EXCESSIVE REASSURANCE SEEKING, SELF ESTEEM, AND DEPRESSIVE SYMPTOMS IN CHILDREN OF AFFECTIVELY ILL PARENTS: AN EXPERIENCE SAMPLING ANALYSIS

Transcription

1 ABELA EXCESSIVE ET AL. REASSURANCE SEEKING Journal of Social and Clinical Psychology, Vol. 26, No. 7, 2007, pp EXCESSIVE REASSURANCE SEEKING, SELF ESTEEM, AND DEPRESSIVE SYMPTOMS IN CHILDREN OF AFFECTIVELY ILL PARENTS: AN EXPERIENCE SAMPLING ANALYSIS JOHN R. Z. ABELA, ELESHIA MORRISON, AND CLAIRE STARRS The current study examined whether excessive reassurance seeking serves as a vulnerability factor to depressive symptoms in a sample of children of affectively ill parents using a multiwave longitudinal design and experience sampling methodology. In addition, we examined whether self esteem moderates the association between excessive reassurance seeking and increases in depressive symptoms following increases in negative events. At Time 1, 56 children (ages 7 14) completed measures assessing excessive reassurance seeking, low self esteem, and depressive symptoms. Subsequently, children were given a handheld personal computer (HP Jornada 690), which signaled them to complete measures assessing depressive symptoms and negative events at six randomly selected times over an 8 week follow up interval. In line with hypotheses, higher levels of reassurance seeking were associated with greater increases in depressive symptoms following increases in negative events. Contrary to hypotheses, however, the strength of this association was not moderated by self esteem. Coyne s (1976) interpersonal theory of depression posits that depressed individuals perpetuate a cycle of negative interpersonal exchanges that triggers increases in their depressive symptoms. More specifically, Coyne proposes that initially nondepressed but mildly dysphoric individuals seek reassurance from significant others in order to assuage The research reported in this article was supported by a Young Investigator Award from the National Alliance for Research on Schizophrenia and Depression (NARSAD) and a New Opportunities Award from the Canada Foundation from Innovation (CFI) awarded to John R. Z. Abela. Correspondence may be addressed to John R. Z. Abela, Department of Psychology, McGill University, Stewart Biological Sciences Building, 1205 Dr. Penfield Avenue, Montreal (Quebec) Canada H3A 1B1. E mail: abela@ego.psych.mcgill.ca. 849

2 850 ABELA ET AL. their doubts about their own self worth and lovability. Others initially respond with genuine concern and support. The individual, however, perceives this support as inadequate and consequently escalates his or her symptoms in an effort to secure more reassurance and acceptance. Although others continue to provide support, they begin to experience feelings of irritation and guilt, leading to a separation between the content and affective quality of their reassuring statements. This discrepancy increases the individual s fear of rejection, which in turn leads to a further escalation of symptoms in order to restore the feeling of security. This downward spiral continues until either others withdraw from the individual or the individual seeks treatment. Joiner and colleagues (e.g., Joiner, Metalsky, Katz, & Beach, 1999a) posit that individual differences exist in the tendency to seek reassurance from others. Excessive reassurance seeking is defined as a relatively stable tendency to excessively and persistently seek assurances from others that one is loveable and worthy, regardless of whether such assurances have already been provided (Joiner et al., 1999a, p. 270). Excessive reassurance seeking behaviors are hypothesized to be motivated by highly accessible, maladaptive, cognitive interpersonal scripts (e.g., If I feel bad, then I ask my parents if they love me ) that become activated when individuals experience concerns about their self worth and/or future (Van Orden, Wingate, Gordon, & Joiner, 2005). According to Joiner and colleagues, excessive reassurance seeking constitutes the central feature of Coyne s (1976) theory as it serves as the vehicle through which the distress and desperation of depression is transmitted from one person to another, triggering negative outcomes for all. Joiner and colleagues (e.g., Joiner, Katz, & Lew, 1999) posit that excessive reassurance seeking is particularly likely to lead to increases in depressive symptoms following negative events. Such reasoning is in line with Coyne s (1976) formulation as Coyne hypothesized that reassurance seeking behaviors are initially triggered by mild depressed mood experienced in response to a negative event (e.g., loss of a relationship or other changes in social structure). Expanding upon Coyne s model, Joiner and colleagues propose that negative events trigger concerns about one s worth and/or future. In an attempt to assuage such concerns, individuals seek reassurance from others regarding their own lovability and others degree of dependability. Thus, according to Joiner et al. s extension of Coyne s (1976) theory, excessive reassurance seeking serves as a vulnerability factor (e.g., diathesis) that interacts with negative events (e.g., stress) to predict increases in depressive symptoms. Research with adult populations has generated support for many of the hypotheses derived from Coyne s (1976) model. First, reassurance

