Cogn Ther Res (2007) 31:483–502 DOI 10.1007/s10608-006-9071-4 ORIGINAL ARTICLE
Sociotropic and Autonomous Personality and Stressful Life Events as Predictors of Depressive Symptoms in the Postpartum Period Sandra Masih Æ Susan H. Spence Æ Tian P. S. Oei
Published online: 10 October 2006 Springer Science+Business Media, Inc. 2006
Abstract Measures of stressful life events, sociotropic and autonomous personality, and depressive symptoms were completed by 76 women in the last trimester of pregnancy and 8-weeks postpartum. During pregnancy, women with strong sociotropic or autonomous personality style, or high levels of negative life events (or perceived loss resulting from events) in sociotropic or autonomous domains, tended to report higher levels of depressive symptoms. Cross-sectionally, there was some support for the cognitive diathesis-stress model, but not a congruency model, of depression. Longitudinal results indicated that high levels of sociotropic personality style, sociotropic loss, or autonomous loss resulting from events significantly predicted increases in depressive symptoms from pregnancy to the postpartum period. The interaction between personality styles and life events did not predict depressive symptoms longitudinally, either congruently or incongruently. Negative automatic thoughts mediated the direct effects of personality and life events upon depressive symptoms. Keywords Postpartum Æ Depression Æ Predictors Æ Depressive symptoms Æ Sociotropy Æ Autonomy Æ Personality Æ Life events
Pregnancy and childbirth represents a significant life transition, irrespective of whether it is occurring for the first time. Approximately 10–20% of women experience clinically significant levels of depression in pregnancy, with elevated levels also being evident
S. Masih University of Queensland, Brisbane, Australia S. H. Spence (&) Division of Linguistics and Psychology, Macquarie University, Sydney NSW 2109, Australia e-mail:
[email protected] T. P. S. Oei University of Queensland, Brisbane, Australia
123
484
Cogn Ther Res (2007) 31:483–502
following childbirth (Miller & Torpy, 2002; O’Hara, Neunaber, & Zekoski, 1984). The increased risk for depression during this period provides an ideal opportunity to study its causal factors. It is clear that, over and above the biological factors associated with pregnancy and childbirth, there is a complex interplay with psychological and social variables that increases the risk of developing depression in this period (Bloch, Daly, & Rubinow, 2003; Harris, 2002). Pregnancy and childbirth are accompanied by a variety of stressful events, such as strain upon the couple’s relationship, financial complications, change in employment status, and general feelings of fatigue, which together have been proposed to increase the risk of depressive symptoms (Lane et al., 1997; Marks, Wieck, Checkley, & Kumar, 1992; Murray, 1992; O’Hara, 1995). Diathesis-stress models of depression emerged from the observation that not all individuals who are exposed to stressful life situations develop depression. Rather, it is proposed that the impact of adversity is greater for individuals who are vulnerable in some way, as the result of factors intrinsic or extrinsic to the person. Cognitive diathesisstress models of depression postulate that these vulnerabilities, which include certain personality styles and their associated maladaptive cognitions, play an integral part in determining the impact of negative life events upon the development of depressive symptoms (Abramson, Alloy, & Metalsky, 1988; Alloy, Hartlage, & Abramson, 1988; Beck, 1983). The focus of this paper is the cognitive diathesis-stress model proposed by Beck (1983), who postulated that two distinct personality dimensions, sociotropy and autonomy, act as relatively stable vulnerabilities to depression by sensitizing individuals to certain types of events that occur in a domain that is congruent to their dominant personality type. Sociotropic individuals place high value on positive interchange with others, seeking to gain approval and acceptance. The autonomous person in contrast, strives to preserve independence, mobility, and freedom. According to Beck (1987), sociotropic individuals are more likely to develop depression in response to life events perceived as involving a loss in personal relatedness. In the same manner, autonomous individuals are more likely to develop depression in response to events that produce a loss of independence, control, or accomplishment. Studies investigating Beck’s personality-event congruency theory have produced mixed results. Some studies have provided support for the congruency model, demonstrating that increases in depressive symptoms are significantly predicted by the congruent interaction of sociotropic personality with negative interpersonal events and by the congruent interaction of autonomous personality with negative autonomous events, but not by either of the non-congruent interactions (Hammen, Ellicott, Gitlin, & Jamison, 1989b; Mazure & Maciejewski, 2003; Morse & Robins, 2005; Segal & Dobson, 1992). Others have found that congruency only between sociotropic personality and sociotropic stressors predicts increased risk of depression (e.g. Bartelstone & Trull, 1995; Clark, Beck, & Brown, 1992; Hammen, Ellicott, & Gitlin, 1992; Hammen & Goodman Brown, 1990; Hammen, Marks, Mayol, & DeMayo, 1985; Mazure & Maciejewski, 2003; Raghavan, Le, & Berenbaum, 2002; Robins, 1990; Robins & Block, 1988; Segal, Shaw, & Vella, 1989). Yet others have found congruency between autonomous personality and autonomous events to do so (e.g. Baker, Nenneyer, & Barris, 1997; Brown, Hammen, Craske, & Wickens, 1995; Hammen, Ellicott, & Gitlin, 1989a; Hammen et al., 1989b; Segal, Shaw, Vella, & Katz, 1992). Then, some studies have found that although personality types in combination with negative life events may predict subsequent depression, the interaction does not necessarily need to be congruent in terms of sociotropic and autonomous domains (e.g. Allen, de L Horne, & Trinder,
123
Cogn Ther Res (2007) 31:483–502
485
1996; Fresco, Sampson, Craighead, & Koons, 2001; Robins et al., 1995). Finally, to confuse matters even further, some researchers have found no evidence that personality style in combination with adverse life events, either congruently or incongruently, increases the risk of depression over and above the direct effects of personality and life events alone (Kwon & Whisman, 1998; Mazure, Bruce, Maciejewski, & Jacobs, 2000; Mazure, Raghavan, Maciejewski, Jacobs, & Bruce, 2001; Oates-Johnson & Clark, 2004). The variation in findings may reflect methodological differences. Studies have varied markedly in the measures used to assess personality and life events, the population involved, and the use of correlational versus longitudinal designs. For example, some studies have assessed sociotropic and autonomous personality styles as independent categorical constructs, whereas others have regarded them as continuous dimensions (See Coyne, Thompson, & Whiffen, 2004 for a review). Coyne et al. (2004) argued that sociotropic and autonomous styles are best construed as correlated, continuous dimensions, and there is little validity for a categorical approach. The method of measuring life events has also varied, with some studies recording the number of life events experienced in a particular domain, and others evaluating the level of perceived loss associated with those events. Although many studies have demonstrated that personality and life stress are associated with postpartum depression (e.g. Gotlib, Whiffen, Wallace, & Mount, 1991; Matthey, Barnett, Ungerer, & Waters, 2000; O’Hara et al., 1984; O’Hara, Rehm, & Campbell, 1982; O’Hara, Schlechte, Lewis, & Varner, 1991) very few have examined a personality diathesis-stress model. Grazioli and Terry (2000), in a prospective study, found that the interactions between autonomous personality and parental stressors and between sociotropic personality and infant temperament significantly predicted depression outcomes. The authors did not, however, examine congruency between type of personality style and the type of event (i.e., sociotropic versus autonomous). The transition from pregnancy to the postpartum period would appear to be an ideal time to investigate the cognitive diathesis-stress congruency theory of depression. While acknowledging that many women find