Vous êtes sur la page 1sur 11

Psychol Stud

DOI 10.1007/s12646-010-0034-1

ORIGINAL ARTICLE

Emotional Disclosure in Day-to-Day Living and Subjective


Well Being
Priya Saxena & Seema Mehrotra

Received: 7 April 2010 / Accepted: 7 May 2010


# National Academy of Psychology (NAOP) India 2010

Abstract There is a large body of empirical research on and Susman 1988) and better immune function immediately
emotional disclosure and its relationship to health and well after disclosure (Pennebaker et al. 1988).Verbal disclosure
being. However, emotional disclosure has been examined has been examined in experimental and quasi-experimental
largely in the context of illness or trauma. Beneficial studies that have reported outcomes such as improve-
outcomes of emotional disclosure in day-to-day living have ments in mood (Lutgendorf and Antoni 1999; Mendolia
not been consistently observed in non-experimental studies, and Kleck 1993), enhanced immune system functioning
although a large proportion of individuals do report (Esterling et al. 1994), better role and physical functioning
perceived benefits. The present study aimed at exploring (Kelley et al. 1997), and enhanced well-being (Pantchenko
the association of emotional disclosure with selected et al. 2003; Lyubomirsky et al. 2006).
intrapersonal and interpersonal variables and their role in
prediction of subjective well being in day-to-day living. The
sample comprised of 209 adults who did not report presence Trends in Emotional Disclosure Research
of any major stressor in the recent past. In hierarchical
regression analysis, lower affect intensity, higher emotional A scan of the available literature indicates that emotional
clarity, lower trait rumination, higher perceived support and disclosure has been primarily explored in the context of
higher emotional disclosure predicted higher subjective well written-disclosure-based interventions. On the other hand,
being. In addition, trait rumination emerged as a significant research on verbal disclosure has been mainly in the form
moderator of the relationship between emotional disclosure of experimental/quasi-experimental studies that tend to be
and subjective well being. restricted to college samples. There are only a few studies
that focus on social sharing of emotions in naturalistic
Keywords Emotional disclosure . Need for emotional contexts, especially in non-college adult samples. Also,
disclosure . Subjective wellbeing . Affect intensity . several of these naturalistic studies have been carried out on
Trait-rumination . Emotional clarity . Perceived support samples of participants experiencing major life events like
bereavement, diagnosis of cancer and HIV (e.g. Zakowski
et al. 2001; Lieberman and Goldstein 2006).
Emotional disclosure refers to the process of verbally In recent times, emotional disclosure research has
communicating about one’s emotional experiences to witnessed a trend parallel to the shift in stress and coping
another person (Pennebaker 1995; Zech 2000). Various research (Affleck et al. 1999). This shift pertains to
beneficial effects of written disclosure have been observed examining mood, affect and coping in the context of day-
such as better health in retrospective reports (e.g. Pennebaker to-day living. In a review of eight studies conducted on
social sharing of everyday experiences, Rime et al. (1998)
observed that emotional experiences were shared in 88–
P. Saxena : S. Mehrotra (*)
96% of the cases. These proportions were independent of
Department of Mental Health & Social Psychology, NIMHANS,
Bangalore 560029, India age, culture, gender and type of emotion. Since emotional
e-mail: drmehrotra_seema@yahoo.com experiences are shared with such high frequencies, it is
Psychol Stud

expected that emotional expression would be leading to In the above backdrop, the present study focused on the
beneficial outcomes. Surprisingly, this line of research following objectives: 1. Exploring the patterns of typical
shows that although talking about emotional experiences need for emotional disclosure and extent of emotional
is generally reported to be helpful by people; contrasting or disclosure in day-to-day-living. 2. Examining the associa-
inconsistent results are observed vis-à-vis actual benefits of tion of need for disclosure and emotional disclosure with
emotional disclosure in day-to-day life, across studies (Zech selected intrapersonal and interpersonal variables. 3. Ex-
2000; Rime et al. 1998; Zech and Rime 2005). This amining the role of emotional disclosure along with the
necessitates a detailed examination of emotional disclosure intra personal and interpersonal variables in prediction of
in day-to-day living and its correlates. well being in day-to-day living.
Negative affect intensity, emotional clarity, motivations
underlying withholding expression of negative affect, emo-
Variables Examined in Emotional Disclosure Literature tional control and trait rumination were examined as intraper-
sonal variables whereas perceived social support and social
There has been a growing recognition of individual differ- constraints were the two interpersonal variables considered.
ences that predict the extent of emotional disclosure or The study focused on disclosure of distressing emotions, as is
moderate the relationship between emotional disclosure and the case in the bulk of research emotional disclosure–health.
health outcomes. A few variables that have been examined
as predictors or moderators of disclosure include alexithy-
mia, ambivalence over emotional expression, neuroticism, Method
and emotional processing (e.g. Dickinson 2005; Norman et
al. 2004; Paez et al. 1999). There are several other Pilot Phase
intrapersonal and interpersonal variables in the broader
field of emotional regulation that have relevance for Selection of variables The researchers had taken into
emotional disclosure research. For example, the model of account the gaps in the available literature on disclosure
emotional awareness, understanding and expression, pro- to decide on the variables for the study. But in addition, it
posed by Kennedy-Moore and Watson (1999) points to the was deemed useful to explore community participants’
potential role of variables such as typical intensity of affect conception about important factors associated with emo-
experienced, belief system about disclosure (reasons/ tional disclosure in an urban Indian context. Hence in the
motivations for non-disclosure), emotional clarity, rumina- pilot phase, a focused group discussion was conducted with
tion, perception of social support and social constraints. twelve volunteer adults. In the focus group discussion, the
These variables have been infrequently examined for their participants came up with their views about why they found
association with emotional disclosure in day- to-day living. disclosure helpful, reasons for not being able to share,
Need for emotional disclosure has also been relatively factors that helped them to share and other modes of
ignored as an individual difference variable. Published handling negative emotions. The themes highlighted the
Indian literature on emotional disclosure and its correlates relevance of variables such as social constraints and various
in day-to-day living is sparse. Raval et al. 2007 examined beliefs and motives underlying non-disclosure.
beliefs about acceptability of expression of emotions and
actual expression in young Gujarati children from urban Selection of measures The measures were selected using
and semi- urban schools. It was observed that children the following criteria: brevity, cultural appropriateness of
learnt to control expression of their emotions such as items, documented use across varied samples and nations,
anger and sadness so as to avoid being reprimanded or ease of comprehension as well as availability of Indian data
ridiculed. The authors surmised that Indian children learn on their basic psychometric properties to the extent
from young age to perceive and attach importance to possible. In addition, the pilot phase utilized a series of
prevalent social norms regarding regulation of emotions. steps including the focus group discussion (outlined above),
A study by Bhati and Gunthey in 1999 examined family content validation by three experts and a cognitive
environment and mental health of working women and interviewing procedure to explore the appropriateness of
homemakers. The authors observed that working women those measures for which Indian data on psychometric
were higher on expressiveness and independence but properties were not available (emotional disclosure, rumi-
experienced poorer mental health. There is a need for a nation, social constraints and reasons for withholding
systematic and comprehensive exploration of relevant negative emotions). A few modifications in the wordings/
intrapersonal as well as interpersonal variables together instructions were made based on the results of the above
with emotional disclosure as predictors of subjective steps (Table 1). Subsequently, the convergent validities of
wellbeing in day-to-day living. these measures were re-established in the Indian context by
Psychol Stud