3 EXCESSIVE REASSURANCE SEEKING 851 seeking has been demonstrated to be a valid, replicable, and cohesive construct, distinct from other interpersonal variables such as general dependency, doubt in others sincerity, and dependence on close others (Davila, 2001; Joiner & Metalsky, 2001). Second, individuals who exhibit high levels of reassurance seeking have been shown to exhibit higher levels of depressive symptoms than individuals who exhibit low levels of reassurance seeking (e.g., Joiner, 1994; Joiner, Alfano, & Metalsky, 1992, 1993; Joiner & Metalsky, 1995; Joiner & Schmidt, 1998; Katz & Beach, 1997; Katz, Beach, & Joiner, 1998; Pothoff, Holahan, & Joiner, 1995). Third, excessive reassurance seeking has been shown to interact with negative events to predict increases in depressive symptoms (e.g., Joiner & Metalsky, 2001; Katz et al., 1998; for an exception, see Shahar, Joiner, Zuroff, & Blatt, 2004). Fourth, researchers have demonstrated a link between excessive reassurance seeking and interpersonal rejection (Joiner et al., 1992, 1993; Joiner & Metalsky, 1995; Katz & Beach, 1997). Last, reassurance seeking has been found to play a role in the social transmission of depression, with individuals who exhibit high levels of reassurance seeking more likely than their low reassurance seeking counterparts to develop depression when interacting with a depressed partner or roommate (Joiner, 1994; Katz, Beach, & Joiner, 1999). Thus, this body of research indicates an etiological role of reassurance seeking both in the development of depression and in the generation of negative interpersonal outcomes (Joiner et al., 1999a). Far less research has examined the relationship between excessive reassurance seeking and depressive symptoms in children. Preliminary research has shown that excessive reassurance seeking is associated with depressive symptoms and disorders among both youth psychiatric inpatients (Joiner, 1999; Joiner, Metalsky, Gencoz, & Gencoz, 2001) and children of affectively ill parents (Abela et al., 2005; Abela, Zuroff, Ho, Adams, & Hankin, 2006). Excessive reassurance seeking has also been found to predict increases in depressive symptoms following both the occurrence of negative events (Abela et al., 2006; Prinstein, Borelli, Cheah, Simon, & Aikins, 2005) and the onset of parental depressive symptoms (Abela et al., 2006). Last, excessive reassurance seeking has been found to interact with depressive symptoms to predict interpersonal rejection in youth psychiatric inpatients (Joiner, 1999). Thus, preliminary research supports the applicability of Joiner et al. s extension of Coyne s (1976) theory to children. At the same time, due to the small number of studies conducted with youth, additional research is needed. The first goal of the current study was to provide an examination of the vulnerability stress hypothesis of Joiner and colleagues (Joiner et al., 1999a) extension of Coyne s (1976) theory in a sample of children and adolescents. In order to provide a rigorous test of this hypothesis, we uti-

4 852 ABELA ET AL. lized experience sampling methodology (ESM). More specifically, participants carried handheld personal computers (Jordana 690), which signaled them to complete questionnaires assessing the occurrence of negative events and depressive symptoms at randomly selected times over a 6-week follow up interval. An advantage of ESM is that it allows for the assessment of study variables in a naturalistic manner, minimizing the impact of timing and context on assessments. Further, as participants are cued by signaling devices to complete self report measures in real time, the recall bias associated with diary methods and recall biases associated with depressogenic distortions are minimized. ESM has been shown to be a valid and practical means of assessing mood and cognition (Csikszentmihalyi & Larson, 1987; devries, Dijkman Caes, & Delespaul, 1990) and has been used in a range of community and clinical samples (e.g., Brown & Moskowitz, 1998; Cote & Moskowitz, 1998; devries & Delespaul, 1989; devries, Delespaul, & Dijkman, 1987; Zuroff, Moskowitz, & Cote, 1999). The second goal of the current study was to examine whether the relationship between excessive reassurance seeking and increases in depressive symptoms following increases in negative events is moderated by self esteem. The structure of Coyne s theory naturally allows for the possibility that the relationship between excessive reassurance seeking and depressive symptoms is affected by moderator variables in addition to the occurrence of negative events. Researchers have investigated several interpersonal and cognitive features of depressed persons as moderators of this relationship including self disclosure (Gurtman, 1987), self blame (Gotlib & Beatty, 1985; Hokanson, Loewenstein, Hedeen, & Howes, 1986) and aid seeking behaviors (Stephens, Hokanson, & Welker, 1987). We hypothesized that an individual s level of self esteem is likely to serve as an additional moderator of this relationship. More specifically, individuals with low self esteem are less likely than their high self esteem counterparts to be reassured by the support and concern that they receive from significant others because it is inconsistent with their negative self view (Swann, Rentfrow, & Guinn, 2003). Such individuals are therefore more likely to enter into the vicious cycle described by Coyne (1976) by seeking additional reassurance from others and consequently, over time, are more likely to have their demands for reassurance met with rejection. Only one study to date has examined self esteem as a moderator of the relationship between reassurance seeking and depressive symptoms. Providing partial support for this hypothesis, using a university student sample, Joiner et al. (1992) reported that the combination of depressive symptoms, excessive reassurance seeking, and low self esteem placed individuals at the greatest risk for interpersonal rejection. This finding, however, held only for males

5 EXCESSIVE REASSURANCE SEEKING 853 and not for females. Further, the authors did not examine whether the interaction of excessive reassurance seeking and low self esteem predicted higher levels of depressive symptoms. Self esteem as a moderator variable has not yet been examined in a child population. In order to provide a powerful test of our hypotheses, the study utilized a sample of children of parents with a history of major depressive episodes. Employing such a sample was advantageous because children of parents with a history of major depressive episodes are four to six times more likely than other children to develop depressive symptoms (Goodman & Gotlib, 2002). Thus, the use of such a sample likely maximized the number of children and adolescents who experienced increases in depressive symptoms during the course of the study. The procedure involved an initial laboratory assessment in which children completed measures assessing reassurance seeking and depressive symptoms. The procedure also involved a series of six follow up assessments occurring at randomly selected times over the 8 week follow up interval, in which children were signaled by a handheld personal computer to complete measures assessing depressive symptoms and the occurrence of negative events. The use of a multiwave longitudinal design allowed us to take an idiographic approach toward examining Joiner and colleagues (1999a) vulnerability hypothesis. Previous research examining this hypothesis (i.e., Katz et al., 1998; Joiner & Metalsky, 2001; Shahar et al., 2004) has typically relied on two time point designs in which (1) reassurance seeking and depressive symptoms were assessed during an initial assessment and (2) depressive symptoms and stressors were assessed during a follow up assessment. Such a design necessitates the use of a nomothetic approach toward operationalizing high levels of stress. In other words, children are considered to be experiencing a high level of stress when their level of stress is higher than the sample s average level of stress. In contrast, the use of a multiwave longitudinal design, in which depressive symptoms and stressors are assessed repeatedly throughout the follow up interval, allows for an idiographic approach toward operationalizing high levels of stress. In other words, children are considered to be experiencing a high level of stress when their level of stress is higher than their own average level of stress. This distinction is central to testing Joiner and colleagues hypothesis that the interaction between excessive reassurance seeking and low self esteem will moderate the association between children s hassles and depressive symptoms because Joiner and colleagues hypothesis posits that increases in levels of stress rather than absolute levels of stress will be associated with increases in depressive symptoms in vulnerable youth (for more detailed discussion, see Abela & Hankin, in press). In line with an idiographic perspective, we examined whether the