pregnancy, childbirth, and the early weeks of parenthood to be an enjoyable experience, this period is also one of rapid changes in roles and responsibilities. Women are more likely to experience a range of life events that represent a loss of threat to autonomy and/or interpersonal relationships, such as cessation of regular employment, reduction in income, fewer contacts with colleagues, fatigue, reduced sleep, medical complications, and child care problems, to mention just a few (O’Hara et al., 1982, 1984). Partner relationship conflict and lack of support are also important issues for women during this period (Gotlib et al., 1991). A cognitive diathesis-stress congruency model of depression makes a good deal of intuitive sense for this transitional period. Women with a strong sociotropic personality style, who place high value on positive interpersonal relationships, would be expected to experience greater psychological distress in response to those events that impair the quality of their valued relationships. Similarly, women with a strong autonomous personality style, who place high value on independence and freedom, would be predicted to be more likely to experience distress in response to pregnancy and postpartum events that present a threat to or loss of autonomy. A further component of Beck’s (1976, 1983) cognitive diathesis-stress model of depression is the proposition that the combination of the personality diathesis and congruent stressful life events trigger negative automatic thoughts that mediate the development of depressive symptoms. The mediational role of automatic thoughts has been demonstrated with student populations (Kwon & Oei, 1992; Oei, Goh, & Kwon,
123
486
Cogn Ther Res (2007) 31:483–502
1996) and clinically depressed patients (Kwon & Oei, 2003). It is important to examine the potential mediating role of automatic thoughts in the development of depression during pregnancy and the postpartum period, as this will inform the design of preventive interventions. Such a finding would support the use of cognitive therapy techniques to prevent the development of depression in women at increased risk as a consequence of adverse life events in combination with congruent personality style. The present study examined the cognitive diathesis-stress congruency model using a longitudinal design with a sample of women tracked from the last trimester of pregnancy through to 8 weeks following childbirth. Thus, it provided an opportunity to examine the model cross-sectionally (during late pregnancy) and longitudinally (from pregnancy to the 8 weeks following childbirth). In keeping with Clark, Beck, & Alford (1999) and Coyne et al. (2004) a dimensional approach to the constructs of sociotropic and autonomous personality was taken, rather than defining them as categorically distinct latent personality types, enabling all participants to be retained in the analyses. The contribution of negative life events was examined in two ways, first using the number of occurring events that had been pre-categorized as either sociotropic or autonomous in terms of the type of loss generally experienced and second using participants’ perceptions of sociotropic or autonomous loss for all stressful life events, irrespective of original categorization. It was hypothesized that women who reported high levels of sociotropic or autonomous personality style during pregnancy, in combination with high negative life events of a congruent nature during late pregnancy and the 8 weeks following childbirth, would be at increased risk of developing depressive symptoms in the postnatal period. Furthermore, it was predicted that the impact of this interaction between personality style and congruent stressful life events upon depression would be mediated by negative automatic thoughts.
Method Participants Participants were women in the 3rd trimester of pregnancy, recruited from antenatal classes at one of four Queensland hospitals. The hospitals were chosen to represent a broad socioeconomic spectrum of their client-base. Of the 230 individuals invited to participate in the study, 103 women gave consent to participate (44.8%). Of the 103 women who agreed to take part, 82 (79.6%) completed the full questionnaire battery in the last trimester of pregnancy (Time 1: T1), with 52 being public and 30 private hospital patients. At approximately 8-weeks postpartum (Time 2: T2), 76 (93% of the initial sample) provided Time 2 data, and 6 (7%) women failed to return completed questionnaires or could not be contacted. At entry to the study, mother’s ages ranged from 18 to 42 years with a mean age of 30.8 years (SD = 5.21 years). Gestational age ranged between 22 and 39 weeks with a mean gestational age of 34.4 weeks (SD = 4.23 weeks). The majority of the sample were married (70.7%) or living in a defacto relationship (25.6%). The majority of women were of Caucasian ethnicity (94%). Their highest level of education included high school (Grade 10 to 12–40%), vocational training college (22%), university undergraduate (Bachelor degree—17%), and university postgraduate (Master or doctoral degree—13%). Mother’s occupations were coded using the Australian Standard
123
Cogn Ther Res (2007) 31:483–502
487
Classifications of Occupations (ASCO) and Socio Economic Status (SES) was coded using income categories for combined couple income as defined by the Australian Bureau of Statistics (ABS, 1997). Comparisons indicated that the sample was representative of both the local and national population from which it was drawn in terms of level of education, occupation, and SES. Exclusion criteria were the inability to speak and read English fluently and severe intellectual impairment, although no volunteers had to be excluded from the study. The majority of women (94%) reported that they were in either ‘‘good’’ current health, or ‘‘excellent’’ current health status, 97.6% of women reported no prescribed medicines, 62.2% were pregnant for the first time and 70.7% reported no past psychiatric history. Procedure Following ethical approval, the researcher attended the antenatal classes to outline the proposed research (but not the specific theoretical content) and invite volunteers. Following receipt of written, informed consent, interested individuals were provided with a questionnaire package to complete during the following week (T1). The measures were presented in random sequence to reduce any bias associated with the order of presentation. Participants were offered a place in a free parenting skills workshop in recognition of their involvement in the study. This workshop was held after completion of all follow-up assessments. The assessment package was collected approximately one week following distribution. Participants were re-contacted approximately 8-weeks postpartum (T2). Questionnaire packages were mailed out and returned by self-addressed stamped envelope. Measures Edinburgh Postpartum Depression Scale (EPDS: Cox, Chapman, Murray, & Jones, 1996; Cox, Holden, & Sagovsky, 1987). The EPDS is a 10 item self-rating measure on which women select one of four responses, 0 (no, not at all) to 3, (yes, very often), relating to depressive symptoms during the previous 7 day period. It was developed specifically to assess depressive symptoms during the postpartum period and avoids items relating to physiological symptoms, such as sleep disturbance and weight change associated with the puerperium. Overall, the psychometric properties of the scale are reported to be good (Schaper, Rooney, Kay, & Silva, 1994). Studies have found the EPDS is a more sensitive instrument than the BDI for the detection of depression in the postnatal period (Beck & Gable, 2001; Cox et al., 1987; Murray & Carothers, 1990). It has also been validated as a measure of depression during pregnancy (Green & Murray, 1994). Research has indicated that an EPDS cut-off point of 13 provides high levels of sensitivity and specificity in the detection of cases of postpartum depression (Cox et al., 1996; Harris et al., 1989). Internal consistency of the scale in the present study was .90 and test–retest reliability .43. The Beck Depression Inventory (BDI: Beck & Steer, 1987; Beck & Gable, 2001). The BDI was used in order to provide a validation check for the results using the EPDS. The alpha coefficient for the current study at T1 was .87 and test–retest reliability .46 suggesting moderate stability over time.