Table 1 Description of intrapersonal and interpersonal variables

I. Intrapersonal variables Operational definition Measure


1. Negative affect Intensity Typical intensity with which negative emotions Affect Intensity Measure—Negative intensity scale
are experienced Bryant et al. 1996
Six items with six-point response format
2. Emotional Clarity Perceived ability to identify and describe Emotional clarity scale of trait meta-mood measure
specific emotions one experiences Salovey et al. 1995
11 items with five-point response format
3. Self regulation of withholding Motivations for withholding expression of Self regulation of withholding negative emotions
negative emotions negative affect: Controlled regulation questionnaire
(CR)—Not expressing due to external Kim et al. 2002
reason (e.g. avoid censure/gain, approval)
Present study response format changed from
Autonomous-regulation(AR):Non-expression
seven-point to five-point and instructions simplified
in consonance with personal values and choice
based on pilot phase data. Six items labeled
integrated regulation had to be omitted due to
low internal consistency
No of items
Controlled regulation: 15
Autonomous regulation: 7
4. Emotional Control Chronic tendency to control/suppress Courtald Emotional Control scale
negative emotions Watson and Greer 1983
21 items with four-point response format
5. Trait Rumination Self attentiveness with a style of repeated, Rumination scale of the rumination–reflection
passive dwelling on perceived threats, losses questionnaire
or injustice to self Trapnell and Campbell 1999
12 items with five-point response format
II. Inter-personal variables Operational definition Measure
1. Social support Perceived availability of emotional, Social Support Questionnaire
informational and instrumental support Pillay and Rao 2002
One item dropped from original version: was
difficult to differentiate from another one as
observed in the pilot phase
Used: 11 items with five-point response format
2. Social constraints Perceptions that one must inhibit from disclosing Social constraints scale
stressful thoughts and feelings to a significant Lepore and Ituarte 1999
other due to certain behaviors of the significant
14 items, four-point response format
other, e.g. trivialization
Wordings modified to reflect constraints in the
context of day-to-day stressors rather than any
one particular stressor

examining their correlations with theoretically relevant was currently experiencing or had experienced in the recent
measures that have known psychometric properties in local past (past 3 months) any major stressful life event.
samples. Hindi versions of the measures were also Definition and examples of daily stressors/hassles, moder-
developed in this phase through standard translation and ate stressors and major stressors were provided to the
back translation procedure. The details of the pilot phase participants to help them differentiate the nature of stress
results are not discussed owing to space constraints and are they might be experiencing. Data were included only from
available from the authors. those participants who did not report a major stressor either
currently or in the past 3 months.

Measures Emotional disclosure To assess the general frequency of


disclosure of negative emotions, the researchers used the
Basic datasheet In addition to the socio-demographics, the Emotional disclosure Scale, which was developed based on
data sheet included an item enquiring whether a participant the pattern of Snell’s Emotional Self-Disclosure Scale
Psychol Stud