6 854 ABELA ET AL. slope of the relationship between negative events and depressive symptoms within children varied across children as a function of excessive reassurance seeking and self esteem. More specifically, we hypothesized that high levels of reassurance seeking would be associated with greater increases in depressive symptoms following increases in negative events in youth possessing low, but not high, levels of self esteem. Given that negative events and depressive symptoms were assessed contemporaneously in our analyses, in order to provide a stringent examination of our central hypothesis, we also examined the reverse model, a stress exposure model (Pothoff et al., 1995; Shahar et al., 2004), in which it was hypothesized that high levels of reassurance seeking would be associated with greater increases in negative events following increases in depressive symptoms in youth possessing low, but not high, levels of self esteem. METHOD PARTICIPANTS Participants in the current study were recruited from a sample participating in a larger project examining vulnerability to depression in children of parents with a history of major depressive episodes (e.g., Abela, Skitch, Adams, & Hankin, 2006; Abela, Skitch, Auerbach, & Adams, 2005). Participants were recruited through ads placed in local English newspapers and posters placed in the Montreal area. Both posters and newspaper ads targeted parents with a history of major depressive disorder and children between the ages of 6 and 14. Respondents were screened during phone interviews using the Structured Clinical Interview for the DSM IV (SCID I; First, Gibbon, Spitzer, & Williams, 1995). Parents who met criteria for current or past major depressive disorder were invited to participate in the study. The initial sample consisted of 102 parents (88 women) with 140 children (71 girls). Thus, 38 sibling pairs were included in the initial sample. All children were given the opportunity to participate in this second ESM portion of the study, and 56 consented to do so (45% female). For these 56 children, ages ranged from 7 to 14, with a mean age of 10.6 years. No siblings participated in the ESM portion of the study. The sample was 78.8% Caucasian, 1.9% Native American, 1.9% Asian, and 17.3% of other descent. Participants mother tongue included English (72.2%), French (5.6%), and others (22.2%). Parents marital statuses were 56.9% married, 25.5% divorced, 9.8% separated, 2.0% single, and 5.9% other. The uppermost level of education completed by the parents was some high school for 3.9%, a high school diploma for 3.9%, some community

7 EXCESSIVE REASSURANCE SEEKING 855 college for 7.8%, a community college diploma for 11.8%, some university for 23.5%, a university diploma for 25.5%, and a graduate school diploma for 23.5%. The median family income ranged from $30,000 to $45,000. Children who chose to participate in the ESM portion of the study did not differ significantly from children who chose not to participate in age (t(103) = 1.31, ns), gender (t(101) =.48, ns), parental education (t(97) =.53, ns), family income (t(94) = 1.03, ns), level of child depressive symptoms at baseline (t(102) = 1.40, ns), or level of child reassurance seeking (t(102) = 0.09, ns). PROCEDURE Participants were recruited during the second and final in lab assessment of the larger project described above. During this assessment, children and parents signed consent to participate in the ESM portion of the study, and children were then verbally administered the following questionnaires: (1) the Children s Depression Inventory (CDI; Kovacs, 1981); (2) Reassurance Seeking Scale for Children (RSSC; Joiner et al., 2001); and (3) Self Esteem Questionnaire (SEQ; Rosenberg, 1965). Subsequently, children were provided with HP Jornada 720 handheld computers for the ESM portion of the study. These computers were programmed to signal children to complete self report measures assessing the occurrence of negative events and the presence of depressive symptoms every 5 to 9 days (~ once a week) at randomly selected times outside of class periods. When the alarm sounded, the child opened up the computer and the question sets appeared automatically on the screen. One question was presented at a time, and all questions were read aloud by the computer while responses appeared on the screen. The child touched the response that best described him or her at that time. The following questionnaires were completed at each of these times: (1) CDI and (2) Hassles Scale for Children (Kanner, Feldman, Weinberger, & Ford, 1987). MEASURES The Structured Clinical Interview for the DSM IV (SCID I; First et al., 1995). The SCID I is a semistructured clinical interview designed to indicate current and lifetime DSM IV diagnoses. To allow for the diagnosis of all DSM IV mood disorders, the current study applied the affective disorders and psychotic screen modules. Diagnostic interviewers were required to complete an intensive training program in order to administer the SCID I interview and for assigning DSM IV diagnoses. The training program involved 40 hours of instruction, listening to