123
488
Cogn Ther Res (2007) 31:483–502
Life Events Record (LER: Clark et al., 1992). An abbreviated LER was used to assess the occurrence of 56 life events and associated perceptions of loss, relating to work, income, education, physical health, close relationships, children, and family issues, social activities, legal matters, and death/injury/illness of friends or relatives. Events represented 28 sociotropic and 28 autonomous themes as outlined by Clarke et al. (1992). After noting the occurrence/non-occurrence of each event women completed three, 1–5 ratings of (1) degree of general impact—‘‘how upsetting was this event for you?’’; (2) perceived sociotropic loss associated with that event ‘‘how much did this event result in a disruption to your relationship with others?’’; and (3) perceived loss of autonomy ‘‘how much did this event result in a loss or your independence, freedom, or life goals?’’. Scores were produced for the total number of stressful life events (autonomous, sociotropic or combined), level of sociotropic and autonomous loss, and level of upset associated with these events. Examples of negative life events in the autonomy domain included ‘‘became unemployed (fired, suspended, laid off, quit)’’, ‘‘serious illness requiring medical treatment’’, ‘‘major financial difficulties’’, and ‘‘decrease in the amount of time spent on personal activities’’. Examples of negative sociotropic events included ‘‘serious arguments with spouse/partner’’, ‘‘increase in spouse time away from the family’’, ‘‘increased conflict with in-laws or relatives’’ and ‘‘no contact with close friends’’. The item ‘‘pregnancy’’ at T1 was replaced with ‘‘birth of a child’’ at T2. Pregnancy was categorized as an autonomous stressor, as participants’ ratings indicated a greater level of perceived loss of autonomy (M = 1.70, SD = 1.03) compared to sociotropic loss for this item (M = 1.38, SD = .85), t(80) = 3.01, P = .003. Similarly, birth of a child was categorized as an autonomous stress, as this item was also rated as producing a higher level of loss of autonomy (M = 1.91, SD = 1.14) compared to sociotropic loss (M = 1.17, SD = .53), t(76) = 5.90 P < .001. It should be noted, however, that although both pregnancy and childbirth were rated as resulting in greater autonomous than sociotropic loss, the ratings indicated that the majority of women did not regard these events as being upsetting or producing a high degree of loss in either domain. The Personality Style Inventory-Revised (PSI-II: Robins et al., 1994) was developed and psychometrically evaluated as a measure of interpersonal and autonomous personality construct vulnerabilities. The 24 items reflect two dimensions of personality (interpersonal and autonomous), each assessing 3 constructs. The three interpersonal constructs include: excessive concerns about what others think, dependency, and desire to please others. The three autonomous constructs include: perfectionism/self-criticism, need for control, and defensive distance from others. Each item is rated for the level of agreement on a 1–6 scale. The PSI-II is reported to have good content validity, convergent and discriminant validity, good factor structure, and temporal stability (see Robins et al., 1994). Coefficient alphas in the current study were .92 for autonomous personality, and .90 for sociotropic personality subscales. Test–retest reliability for sociotropic personality was .77, and autonomous personality .78 suggesting high stability over time. Automatic Thoughts Questionnaire (ATQ: Hollon & Kendall, 1980). The ATQ is a 30-item cognitive inventory designed to assess the frequency of occurrence of automatic negative thoughts associated with depression. Respondents are asked to rate each item on a five-point scale indicating frequency of occurrence of each negative self-statement (1 = not at all, 5 = all the time). Higher scores indicate greater frequency of depressotypic cognitive rumination. The psychometric properties are reported to be strong
123
Cogn Ther Res (2007) 31:483–502
489
(Harrell & Ryon, 1983; Hollon & Kendall, 1980). Test-test reliability for the ATQ for this study was calculated (r = .45), with internal consistency of .95
Results Descriptive characteristics of the sample Table 1 shows the mean values for the sample for all measures at T1 and T2. Analyses using the BDI revealed results that closely mirrored those for the EPDS, thus only the findings for the EPDS are reported here. Depressive symptoms No significant difference was found between the level of EPDS depressive symptoms during the third trimester of pregnancy and the first 8-weeks postpartum. The EPDS means of 7.2 at T1 and 6.5 at T2 are consistent with levels reported by Grazioli and Terry (2000) during the pre- and post-natal periods respectively. Using the EPDS cutoff of 13 or more, 14 (17.3%) women were classified as experiencing clinically significant depressive symptoms during the third trimester of pregnancy (T1) and 14 (18.6%) at the postpartum assessment (T2). Six women did not provide EPDS data at T2, 3 who initially reported elevated EPDS scores and 3 who did not. Of those present on both occasions, 4 of 11 (36.3)%) women who reported elevated depression scores at T1 also did so at T2, compared to 7 of 11 (63.7%) who reported high scores at T1 but not at T2. Of the 64 women who reported scores below the EPDS cut-off at T1, 10 (15.6%) reported elevated scores at T2, compared to 54 (84.4%) whose subsequent scores were again below the cut-off. Life events Ratings of autonomous and sociotropic loss for autonomous and sociotropic events at T1 confirmed that in general women reported significantly higher ratings of sociotropic loss (M = 3.23, SD = 4.10) than autonomous loss (M = 2.63, SD = 3.62) for events that had been categorized as sociotropic (t = 3.66, df = 80, P < .001). Similarly, ratings of autonomous loss (M = 5.37, SD = 5.13) were significantly higher than sociotropic loss (M = 4.01, SD = 4.31) for autonomous events (t = 6.03, df = 80, P < .001). Although these findings are consistent with the sociotropic or autonomous categorization of events, it is clear that both sociotropic and autonomous events were associated with perceptions of loss in both domains. This justified examination of the impact of life events in two ways: first for the number of events occurring in each domain and second in terms of level of sociotropic and autonomous loss for all stressful life events, irrespective of original categorization. The most commonly occurring events in pregnancy were becoming unemployed (32.9%), serious arguments with spouse/partner (22%), increased conflict with in-laws or relatives (14.6%), decrease in the amount of time spent on personal activities (13.4%), increase in spouse time spent away from family (12%), and major financial difficulties (11%). At the postpartum assessment, the most commonly reported events were decreased amount of time spent on personal activities (41.5%), major surgery (22%), becoming unemployed (18.3%), and serious arguments with spouse/partner (18.3%).
