggested that dreams can be equivalent to other narrative and clinical data in their ability to quickly and accurately portray significant aspects of the dreamer’s personality and salient issues in the dreamer’s waking life. Two studies have empirically demonstrated the potential of dreams to provide unique and relevant interpersonal infor- mation. In one study, Krohn and Mayman (1974) found external judgess, using a quantitative Object Representation Scale (ORS) for Dreams, could rate patients’ dreams with “good” to “excellent” interrater reliability (58% of exact agreement between judges and 83% agreement within 1 scale point). Krohn and Mayman found adequate construct validity through comparisons of independent scale ratings of patients’ dreams to therapists’ ratings of patient level of object relations (correlations with modal dream score, r = .53). In addition, Krohn and Mayman determined concurrent validity through significant correlations between ORS ratings of dreams and ORS ratings of other measures (e.g., Rorschach [Rorschach, 1942] and early memories responses; correlations with modal dream scores were r = .46 and r = .62, respectively). In other words, patients’ levels of object representations, which Krohn and Mayman (1974) described as a “central dimension” (p. 464) in personality, could be concurrently identified in dream narratives as well as in Rorschach, Early Memory responses, and from treatment interactions with patients’ therapists. Downloaded by [UNAM Ciudad Universitaria] at 22:43 24 December 2014 326 EUDELL-SIMMONS, STEIN, DEFIFE, HILSENROTH A second study (Popp et al., 1996) was described by the authors as the “first report of the application of a more psychometrically developed measure of central relationship pattern [i.e., the Core Conflictual Relationship Theme (CCRT) method; Luborsky, 1984] to a large body of dreams in psychotherapy” (p. 1074). The CCRT is a method of identifying and quantifying cyclical relationship patterns, consisting of wishes of the individual, typical response of others, and subsequent responses of the individual as they appear in narratives. Applying this measure to dream narratives, Popp et al. found good to excellent interrater reliability in scoring CCRTs from the dreams of 13 patients in psychoanalysis (kappas of .58 to .83), similar to the results obtained when scoring CCRTs from waking narratives of the same sample (kappas of .67 to .75). Similar themes were found in dreams and waking narratives of specific patients, even when narratives were taken from different sessions of therapy. In other words, Popp et al.s’ study found that patients’ core relational patterns could be reliably identified in their dreams and that these patterns corresponded to patterns seen in other relational narratives told in therapy. The findings of these two studies provide empirical support for the notion that quantitative content analysis of dreams by external raters can provide reliable and valid judgments regarding dreamers’ functioning. This type of analysis, which can enhance the understanding of an individual’s experience, might be especially useful following a type of event that typically results in frequent, intense, and memorable dreams. The presence of disturbing and intrusive dreams following a traumatic event is well documented, particularly when the dreamer is experiencing other symptoms of traumatic stress. For example, a study by Mellman, David, Bustamante, Torres, and Fins (2001) examined stress responses and dream content reported soon after a variety of lifethreatening events including motor vehicle accidents, industrial accidents, and impersonal assaults (i.e., nonsexual, unfamiliar assailant) by participants injured in these events. Mellman et al. reported that of the 30% of their sample to recall at least one dream, 46% of these were trauma dreams and were reported by those participants who exhibited more severe posttraumatic stress disorder (PTSD) symptoms than those who reported less disturbing dreams. Other studies have documented similar findings of disturbing and intense dreams even some time after a traumatic event. For example, 6 to 12 months following Hurricane Andrew, Mellman, David, Kulick-Bell, Hebding, and Nolan (1995) used sleep laboratory assessment to find increased dream recall, more negative dream content, and more pervasive sleep disturbances among those individuals suffering from PTSD than among those not reporting such symptoms. Intrusive and disturbing dreams have also frequently been associated with the traumatic effects of combat experience, often in conjunction with other symptoms of distress. For instance, Esposito, Benitez, Barza, and Mellman (1999) found spontaneously recalled dreams reported by combat veterans suffering from chronic PTSD to contain threat, reality distortion, and replications of the trauma. Several studies specifically examining symptoms of combat-related PTSD have indicated that frequent anxiety dreams as well as nightmares that involve intense emotions and reliving the trauma seem to accompany other PTSD symptoms (Inman, Silver, & Doghramji, 1990; see Ross, Ball, Sullivan, & Caroff, 1989, p. 698). Kramer (1993) found that 50% of the dreams of Vietnam veterans suffering from chronic PTSD contained combat related intrusive imagery. In addition, increases in dreams and nightmares including aggressive, persecutory, and frightening content were found in Palestinian adolescents and children living in violent, traumatic conditions (Punamäki, 1998). In general, these studies of dream content during stressful life experiences have seemed to indicate that not only are frequent and disturbing dreams common in response to a traumatic experience but that these dreams also serve to descriptively and accurately reflect individuals’ experiences of the trauma. Studies of dream content as part of a stress response suggest that dreams can reflect individuals’ experiences of significant and emotionally intense current life events, particularly traumatic incidents, interpersonal difficulties, and ongoing stressful situations. Hartmann (1998) suggested that dreams following a traumatic event can be useful illustrations of the individual’s affective experience of