8 856 ABELA ET AL. audiotaped interviews, conducting practice interviews, and passing regular exams, where 85% or above had to be achieved. In order to qualify for conducting interviews, interviewers were required to listen to three audiotaped interviews and demonstrate 100% agreement with the principal investigator on the presence or absence of a diagnosis of a depressive disorder. There also had to be a minimum of 85% agreement on severity ratings of individual depressive symptoms. The principal investigator held weekly supervision sessions for the interviewers where the interviewers notes were reviewed to confirm the presence or absence of a diagnosis. In order to address any questions regarding the accuracy of symptom ratings or diagnoses, interviewer audiotapes were also reviewed. Children s Depression Inventory (CDI; Kovacs, 1981). The CDI is a 27 item self report questionnaire that measures the cognitive, affective, and behavioral symptoms of depression as displayed in children. For each item, children were asked to report whether it described how they were thinking and feeling over the past week. Questionnaire items are scored from 0 2, where a higher score indicates greater symptom severity. Total scores on the CDI questionnaire range from 0 to 52 points. As demonstrated in previous studies, the CDI possesses a high level of internal consistency and distinguishes children with major depressive disorders from their nondepressed counterparts (Saylor, Finch, Spirito, & Bennett, 1984). The Reassurance Seeking Scale for Children (RSSC; Joiner & Metalsky, 1995). The RSSC is a modified version of the Reassurance Seeking Scale (RSS; Joiner & Metalsky, 1995) that was reworded to be comprehensible to children. The revised scale consists of four items (e.g., Sometimes when I ask people if they like me, they tell me to stop asking). Children are asked to rate each statement on a 3 point scale (e.g., not very much, kind of a lot, and very much). Total scores range from 0 to 8, where higher scores correspond to higher levels of reassurance seeking. The reliability and validity of the original RSS for adults have been well documented in various studies (e.g., Joiner, 1994; Joiner et al., 1992). Studies using the RSSC in samples of children between the ages of 6 and 17 also report high degrees of reliability and validity (Abela, Hankin, et al., 2005; Abela, Zuroff, et al., 2006; Joiner, 1999; Joiner et al., 2001). Self Esteem Questionnaire (SEQ, Rosenberg, 1965). The SEQ asks children to rate 10 statements about self perceptions on a 1 5 scale. Children are asked to indicate whether they strongly agree, agree, disagree, or strongly disagree with each item (e.g., I feel that I have a number of good qualities"). Items are scored from 0 to 3 with higher scores indicating lower levels of self esteem. Scores for each item are added, yielding a composite score ranging from 0 to 30.

9 EXCESSIVE REASSURANCE SEEKING 857 TABLE 1. Means, Standard Deviations, and Intercorrelations between Time 1 Measures CDI 2. RSSC.36** 3. CSEQ.51***.34* 4. AGE.05.37** GENDER Mean SD Note. CDI = Children s Depression Inventory. RSSC = Reassurance Seeking Scale for Children. CSEQ = Children s Self Esteem Questionnaire. AGE = Children s age. GENDER = Children s gender (0 = Boy; 1 = Girl). ***p <.001. **p <.01. *p <.05. Hassles Scale for Children (HASC; Kanner et al., 1987). The HASC is a list of 39 negative events that children may experience (e.g., parents separate or divorce, a close friend moves away). Children are asked to rate how often each event happened to them during the past week using a scale of 0(never)to4(all the time). A total score is obtained by summing responses on all items and can range from 0 to 156, whereby higher scores indicate the occurrence of a greater number of hassles. RESULTS DESCRIPTIVE STATISTICS Means, standard deviations, and intercorrelations between all Time 1 measures are presented in Table 1. Several findings warrant additional attention. First, higher levels of depressive symptoms were associated with higher levels of reassurance seeking and lower levels of self esteem. Second, higher levels of reassurance seeking were associated with lower levels of self esteem. Last, older children reported lower levels of reassurance seeking behaviors than younger children. Means and standard deviations for CDI and HASC scores across the six follow up assessments are presented in Table 2. As each child completed the CDI at multiple follow up assessments, each child has his or her own mean level of depressive symptoms (i.e., his or her her average CDI score during the follow up interval) as well as his/her own degree of variation in depressive symptoms during the follow up interval (i.e., his or her SD on the CDI across administrations). Within subject means on the CDI ranged from 0.00 to (μ = 7.64; SD = 5.87) while

10 858 ABELA ET AL. TABLE 2. Means and Standard Deviations of Child Depressive Symptoms and Hassles over Six Follow up Assessments CDI Mean SD HASC Mean SD Note. CDI = Children s Depression Inventory. HASC = Hassles Scale for Children. within subject standard deviations ranged from 0.55 to (μ = 3.00; SD = 2.76). With respect to the children s levels of hassles, within subject means on the HASC ranged from 3.00 to (μ = 32.91; SD = 18.62) while within subject standard deviations ranged from 1.41 to (μ = 9.76; SD = 8.78). TEST OF THE VULNERABILITY STRESS HYPOTHESIS To test our hypothesis that children who possess both high levels of excessive reassurance seeking and low levels of self esteem would report greater increases in depressive symptoms following increases in stress than children who possess only one or neither of these vulnerability factors, multilevel modeling was employed. Analyses were carried out using the SAS (version 8.1) MIXED procedure as well as the maximum likelihood estimation. Our dependent variable was within subject fluctuations in CDI scores during the follow up interval (FU_CDI). Because FU_CDI is a within subject variable, CDI scores were centered at each participant s mean such that FU_CDI reflects fluctuations in a child s levels of depressive symptoms as compared to his or her mean level of depressive symptoms. Our primary predictors of changes in depressive symptoms (FU_CDI) were excessive reassurance seeking (RSSC), self esteem (SEQ), and fluctuations in hassle (HASC) scores during the follow up interval (FU_HASSLES). As RSSC and CSEQ are between subject predictors, RSSC scores and CSEQ scores were standardized prior to analyses. Because FU_HASSLES is a within subject predictor, HASC scores were centered at each participant s mean prior to analyses so that FU_HASSLES may reflect fluctuations in a child s level of hassles as compared to his or her mean level of hassles. Preliminary analyses were conducted examining whether children s age or gender served as a moderator of any relationships. No significant