123
123 1
2
Time 2 variable 3
4
5
7
8
9
.28** .26** .36***
.09 .25* .30** .37** .41** .42***
.34** .45*** .49***
.20 .30** .42*** .47*** .48*** .47***
.38** .55*** .58*** .57***
.22 .38***
.41*** .50*** .54***
.35** .59*** .61*** .61***
.26* .40***
.45*** .50*** .55***
.82 (1.09) 2.55 (1.37) 5.76 (5.74) 7.92 (7.49)
6
Note: *P < .05; **P < .01; ***P < .001 EPDS = Edinburgh Postnatal Depression Scale; BDI = Beck Depression Inventory; ATQ = Negative automatic thoughts; LE Soc Number = total negative sociotropic events; LE Aut Number = total negative events; LE Sociotropic Loss = sociotropic loss perception for total life events; LE Autonomous loss = autonomous loss perceptions for total life events
Mean (SD) 6.54 (5.52) 7.99 (6.96) 40.17 (13.08) 84.14 (20.26) 75.43 (19.55) Time 1 Variables Dependent 1. EPDS 7.17 (5.00) .43*** .49*** .47*** .45*** .32*** 2. BDI 8.32 (6.38) .39*** .46*** .42*** .36*** .35*** 3. ATQ 39.83 (12.84) .35*** .42*** .45*** .45*** .37*** Independent Personality 4. PSI sociotropy 87.64 (19.90) .41*** .38*** .43*** .77*** .49*** 5. PSI autonomy 75.85 (16.75) .34** .32** .36*** .59*** .78*** Life events 6. LE Soc number 1.40 (1.41) .26* .32** .32** .34** .27* 7. LE Aut number 2.44 (1.57) .25* .46*** .39*** .34** .30** 8. LE sociotropic loss 7.25 (7.53) .35** .52*** .48*** .45*** .38** 9. LE autonomous loss 7.99 (7.73) .34** .52*** .47*** .44*** .35**
Mean (SD)
Table 1 Means, standard deviations and correlations for time 1 and time 2 variables
490 Cogn Ther Res (2007) 31:483–502
Cogn Ther Res (2007) 31:483–502
491
Bivariate correlations None of the demographic variables correlated significantly with depressive symptoms or the independent variables and were not therefore included as control variables in subsequent regression analyses. Inter-correlations among the independent and dependent variables from T1 to T2, are shown in Table 1. Significant associations were evident between EPDS depression scores at T2 and all T1 measures of personality style and negative life events in pregnancy. Regression analyses: cross-sectional findings during pregnancy Analyses were conducted separately for sociotropic and autonomous personality style, with the centered personality variable entered at step 1, followed by the centered number of life events in either the congruent, or incongruent domain at step 2 and finally the interaction term at step 3. As can be seen from Table 2, at step 1 both sociotropic and autonomous personality styles were significantly associated with EPDS depressive symptoms during pregnancy. Both sociotropic and autonomous events then further contributed to the variance in depression scores at step 2. However, the only significant interaction at step 3 was for the incongruent interaction between autonomous personality and sociotropic life events, b = .21, t = 2.25, P = .03. The nature of the association between depression and this interactive composite was examined as outlined by Aiken and West (1991). Depression scores as predicted by the regression equation were computed for high and low values of autonomous personality and sociotropic life events (1 SD above and 1 SD below the mean) and these findings are shown in Fig. 1. The results revealed that higher EPDS scores were associated with a combination of high autonomous personality and high sociotropic life events during pregnancy (9.32) whereas women with low autonomous personality and low sociotropic life events tended to report particularly low depression scores (2.03). The combination of high autonomous personality and low sociotropic life events, or low autonomous personality and high sociotropic life events predicted EPDS scores of 8.09 and 8.02 respectively. Given that sociotropic life events were associated with a degree of autonomous loss, in addition to sociotropic loss as noted above, we examined whether this finding could perhaps reflect an interaction between autonomous personality and the autonomous loss associated with sociotropic events (which would theoretically be a congruent interaction). However, this interaction did not significantly predict EPDS scores. The proposed role of negative automatic thoughts in mediating the effects of personality style and life events upon depression was then examined following the procedure outlined by Baron and Kenny (1986). To test for mediation, a series of regression analyses are first conducted to demonstrate significant direct effects between the distal variable (personality style or life events or their interaction), the outcome variable (depression) and the proposed mediator (negative automatic thoughts). If all direct effects between these variables are significant, then it is appropriate to examine whether the proposed mediator significantly reduces the beta coefficient in the relationship between the outcome variable and the distal variable when the mediator is entered first into the prediction equation. The Sobel test provides an indictor of the significance of mediation (Preacher & Leonardelli, 2001). As noted above, significant direct effects were evident between T1 EPDS scores and T1 PSI sociotropic personality, b = .55, t = 5.72, P < .001, PSI autonomy personality, b = .42, t = 4.09, P < .001, sociotropic
123
123 .44*** .32***
.46*** .29** .10
Step 2 PSI Soc Sociotropic events
Step 3 PSI Soc Sociotropic events PSI Soc · Sociotropic events
.45*** .33** .02
Step 3 PSI Soc Autonomous events PSI Soc · Autonomous events 4.86 2.83 .19
4.89 3.75
5.72
4.77 2.92 1.05
4.65 3.40
5.72
t
.00
.11
.40
.01
.09
.30
R2 D
.04
14.03**
32.67***
1.11
11.57***
32.67***
F of R2 D
(1,75)
(1,76)
(1,77)
(1,75)
(1,76)
(1,77)
df
Step 3 PSI Aut Sociotropic events PSI Aut x Sociotropic events
Step 2 PSI Aut Sociotropic events
(d) Incongruent Step 1 PSI Aut
Step 3 PSI Aut Autonomous events PSI Aut x Autonomous events
Step 2 PSI Aut Autonomous events
(c) Congruent Step 1 PSI Aut
Autonomy
Note: *P < .05; **P < .01; ***P < .001 PSI Soc = Personality Style Inventory-Revised Sociotropy; PSI Aut = Personality Style Inventory-Revised Autonomy
.45*** .34***
Step 2 PSI Soc Autonomous events
.55***
.55***
(a) Congruent Step 1 PSI Soc
(b) Incongruent Step 1 PSI Soc
b
Sociotropy
.37*** .36*** .21*
.34*** .39***
.42***
.31** .34** .06
.30** .37***
.42***
b
3.82 2.07 2.25
3.47 4.05
4.09
3.02 2.80 .50
3.01 3.69
4.09
t
.04
.15
.18
.002
.12
.18
R2 D
5.08*
16.41***
16.74***
.25
13.66***
16.74***
F of R2 D
(1,75)
(1,76)
(1,77)
(1,75)
(1,76)
(1,77)
df
Table 2 Cross-sectional regression analyses: sociotropic and autonomous personality style, autonomous and sociotropic life events and congruent and incongruent interactions as predictors of depressive symptoms (EPDS) in pregnancy
492 Cogn Ther Res (2007) 31:483–502
Cogn Ther Res (2007) 31:483–502 Fig. 1 EPDS depression scores as a function of autonomous personality style and sociotropic life events
493
10 9 8 7 6 5 4 3 High autonomous personality
2
Low autonomous personality
1 0
Low
High Sociotropic Life Events
events, b = .47, t = 4.67, P < .001 and autonomy events, b = .47, t = 4.71, P < .001. Direct effects were evident between T1 negative automatic thoughts and T1 depressive symptoms, b = .73, t = 9.37, P < .001. Also, T1 negative automatic thoughts were significantly associated with T1 sociotropic personality, b = .58, t = 6.34, P < .001, autonomous personality, b = .54, t = 5.80, P < .001, sociotropic events, b = .34, t = 3.22, P = .002, and autonomous events, b = .54, t = 5.66, P < .001. When negative automatic thoughts were entered into the equation at the same step as the distal variable (personality or life events) in the prediction of EPDS scores, the b values dropped (in comparison to the b values for the direct effect without the mediator included in the equation) for PSI sociotropy, b = .18, t = 1.94, P = .06, PSI autonomy, b = .03, t = .34, P = .74, sociotropic events, b = .25, t = 3.23, P = .002, and autonomous events, b = .11, t = 1.25, P = .21. Sobel tests indicated significant mediation by negative automatic thoughts in the relationship between depressive symptoms and the distal variables of sociotropic personality, z = 4.58, P < .001, autonomous personality style, z = 4.63, P < .001, sociotropic events, z = 3.00, P = .003, and autonomous events, z = 4.48, P < .001. Also, the interaction of autonomous personality and sociotropic events was no longer significantly associated with depression scores once automatic thoughts were entered into the equation, again consistent with a mediational model. Figure 2 summarizes the mediational results for the cross-sectional model in pregnancy. The effects of perceived sociotropic and autonomous loss for all life events (irrespective of autonomous or sociotropic categorization) were also examined in the prediction of EPDS scores in pregnancy. Sociotropic loss significantly predicted EPDS scores when entered after sociotropic personality, b = .42, t = 4.38, P < .001. Similarly autonomous loss predicted EPDS scores at step 2 after autonomous personality, b = .51, t = 5.33, P < .001. However, neither the congruent or incongruent interactions of sociotropic or autonomous loss with sociotropic or autonomous personality significantly predicted EPDS scores in pregnancy. Similarly, the interactions between total number of life events (for both categories combined) and personality styles did not significantly predict EPDS depressive symptoms in pregnancy.