the event given that “contextualizing images” of fear and helplessness are common and clear reflections in trauma dreams of the emotional activation caused by the trauma. A contextualizing image is defined as a dominant emotion or emotional concern of the dreamer (Hartmann, 2001). Hartmann’s (1998, 2001) research is supported by other examinations on the effects of sexual assault, sexual abuse, and incest, which have found excessive themes of fear, aggression, and hostility within dream content (Belicki & Cuddy, 1996; King & Sheehan, 1996). Wood and Bootzin (1992, p. 223), in their study examining dreams immediately following a severe earthquake, concluded that dream content can reflect “personal concerns” and recent threatening events. In addition, those individuals who experienced more stress and discomfort as a result of a trauma (i.e., those with PTSD symptoms or depression) appear to experience more frequent, intense, and disturbing dreams regarding the trauma (Hartmann, 2001; Siegel, 1996; Wood & Bootzin, 1992). To identify and describe dream content in a clinically useful way, an externally rated measure of interpersonal narrative content was used to score dream reports following the September 11th World Trade Center (WTC) attacks. The Social Cognition and Object Relations Scale (SCORS; Hilsenroth, Stein, & Pinsker, 2004; Westen, 1991, 1995) was designed to quantify patient narratives generated from the Thematic Apperception Test (TAT; Murray, 1943), Early Memories, and psychotherapy sessions in terms of internal representations and expectations of social and inter- !!! Downloaded by [UNAM Ciudad Universitaria] at 22:43 24 December 2014 SCORS AND DREAMS personal functioning. The SCORS allows therapists to identify thoughts and feelings that underlie patterns of in!!!!! terpersonal behavior (Westen, 1991) and has been shown to be clinically useful in characterizing interpersonal functioning from psychotherapy narratives (Peters, Hilsenroth, Eudell-Simmons, Blagys, & Handler, in press; Westen, 1995), TAT narratives (Ackerman, Clemence, Weatherill, & Hilsenroth, 1999), and Early Memories (Fowler, Hilsenroth, & Handler, 1995, 1996). The premise of this study is that dreams are examples of qualitatively similar narrative data and that the SCORS, which has been supported as clinically useful when applied to waking interpersonal narratives, could have a similar result when applied to dreams. Despite the view that dream narratives often contain interpersonal and clinically relevant content that is similar to waking narratives, no study has yet examined the application of SCORS ratings to dream narratives. In this study, we examined the validity of using dream content to assess internally represented patterns of relational and affective functioning following trauma (the September 11th attacks on the WTC). First, we attempted to evaluate the interrater reliability of the SCORS (Hilsenroth et al., 2004; Westen, 1995) as applied to dream narratives. We hypothesized that dreams rated using the SCORS would achieve at least good interrater reliability (intraclass correlation coefficients [ICC] > .6 as defined by Fleiss, 1981). Second, we conducted correlational analyses of an independently rated measure of dream distortion to extend validity for the use of the SCORS in rating dreams. We hypothesized that dream distortion would be significantly and inversely related to the more cognitive variables of the SCORS (Complexity, Morals, Causality, and Identity). Finally, we compared ratings of the SCORS variables to participant ratings of positive and negative affect present in their dreams. These analyses were an investigation of validity, the extent to which external raters could perceive dreams with the same affective qualities as the dreamers themselves. We believed that adaptive and maladaptive ratings on the SCORS, particularly on more affective SCORS variables (Affect, Relationships, Aggression, and Self-Esteem), would correspond to positive and negative affective ratings of dream content. METHOD Participants !!! Die Informationen haben wir leider nicht —> Adjective checklist We collected data from 80 undergraduate student participants enrolled in introductory psychology classes at a private university in the New York City metropolitan area approximately 30 miles from ground zero. In addition, many of these participants had family members, friends, or acquaintances affected by, and in some cases involved in, the attacks and their aftermath. The proximity to the attacks and relational involvement of the participants with those affected suggests 327 the sample would have seen the attacks as having happened to individuals similar to themselves and consider themselves more threatened, thereby experiencing the attacks as traumatic. Participant dreams were collected from the same sample of participants as part of a longitudinal study examining the effects of the September 11th attacks reported elsewhere (Callahan, Hilsenroth, Jonay, & Waehler, 2005). Although the participant sample is the same, the measures we utilized in this investigation are independent of the data analyzed for the Callahan et al. study. The mean age of the students was 18.25 years (SD = 0.88). Of the participants, 20% were male, and 80% were female. All participants were single. Participants were offered extra credit in their classes for completing the study protocol. Although no student refused to participate in the study, data used were only from those participants who completed the Dream Log at both collection times, resulting in 160 dreams. Participants completed measures at two times: 1 month following the attacks (Month 1; week of October 8, 2001) and 3 months following the attacks (Month 3; week of December 10, 2001). At each time, each participant completed a Dream Log based on one recent recalled dream. All participants signed an informed consent form regarding the completion of measures in relation to the events of September 11th. Measures Dream log and adjective checklist. Participants completed a dream log for each of the two dreams (at Month 1 and Month 3) they were asked to recall in the study. Dreams were completed at home, and we asked participants to fill out the following