(Snell et al. 1988). In the present study, participants were consistency in the present sample are shown in Table 1 and
enquired about the typical frequency of disclosure of Table 2.
negative emotions to significant others. There were three
items, one each for sadness, anxiety and anger, each to be Sample
answered on a five-point likert-type rating scale. Higher
total score indicates typically higher extent of emotional The study included males and females between 21 and
disclosure. The internal consistency of the scale was 0.87. 60 years of age with minimum 12 years of formal education
Higher use of coping through support seeking is expected who could speak English or Hindi. To begin with, a pool of
to be associated with higher disclosure. In the pilot phase of 229 participants gave consent and completed all the
the study, scores on the emotional disclosure measure were measures. Twenty participants in this initial pool (9%)
positively correlated(r=0.34) with social support subscale reported facing a major stressor in the past 3 months as
of Coping Checklist (Rao et al. 1989) lending support to the operationalized in the study. Death or poor health of a
convergent validity of this measure. Typical intensity of significant other were the most often cited major stressors
need for emotional disclosure was assessed using an item to in this subgroup. The final sample (N=209) consisted of
be rated on a five point likert type scale ranging from ‘quite only those who did not report major stressors. This criterion
strongly’ to ‘not at all’. The researcher developed and was used at the level of analysis because the purpose of the
appended additional items to assess general aspects of research was to determine the contribution of emotional
emotional disclosure. There were items enquiring whether disclosure and other variables in the context of day-to-day
emotional disclosure was perceived to be generally helpful living rather than in the presence of major stressors. There
or unhelpful. Those who perceived it to be generally helpful was a fairly equal distribution of participants in the age
answered another item on perceived reasons for helpfulness groups of 21–30 years, 31–40 years, 41–50 years and 51–
of emotional disclosure. Similarly, those who perceived 60 years. Forty eight percent of the participants were men
disclosure to be generally unhelpful indicated reasons for and 52% were women. Majority (65%) were married. Fifty
the same. For each of these items, there was a list of reasons two percent were graduates; 40% had qualifications above
provided as a checklist and the participants could endorse graduate levels, while less than one-tenth of the participants
as many reasons as were applicable to them. They also had were educated up to standard 12. Majority of the sample
the option to cite any other unlisted reason (60%) were professionals, a quarter were students; while
homemakers/not-employed formed 15% of the sample.
Wellbeing The data were collected after obtaining written informed
consent. The measures were administered in a single
World Health Organization Five Well-being Index (WHO- session for each participant. The order of presentation of
Five) (Bech 1998) This scale was derived from a larger measures was constant across participants. In 76% of the
rating scale developed for a WHO project and has since cases, one to one administration was carried out whereas
been used across the globe with varied samples. Initially for remaining student sub-sample, the measures were
ten items were selected from a larger pool of 18 items administered in small groups of 15 to 20 participants each,
based on the criteria of homogeneity among samples. under close supervision of the first author.
Subsequently only the positively worded items were
retained. The five positively worded items tap positive
psychological well-being, involving relaxation, vitality and Results and Discussion
general interests. The raw score is calculated by totaling
the responses across five items. The scores range from 0 The descriptive statistics on all the variables in term of
to 25, zero representing ‘worst possible’ and 25 represent- possible and obtained range of scores, means, medians,
ing ‘best possible’ subjective well-being. The internal standard deviations, information on normality of distribu-
consistency and convergent validity of this measure have tions and internal consistency reliabilities are presented in
been demonstrated in other Indian studies (e.g. Prathima Table 2. All the measures had high internal consistencies,
2005). the values ranging between 0.80 and 0.92. Using KS-Z
statistic, a stringent measure of normality, it was observed
that the distributions of scores on emotional disclosure,
Intrapersonal and Interpersonal Variables need for disclosure, affect intensity and subjective well
being were non-normal. The analyses were carried out
The intrapersonal and interpersonal variables that using appropriate non-parametric statistics. Hierarchical
were examined, their capsule operational definition, the regression analysis was carried out after checking for
measures used to assess the same along with their internal normal distribution of errors.
Psychol Stud

Table 2 Descriptive statistics on all the measures

Variables/Scale Min–Max possible Min–Max obtained Mean Median S.D. KS-Za Cronbach’s Alpha
(present study)

Need for emotional disclosure 0–4 0–4 2.70 03 1.02 2.81b –


Emotional disclosure 0–12 0–12 6.96 07 3.35 1.54c 0.88
Intrapersonal variables
Controlled regulation 15–75 15–74 34.2 32 13.0 1.27 0.91
Autonomous regulation 7–35 7–34 22.3 23 5.96 1.23 0.80
Emotional control 21–84 25–84 48.3 48 10.78 0.82 0.92
Emotional clarity 11–55 19–55 42.1 43 8.56 1.13 0.91
Affect intensity 6–36 6–36 20.1 19 5.94 1.47c 0.84
Trait rumination 12–60 19–60 39.3 40 8.98 0.96 0.86
Interpersonal variables
Social support 14–70 29–70 53.8 55 10.10 1.04 0.93
Social constraints 14–56 14–49 28.7 29 8.20 1.25 0.89
Outcome variable
Subjective well-being (SWB) 0–25 3–25 17.50 19 4.64 2.68b 0.92
a
Kolmogorov-Smirnov test for testing departure from normality
b
Significant at 0.001 level
c
Significant at 0.05 level

Patterns of Need for Emotional Disclosure and Extent moderate positive correlation between need for emotional
of Emotional Disclosure and typical extent of emotional disclosure. This suggests
that experiencing a need for disclosure may not translate
More than half of the participants (56%) reported a high into actual disclosure and high disclosure (full/almost full)
need for emotional disclosure, by endorsing their general is not typical for a large proportion of Indian adults. This
need for emotional disclosure to be ‘quite strong’ or pattern also implies that other variables (.e.g. supportive
‘strong’. A moderate need for disclosure was reported by context) may operate to determine the extent to which need
roughly one-third (32%) of the participants. Twelve percent for disclosure may result in actual disclosure.
of the participants reported a low need for disclosure. As far On the basic data sheet- item enquiring perceived
as the general extent of emotional disclosure is concerned; helpfulness of emotional disclosure in general, majority of
nearly half of the participants (46%) reported sharing of the participants (90%) reported that they perceived sharing
negative emotions ‘fully-almost fully’ with a significant negative emotions to be generally helpful, while 10% of the
other. Roughly a quarter (26%) generally shared such participants did not perceive it to be helpful. The reasons
negative emotions to ‘some extent’. On the other hand, one for perceived helpfulness or non-helpfulness were also
fifth of the participants (19%) reported generally sharing elicited as described earlier. The reason that was endorsed
negative emotions only to a ‘little extent’. Nine percent of the by the participants with maximum frequency (30%) was
participants reported that generally they did not share their ‘feeling a sense of relief after sharing’. This was followed
negative emotions at all. There was a moderate positive by the reason that ‘sharing is helpful because one feels
correlation(r=0.43) between need for emotional disclosure understood’ (21%), ‘one gets ideas on problem-solving’
and typical extent of emotional disclosure. On the whole, (18%), ‘gets ideas on handling feelings’ (16%), and ‘one
emotional disclosure in the naturalistic context seemed to feels connected/close to the significant other’(11%). A
be a quite prevalent phenomenon and this is in line with the minority of participants (4%) added their own reasons, like
review of studies on social sharing of everyday experiences communication of distress to make the other person aware,
(Rime et al. 1998). However, the present study highlights seeking justification/validation for what one is feeling. The
individual differences in need for emotional disclosure as ten percent of participants (n=21) who reported that they
well as actual extent of emotional disclosure, with a perceived emotional disclosure to be generally ‘not helpful’
substantial minority reporting minimal disclosure. It is were asked to indicate reasons for the same. ‘Not being able
observed that even though 56% of the participants reported to talk freely’ (17%), ‘being criticized’ (17%) and ‘signif-
a strong need to disclose, yet only 46% were actually able icant other not listening completely’ (17%). were most
to disclose fully. This finding is further reflected in the often cited as reasons for non-usefulness of emotional
Psychol Stud