123
494
Cogn Ther Res (2007) 31:483–502
Sociotropic Personality Autonomous Personality
Negative Automatic Thoughts
Depressive Symptoms
Sociotropic Life Events
Autonomous Life Events
Fig. 2 Mediational role of negative automatic thoughts in the relationship between personality, life events and depressive symptoms in pregnancy
Regression analyses: cross-sectional findings at the postpartum period The cross-sectional results for pregnancy were replicated in full for the cross-sectional analyses using the postpartum data, although to save space these results are not reported in full. Again, women who in the postpartum period reported a high sociotropic or autonomous personality style, or high sociotropic or autonomous negative life events, were more likely to report high depression scores. Also, the cross-sectional results at T2 replicated the effect for the significant interaction between autonomous personality style and sociotropic events in the prediction of depression scores. Regression analyses: longitudinal findings after childbirth The longitudinal test of the cognitive diathesis-stress congruency model predicted depressive symptoms in the postpartum period from personality style assessed in late pregnancy and life events occurring in the transitional period from late pregnancy through the postpartum period. It is debatable whether a longitudinal test of the diathesis-stress congruency model should examine historical life events occurring prior to the baseline assessment (T1—pregnancy) or those occurring in the intervening period between baseline and the subsequent assessment (T1 to T2—pregnancy to the postpartum). The problem with this latter approach is the risk of bias in reporting of life events, reflecting concurrent depressive symptoms. However, while acknowledging this limitation, it was considered that the most appropriate test of the model was to examine the interaction between pre-existing personality style in pregnancy (T1) and the occurrence of events in the transitional months between late pregnancy and the postpartum period, assessed at T2 in predicting postnatal depressive symptoms. The results of these analyses are shown in Table 3. At step 1, the analysis controlled for levels of T1 depressive symptoms, and indicated that depression during pregnancy significantly predicted postpartum depressive symptoms, b = .43, t = 4.10, P < .001. At step 2, T1 sociotropic personality, b = .25, t = 2.05, P = .04, significantly predicted T2 EPDS scores. However, T1 autonomous personality, T2 autonomous events, and T2 sociotropic events did not significantly predict T2 EPDS scores, after controlling for T1 depression. None of the congruent or incongruent interactions between T1 personality style and T2 life events significantly predicted T2 depressive symptoms.
123
t
1.94 2.04 1.85 .035 2.05 2.05 1.85 1.16 .014
.24 .24 .20 .25* .24* .20 .12
2.53 2.05 .045 2.07 2.16 1.22 .016 2.10 2.15 1.19 .40 .002
.30* .25* .26* .26* .13 .27* .26* .13 .04
.43*** 4.11 .188
2.53 2.05 .045
.30* .25*
.16
1.48
4.20*
16.87***
1.34
3.42
4.20*
16.87***
(1,70)
(1,71)
(1,72)
(1,73)
(1,70)
(1,71)
(1,72)
(1,73)
R2 D F of R2 D df
.43*** 4.11 .188
b
Note: *P < .05; **P < .01; ***P < .001
(a) Congruent Step 1 EPDS1 Step 2 EPDS1 PSI Soc 1 Step 3 EPDS1 PSI Soc 1 Sociotropic events 2 Step 4 EPDS1 PSI Soc 1 Sociotropic events 2 PSI Soc · Sociotropic events 2 (b) Incongruent Step 1 EPDS1 Step 2 EPDS1 PSI Soc 1 Step 3 EPDS1 PSI Soc 1 Autonomous events 2 Step 4 EPDS1 PSI Soc 1 Autonomous Events 2 PSI Soc · Autonomous events 2
Sociotropy (c) Congruent Step 1 EPDS1 Step 2 EPDS1 PSI Aut 1 Step 3 EPDS1 PSI Aut 1 Autonomous Events 2 Step 4 EPDS1 PSI Aut 1 Autonomous events 2 PSI Aut x Autonomous Events 2 (d) Incongruent Step 1 EPDS1 Step 2 EPDS1 PSI Aut 1 Step 3 EPDS1 PSI Aut 1 Sociotropic Events 2 Step 4 EPDS1 PSI Aut 1 Sociotropic events 2 PSI Aut x Sociotropic Events 2
Autonomy
t
2.96 1.53 .82 .51 .003
2.94 1.65 .73 .006
3.18 1.79 .035
.32** .17 .20 .11
.31** .17 .17
.36** .20
2.78 1.47 1.73 .99 .011
2.71 1.47 1.55 .025
3.18 1.79 .035
.43*** 4.11 .188
.34** .18 .09 .06
.34** .19 .08
.36** .20
.99
2.39
3.21
16.87***
.261
.54
3.21
16.87***
(1,70)
(1,71)
(1,72)
(1,73)
(1,70)
(1,71)
(1,72)
(1,73)
R2 D F of R2 D df
.43*** 4.11 .188
b
Table 3 Longitudinal regression analyses: time 1 sociotropic and autonomous personality style, transitional autonomous and sociotropic life events and congruent and incongruent interactions as predictors of time 2 postpartum depressive symptoms
Cogn Ther Res (2007) 31:483–502 495
123
496
Cogn Ther Res (2007) 31:483–502
All longitudinal analyses were repeated for perceived sociotropic and autonomous loss associated with total life events. Both T2 sociotropic loss, b = .29, t = 2.62, P = .01, and T2 autonomous loss, b = .32, t = 3.12, P = .003, significantly predicted T2 EPDS scores after controlling for T1 EPDS and congruent personality style. Again none of the interactions, either congruent or incongruent, between T1 personality style and T2 loss associated with events significantly predicted T2 EPDS scores. The potential mediational role of negative automatic thoughts was then examined for those variables that significantly predicted change in EPDS scores, namely T1 sociotropic personality and T2 sociotropic and T2 autonomous loss associated with events. Negative automatic thoughts at T2 significantly predicted T2 EPDS scores after controlling for T1 symptoms, b = .75, t = 9.05, P < .001. Also, T2 negative automatic thoughts were significantly predicted by T1 sociotropic personality, b = .43, t = 4.09, P < .001, T2 sociotropic loss, b = .45, t = 4.29, P < .001, and T2 autonomous loss, b = .51, t = 5.03, P < .001, thus fulfilling all criteria for direct effects required prior to testing for mediation. When T2 ATQ scores were entered into the equation in the prediction of T2 EPDS scores, the b values were no longer significant for T1 PSI sociotropic personality, b = .07, t = .78, P = .44, T2 sociotropic loss, b = .12, t = 1.47, P = .15, or T2 autonomous loss, b = .14, t = 1.56, P = .15. The Sobel test indicated that T2 negative automatic thoughts significantly mediated the association between T2 depressive symptoms and T1 sociotropic personality, z = 3.55, P = .001, T2 sociotropic loss, z = 3.77, P < .001, and T2 autonomous loss, z = 4.26, P < .001. Figure 3 summarizes the mediational results for the longitudinal model. Finally, we repeated the analyses taking T1 reports of life events as the indicator of life events. However, none of the interactions between T1 events and T1 personality style significantly predicted T2 EPDS scores, after controlling for T1 levels of depressive symptoms. This was the case for number of sociotropic and autonomous events and for sociotropic or autonomous loss perceptions, both congruently and incongruently with personality style. Influence of initial level of depressive symptoms It was important to determine whether the significant direct effects of T1 sociotropic personality and T2 sociotropic life events in the prediction of T2 depression scores Sociotropic Personality in Pregnancy
Sociotropic Loss in Transition Pregnancy to Postpartum
Postpartum Negative Automatic Thoughts
Changes in Depressive Symptoms from Pregnancy to Postpartum
Autonomous Loss in Transition Pregnancy to Postpartum
Fig. 3 Mediational role of negative automatic thoughts in the relationship between personality, life events and changes in depressive symptoms, from pregnancy to the postpartum period
123
Cogn Ther Res (2007) 31:483–502
497
applied equally irrespective of initial level of depression (Joiner, 2000). This issue was examined by determining whether baseline depression scores (the covariate), interacted with these predictor variables in the prediction of follow-up depression scores. If baseline depression scores interact significantly with other variables in the prediction of future depressive symptoms, then this suggests that the predictors have their influence only at certain levels of initial depression. However, the interaction terms between baseline EPDS scores and the predictor variables were not found to significantly predict T2 depression scores. Thus, it can be concluded that the significant predictive effects have their influence upon changes in depression, irrespective of initial levels of depression.