11 EXCESSIVE REASSURANCE SEEKING 859 interactions involving either children s age or gender were found. Consequently, for the sake of simplicity, results are presented only for models including RSSC, SEQ, FU_HASSLES, and all two and three way interactions. When fitting hierarchical linear models, one must specify appropriate mean and covariance structures. It is important to note that mean and covariance structures are not independent of one another. Rather, an appropriate covariance structure is essential in order to obtain valid inferences for the parameters in the mean structure. Overparametrization of the covariance structure can lead to inefficient estimation and poor assessment of standard errors (Altham, 1984). On the other hand, too much restriction of the covariance structure can lead to invalid inferences when the assumed structure does not hold (Altham, 1984). Diggle, Liang, and Zeger (1994) recommend that one use a saturated model for the mean structure while searching for an appropriate covariance structure. In the current study, we were interested in examining the effects of RSSC, CSEQ, and FU_HASSLES on children s CDI scores during the 6 week follow up interval. Consequently, we chose a mean structure that included RSSC, CSEQ, and FU_HASSLES, and all two and three way interactions. Three additional effects were also included in this initial mean structure. First, in order to control for individual differences in baseline levels of depressive symptoms, children s Time 1 CDI scores were included in the model. Second, because different children are likely to have different levels of depressive symptoms when experiencing their own average level of hassles, a random intercept was included in the model. Last, given that fluctuations in negative events is a within subject predictor whose effect is expected to vary from participant to participant, a random effect for slope was included in the model. Compound symmetry, autoregressive, and banded Toeplitz are examples of commonly used covariance structures in studies where multiple responses are obtained from the same individual over time, with a consequent correlation with within subject residuals. In order to select one of these covariance structures for analysis of the data, models were fitted utilizing each structure and the best fit was chosen based on the Akaike information criterion (AIC and AICC) and the Schwarz Bayesian criterion (BIC). The best fit was found to be an autoregressive structure. After choosing the appropriate covariance structure, we next examined the random effects component of our model. Nonsignificant random effect parameters were deleted from the model prior to examining the fixed effects component. With respect to random effects, the random intercept (p <.01) was significant and thus was retained in the model. Further, the AR(1) parameter was significant (r = 0.28, p <.05)

12 860 ABELA ET AL. TABLE 3. Hierarchical Linear Modeling Analyses: Predicting Fluctuations in Depressive Symptoms during the Follow up Interval Predictor b SE F df Time 1 CDI *** 45 RSSC SEQ FU_HASSLES ** 202 RSSC FU_HASSLES * 202 SEQ FU_HASSLES RSSC SEQ RSSC SEQ FU_HASSLES Note. CDI = Children s Depression Inventory. RSSC = Reassurance Seeking Scale for Children. SEQ = Self Esteem Questionnaire. FU_HASSLES. = Within subject fluctuations in Hassles Scale for Children (HASC) scores during the follow up interval. ***p <.001. **p <.01. *p <.05. and thus was retained in our model. The random slope, however, was not significant and thus was deleted from the model prior to examining the fixed effects. The final results relating to the fixed effects component of the model are presented in Table 3. In line with Joiner and colleagues vulnerability stress hypothesis, a significant two way, cross-level interaction emerged between RSSC and FU_HASSLES. In order to examine the form of this interaction, we used the model summarized in Table 1 to calculate predicted CDI scores for children possessing either low or high levels of excessive reassurance seeking (± 1.5 between subject SD) and who are experiencing either low or high levels of stress in comparison to their own average level of stress (± 1.5 mean within subject SD). The results of these calculations are presented in Figure 1. Because both FU_CDI and FU_HASSLES are within subject variables centered at each participant s mean, slopes are interpreted as the increase in a child s CDI score that would be expected in the case where he or she scored one point higher on the HASC. Analyses were conducted for each RSSC condition, examining whether the slope of the relationship between negative life events and depressive symptoms significantly differed from 0. Analyses demonstrated that children who possessed high levels of excessive reassurance seeking reported higher levels of depressive symptoms when experiencing high levels of negative events than when experiencing low levels of negative events (slope = 0.22; t(202) = 3.61, p <.001). Level of depressive symptoms, however, did not vary as a function of level of negative events for children possessing low levels of excessive reassurance seek-

13 EXCESSIVE REASSURANCE SEEKING 861 High Reassurance Seeking Low Reassurance Seeking Predicted CDI Score Low Intercept High Level of Negative Events FIGURE 1. Predicted level of depressive symptoms in children as a function of reassurance seeking and hassles. ing (slope = 0.08, t(202) = 1.19, ns). A planned comparison of the slope of the relationship between depressive symptoms and negative events for children possessing high and low levels of RSSC indicated that these two slopes differed significantly from one another (t(202) = 2.47, p <.05). We next examined whether the strength of the association between the RSSC FU_HASSLES interaction and FU_CDI was moderated by self esteem. Contrary to our integrative model, however, the SEQ RSSC FU_HASSLES interaction was not significant. TEST OF ALTERNATIVE MODEL Given that negative events and depressive symptoms were assessed contemporaneously in the above described analyses, we also examined an alternative model in which it was hypothesized that high levels of reassurance seeking would be associated with greater increases in negative events following increases in depressive symptoms in youth possessing low, but not high, levels of self esteem. Analyses were similar to those described above with the exception of (1) our dependent variable being fluctuations in HASC scores during the follow up inter-