disclosure. Not getting advice on how to handle feelings habitually suppressing their negative emotions, had lower
(15%) and that by sharing, one does not get help on need for disclosure as well as low disclosure with
problem-solving (12%), were also applicable in this sub- significant others. Despite the overlap between emotional
sample. Apart from these reasons provided in the checklist, disclosure and emotional control, it is noteworthy that the
a few participants spontaneously mentioned additional emotional control measure assessed the frequency with
reasons (22%) such as ‘not being able to trust others’, which a person habitually suppressed one’s negative
‘not feeling the need to share’ or ‘feeling more confused by emotions across contexts whereas the emotional disclosure
sharing’. On the whole, the perceptions of helpfulness of measure directly assessed typical extent of disclosure with a
emotional disclosure in to day-to-day life appear to be ‘significant other’. Barr et al. (2008) reported in their study
linked to multiple reasons though sense of relief and that emotional constraints and emotional expression were
validation emerged as the most prominent reasons. Other two empirically distinct higher order factors despite their
reasons reported in this sample were related to assistance in moderate positive correlation. It would also be relevant here
problem solving or emotional regulation. Perceived non- to examine the reasons for not expressing one’s negative
helpfulness of disclosure was most commonly reported to feelings. Controlled regulation refers to holding beliefs that
be due to social constraints. Other reasons included lack of inhibit expression due to feared negative interpersonal
assistance in problem solving or emotional regulation, consequences of sharing. Autonomous regulation refers to
difficulties in trusting and low need for disclosure. the strength of well internalized motives and beliefs for
non-expression that is perceived as a matter of choice.
Relationship of Need for Disclosure and Extent Controlled regulation was uncorrelated to need for emo-
of Disclosure with Other Variables tional disclosure but was significantly negatively associated
with emotional disclosure. This suggests that individuals,
Socio-Demographic Variables who hold beliefs reflecting negative interpersonal repercus-
sions of disclosure, may be a heterogeneous group in terms
Men and women were found to be significantly different on of their actual need for disclosure. Some of the participants
need for emotional disclosure, with women reporting a with high controlled regulation could be experiencing and
higher need to share their negative emotions. Homemakers reporting high unmet need for disclosure whereas some
reported a significantly higher need for emotional disclo- others in this group might be able to downplay or deny the
sure in comparison to the employed participants. But a existence of such a need. But irrespective of need,
supplementary analysis indicated that this was attributable withholding one’s emotions due to controlled reasons tends
to gender, rather than occupational status, the homemakers- to be associated with lower frequency of actual disclosure.
group consisting of only women. In this supplementary In contrast to controlled regulation, autonomous regulation
analysis, there was no difference between employed women was negatively correlated with low need for emotional
and homemaker on need for disclosure (Mann Whitney U = disclosure as well as low actual disclosure. These findings
574.00, NS). Gender differences on extent of disclosure imply that when one experiences a sense of autonomy with
mirrored the finding on need for disclosure, with women reference to reasons for not sharing, one also experiences a
reporting higher disclosure than men. These results are in reduced need to share and lower disclosure levels are in
line with available literature (e.g. Goldshmidt and Weller consonance with lower experience of need.
2000; Hess et al. 2000) and may be at least partly Higher affect intensity was associated with higher need
attributable to gender role socialization. Need for emotional for emotional disclosure but it was uncorrelated with actual
disclosure and typical extent of emotional disclosure were disclosure. Thus experience of high need for disclosure did
not found to be significantly different in sub-groups of not translate into high disclosure for those who typically
participants varying on age, educational qualification, experienced strong intensity of emotions. Lynch et al.
religion and marital status. (Table 3) Although some studies (2001) reported that higher affect intensity was linked to
have reported differences between older and younger higher suppression of feelings. This could be due to the
participants on use of emotional regulation strategies and perceived negativity from others if an individual were to
social support utilization in general; these reports refer to express his/her intense negative feelings. To what extent the
adults above 60 years of age (e.g. Phillips et al. 2006). nature of relationship between affect intensity and emo-
tional disclosure may be moderated by other variables
Intrapersonal and Interpersonal Variables needs to be examined in further research. Emotional clarity
was found to be significantly negatively correlated with
Need for emotional disclosure and extent of emotional need for emotional disclosure, but had a modest positive
disclosure had low to moderate correlations with the correlation with typical extent of emotional disclosure. This
intrapersonal variables (Table 4). Participants who reported pattern can be understood in the light of the existing
Psychol Stud

Table 3 Examining need for emotional disclosure and extent of emotional disclosure in subgroups of participants varying on socio-demographic
characteristics (N=209)

Socio-demographic Number Need for Emotional Mann–Whitney (U) Extent of Emotional Mann–Whitney (U)
characteristics disclosure Mean (S.D.) disclosure Mean (S.D.)