Discussion This study produced some interesting findings regarding the relationships between personality style, life events and depressive symptoms in pregnancy and the postpartum period. The cross-sectional analyses for data collected during the last trimester of pregnancy indicated that women who reported a strong sociotropic or autonomous personality style, or a higher number of sociotropic or autonomous stressful life events (or sociotropic or autonomous loss for events) in the past 3 months, were more likely to report high levels of depressive symptoms. However, contrary to predictions the results did not support the cognitive diathesis-stress congruency model of depression. Only the non-congruent interaction between autonomous personality style and number of sociotropic life events was significantly associated with depression scores in pregnancy. Although this finding does not support the congruency theory of depression, it does at first sight appear to be consistent with a cognitive diathesis-stress model. However, when the nature of this significant interaction effect was examined, the results showed that it could mainly be explained by women who reported a low autonomous personality style in combination with a low number of sociotropic events, who tended to report very low levels of depressive symptoms in pregnancy. This finding was replicated for the cross-sectional analyses using the postpartum data and may be best interpreted as indicating that the interaction of low autonomous personality and low sociotropic events is associated with better emotional well-being, rather than being consistent with a diathesis-stress model of depression. We can only speculate as to the reasons for this effect. Perhaps women who place less value on independence, perfectionism, control, and who have very low levels of relationship difficulty or loss of social support during pregnancy and following childbirth are particularly happy during this period of their lives? The longitudinal analysis from pregnancy to the postpartum period revealed that, as hypothesized, depression during pregnancy significantly predicted depression in the postpartum period. Contrary to expectations, there was no support in the longitudinal analysis for the cognitive diathesis-stress model of depression during the postpartum period, either congruently or incongruently. The interactions between personality style and life events, of congruent or incongruent combinations, did not significantly predict changes in depression. Rather than supporting a cognitive diathesis-stress model, the results identified a number of direct effects. Although the number of life events in sociotropic or autonomous domains did not significantly predict changes in depressive symptoms over time, irrespective of their personality style, women who reported high levels of sociotropic or autonomous loss associated with events in the transition from
123
498
Cogn Ther Res (2007) 31:483–502
pregnancy to the postpartum period were significantly more likely to experience increases in depressive symptoms. Thus, it is the perception of events as producing either autonomous or sociotropic loss, rather than the occurrence of sociotropic or autonomous events per se, that placed women at risk for increased depressive symptoms. This finding emphasizes the importance of examining loss perceptions in addition to the occurrence of events, and also the need to consider that most events have an impact in terms of loss in both sociotropic and autonomous domains. The most commonly occurring events in the postpartum period were categorized as autonomous stressors, and included major surgery, decreased amounts of time spent on personal activities, and becoming unemployed. In the sociotropic domain, having serious arguments with one’s spouse and increase conflict with in-laws or relatives were the most frequently endorsed events. Although the level of loss associated with events was higher in the domain to which events were pre-categorized, it was noted these events were associated with perceptions of loss in both domains. For example, the experience of becoming unemployed or having less time for personal activities is likely to lead to reduced social support and loss of relationships, in addition to loss of autonomy and independence. Similarly, conflict with in-laws and relatives may result in negative impact upon autonomy and independence (e.g. loss of support in child care), in addition to the negative impact upon relationships and social support. Thus, it is important to examine perceptions of loss in sociotropic and autonomous domains for all events, not just the occurrence of events in pre-categorized domains, when examining the impact of life events upon emotional adjustment. In terms of personality, consistent with predictions, women with a strong pre-existing sociotropic style were at risk for increased postnatal depressive symptoms irrespective of the life events that they experienced. It appears that women with a strong sociotropic style, who place strong emphasis on personal relationships and social support, may be vulnerable to depression in the transition from pregnancy to childbirth. The dimension of sociotropic personality includes elements such as the effort to please others, strong concern about what others think and dislike of being alone. These women may feel that they are unable to satisfy the needs of the varying individuals in their lives (their partner, friends, other family members and the new baby), worry about meeting the expectations of others, and find it hard to be alone with a new baby for extended periods of time, as is frequently the case for many new mothers. Contrary to expectations, however, autonomous personality style did not predict changes in depression across the transition period, after controlling for initial levels of depressive symptoms. The differences between the cross-sectional and longitudinal results relating to autonomous personality style warrant discussion. Whereas sociotropic personality was associated with concurrent and future levels of depressive symptoms, autonomous personality style was only associated with depressive symptoms crosssectionally and not longitudinally. It seems possible that autonomous personality may be associated more with immediate mood states, whereas sociotropic personality may have a more enduring impact upon depressive symptoms. It has also been suggested that autonomous personality style is a complex construct containing some elements that may play an event-buffering role, or may reduce rather than increase the risk for development of depression (Robins & Block, 1988). It is possible that some elements of autonomous personality, such as the ability to cope independently in the absence of social support, may provide a protective function during the transition from pregnancy to postpartum, whereas other aspects, such as the need for control and time for oneself, may increase the risk of depression during this period. Future studies should attempt to
123
Cogn Ther Res (2007) 31:483–502
499
unravel the complexities regarding to construct of autonomous personality in order to clarify those aspects are associated with mental health outcomes (Sato & McCann, 2000; Sato, McCann, & Ferguson-Isaac, 2004). The present study also examined the proposition of Beck’s cognitive model of depression that the cognitive diathesis-stress interaction impacts upon depressive symptoms through the activation of negative automatic thoughts. Given that the interactions between personality style and life events did not significantly predict changes in depression longitudinally, this proposition was not supported in terms of the development of depression. Cross-sectionally, however, negative automatic thoughts mediated the relationship between depression and the autonomous personality and sociotropic events interaction in both pregnancy and the post-partum period. Negative automatic thoughts also mediated the direct effects upon depression of sociotropic and autonomous personality and life events cross-sectionally, and the longitudinal impact of sociotropic personality, sociotropic loss and autonomous loss upon changes in depression from pregnancy to the postpartum period. Strengths and limitations The study has several strengths that are worthy of mention. The study is novel in that it examines the validity of the cognitive diathesis-stress congruency model in a sample and context in which it has not previously been explored. Its longitudinal design is also an important feature. Nevertheless, there are also several methodological issues that limit the conclusions that can be drawn. The sample size was relatively small leading to issues of power, although this problem is evident in the majority of studies in this area. The measurement of life events also presented a challenge, in terms of the most appropriate time at which to assess the occurrence of events, and the potential bias associated with self-report. The perception of some events, such as relationship conflict or illness, is open to subjective interpretation and bias may occur in the reporting of life events and depressive symptoms when these variables are measured simultaneously. This problem, however, is manifest throughout the literature in this area, as is the difficulty in categorizing events into sociotropic or autonomous domains (Kwon & Whisman, 1998). Although the results supported the categorization of events as stronger in either autonomous or sociotropic domains consistent with the original categorizations by Clarke et al. (1992), the majority of events were associated with loss in both domains. It can be argued that individuals vary in their perceptions of the sociotropic or autonomous impact of events and that it is these perceptions rather than the events per se that determine their influence upon depression. The present study took this proposition into account by assessing the loss perceptions in both domains for each event, in addition to the number of events experienced. The timing of assessments is also an issue in research of this type. In order to demonstrate directions of causality one would ideally have regular assessments of all predictor, mediating and outcome variables, to permit time series analyses that identify causal patterns and reciprocal relationships over varying time periods. The two time points used in the present study present only a simplistic picture of the way in which personality, life events and depressive symptoms influence each other. Similarly, the present study did not provide information about prior experience of depression, or the clinical nature of depressive disorder. It should also be noted that the assessment of women at 8-weeks postpartum would not have captured those postpartum depression symptoms that occurred 3 to 4 months after childbirth.