information on waking after a dream (Wellman, 2002). The dream log included a space for the dream narrative as well as an adjective checklist consisting of 24 possible adjectives that participants could rate as being present in their dream(s). The adjectives were then divided into scales representing Positive Affect (Comfortable, Exciting, Happy, Loving, Sensual, and Sexy) and Negative Affect (Confusing, Deadly, Depression, Dreadful, Hateful, NerveRacking, Painful, Sad, Scary, and Unpleasant). Measurement of positive and negative affect constructs is an important area of evaluation in clinical psychology, as these constructs have exhibited significant relations to clinical conditions (Brown, Chorpita, & Barlow, 1998; Clark, Watson, & Mineka, 1994; Watson & Clark, 1984; Watson, Clark, & Harkness, 1994). We computed coefficient alphas for the Positive Affect scale at Month 1 (α = .61) and Month 3 (α = .74) and for the Negative Affect scale at Month 1 (α = .77) and Month 3 (α = .75). With the exception of the Positive Affect scale at Month 1, these alphas surpassed the generally accepted lower bound of .70 (Nunnally & Bernstein, 1994). Analyses using the Positive Affect scale at Month 1 were still conducted, as its coefficient alpha of .61 surpassed the .50 to .60 range originally recommended by Nunnally (1967; Streiner, 2003) for the early stages of research, it had a modest mean item-total cor- 328 EUDELL-SIMMONS, STEIN, DEFIFE, HILSENROTH relation of .35, and it had a mean interitem correlation of .22 that surpasses the range of .15 to .20 recommended by Clark and Watson (1995; Streiner, 2003) for scales that measure broad characteristics. Positive and Negative Affect scales were significantly inversely correlated at both Month 1 (r = –.53, p < .01) and Month 3 (r = –.47, p < .01). SCORS. The SCORS (Hilsenroth et al., 2004; Westen, 1995) is a clinician-rated assessment tool used to characterize representations of interpersonal functioning seen in narratives. In this study, we utilized the most recent, global rating version of the SCORS (Hilsenroth et al., 2004; Westen, 1995), which is made up of eight variables: Complexity of Representation (Complexity), Affective Quality of Representation (Affect), Emotional Investment in Relationships (Relationships), Emotional Investment in Values and Moral Standards (Morals), Understanding of Social Causality (Causality), Experience and Management of Aggressive Impulses (Aggression), Self-Esteem, and Identity and Coherence of Self (Identity). Each of these eight variables is now scored on a 7-point anchored rating scale ranging from 1 to 7. Lower scores (e.g., 1, 2, or 3) indicate the presence of more pathological responses, whereas higher scores (e.g., 5, 6, or 7) indicate healthy responses. Ackerman et al. (1999, p. 430) provided descriptions of the eight global SCORS variables. The Complexity variable assesses the richness of an individual’s representations of !!! self and others including the individual’s abilities to integrate both positive and negative attributes of the self and others and to take others’ points of view. The Affect variable assesses an individual’s positive and/or negative expectations from others in relationships and how the individual describes relationships. The Relationships variable identifies the individual’s level of commitment and emotional sharing in relationships. The Morals variable distinguishes between individuals who show no remorse for “inconsiderate or selfish actions” (p. 30) and those who “think about moral questions in a way that combines abstract thought, a willingness to challenge or question convention, and genuine compassion and thoughtfulness in actions (i.e., not just intellectualMentalisierungs ization)” (p. 30; Westen, 1995). The Causality variable assesses how well an individual understands why people do fähigkeit what they do. The Aggression variable assesses an individual’s ability to control and appropriately express his or her aggression and anger. The Self-Esteem variable assesses the affective quality of self-representation. The Identity variable assesses an individual’s level of fragmentation and integration (Westen, 1995). These variables can be divided into those with more affective and object relational content (i.e., Affect, Relationships, Aggression, and Self-Esteem) and those with more cognitive content (i.e., Complexity, Morals, Causality, and Identity; Cogan & Porcerelli, 1996; Fowler et al., 1995; Hibbard, Hilsenroth, Hibbard, & Nash, 1995; Porcerelli, Cogan, & Hibbard, 1998; and Porcerelli, Hill, & Dauphin, 1995). Ackerman et al. (1999) found good to excellent interrater reliability on this version of the SCORS (Spearman–Brown corrected ICCs ranging from .70 to .95) when used on TAT narratives, and reported on a number of other studies with similar results. Other studies have found good to excellent interrater reliability of SCORS variables when rating psychotherapy narratives and Early Memories (Fowler et al., 1995, 1996; see Peters et al., in press; Westen, 1991). Downloaded by [UNAM Ciudad Universitaria] at 22:43 24 December 2014 !!! Dream Distortion Scale. The Dream Distortion Scale (Zepelin, 1979) was used to characterize the level of bizarreness evident in each dream narrative. This scale is a 6-point ordinal scale ranging from 0 (The events closely resemble the recent waking experience of subject or the common, everyday experience of people in general) to 5 (Important aspects of the dream are impossible, or the dream contains a combination of illogical or improbable elements which make the dream narrative as a whole seem impossible) designed to measure the extent to which dream content differs from the dreamer’s waking life. A score of 0 indicates the dream content reflects “common, everyday experience of people in general.” A score of 5 indicates the dream contains elements that seem “impossible.” Each point on the scale is extensively described and includes examples for ease of scoring. Zepelin (1979) found excellent interrater reliability when using this scale to rate dreams reported in the laboratory or from the home, with Pearson correlation coefficients of .84 for both sets of dream narratives. Dream length. We examined the