Gender
Male 101 2.42 (0.93) 3752.00a 6.40 (3.03) 4316.50a
Female 108 2.96 (1.04) 7.47 (3.57)
Marital status
Single 73 2.58 (1.06) 4011.00 6.79 (3.45) 4801.50
Married 136 2.76 (0.99) 7.04 (3.32)
Religion
Hindu 155 2.68 (1.03) 4005.00 6.88 (3.43) 3910.50
Others 54 2.76 (1.00) 7.18 (3.16)
Number Mean (S.D.) Kruskal-Wallis (χ2) Mean (S.D.) Kruskal-Wallis (χ2)
Age (in years)
21–30 60 2.65 (1.07) 1.02 6.90 (3.51) 1.93
31–40 52 2.82 (0.96) 7.46 (3.50)
41–50 47 2.61 (1.09) 6.68 (3.25)
51–60 50 2.72 (0.97) 6.76 (3.14)
Education
Standard XII 17 2.47 (1.37) 0.56 6.41 (4.15) 1.40
Graduation 109 2.70 (1.03) 6.75 (3.46)
Post-graduation 83 2.74 (0.93) 7.34 (3.03)
Occupation
1. Students 53 2.62 (1.07) 6.18b (2 & 3) significantly 7.07 (3.50) 0.53
2. Employed 125 2.63 (1.00) different groups in post 6.91 (3.17)
hoc analysis
3. Homemaker/not 31 3.12 (0.92) 6.93 (3.89)
employed
a
Significant at 0.01 level
b
Significant at 0.05 level

literature which suggests that individuals high in emotional extent of emotional disclosure than with need for emotional
clarity employ more negative emotional regulation strate- disclosure. The degree of correlation between social
gies, especially active strategies (Barrett et al. 2001; Gohm support and constraints (r=0.31) indicate that these two
and Clore 2002) Their lower need to disclose their negative are not completely overlapping constructs and measurement
emotions, could be related to their perceived ability to of one does not make the measurement of the other
manage their emotions by themselves. Though trait rumi- redundant. Several responses to the item on perceived
nation was not consistently linked to higher/lower levels of non-helpfulness of disclosure reflect the presence of social
need for disclosure as evident in absence of a significant constraints such as not being listened to completely,
correlation, it was associated with lower actual extent of receiving a critical/trivializing response etc. Social con-
disclosure. Trait rumination has been found to be signifi- straints as a variable has been relatively ignored in studies
cantly associated with neuroticism dimension of personality on emotional disclosure in day-to-day living, but the
and individuals high on neuroticism show ambivalence in patterns observed on interpersonal variables largely mirror
their ability to deal with emotions. (Luminet et al. 2000). those in other studies (e.g. Figueiredo et al. 2004, Schmidt
The patterns of correlations observed indicate that the and Andrykowski 2004). The reports of social constraints
availability of perceived support tends to go hand in hand were elicited in the context of a significant other with
with increased experience of need for disclosure and whom the respondents generally wished to disclose prob-
increased disclosure; however perceptions of social con- lems and associated emotions. The pattern of utilization of
straints may not dampen the need for disclosure despite other supportive relationships for emotional disclosure in
being associated with decreased disclosure. (Table 4) The the presence of constraints, and its potential buffering
association of interpersonal variables was stronger with impact forms an important area for further enquiry.
Psychol Stud

Table 4 Intrapersonal & inter-


personal correlates of need for Need for emotional disclosure Extent of emotional disclosure
emotional disclosure & typical
extent of emotional disclosure Affect intensity 0.25a 0.05
Clarity −0.15b 0.14b
Emotional control −0.39a −0.65a
Controlled regulation −0.10 −0.29a
Spearman rank order correla- Autonomous regulation −0.22a −0.31a
tions were used (two-tailed Trait rumination 0.08 −0.15b
tests)
a
Perceived social support 0.16b 0.36a
Significant at 0.01 level
b
Social constraints −0.07 −0.32b
Significant at 0.05 level

Hierarchical Regression Analysis: Joint Predictors would vary depending on gender, levels of perceived
of Subjective Well-Being support, social constraints, typical intensity of negative
affect, trait rumination and the need for disclosure itself.
A hierarchical regression analysis was carried out to In the hierarchical regression analysis, the total variance
determine the predictors of subjective well-being. The in subjective wellbeing predicted by all the variables was
predictor variables were entered in predetermined steps 44%. The first step involved introducing the socio-
(Table 5). The variables were selected based on the demographic variables (age, marital status, occupational
significant relationships with psychological outcomes categories, and gender) that contributed to predicting 08%
obtained in bi-variate analyses, in order to keep the ratio of the total variance in subjective well-being. In the second
between sample size and number of variables as optimum. step, adding intrapersonal variables, contributed to predict-
Standardization was done for all the predictor variables to ing additional 23% of the total variance. Entering need for
reduce multi-collinearity (Cohen et al. 2003). Categorical emotional disclosure in the third step did not add
variables were dummy-coded. Socio-demographic variables significantly to the prediction of variance in subjective
were entered in the first step, so as to control for their well-being. Interpersonal variables, i.e. social constraint and
effect. This was followed by intrapersonal variables as perceived social support were entered in the fourth step,
these are intrinsic to individuals, irrespective of contextual and these significantly contributed to predicting 05% of the
factors. Need for emotional disclosure was entered in the total variance. Finally in the fifth step, emotional disclosure
third step to determine whether it predicted psychological was added to the equation and it predicted an additional
outcomes after controlling for the intrapersonal variables. In 03% of the total variance, over and above all the other
the fourth step, interpersonal variables were added. Emo- study variables. Adding the moderator variables in the last
tional disclosure was entered in the fifth step to determine step also contributed to significantly predicting an addi-
whether it predicted psychological outcomes over and above tional variance of 05% in subjective well-being. Need for
all the other study variables. In the final step, potential emotional disclosure, social constraints and trait rumination
moderators of the relationship between emotional disclosure emerged as moderators of the relation between emotional
and psychological outcomes were entered. This was to disclosure and subjective well-being in the last step.
explore if the impact of emotional disclosure on wellbeing Perceived social support showed a trend as a moderator of

Table 5 Hierarchical regression: predictors and moderators of subjective well-being