123
500
Cogn Ther Res (2007) 31:483–502
Summary In summary, the cognitive diathesis-stress congruency model does not appear to have been examined previously in the transitional period between pregnancy and the postpartum period. The findings of the present study thus provide valuable insights into factors that could be identified during late pregnancy as risk factors for depression in the postpartum period. Women with a strong sociotropic personality style, or who experience a high level of sociotropic or autonomous loss relating to life events from late pregnancy to the postpartum period, were at particular risk for increases in depressive symptoms. Knowledge of these risk factors could enable women at risk to be monitored carefully in the postpartum period and may guide development of brief interventions for preventive purposes.
References Abramson, L. Y., Alloy, L. B., & Metalsky, G. I. (1988). The cognitive diathesis-stress theories of depression: Toward an adequate evaluation of the theories’ validities. In L. B. Alloy (Eds.), Cognitive processes in depression. New York: Guilford Press. Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions. Newbury Park: Sage. Allen, N. B., de L Horne, D. J., & Trinder, J. (1996). Sociotropy, autonomy, and dysphoric emotional responses to specific classes of stress: A psychophysiological evaluation. Journal of Abnormal Psychology, 105(1), 25–33. Alloy, L. B., Hartlage, S., & Abramson, L. Y. (1988). Testing the cognitive diathesis-stress theories of depression: Issues of research design, conceptualization, and assessment. In L. B. Alloy (Eds.), Cognitive processes in depression. New York: Guilford Press. Australian Bureau of Statistics. (1997). ASCO Australian standard classification of occupations. Canberra: Australian Bureau of Statistics. Baker, K. D., Nenneyer, R. A., & Barris, B. P. (1997). Cognitive organization in sociotropic and autonomous inpatient depressives. Journal of Cognitive Psychotherapy, 11(4), 279–297. Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable distinction in social and psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51(6), 1173–1182. Bartelstone, J. H., & Trull, T. J. (1995). Personality, life events, and depression. Journal of Personality Assessment, 64(2), 279–294. Beck, A. T. (1976). Cognitive therapy and the emotional disorders. Oxford, England: International Universities Press. Beck, A. T. (1983). Cognitive therapy of depression: New perspectives. In J. E. Barrett (Eds.), Treatment of depression: Old controversies and new approaches (pp. 265–294). New York, US: Raven Press. Beck, A. T. (1987). Cognitive models of depression. Journal of Cognitive Psychotherapy, 1(1), 5–37. Beck, A. T., & Steer, R. A. (1987). Manual of the revised Beck Depression Inventory. San Antonio: Psychological Corporation. Beck, C. T., & Gable, R. K. (2001). Comparative analysis of the performance of the Postpartum Depression Scale with two other depression instruments. Nursing Research, 50(4), 242–250. Bloch, M., Daly, R. C., & Rubinow, D. R. (2003). Endocrine factors in the etiology of postpartum depression. Comprehensive Psychiatry, 44(3), 234–246. Brown, G. P., Hammen, C. L., Craske, M. G., & Wickens, T. D. (1995). Dimensions of dysfunctional attitudes as vulnerabilities to depressive symptoms. Journal of Abnormal Psychology, 104(3), 431–435. Clark, D. A., Beck, A. T., & Alford, B. A. (Eds.) (1999). Scientific foundations of cognitive theory and therapy of depression. New York: Wiley. Clark, D. A., Beck, A. T., & Brown, G. K. (1992). Sociotropy, autonomy, and life event perceptions in dysphoric and nondysphoric individuals. Cognitive Therapy and Research, 16(6), 635–652. Cox, J. L., Chapman, G., Murray, D., & Jones, P. (1996). Validation of the Edinburgh Postnatal Depression Scale (EPDS) in non-postnatal women. Journal of Affective Disorders, 39(3), 185–189.
123
Cogn Ther Res (2007) 31:483–502
501
Cox, J. L., Holden, J., & Sagovsky, R. (1987). Detection of postnatal depression development of the 10-item Edinburgh Postnatal Depression Scale. British Journal of Psychiatry, 150, 782–786. Coyne, J. C., Thompson, R., & Whiffen, V. (2004). Is the promissory note of personality as vulnerability to depression in default? Reply to Zuroff, Mongrain, and Santor (2004). Psychological Bulletin, 130(3), 512–517. Fresco, D. M., Sampson, W. S., Craighead, L. W., & Koons, A. N. (2001). The relationship of sociotropy and autonomy to symptoms of depression and anxiety. Journal of Cognitive Psychotherapy, 15(1), 17–31. Gotlib, I. H., Whiffen, V. E., Wallace, P. M., & Mount, J. H. (1991). Prospective investigation of postpartum depression: Factors involved in onset and recovery. Journal of Abnormal Psychology, 100(2), 122–132. Grazioli, R., & Terry, D. J. (2000). The role of cognitive vulnerability and stress in the prediction of postpartum depressive symptomatology. British Journal of Clinical Psychology, 39(4), 329–347. Green, J. M., & Murray, D. (1994). The use of the Edinburgh postnatal depression scale in research to explore the relationship between antenatal, postnatal dysphoria. In: J. Holden (Ed.), Pernatal psychiatry – Use and misuse of the Edinburgh Postnatal Depression Scales (pp. 180–198). London: Gaskell. Hammen, C., Ellicott, A., & Gitlin, M. (1989a). Vulnerability to specific life events and prediction of course of disorder in unipolar depressed patients. Canadian Journal of Behavioural Science, 21(4), 377–388. Hammen, C., Ellicott, A., Gitlin, M., & Jamison, K. R. (1989b). Sociotropy/autonomy and vulnerability to specific life events in patients with unipolar depression and bipolar disorders. Journal of Abnormal Psychology, 98(2), 154–160. Hammen, C., Ellicott, A., & Gitlin, M. J. (1992). Stressors and sociotropy/autonomy: A longitudinal study of their relationship to the course of bipolar disorder. Cognitive Therapy and Research, 16(4), 409–418. Hammen, C., & Goodman Brown, T. (1990). Self-schemas and vulnerability to specific life stress in children at risk for depression. Cognitive Therapy and Research, 14(2), 215–227. Hammen, C., Marks, T., Mayol, A., & DeMayo, R. (1985). Depressive self-schemas, life stress, and vulnerability to depression. Journal of Abnormal Psychology, 94(3), 308–319. Harrell, T. H., & Ryon, N. B. (1983). Cognitive-behavioral assessment of depression: Clinical validation of the Automatic Thoughts Questionnaire. Journal of Consulting and Clinical Psychology, 51(5), 721–725. Harris, B. (2002). Postpartum depression. Psychiatric