length of the dream narrative as a possible confounding variable in the rating process. As such, we counted the words in each dream narrative by computer word processor to provide an accurate measure of the length of the dream report. Procedure for Raters The three raters we used in the study were two graduate students enrolled in an American Psychological Association approved Clinical Psychology PhD program and one graduate student enrolled in a psychology masters program at the same university. Raters received extensive training on the SCORS, using both TAT narratives and nonstudy dreams, and were supervised by a licensed Clinical Psychologist with experience in using the measure. Raters met weekly for 2-hr meetings for 16 weeks. We evaluated raters for reliability on both the SCORS and the Dream Distortion Scale prior to applying the measures to dream narratives in the study. Each rater achieved good to excellent interrater reliability on all SCORS variables, ICC (1, 3) on practice TAT narratives from .71 to .90 and excellent interrater reliability on the Dream Distortion Scale, ICC (1, 3) = .88 on 30 nonstudy dreams. We transcribed and printed each dream narrative to avoid inconsistencies in scoring due to handwriting quality. Each of the three raters independently rated all dream narratives using 329 SCORS AND DREAMS each measure (i.e., SCORS and Dream Distortion). Each measure was used in turn to rate all dream narratives to avoid carryover effects. More specifically, we did not want ratings of the SCORS to affect the ratings of the Dream Distortion Scale. All dreams were first scored on the Dream Distortion Scale followed by the SCORS. Each transcribed dream narrative was rated in random order and independently by all raters. Raters were blind to participant identity, month of the dream, and all other subject material such as adjective ratings, demographics, and other assessment materials. Raters scored each dream in isolation from each other. Downloaded by [UNAM Ciudad Universitaria] at 22:43 24 December 2014 RESULTS Interrater Reliability Table 1 reports the means and standard deviations of all variables used in the analyses. The interrater reliability of dream ratings on the eight SCORS variables and the Dream Distortion Scale were evaluated via one-way random effects model ICC (1) (Shrout & Fleiss, 1979, see Table 2). This ICC model was chosen, as it is seen as the most conservative and generalizable calculation of ICCs (Meyer et al., 2002). Mean SCORS and Dream Distortion ratings across the three raters were used for the analyses, as all scored each dream. As such, the Spearman–Brown correction for the one-way random effects model, ICC (1, 3), was calculated to examine the reliability of the mean score for each of the SCORS variables and the Dream Distortion score for each dream (see Table 2). ICCs are considered to be excellent if greater than .74, good TABLE 1 Means and Standard Deviations for Dream Measures Month 1 Dreams SCORS Measure Complexity Affect Relationships Morals Causality Aggression Self-Esteem Identity Dream Distortion Scale Dream length Adjective Checklist Negative Affect Positive Affect Month 3 Dreams M SD M SD 3.75 3.66 3.62 3.91 3.55 3.79 3.94 4.09 2.48 95.40 0.67 1.19 1.06 0.54 0.65 0.58 0.41 0.54 1.50 72.62 3.83 3.91 3.67 3.86 3.55 3.83 3.93 4.20 1.93 66.19 0.51 1.03 1.07 0.53 0.68 0.56 0.43 0.43 1.46 59.48 3.11 1.29 2.43 1.42 2.78 1.08 2.10 1.29 Note. N=80. SCORS = Social Cognition and Object Relations Scale; Complexity = complexity of representation; Affect = affective quality of representation; Relationships = emotional investment in relationships; Morals = emotional investment in values and moral standards; Causality = understanding of social causality; Aggression = experience and management of aggressive impulses; Identity = identity and coherence of self. TABLE 2 ICC for Ratings on SCORS Variables and Dream Distortion Scale Month 1 ICC Month 3 ICC SCORS Variable (1) (1, 3) (1) (1, 3) Complexity Affect Relationships Morals Causality Aggression Self-Esteem Identity M SCORS variable Dream Distortion Scale .40 .72 .59 .30 .26 .32 .23 .39 .40 .84 .67 .89 .81 .56 .52 .59 .47 .66 .65 .94 .25 .71 .68 .46 .40 .55 .38 .19 .45 .77 .50 .88 .86 .72 .67 .79 .65 .42 .69 .91 Note. N=80. ICC = intraclass correlation coefficients; ICC(1) = one-way random effects model ICC; ICC(1, 3) = Spearman–Brown correction for three raters (Shrout & Fleiss, 1979); Complexity = complexity of representation; Affect = affective quality of representation; Relationships = emotional investment in relationships; Morals = emotional investment in values and moral standards; Causality = understanding of social causality; Aggression = experience and management of aggressive impulses; Identity = identity and coherence of self. if ranging from .60 to .74, fair if ranging from .40 to .59, and poor if under .40 (Fleiss, 1981). We hypothesized that one would be able to rate dreams using the SCORS with good interrater reliability (ICC > .6). Although uncorrected ICC values were generally in the fair range of reliability, the Spearman–Brown corrected ICCs for the eight SCORS variables and the Dream Distortion Scale were considered to be of good reliability. Table 2 illustrates Spearman–Brown ICCs for the Month 1 (.47 to .89, M = .65) and Month 3 (.42 to .88, M = .69); SCORS variables ranged from fair to excellent. The Dream Distortion Scale demonstrated excellent interrater reliability with ICCs of .84 and .94 at Month 1 and .77 and .91 at Month 3. In all subsequent analyses, therefore, the mean of the three raters’ scores on the SCORS variables and Dream Distortion Scale was used. SCORS Ratings and the Dream Distortion Scale The second hypothesis stated that independent ratings of dream distortion would be negatively correlated with SCORS variables—particularly Complexity, Morals, Social Causality, and Identity—as they intend to address more cognitive components of narratives. To test this hypothesis, we examined independent ratings of the Dream Distortion Scale and SCORS variables at Month 1 and Month 3 using Pearson product–moment correlation coefficients (see Table 3). As dream length was significantly correlated with Dream Distortion scores at both Month 1 (r = .37, p < .01) and at Month 3 (r = .45, p < .01), the effects of dream length were controlled for using partial correlations as can be seen in parentheses in Table 3. 