Steps Predictors entered/added R R square R square Adjusted F change Significant


change R square F change

Step 1 Age, marital status, gender, occupational status 0.33 0.11 0.11 0.08 4.83 0.000
Step 2 Emotional control, controlled regulation, affect 0.59 0.35 0.24 0.31 14.44 0.000
intensity, emotional clarity, trait rumination
Step 3 Need for emotional disclosure 0.59 0.35 0.00 0.31 0.32 0.575
Step 4 Perceived social support, social constraints 0.63 0.40 0.05 0.36 8.85 0.000
Step 5 Emotional disclosure 0.65 0.43 0.03 0.39 8.52 0.004
Step 6 ED × Need for emotional disclosure; ED × Social 0.70 0.49 0.07 0.44 4.05 0.001
constraints; ED × Perceived social support; ED ×
Gender; ED × Trait rumination; ED × Affect intensity

R square=0.49; F=9.01; p=0.000; Adjusted R square=0.44


Psychol Stud

the relation between emotional disclosure and subjective found to benefit with external help (Davis and Nolen-
well-being. However, further post hoc probing for signifi- Hoeksema 2000).In the current study, high ruminators
cance of moderation effects was carried out using the steps could be benefiting more from emotional disclosure
recommended by Aiken and West 1991. This revealed that because sharing might provide an avenue to break the
only trait rumination remained as a significant moderator, repetitive chain of thoughts as well as a more coherent
with emotional disclosure predicting higher well being structure to the thought patterns. More research is needed to
among those scoring higher on trait rumination. unravel the complex relationships between trait rumination,
Final model: Examination of beta coefficients in the final need for disclosure and extent of disclosure in view of the
model, (after entering all the above mentioned variables and possibility that that participants high on trait rumination
moderators) indicated that among the socio-demographic may not find emotional disclosure to be of value to them
variable, the ‘student-status’ predicted lower wellbeing due to the feared/actual negative responses from significant
compared to ‘employed’ or ‘homemaker’ status. This others(e.g. Nolen-Hoeksema and Davis 1999). In the
pattern could be linked to the demands related to the present study too, in a supplementary analysis, trait
development phase in the student participants who were at rumination had a moderate positive correlation with social
the threshold of completing their college-education and constraints(r=0.43) which might explain their typically low
were equipping themselves to work towards their aspired disclosure. Whether such negative expectancies in social
and yet uncertain career tracks. In the final model, lower situation could partly be a function of their style and quality
affect intensity, higher emotional clarity, lower trait rumi- of disclosure needs further exploration. This in turn would
nation higher perceived social support and higher emotional imply that merely facilitating the extent of disclosure might
disclosure significantly predicted higher subjective well- not be beneficial for high trait ruminators, unless the
being. Flett et al. (1996), in their study, had reported that manner and quality of disclosure is also taken into account.
higher affect intensity was strongly associated with greater Higher perceived social support added significantly to the
use of emotion-oriented coping and lower negative mood variance in wellbeing predicted by socio-demographic and
regulation expectancy, both of which mediated the link intrapersonal variables. In the bi-variate analysis, emotional
between affect intensity and depressive symptoms. In the control (compared to emotional disclosure) had a small
present study, high affect intensity co-occurred with high positive correlation with subjective well being; and it was
felt need for disclosure but was uncorrelated with actual emotional disclosure rather than control that emerged as a
disclosure. The observed pattern in the background of the significant predictor in the regression analysis, over and
available literature raises the hypothesis that low well being above other variables.
in high affect intensity participants may be attributable
partly to perceived barriers in disclosure coupled with low Further Observations
efficacy in negative mood regulation. Participants higher on
emotional clarity, found it easier to achieve desired On the whole, the study variables predicted 44% of
cognitive states indicative of high wellbeing. And this is variance in subjective wellbeing Emotional disclosure
in line with the findings in other studies (e.g. Gohm and predicted a small proportion of variance in the well being
Clore 2002). Thomsen (2006) in a review on rumination over and above all the other study variables. The small
reported that most of the studies had found a positive contribution of emotional disclosure observed in the present
association between rumination and distress/negative affect study deserves a closer scrutiny. Most of the studies on
and in the present study lower trait rumination predicted emotional disclosure, as mentioned earlier are on partic-
higher well being. Higher trait rumination generally went ipants suffering from major illnesses or those having
hand in hand with lower disclosure and yet, ironically, experienced a significant trauma. Moreover several of these
being higher on emotional disclosure emerged as potential- have used quasi-experimental designs. Naturalistic studies
ly useful for wellbeing in individuals high on trait on disclosure of negative emotions in community partic-
rumination. This moderation effect observed in the com- ipants have not consistently demonstrated psychological
munity sample is noteworthy as the previous literature has benefits, even though participants tend to report perceived
documented the relation between trait rumination and benefits (Zech and Rime 2005; Rime et al. 1998). In the
emotional disclosure mostly in the context of major present study too, it was interesting to contrast the
stressors. Nolen-Hoeksema and Davis (1999), in their study observation that an overwhelming majority of participants
on participants facing the loss of a significant other, found reported emotional disclosure to be useful, with the finding
that ruminators were more distressed in the absence of of relatively smaller contribution of emotional disclosure to
support than non-ruminators. Ruminators have also been well being scores. Sense of emotional relief and feeling
shown to be high on cognitive inflexibility, finding it understood were most often cited as reasons for global
difficult to shift/change patterns of thinking and have been perception of its utility. It appears that the immediate sense
Psychol Stud