14 862 ABELA ET AL. val rather than fluctuations in CDI scores (2) and our within subject predictor variable being fluctuations in CDI scores during the follow up interval rather than fluctuations in HASC during the follow up interval. The results of our final model are reported in Table 4. Of primary importance, none of the two way or three way cross level interaction terms were significant predictors of fluctuations in hassles scores during the follow up interval. 1 DISCUSSION The results of the current study provide partial support for our hypotheses. More specifically, in line with Joiner and colleagues (1999a) extension of Coyne s (1976) interpersonal theory of depression, higher levels of reassurance seeking were associated with higher levels of current depressive symptoms. Further, in line with the vulnerability stress component of Joiner and colleagues model, higher levels of reassurance seeking were associated with greater increases in depressive symptoms following increases in negative events. It is important to note that such findings are consistent with those obtained in past cross sectional (Abela, Hankin, et al., 2005; Joiner, 1999; Joiner et al., 2001) and longitudinal (Abela, Zuroff et al., 2006; Prinstein et al., 2005) research with youth, thus adding to a growing body of literature suggesting that excessive reassurance serves as a vulnerability factor to depressive symptoms in both children and early adolescents. In addition, the current results expand upon those obtained in past prospective research (Joiner & Metalsky, 2001; Katz et al., 1998; Prinstein et al., 2005) by demonstrating that excessive reassurance seeking interacts with within subject 1. Given that past research examining the stress exposure model (e.g., Potthoff et al., 1995) has examined the effect of reassurance seeking on hassles within a main effect framework, we also examined the hierarchical linear model reported in Table 4 in a stepwise fashion. In doing so, we first examined the main effects of RSSC and SEQ on FU_HASSLES. Next, we examined the cross-level interactions between RSSC/SEQ and FU_CDI. Finally, we examined the three way interaction between RSSC, SEQ, and FU_CDI. In our first model, after controlling for initial levels of depressive symptoms, neither reassurance seeking (b=1.40, SE = 2.91, F(1, 49) = 0.23, ns) nor self esteem (b=3.44, SE = 2.94, F(1, 49) = 1.37, ns) were associated with children s level of hassles during the 8-week follow up interval. In our second model, although fluctuations in children s levels of depressive symptoms were significantly associated with fluctuations in their levels of hassles (b=0.60, SE = 0.21, F(1, 203) = 7.68, p <.01), the strength of this association was not moderated by either reassurance seeking (b=0.33, SE = 0.30, F(1, 203) = 1.24, ns) or self esteem (b= 0.02, SE = 0.32, F(1, 203) = 0.00, ns). In our final model, neither RSSC SEQ (b=2.31, SE = 2.34, F(1, 45) = 0.98, ns) nor its interaction with FU_CDI (b= 0.37, SE = 0.21, F(1, 203) = 3.05, ns) were significant predictors of children s levels of hassles during the follow up interval.

15 EXCESSIVE REASSURANCE SEEKING 863 TABLE 4. Hierarchical Linear Modeling Analyses: Predicting Fluctuations in Negative Events During the Follow up Interval Predictor b SE F df Time 1 CDI RSSC SEQ FU_CDI ** 202 RSSC CDI SEQ CDI RSSC SEQ RSSC SEQ CDI Note. CDI = Children s Depression Inventory. RSSC = Reassurance Seeking Scale for Children. SEQ = Self Esteem Questionnaire. FU_CDI = Within subject fluctuations in CDI scores during the follow up interval. ***p <.001. **p <.01. *p <.05. fluctuations in stress, as opposed to the between subject effect of stress, to predict within subject fluctuations in depressive symptoms. Given that Joiner and colleagues model posits that increases in stress, rather than absolute levels of stress, trigger reassurance seeking behavior in vulnerable individuals, future research should continue to examine this hypothesis from an idiographic, as opposed to a nomothetic, perspective. Doing so is likely to lead to more consistent and powerful support for the vulnerability hypothesis of Joiner and colleagues extension of Coyne s theory. 2 Contrary to our hypotheses, the association between excessive reassurance seeking and greater increases in depressive symptoms following increases in negative events was not moderated by self esteem. One possible explanation for the lack of support for our integrative model is that we examined trait self esteem as a moderator rather than self esteem lability. More specifically, research investigating self esteem has typically used a trait conceptualization in which an individual s level is measured at a single time point and day to day fluctuations are dis- 2. In order to examine whether similar results to those reported in the current article would be obtained using a nomothetic approach to analysis, we examined a hierarchical linear model that included Time 1 CDI, RSSC, SEQ, the between subject effect of stress (MN_HASSLES), and all two and three way interactions involving RSSC, SEQ, and MN_HASSLES. Although youth experiencing a level of stress that was high in comparison to the sample s average level of stress reported higher levels of depressive symptoms (b= 0.15, SE = 0.02, F(1,41) = 40.04, p <.001), the strength of this association was not moderated by RSSC (b=0.05, SE = 0.03, F(1, 41) = 0.72, ns), SEQ (b=0.02, SE = 0.04, F(1, 41) = 0.32, ns), or the RSSC SEQ interaction (b = 0.01, SE = 0.03, F(1, 41) = 0.04, ns).