Annals, 32(7), 405–415. Harris, B., Huckle, P., Thomas, R., Johns, S., et al. (1989). The use of rating scales to identify post-natal depression. British Journal of Psychiatry, 154, 813–817. Hollon, S. D., & Kendall, P. C. (1980). Cognitive self-statements in depression: Development of an automatic thoughts questionnaire. Cognitive Therapy and Research, 4(4), 383–395. Joiner, T. E. Jr. (2000). A test of the hopelessness theory of depression in youth psychiatric inpatients. Journal of Clinical Child Psychology, 29(2), 167–176. Kwon, P., & Whisman, M. A. (1998). Sociotropy and autonomy as vulnerabilities to specific life events: Issues in life event categorization. Cognitive Therapy and Research, 22(4), 353–362. Kwon, S. M., & Oei, T. P. (1992). Differential causal roles of dysfunctional attitudes and automatic thoughts in depression. Cognitive Therapy and Research, 16(3), 309–328. Kwon, S. M., & Oei, T. P. S. (2003). Cognitive change processes in a group cognitive behavior therapy of depression. Journal of Behavior Therapy & Experimental Psychiatry, 34(1), 73–85. Lane, A., Keville, R., Morris, M., Kinsella, A., Turner, M., & Barry, S. (1997). Postnatal depression and elation among mothers and their partners: Prevalence and predictors. British Journal of Psychiatry, 171, 550–555. Marks, M. N., Wieck, A., Checkley, S. A., & Kumar, R. (1992). Contribution of psychological and social factors to psychotic and non-psychotic relapse after childbirth in women with previous histories of affective disorder. Journal of Affective Disorders, 24(4), 253–263. Matthey, S., Barnett, B., Ungerer, J., & Waters, B. (2000). Paternal and maternal depressed mood during the transition to parenthood. Journal of Affective Disorders, 60(2), 75–85. Mazure, C. M., Bruce, M. L., Maciejewski, P. K., & Jacobs, S. C. (2000). Adverse life events and cognitive-personality characteristics in the prediction of major depression and antidepressant response. American Journal of Psychiatry, 157(6), 896–903. Mazure, C. M., & Maciejewski, P. K. (2003). The interplay of stress, gender and cognitive style in depressive onset. Archives of Women’s Mental Health, 6(1), 5–8.
123
502
Cogn Ther Res (2007) 31:483–502
Mazure, C. M., Raghavan, C., Maciejewski, P. K., Jacobs, S. C., & Bruce, M. L. (2001). Cognitivepersonality characteristics as direct predictors of unipolar major depression. Cognitive Therapy and Research, 25(2), 215–225. Miller, L. J., & Torpy, J. M. (2002). Postpartum depression. JAMA: Journal of the American Medical Association, 287(6), 762–765. Morse, J. Q., & Robins, C. J. (2005). Personality – life event congruence effects in late-life depression. Journal of Affective Disorders, 84(1), 25–31. Murray, J. (1992). Prevention and the identification of high risk groups. International Review of Psychiatry, 4(3–4), 281–286. Murray, L., & Carothers, A. D. (1990). The validation of the Edinburgh Post-natal Depression Scale on a community sample. British Journal of Psychiatry, 157, 288–290. O’Hara, M. W. (1995). Postpartum depression: Causes and consequences. New York: Spinger-Verlag. O’Hara, M. W., Neunaber, D. J., & Zekoski, E. M. (1984). Prospective study of postpartum depression: Prevalence, course, and predictive factors. Journal of Abnormal Psychology, 93(2), 158–171. O’Hara, M. W., Rehm, L. P., & Campbell, S. B. (1982). Predicting depressive symptomatology: Cognitive-behavioral models and postpartum depression. Journal of Abnormal Psychology, 91(6), 457–461. O’Hara, M. W., Schlechte, J. A., Lewis, D. A., & Varner, M. W. (1991). Controlled prospective study of postpartum mood disorders: Psychological, environmental, and hormonal variables. Journal of Abnormal Psychology, 100(1), 63–73. Oates-Johnson, T., & Clark, D. A. (2004). Sociotropy, body dissatisfaction and perceived social disapproval in dieting women: A prospective diathesis-stress study of dysphoria. Cognitive Therapy and Research, 28(6), 715–731. Oei, T. P. S., Goh, Y. W., & Kwon, S. M. (1996). A cross-cultural study of the integrated cognitive model of depression. Cahiers de Psychologie Cognitive/Current Psychology of Cognition, 15(3), 265–281. Preacher, K. J., & Leonardelli, G. L. (2001). Calculation for the Sobel Test: An interactive calculation tool for mediation tests. Retrieved 2002, from http://www.quantum2.psy.ohio-state.edu/kris/sobel/sobel.htm. Raghavan, C., Le, H. N., & Berenbaum, H. (2002). Predicting dysphoria and hostility using the diathesisstress model of sociotropy and autonomy in a contextualized stress setting. Cognitive Therapy and Research, 26(2), 231–244. Robins, C. J. (1990). Congruence of personality and life events in depression. Journal of Abnormal Psychology, 99(4), 393–397. Robins, C. J., & Block, P. (1988). Personal vulnerability, life events, and depressive symptoms: A test of a specific interactional model. Journal of Personality and Social Psychology, 54(5), 847–852. Robins, C. J., Hayes, A. M., Block, P., Kramer, R. J. et al. (1995). Interpersonal and achievement concerns and the depressive vulnerability and symptom specificity hypotheses: A prospective study. Cognitive Therapy and Research, 19(1), 1–20. Robins, C. J., Ladd, J., Welkowitz, J., Blaney, P. H. et al. (1994). The personal style inventory: Preliminary validation studies of new measures of sociotropy and autonomy. Journal of Psychopathology and Behavioral Assessment, 16(4), 277–300. Sato, T., & McCann, D. (2000). Sociotropy-autonomy and the Beck Depression Inventory. European Journal of Psychological Assessment, 16(1), 66–76. Sato, T., McCann, D., & Ferguson-Isaac, C. (2004). Sociotropy-autonomy and situation-specific anxiety. Psychological Reports, 94(1), 67–76. Schaper, A. M., Rooney, B. L., Kay, N. R., & Silva, P. D. (1994). Use of the Edinburgh Postnatal Depression Scale to identify postpartum depression in a clinical setting. Journal of Reproductive Medicine, 39, 620–624. Segal, Z. V., Shaw, B. F., & Vella, D. D. (1989). Life stress and depression: A test of the congruency hypothesis for life event content and depressive subtype. Canadian Journal of Behavioural Science, 21(4), 389–400. Segal, Z. V., & Dobson, K. S. (1992). Cognitive models of depression: Report from a consensus development conference. Psychological Inquiry, 3(3), 219–224. Segal, Z. V., Shaw, B. F., Vella, D. D., & Katz, R. (1992). Cognitive and life stress predictors of relapse in remitted unipolar depressed patients: Test of the congruency hypothesis. Journal of Abnormal Psychology, 101(1), 26–36.
123