of relief and validation accompanying disclosure may be disclosure along with other related intrapersonal and
powerful reasons that lead individuals to report that interpersonal variables on subjective wellbeing in day-to-
disclosure is useful. However this may not necessarily day-life, using a rigorous analytical strategy. Participants
translate into correspondingly higher levels of overall who could be possibly facing moderate levels of chronic
subjective well being. Also, it is speculated that the stressors and daily stressors were not excluded to maximize
quality of disclosure would be an important determinant the chance that the sample is a close approximation of the
of its impact whereas the present paper is restricted to larger population wherein moderate levels of chronic
examining an index of the typical extent to emotional stressors and daily hassles are common. It selectively
disclosure. Social sharing of emotions in daily living may examined those intrapersonal and interpersonal correlates
also be linked to other beneficial outcomes (Rime et al. of emotional disclosure that have hitherto been less-
1998) such as interpersonal bonding, social integration researched. Need for emotional disclosure is a variable
and health which were not the focus of attention in the that is often conspicuous by its absence in research studies
present study. and the role of social constraints has also been relatively
Frattaroli (2006), in his meta-analytic review observed ignored in the emotional disclosure research in community
that studies on emotional disclosure and its beneficial samples. The study attempted to address this gap while
outcomes carried out on participants having illness/ also examining the role of selected moderators in
trauma had higher effect sizes than those involving disclosure–wellbeing relationship. The findings raise
normal participants, the latter already scoring somewhat important hypotheses for further research besides provid-
lower on distress. The current study not just drew ing pointers for developing and testing preventive and
participants from general community, but also included promotive intervention modules that could incorporate a
data only from those participants who were not experi- focus on intrapersonal and interpersonal processes related
encing any major stressor. This strategy was in keeping to emotional disclosure.
with the study objective and may partly explain the
modest association of emotional disclosure with subjec-
tive wellbeing. Although cross cultural studies on social
sharing of emotions point to its wide prevalence, it has References
also been observed that in Asian cultures, more number
of emotion episodes may go unshared, there may be less Affleck, G., Zautra, A., Tennen, H., & Armeli, S. (1999). Multilevel
repetitive sharing and longer delays between the event daily process designs for consulting and clinical psychology: a
and initiation of sharing (Rime et al. 1998).The extent to preface for the perplexed. Journal of Consulting and Clinical
Psychology, 67(5), 746–754.
which these dimensions of sharing may impact on affect
Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and
and well being in the Indian culture requires in-depth interpreting interactions. Newbury Park: Sage.
exploration perhaps using an experience sampling Barr, L. H., Kahn, J. H., & Schneider, W. J. (2008). Individual
methodology. differences in emotional expression: hierarchical structure and
relation with psychological distress. Journal of Social and
Clinical Psychology, 27(10), 1045–1077.
Limitations and Strengths of the Study Barrett, L. F., Gross, J., Conner, T., & Benvenuto, M. (2001).
Emotion differentiation and regulation. Cognition and Emotion,
A few limitations of the study include the use of a cross 15, 713–724.
Bech, P. (1998). Quality of life in the psychiatric patient. London:
sectional, design, lack of representation of the elderly
Mosby, Wolfe.
group, relatively poorer representation of individuals below Bhati, H., & Gunthey, R. (1999). Working women: family environ-
graduation level and insufficient sample-size needed for ment and mental health. Indian Journal of Clinical Psychology,
detailed analysis of moderation effects through structural 26(2), 246–249.
Bryant, F. B., Yarnold, P. R., & Grimm, L. G. (1996). Towards a
equation modeling.
measurement model of the affect intensity measure: a three factor
The limitations notwithstanding, the study is an impor- research. Journal of Research in Personality, 30(2), 223–247.
tant addition to the field in more ways than one. It is one of Cohen, J., Cohen, P. C., West, S. G., & Aiken, L. S. (2003). Applied
the few studies on emotional disclosure that sampled multiple regression/correlation analysis for the behavioral
sciences (3rd ed.). New Jersey: Erlbaum.
participants representing a broad age range of 21-to 60 years
Davis, R. N., & Nolen-Hoeksema, S. (2000). Cognitive inflexibility
from the community. Knowledge about the impact of among ruminators and non-ruminators. Cognitive Therapy and
emotional disclosure on psychological outcomes is well research, 24(6), 699–711.
established in the context of major life stressors but the Dickinson PE (2005) Written emotional expression in spousal care-
givers of cancer patients. Unpublished PhD thesis, California
impact of emotional disclosure while negotiating day-to-
School of Professional Psychology, California
day stressors is less understood. The present study Esterling, B. A., Antoni, M. H., Fletcher, M. A., Margulies, S., &
demonstrates the role of typical extent of emotional Schneiderman, N. (1994). Emotional disclosure through writing
Psychol Stud