16 864 ABELA ET AL. missed as measurement error (e.g., Abela & Taylor, 2003; Abela, Webb, Wagner, Ho, & Adams, 2006). In line with this reasoning, we measured children s level of self esteem only during the initial assessment. Recent research however, suggests that self esteem lability may play a stronger role in vulnerability to depression than does trait level of self esteem (Butler, Hokanson & Flynn, 1994; Roberts & Kassel, 1997). The term lability signifies fluctuation, whereby an individual s level of self esteem is continuously influenced by daily threats and boosts. Irrespective of whether or not they possess low or high trait levels of self esteem, youth with high levels of self esteem lability may be especially reactive to negative events as disturbances in their social worlds trigger disturbances in their self perception. When coupled with excessive reassurance seeking, self esteem lability may be particularly problematic because negative events are likely to lead to reassurance seeking behaviors that are apt to be met with rejection (Joiner, 1999; Joiner et al., 1992, 1993; Joiner & Metalsky, 1995; Katz & Beach, 1997) further amplifying instability in sense of self and consequently depressive symptoms. In light of our findings, we would suggest that future research should examine whether the association between reassurance seeking and fluctuations in depressive symptoms following fluctuations in negative events is moderated by self esteem lability. The current study advances research examining Coyne s (1976) interpersonal theory of depression as extended by Joiner and colleagues in several ways. First, the current study is one of the first prospective studies to examine whether excessive reassurance seeking serves as a vulnerability factor to depressive symptoms in a youth sample (for additional prospective studies, see Abela, Zuroff, et al., 2006; Prinstein et al., 2005). Second, the current study is one of the first to examine whether the association between excessive reassurance seeking and depressive symptoms is moderated by the occurrence of negative events (for additional studies, see Abela, Zuroff, et al., 2006; Joiner & Metalsky, 2001; Katz et al., 1998; Prinstein et al., 2005). Last, the current study is one of the first to test the vulnerability stress hypothesis of Joiner and colleagues extension of Coyne s interpersonal theory of depression using a multiwave longitudinal design and an idiographic approach to analysis (see Abela, Zuroff, et al., 2006) as well as the first to do so using experience sampling methodology. At the same time, several limitations of the current study should be noted. First, self report measures were used to assess depressive symptoms. Although both the CDI and the BDI possess high degrees of reliability and validity, one cannot draw conclusions about clinically diagnosed depression based on self report questionnaires. Second, self report measures were used to assess stress. Although measures of

17 EXCESSIVE REASSURANCE SEEKING 865 life events that require participants only to indicate whether or not an event occurred are probably less likely to be influenced by informant bias than those that ask subjects to rate the subjective impact of each event, more sophisticated methods of analysis such as interviewing procedures that assess contextual threat may provide better assessments of stress (Brown & Harris, 1978). Third, although the current study examined, and failed to provide support for, a stress exposure model (i.e., children with high levels of reassurance seeking are exposed to a greater number of negative events), the design of the current study did not allow us to examine a stress generation model (i.e., children with high levels of reassurance seeking contribute to higher levels of dependent interpersonal, but not dependent noninterpersonal and independent negative, events; Hammen, 1991). A true examination of such a model requires an interview-based assessment of stressors in which a team of coders blind to each youth s vulnerability status rates stressors as dependent or independent of youths behaviors and characteristics (Hammen, 2006). Thus, future research is needed to exame the degree to which youth with excessive reassurance seeking tendencies contribute to the negative events that occur in their lives particularly those in the interpersonal domain. Fourth, the current study examined only the relationship between excessive reassurance seeking and depressive symptoms. Thus, we were unable to identify whether excessive reassurance seeking served as a vulnerability factor specifically to depressive symptoms. Future research should assess a broader range of psychological symptoms in order to examine the specificity of excessive reassurance seeking to depressive symptoms in early adolescents. Fifth, the current study did not control for additional variables that could potentially account for the relationship between reassurance seeking and depressive symptoms. Consequently, we cannot rule out the possibility that a third variable could account for the pattern of findings obtained. At the same time, it is important to note that research with adult populations has obtained support for the relationship between reassurance seeking and depressive symptoms even after controlling for self esteem (e.g., Joiner et al., 1992), interpersonal difficulties (e.g., Potthoff et al., 1995), attachment insecurity (Davila, 2001), and neuroticism (Joiner, Metalsky, Katz, & Beach, 1999b). Last, the current study utilized a sample of self identified, high risk parents (predominantly mothers) and their children. Although such a design leads to a powerful test of the vulnerability hypothesis, results cannot be generalized to other populations (e.g., samples not recruited specifically on the basis of parental depression). Future research is needed to examine whether similar results are obtained using a community sample of children and their parents. In sum, the results of the current study indicate that excessive reassur-