330 EUDELL-SIMMONS, STEIN, DEFIFE, HILSENROTH TABLE 3 Pearson Product Moment Correlation Coefficients Comparing Mean Dream Distortion Ratings and Mean SCORS Ratings SCORS Variable Downloaded by [UNAM Ciudad Universitaria] at 22:43 24 December 2014 Complexity Affect Relationships Morals Causality Aggression Self-Esteem Identity Dream Distortion Scale Month 1 Month 3 –.33** (–.51**) –.38** (–.37**) –.26* (–.37**) –.16 (–.23*) –.33** (–.50**) –.20 (–.29*) –.17 (–.19) –.46** (–.45**) –.27* (–.33**) –.20 (–.13) .03 (–.08) .12 (.11) –.31** (–.34**) –.20 (–.09) –.16 (–.21) –.24* (–.37**) Note. N = 80. Correlations in parentheses represent partial correlations between SCORS variables and Dream Distortion Scale controlling for dream length. SCORS = Social Cognition and Object Relations Scale; Complexity = complexity of representation; Affect = affective quality of representation; Relationships = emotional investment in relationships; Morals = emotional investment in values and moral standards; Causality = understanding of social causality; Aggression = experience and management of aggressive impulses; Identity = identity and coherence of self. *p < .05. **p < .01. After controlling for dream length, more distorted dreams were related to less adaptive affective states as measured by the SCORS at 1 month following the September 11th attacks (correlations with Affect, Relationships, and Aggression were negative and significant at the p < .05 level). These findings did not hold up with more temporal distance from the traumatic event, and more affective SCORS variables were not found to be significantly related to dream distortion at Month 3. In support of our hypothesis, it is important to note that dream distortion was negatively and significantly correlated with three of the four more cognitive SCORS variables at Month 1 and Month 3 both before and after controlling for dream length. More distorted dreams exhibited significantly lowered Complexity of Representations (r = –.51 and –.33), less understanding of Social Causality (r = –.50 and –.34), and less Identity and Coherence of Self (r = –.45 and –.37) at Month 1 and Month 3, respectively. SCORS Ratings and Participant-Rated Affect Our final hypothesis stated that ratings on SCORS variables, particularly on affective variables (i.e., Affect, Relationships, Aggression, and Self-Esteem), would be positively correlated to participant ratings of positive affective dream content and negatively correlated to participant ratings of negative affective dream content. To test this hypothesis, we examined relationships between SCORS ratings and participant ratings of positive affect and negative affect present in their dream at each time using Pearson product–moment correlation coefficients (see Table 4). As dream distortion was significantly correlated with affective SCORS variables at Month 1 and cognitive SCORS variables at both Month 1 and Month 3, correlations of dream distortion with the participant rated affect scales were also included in Table 4. Dream length was significantly correlated to ratings of dream distortion (see previously), SCORS ratings of Complexity (r = .32, p < .01) and Causality (r = .31, p < .01) at Month 1, as well as SCORS ratings of Relationships (r = .22, p < .05) and Aggression (r = –.28, p = .01) at Month 3; thus, the analyses included partial correlations controlling for the effects of dream length on these variables. Partial correlations can be seen in parentheses on Table 4. As Table 4 illustrates, ratings of dream narratives using SCORS variables were related in the expected directions to participant ratings of dream affect on the adjective checklist. More adaptive SCORS ratings positively correlated with the Positive Affect scale and negatively correlated with the Negative Affect scale of the checklist. Examination of the affective component SCORS variables shows that the variables of Affect and Relationships were positively and significantly correlated to the Positive Affect scale at Months 1 (correlations of .59 and .33, respectively) and 3 (correlations of .54 and .31 and .38 [for partial correlation], respectively), with Self-Esteem demonstrating a trend toward significance at Month 3 (r = .21, p = .07). Correlations of Self-Esteem at Month 1 and Aggression at Months 1 and 3 with Positive Affect were in the expected direction but not significant. TABLE 4 Pearson Product Moment Correlation Coefficients Comparing Mean SCORS Ratings and Mean Dream Distortion Scale Ratings With Participant Rated Adjective Checklist Scales Positive Affect Negative Affect Scale Month 1 Month 3 Month 1 Month 3 Complexity Affect Relationships Morals Causality Aggression Self-Esteem Identity Dream Distortion Scale .13 (.21) .59** .33** .12 .13 (.21) .12 .18 .17 .28* .54** .31** (.38**) .07 .19 .16 (.11) .21 .16 –.18 (–.24*) –.66** –.29** –.19 –.22* (–.28*) –.45** –.23* –.23* –.37** –.64** –.16 (–.21) –.01 –.27* –.24* (–.19) –.25* –.25* –.13 (–.06) –.10 (–.01) .30**(.27*) .29** (.24*) Note. N = 80. Correlations in parentheses represent partial correlations between SCORS variables and Dream Distortion Scale controlling for dream length. SCORS = Social Cognition Object Relations Scale; Complexity = complexity of representation; Affect = affective quality of representation; Relationships = emotional investment in relationships; Morals = emotional investment in values and moral standards; Causality = understanding of social causality; Aggression = experience and management of aggressive impulses; Identity = identity and coherence of self. *p < .05. **p < .01. Downloaded by [UNAM Ciudad Universitaria] at 22:43 24 December 2014 SCORS AND DREAMS Similarly, ratings of affective SCORS variables were negatively correlated with experiences of negative affect in the dreams. Again, the Affect variable demonstrated the strongest relationship to participant rated affect and correlated –.66 with the Negative Affect scale at Month 1 and –.64 at Month 3. Self-Esteem was also significantly negatively correlated with the Negative Affect