or speaking modulates latent Epstein-Barr virus antibody titers. disclosure and psychological health. Journal of Personality and
Journal of Consulting and Clinical Psychology, 62(1), 130–140. Social Psychology, 77(3), 630–641.
Figueiredo, M. I., Fries, E., & Ingram, K. M. (2004). The role of Pantchenko, T., Lawson, M., & Joyce, M. R. (2003). Verbal and non-
disclosure patterns and unsupportive social interactions in the verbal disclosure of recalled negative experiences: relation to
well being of breast cancer patients. Psycho-Oncology, 13(2), well being. Psychology and Psychotherapy: Theory Research
96–105. and Practice, 76(3), 251–265.
Flett, G. L., Blankstein, K. R., & Obertynski, M. (1996). Affect Pennebaker, J. W. (1995). Emotion, disclosure and health. Washington
intensity, coping styles, mood regulation expectancies and DC: American Psychological Association.
depressive symptoms. Personality and Individual Differences, Pennebaker, J. W., & Susman, J. R. (1988). Disclosure of traumas and
20(2), 221–228. psychosomatic processes. Social Science and Medicine, 26(3),
Frattaroli, J. (2006). Experimental disclosure and its moderators: a 327–332.
meta-analysis. Psychological Bulletin, 132(6), 832–865. Pennebaker, J. W., Keicolt-Glaser, J. K., & Glaser, R. (1988).
Gohm, C. L., & Clore, G. L. (2002). Four latent traits of emotional Disclosure of traumas and immune function: health implications
experience and their involvement in well being, coping and for psychotherapy. Journal of Consulting and Clinical Psychol-
attributional style. Cognition and Emotion, 16(4), 495–518. ogy, 56(2), 239–245.
Goldshmidt, O. L., & Weller, L. (2000). “Talking emotions”: gender Phillips, L. H., Henry, J. D., Hoise, J. A., & Milne, A. B. (2006). Age,
differences in a variety of conversational contexts. Symbolic anger regulation and well being. Aging and Mental Health, 10(3),
Interactions, 23(2), 117–134. 250–256.
Hess, U., Senecal, S., Kirouac, G., Herrera, P., Philipot, P., & Kleck, R. Pillay, U., & Rao, K. (2002). The structure and function of social
E. (2000). Emotional expressivity in men and women: stereotypes support in relation to help-seeking behavior. Family Therapy, 29
and self perceptions. Cognitions and Emotion, 14(5), 609–642. (3), 153–167.
Kelley, J. E., Lumley, M. A., & Leisen, J. C. (1997). Health effects of Prathima, S. (2005). Care-giving process in cancer: a strength
emotional disclosure in rheumatoid–arthritis patients. Health perspective. In A. Shah & K. Rao (Eds.), Psychological research
Psychology, 16(4), 331–340. in mental health and neurosciences (1957–2007) (pp. 54–55).
Kennedy-Moore, E., & Watson, J. C. (1999). Expressing emotion: Bangalore: NIMHANS.
Myths, realities and therapeutic strategies. New York: Guilford Rao, K., Subbakrishna, D. K., & Prabhu, G. G. (1989). Development
Press. of a coping checklist—a preliminary-report. Indian Journal of
Kim, Y., Deci, E. L., & Zuckerman, M. (2002). The development of the Psychiatry, 31(2), 128–133.
self regulation of withholding negative emotions questionnaire. Raval, V. V., Martini, T. S., & Raval, P. H. (2007). “Would others
Educational and Psychological Measurement, 62(2), 316–336. think it is ok to express my feelings?’ Regulation of anger,
Lepore, S. J., & Ituarte, P. H. G. (1999). Optimism about cancer sadness and physical pain in Gujarati children in India. Social
enhances mood by reducing negative social interactions. Cancer Development, 16(1), 79–105.
Research, Therapy and Control, 8, 165–174. Rime, B., Finkenauer, C., Luminet, O., Zech, E., & Phillipot, P.
Lieberman, M. A., & Goldstein, B. A. (2006). Not all negative (1998). Social sharing of emotion: new evidence and new
emotions are equal: the role of emotional expression in an online questions. European Review of Social Psychology, 9, 145–189.
support group for women with breast cancer. Psycho-Oncology, Salovey, P., Mayer, J., Goldman, S. L., Turvey, C., & Palfai, T. P.
15(2), 160–168. (1995). Emotional attention, clarity and repair. Exploring
Luminet, O., Zech, E., Rime, B., & Wagner, H. (2000). Predicting emotional intelligence using the trait meta mood scale. In J. W.
cognitive and social consequences of emotional episodes: the Pennebaker (Ed.), Emotion, disclosure, & health (pp. 125–154).
contribution of emotional intensity, the five factor model and Washington, D.C.: American Psychological Association.
alexithymia. Journal of Research in Personality, 34(4), 471–497. Schmidt, J. E., & Andrykowski, M. A. (2004). The role of social and
Lutgendorf, S. K., & Antoni, M. H. (1999). Emotional and cognitive dispositional variables associated with emotional processing in
processing in a trauma-disclosure paradigm. Cognitive therapy adjustment to breast cancer: an internet based study. Health
and Research, 23(4), 423–440. Psychology, 23(3), 259–266.
Lynch, T. R., Robins, C. J., Morse, J. Q., & MorKrause, E. D. Snell, W. E., Miller, R. S., & Belk, S. S. (1988). Development of
(2001). A mediational model relating affect intensity, emotional emotional self disclosure scale. Sex Roles, 18(1), 59–73.
inhibition and psychological distress. Behavior Therapy, 32(3), Thomsen, D. K. (2006). The association between rumination and negative
519–536. affect: a review. Cognition and Emotion, 20(8), 1216–1235.
Lyubomirsky, S., Sousa, L., & Dickerhoof, R. (2006). The costs and Trapnell, P. D., & Campbell, J. D. (1999). Private self consciousness
benefits of writing, talking and thinking about life’s triumphs and and the five factor model of personality: distinguishing rumination
defeats. Journal of Personality and Social Psychology, 90(4), from reflection. Journal of Personality and Social Psychology, 76
692–708. (2), 284–304.
Mendolia, M., & Kleck, R. E. (1993). Effects of talking about a Watson, M., & Greer, S. (1983). Development of a questionnaire
stressful event on arousal: does what we talk about make a measure of emotional control. Journal of Psychosomatic Medi-
difference? Journal of Personality and Social Psychology, 64(2), cine, 27, 299–305.
283–292. Zakowski, S. G., Valdimarsdottir, H. B., & Bovbjerg, D. H. (2001).
Nolen-Hoeksema, S., & Davis, C. G. (1999). “Thanks for sharing Emotional expressivity and intrusive cognitions in women with
that”: ruminators and their social support networks. Journal of family histories of breast cancer: application of a cognitive
Personality and Social Psychology, 77(4), 801–814. processing model. British Journal of Health Psychology, 6(2),
Norman, S. A., Lumley, M. A., Dooley, J. A., & Diamond, M. P. 151–165.
(2004). For whom does it work? Moderators of the effects of Zech, E. (2000). The effects of the communication of emotional
written emotional disclosure in a randomized trial among experiences. Unpublished Doctoral Dissertation, University of
women with chronic pelvic pain. Psychosomatic Medicine, 66 Louvain, Louvain-la-Neuve,Belgium.
(2), 174–183. Zech, E., & Rime, B. (2005). Is talking about an emotional experience
Paez, D., Velasco, C., & Gonzalez, J. L. (1999). Expressive writing helpful? Effects of emotional recovery and perceived benefits.
and the role of alexithymia as a dispositional-deficit in self Clinical Psychology and Psychotherapy, 12(4), 270–287.

Vous aimerez peut-être aussi