18 866 ABELA ET AL. ance seeking is associated with greater increases in depressive symptoms following increases in negative events in children and early adolescents. At the same time, contrary to our integrative model, self esteem was not shown to moderate this relationship. As future research examining the role of interpersonal factors in the etiology of depression accumulates, a greater understanding of the mechanisms that underlie the onset of depression in children will likely emerge. Such an enhanced understanding is likely to improve our ability to identify children at risk for depression, to prevent such children from experiencing future depressive episodes, and to provide successful treatment programs for children already experiencing depressive disorders. REFERENCES Abela, J.R.Z., & Hankin, B.L. (in press). Cognitive vulnerability to depression in children and adolescents: A developmental psychopathology perspective. In J.R.Z. Abela & B.L. Hankin (Eds.), Handbook of Child and Adolescent Depression. New York: Guilford Press. Abela, J.R.Z., Hankin, B.L., Haigh, E.A.P., Adams, P.A., Vinokuroff, T., & Trayhern, L. (2005). interpersonal vulnerability to depression in high risk children: The role of insecure attachment and reassurance seeking. Journal of Clinical Child and Adolescent Psychology, 34, Abela, J.R.Z., Skitch, S.A, Adams, P. A., & Hankin, B.L. (2006). The timing of parent and child depression: A hopelessness theory perspective. Journal of Clinical Child and Adolescent Psychology, 35, Abela, J.R.Z., Skitch, S.A., Auerbach, R.P., & Adams, P.A. (2005). The impact of borderline personality disorder on vulnerability to depression in children of affectively ill parents. Journal of Personality Disorders, 19, Abela, J.R.Z, & Taylor, G. (2003). Specific vulnerability to depressive mood reactions in schoolchildren: The moderating role of self esteem. Journal of Clinical Child and Adolescent Psychology, 32, Abela, J.R.Z, Webb, C.A, Wagner, C., Ho, M.H.R., & Adams, P.A. (2006). The role of self criticism, dependency, and hassles in the course of depressive illness: A multi wave longitudinal study. Personality and Social Psychology Bulletin, 32, Abela, J.R.Z., Zuroff, D.C., Ho, M.H.R., Adams, P.A., & Hankin, B.L. (2006). Excessive reassurance seeking, hassles, and depressive symptoms in children of affectively ill parents: A multi wave longitudinal study. Journal of Abnormal Child Psychology, 34, Altham, P.M.E. (1984). Improving the precision of estimation by fitting a model. Journal of the Royal Statistical Society, Series B, 46, Brown, G.W., & Harris, T. (1978). Social origins of depression: A study of psychiatric disorder in women. London: Tavistock Publications. Brown, K.W., & Moskowitz, D.S. (1998). Dynamic stability of behavior: The rhythms of our interpersonal lives. Journal of Personality, 66, Butler, A.C., Hokanson, J.E., & Flynn, H.A. (1994). A comparison of self esteem lability and

19 EXCESSIVE REASSURANCE SEEKING 867 low trait self esteem as vulnerability factors for depression. Journal of Personality and Social Psychology, 66, Cote, S., & Moskowitz, D.S. (1998). On the dynamic covariation between interpersonal behavior and affect: Prediction from neuroticism, extraversion, and agreeableness. Journal of Personality and Social Psychology, 75, Coyne, J.C. (1976). Depression and the response of others. Journal of Abnormal Psychology, 85, Csikszentmihalyi, M., & Larson, R. (1987). Validity and reliability of the experience sampling method. Journal of Nervous and Mental Disease, 175, Davila, J. (2001). Refining the association between excessive reassurance seeking and depressive symptoms: The role of related interpersonal constructs. Journal of Social and Clinical Psychology, 20, devries, M.W., & Delespaul, P.A.E.G. (1989). Time, context, and subjective experiences in schizophrenia. Schizophrenia Bulletin, 15, devries, M.W., Delespaul, P.A.E.G., & Dijkman Caes, C.I.M. (1987). Affect and anxiety in daily life. In G. Racagni (Ed.), Anxious depression: Assessment and treatment (pp ). New York: Raven Press. devries, M.W., Dijkman Caes, C.I.M, & Delespaul, P.A.E.G. (1990). The sampling of experience: A method of co measuring the co occurrence of anxiety and depression in daily life. In J.D. Maser, & C.R. Cloninger (Eds.), Comorbidity of mood and anxiety disorders (pp ). Washington, DC: American Psychiatric Association. Diggle P., Liang, K Y., & Zeger, S.L. (1994) Analysis of longitudinal data. Oxford: Oxford University Press. First, M.B., Gibbon, M., Spitzer, R.L., & Williams, J.B. (1995). Structured Clinical Interview for Axis I DSM IV Disorders (SCID I) Patient edition. (With Psychotic Screen) Version 2.0, 8/98 revision. Washington, DC: American Psychiatric Press. Goodman, S.H., & Gotlib, I.H. (2002). Introduction. In S.H. Goodman & I.H. Gotlib (Eds.), Children of depressed parents: Mechanisms of risk and implications for treatment (pp. 3 9). Washington, DC: American Psychological Association. Gotlib, I.H., & Beatty, M.E. (1985). Negative responses to depression: The role of attributional style. Cognitive Therapy and Research, 9, Gurtman, M.B. (1987). Depressive affect and disclosure as factors in interpersonal rejection. Cognitive Therapy and Research, 11, Hammen, C. (1991). Generation of stress in the course of unipolar depression. Journal of Abnormal Psychology, 100, Hammen, C. (2006). Stress generation in depression: Reflections on origins, research, and future directions. Journal of Clinical Psychology, 62, Hokanson, J.E., Loewenstein, D.A, Hedeen, C., & Howes, M.J. (1986). Dysphoric college students and roommates: A study of social behaviors over a three month period. Personality and Social Psychology Bulletin, 12, Joiner, T.E., Jr. (1994). Contagious depression: Existence, specificity to depressed symptoms, and the role of reassurance seeking. Journal of Personality and Social Psychology, 67, Joiner, T.E., Jr. (1999). A test of interpersonal theory of depression in youth psychiatric inpatients. Journal of Abnormal Child Psychology, 27, Joiner, T.E., Jr., Alfano, M.S., & Metalsky, G.I. (1992). When depression breeds contempt: Reassurance seeking, self esteem and rejection of depressed college students by their roommates. Journal of Abnormal Psychology, 101, Joiner, T.E,. Jr., Alfano, M.S., & Metalsky, G.I. (1993). Caught in the crossfire: Depression, self consistency, self enhancement, and the response of others. Journal of Social and Clinical Psychology, 12,