scale at Month 1 (r = –.23) and at Month 3 (r = –.25). The SCORS variables of Relationships and Aggression demonstrated significant correlations with Negative Affect at Month 1 (rs = –.29 and –.45, respectively) yet only trended toward significance after controlling for dream length at Month 3 (r = –.21, p = .07 for Relationships and r = –.19, p = .09 for Aggression). In a more exploratory analysis, cognitive-based ratings of dream narratives were examined for their relationship to participant-rated dream affect. Ratings of more cognitive narrative variables were not as often or as strongly related to participant experiences of positive affect in their dreams. Only the SCORS variable of Complexity demonstrated a significant relation to the Positive Affect scale of the checklist at Month 3 (r = .28), and Complexity and Causality trended toward significance after controlling for dream length (both rs = .21, ps = .07). However, dream content that was less complex, less coherent, and more distorted was consistently related to participant experiences of negative affect in the dreams. Significant correlations of the Negative Affect scale were found with Complexity (r = –.24 after controlling for dream length at Month 1, and r = –.37 at Month 3), Causality (r = –.28 after controlling for dream length at Month 1, and Month 3 r = –.27), Identity (Month 1 r = –.23, and Month 3 r = –.25), and dream distortion (after controlling for dream length at both points; Month 1 r = .27, and Month 3 r = .24). DISCUSSION Research has documented the presence of disturbing and intrusive dreams following many types of traumatic experiences (Belicki & Cuddy, 1996; Esposito et al., 1999; Hartmann, Zborowski, Rosen, & Grace, 2001; Horowitz, 2001; Inman et al., 1990; King & Sheehan, 1996; Mellman et al., 2001, 1995; Punamäki, 1998; Ross et al., 1989; Siegel, 1996; Wood & Bootzin, 1992). In fact, due to the easily accessible and frequently disturbing nature of dreams following a trauma, these dreams may provide an important source of information regarding the affective and interpersonal experiences of individuals following such an event. In addition, research examining the reliability and validity of interpersonal patterns identified from dreams suggests that the current status of an individual’s perceptions of his or her support system and his or her role in the world may be portrayed not only by stories they tell (Luborsky, Barber, & Diguer, 1992) but also by dreams they report. In this study, we aimed to explore the potential use of dreams in the assessment of interpersonal representations and subjective affective experience. 331 It is also the first study to attempt to quantify dreams in terms of their representation of interpersonal and social relationships using the SCORS. Our first goal in the study was to determine if SCORS ratings could be reliably applied to dream narratives. Results indicate that the 160 dreams we collected in this study were generally scored with fair to excellent interrater reliability. At both Month 1 and Month 3, the mean Spearman–Brown ICCs of SCORS ratings were in the good range, with ICCs for two SCORS variables—Affect and Relationships—located in the excellent range across the two groups of dreams. The Affect and Relationships variables are the most salient to this study, as they examine an individual’s affective experience and ability to invest in relationships. It may be that these two variables are the clearest variables to see in dreams following a traumatic experience. Findings of good to excellent interrater reliability coefficients using the SCORS are similar to other research that has used mean SCORS ratings from multiple narrative items (e.g., multiple TAT stories and early memories; Ackerman, et al., 1999; Fowler et al., 1995, 1996; see Westen, 1991). Our second goal in this study was to extend the construct validity of the SCORS as applied to dream narratives. To achieve this, the Dream Distortion Scale was rated independently from the SCORS by trained raters. In agreement with prior research (Zepelin, 1979), interrater reliability for the Dream Distortion Scale was found to be in the excellent range, further supporting the reliable application of this scale to narratives. The Dream Distortion Scale, designed as a measure to score bizarreness in dream content versus typical descriptions of everyday waking life, was hypothesized to be related to cognitive variables of the SCORS that evaluate a narrative’s complexity, causality, and coherence in representations of self and others. Although representations of morality were not significantly related to dream distortion, more distorted dreams did exhibit less complexity of self/other representations, more gaps in interpersonal narrative accounts, and less stability/coherence in self-representations. Furthermore, dream distortion covaried with participantrated dream affect in the same way that the SCORS cognitively based variables did; dreams rated as more distorted and less cogent were not systematically related to reports of positive dream affect but were experienced by participants as containing more distressing affective states. The relationship between the Dream Distortion Scale and cognitively based SCORS variables not only extends the construct validity for assessing structural aspects of dream narratives but also, by their systematic relation to negative affective states, supports the practical utility of the SCORS measure for application to dream narratives. Our final goal in this study was to examine the criterion validity of the SCORS, that is, the relationship between external ratings of dream representational quality to the dreamers’ own ratings of affect present in their dreams. Several significant correlations and trends supported the clinically based premise Downloaded by [UNAM Ciudad Universitaria] at 22:43 24 December 2014 332 EUDELL-SIMMONS, STEIN, DEFIFE, HILSENROTH that external ratings of adaptive–maladaptive interpersonal dream content corresponded to participant ratings of more positive–negative affective content. The relationships between affectively based SCORS variables (i.e., Affect, Relationships, Aggression, and Self-Esteem) and adjective checklist ratings were as expected in that these SCORS variables are more targeted toward the various affective qualities in representations of the self or others. The variables affect and relationships were most globally related to the dreamers’ affective ratings, as they were significantly bidirectionally correlated with participant ratings of both positive and negative affect. As hypothesized, dreams judged as portraying positive expectations of, commitment to, and emotional sharing in relational interactions were experienced by the dreamer as having significantly more positive affects. Dreams judged as portraying caustic, shallow, and distant relationships were experienced by the dreamer as manifesting more negative affects. The remaining affective SCORS variables did not appear to be systematically related to positive affect yet did exhibit significant relationships to ratings of negative dream affect. Thus, dreams experienced by participants as having more distressing affects were also rated as containing self/other representations with lower self-esteem and less adaptive management of hostile/aggressive impulses. Some limitations of note include considerations of the lack of premorbid data, absence of a control group, and homogeneity of the participant sample. Due to the unanticipated and widespread nature of the WTC attacks, there were no comparable data regarding participants’ interpersonal representations prior to September 11th; nor was a control group of unaffected individuals pragmatically available. However, the focus of this study was on reliable and valid application of SCORS ratings to a sample of collected dream narratives and not on changes of representations through time/events or quantitative comparisons of SCORS ratings among groups; thus, premorbid assessment or a participant control group would be of little utility to the main aims of the study. Concerns about the college-age sample of limited diversity used in this study may raise questions regarding generalizability of results. Finally, collecting a series of dreams yielding one average rating score on each of the eight SCORS variables as opposed to scoring one dream per participant at each collection time may provide some beneficial effects in terms of reliability and validity. However, this method is more appropriate for an investigation of the participants’ stability of representations between dream life and waking life. Again, our aims in this study were to investigate the reliable and valid application of the SCORS to single dream narratives and not an investigation into the stability of representational styles or personality characteristics. Future research may move toward investigations of cognitive/affective representations from dreams versus waking life, over the course of development, in relation to symptomatic/interpersonal/personality characteristics, and through the course of psychotherapy. In conclusion, these findings suggest that external raters can fairly reliably apply a narrative-based measure of interpersonal representations to dream narratives. Furthermore, raters may accurately identify participants’ affective valence toward their dreams, particularly when directly rating the affect tone (i.e., affect) of those representations. Demonstration of validity for SCORS ratings of dream narratives may prove especially relevant toward applications for dream use in psychotherapy (see Eudell-Simmons & Hilsenroth, 2005). The SCORS, as a measure of social cognition and object relations, is meant to reflect meaningful aspects of personality functioning. As SCORS ratings can be reliably applied to dream narratives and demonstrate meaningful relationships to dream affects, future research should examine the practical application of SCORS ratings to dreams and narratives told in psychotherapy. Popp et al. (1996) illustrated in their application of the CCRT method to dreams and psychotherapy narratives that relational patterns found in dream narratives are thematically similar to relational patterns described by patients in their psychotherapy sessions. Thus, dream narratives may provide an additional source of data for therapists and patients to examine in collaborative endeavors to recognize, understand, and work through maladaptive patterns of cognitive/affective representations. Active and collaborative use of the SCORS through assessment and treatment may guide clinicians working with their patients in efforts to establish more complex, benevolent, and adaptive representations of self and others. REFERENCES Ackerman, S., Clemence, A. J., Weatherill, R., & Hilsenroth, M. J. (1999). Use of the TAT in the assessment of DSM–IV cluster B personality disorders. Journal of Personality Assessment, 73, 422–448. Belicki, K., & Cuddy, M. (1996). Identifying sexual trauma histories from patterns of sleep and dreams. In D. Barrett (Ed.), Trauma and dreams (pp. 46–57). Cambridge, MA: Harvard University Press. Brown, T. A., Chorpita, B. F., & Barlow, D. H. (1998). Structural relationships among dimensions of the DSM–IV anxiety and mood disorders and dimensions of negative affect, positive affect, and autonomic arousal. Journal of Abnormal Psychology, 107, 179–192. Callahan, K., Hilsenroth, M., Jonay, T., & Waehler, C. (2005). Longitudinal stress responses to the 9/11 terrorist attacks in a New York metropolitan college sample. Stress, Trauma, and Crisis: An International Journal, 8, 45–60. Clark, L. A., & Watson, D. (1995). Constructing validity: Basic issues in objective scale development. Psychological Assessment, 7, 309–319. Clark, L. A., Watson, D., & Mineka, S. (1994). Temperament, personality, and the mood and anxiety disorders. Journal of Abnormal Psychology, 103, 103–116. Cogan, R., & Porcerelli, J. H. (1996). Object relations in abusive partner relationships: An empirical investigation. Journal of Personality Assessment, 66, 106–115. Esposito, K., Benitez, A., Barza, L., & Mellman, T. (1999). Evaluation of dream content in combat-related PTSD. Journal of Traumatic Stress, 12, 681–687. Eudell-Simmons, E. M., & Hilsenroth, M. J. (2005). A review of empirical research supporting four conceptual uses of dreams in psychotherapy. Clinical Psychology and Psychotherapy, 12, 255–269. Downloaded by [UNAM Ciudad Universitaria] at 22:43 24 December 2014 SCORS AND DREAMS Fleiss, J. L. (1981). Statistical methods for rates and proportions (2nd ed.). New York: Wiley. Fowler, C., Hilsenroth, M. J., & Handler, L. (1995). Early memories: An exploration of theoretically derived queries and their clinical utility. Bulletin of the Menninger Clinic, 52, 31–52. Fowler, C., Hilsenroth, M. J., & Handler, L. (1996). Two methods of Early Memories data collection: An empirical comparison of the projective yield. Assessment, 3, 63–71. Hartmann, E. (1998). Nightmare after trauma as paradigm for all dreams: A new approach to the nature and functions of dreaming. Psychiatry, 61, 223–238. Hartmann, E., Zborowski, M., Rosen, R., & Grace, N. (2001). Contextualizing images in dreams: More intense after abuse and trauma. Dreaming, 11, 115–126. Hibbard, S., Hilsenroth, M. J., Hibbard, J. K., & Nash, M. R. (1995). A validity study of two projective object representations measures. Psychological Assessment, 7, 432–439. Hill, C. E. (1996). Working with dreams in psychotherapy. New York: Guilford. Hilsenroth, M., Stein, M., & Pinsker, J. (2004). Social Cognition and Object Relations Scale: Global Rating Method (SCORS–G). Unpublished manuscript, The Derner Institute of Advanced Psychological Studies, Adelphi University, Garden City, NY. Horowitz, M. J. (2001). Stress response syndromes: Personality styles and interventions (4th ed.). Northvale, NJ: Aronson. Inman, D. J., Silver, S. M., & Doghramji, D. (1990). Sleep disturbance in post-traumatic stress disorder: A comparison with non-PTDS insomnia. Journal of Traumatic Stress, 3, 429–437. King, J., & Sheehan, J. R. (1996). The use of dreams with incest survivors. In D. Barrett (Ed.), Trauma and dreams (pp. 56–67). Cambridge, MA: Harvard University Press. Kramer, M. (1993). The selective mood regulatory function of dreaming: An update and revision. In A. Moffitt, M. Kramer, & R. Hoffman (Eds.), The functions of dreaming (pp. 139–195). Albany: State University of New York Press. Krohn, A., & Mayman, M. (1974). Object representations in dreams and projective tests. Bulletin of the Menninger Clinic, 38, 445–466. Luborsky, L. (1984). Principles of psychoanalytic psychotherapy: A manual for supportive-expressive therapy. New York: Basic. Luborsky, L., Barber, J. P., & Diguer, L. (1992). The meanings of narratives told during psychotherapy: The fruits of a new observational unit. Psychotherapy Research, 2, 277–290. Mayman, M. (1968). Early memories and character structure. Journal of Projective Techniques and Personality Assessment, 32, 303–316. Mellman, T. A., David, D., Bustamante, V., Torres, J., & Fins, A. (2001). Dreams in the acute aftermath of trauma and their relationship to PTSD. Journal of Traumatic Stress, 14, 241–247. Mellman, T. A., David, D., Kulick-Bell, R., Hebding, J., & Nolan, B. (1995). Sleep disturbance and its relationship to psychiatric morbidity after Hurricane Andrew. American Journal of Psychiatry, 152, 1659–1663. Meyer, G. J., Hilsenroth, M. J., Baxter, D., Exner, J. E., Jr., Fowler, J. C., Piers, C. C., et al. (2002). An examination of interrater reliability for scoring the Rorschach Comprehensive System in eight data sets. Journal of Personality Assessment, 78, 219–274. Murray, H. E. (1943). Thematic Apperception Test manual. Cambridge, MA: Harvard University Press. Nunnally, J. C. (1967). Psychometric theory. New York: McGraw-Hill. Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric theory (3rd ed.). New York: McGraw-Hill. 333 Peters, E., Hilsenroth, M., Eudell-Simmons, E., Blagys, M., & Handler, L. (in press). Reliability and validity of the Social Cognition and Object Relations Scale (SCORS) in clinical use. Psychotherapy Research. Porcerelli, J. H., Cogan, R., & Hibbard, S. (1998). Cognitive and affective representations of people and MCMI–II personality psychopathology. Journal of Personality Assessment, 70, 535–540. Porcerelli, J. H., Hill, K. A., & Dauphin, V. B. (1995). Need-gratifying object relations and psychopathology. Bulletin of the Menninger Clinic, 59, 99–104. Popp, C., Diguer, L., Luborsky, L., Faude, J., Johnson, S., Morris, M., et al. (1996). Repetitive relationship themes in waking narratives and dreams. Journal of Consulting and Clinical Psychology, 64, 1073–1078. Punamäki, R. L. (1998). The role of dreams in protecting psychological well-being in traumatic conditions. International Journal of Behavioral Development, 22, 559–588. Rorschach, H. (1942). Psychodiagnostics (5th ed.). Berne, Switzerland: Hans-Huber. 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, 697–707. Shrout, P. E., & Fleiss, J. L. (1979). Intraclass correlations: Uses in assessing rater reliability. Psychological Bulletin, 86, 420–428. Siegel, A. (1996). Dreams of firestorm survivors. In D. Barrett (Ed.), Trauma and dreams (pp. 159–176). Cambridge, MA: Harvard University Press. Streiner, D. L. (2003). Starting at the beginning: An introduction to coefficient alpha and internal consistency. Journal of Personality Assessment, 80, 99–103. Watson, D., & Clark, L. A. (1984). Negative affectivity: The disposition to experience aversive emotional states. Psychological Bulletin, 96, 465–490. Watson, D., Clark, L. A., & Harkness, A. R. (1994). Structures of personality and their relevance to psychopathology. Journal of Abnormal Psychology, 103, 18–31. Wellman, P. (2002). In Instructor’s manual to accompany Psychology (3rd ed., p. 131). New York: Wiley. Westen, D. (1991). Social cognition and object relations. Psychological Bulletin, 109, 429–455. Westen, D. (1995). Social Cognition and Object Relations Scale: Q-Sort for projective stories (SCORS–Q). Unpublished manuscript, Cambridge Hospital and Harvard Medical School, Cambridge, MA. Wood, J., & Bootzin, R. R. (1992). Effects of the 1989 San Francisco earthquake on frequency and content of nightmares. Journal of Abnormal Psychology, 101, 219–224. Zepelin, H. (1979). Dream distortion. In C. Winget & M. Kramer (Eds.), Dimensions of dreams (pp. 140–143). Gainesville: University Presses of Florida. Erin M. Eudell-Simmons U.S. Naval Hospital, Okinawa PSC 482, Box 2944 FPO AP 96362 Email: [email protected] Received August 30, 2004 Revised February 10, 2005