Journal of Research in Personality 44 (2010) 70–77

Contents lists available at ScienceDirect

Journal of Research in Personality

journal homepage: www.elsevier.com/locate/jrp

The moderating effect of self-concept valence on the relationship between

self-focused attention and mood: An experience sampling study
Nigel P. Field a,*, Raeda Joudy a, Daniel Hart b
a
Ph.D. Program in Clinical Psychology, Palo Alto University, Palo Alto, CA, USA
b
Department of Psychology, Rutgers University, Camden, NJ, USA

a r t i c l e i n f o a b s t r a c t

Article history: This study examined the moderating role of self-esteem and neuroticism on the relationship between
Available online 10 November 2009 self-focused attention and mood using experience sampling methodology. 107 undergraduate partici-
pants were paged seven times each day over the course of a week and completed a set of measures that
Keywords: included items assessing current mood, focus of attention, and self-evaluative thoughts during these
Self-focused attention times. A significant within-person relationship was found between focus of attention and mood such that
Experience sampling self-focus samplings were associated with a shift toward negative mood compared with samplings
Self-esteem
involving non-self-focus. As hypothesized, those with higher self-esteem and lower neuroticism showed
Neuroticism
Mood
less of a shift toward negative mood during self-focused attention relative to those with lower self-
esteem and higher neuroticism. Moreover, the valence of self-evaluative thoughts during states of self-
focused attention was shown to mediate the impact of these personality traits on mood.
Ó 2009 Elsevier Inc. All rights reserved.

1. Introduction
There is evidence that whether self-focused attention incites
negative mood is contingent on the evaluative content of the
self-concept (e.g., Phillips & Silvia, 2005; Sedikides, 1992). Under
the assumption that the most salient and chronically accessible
self-aspects are likely to become the focus of attention under
self-aware conditions (Carver, 1979), the nature of their evaluative
content should have direct bearing on whether self-focused atten-
tion induces negative mood (Carver & Scheier, 1998; Sedikides,
1992). Specifically, individuals who possess more negatively valen-
ced self-aspects should show a greater increase in negative mood
relative to individuals with more positively-valenced self-aspects
under self-aware conditions. The present study examines this in
the context of daily life through the use of experience sampling
methodology. It thereby extends previous work on the affective
consequences of experimentally manipulated self-focused atten-
tion in addressing its spontaneous occurrence and effect on mood
as it occurs naturalistically.
1.1. Background
Studies examining the impact of self-focused attention on mood
have typically relied on experimental manipulation of self-focused
* Corresponding author. Address: Palo Alto University, 1791 Arastradero Rd., Palo
Alto, CA 94304, USA.
E-mail address: nfield@pgsp.edu (N.P. Field).
attention, such as through the use of mirrors (e.g., Duval & Wickl-
und, 1972; Phillips & Silvia, 2005; Sedikides, 1992). These studies
characteristically employ between-group designs in which one
group is administered a self-focus manipulation while participants
in a control condition are given instructions promoting attentional
focus away from the self. Mood ratings comparing the two groups
obtained after the manipulation provide the means for examining
the effect of self-focus versus non-self-focus on mood.
Studies addressing the linkage between self-focused attention
and mood have obtained mixed results; while some studies have
shown that self-focused attention is related to more negative mood
(e.g., Ingram, Lumry, Cruet, & Sieber, 1987; Pyszczynski, Holt, &
Greenberg, 1987), in other studies self-focused attention is associ-
ated with more positive mood (e.g., Ickes, Wicklund, & Ferris, 1973;
Salovey, 1992). These discrepant findings informed the present
study hypothesis that the valence of the self-concept may play a
role in whether self-focused attention is likely to induce negative
or positive mood. Indeed, previous studies provide some support
for this position. In a study examining the relationship between
self-focused attention, the evaluative tone of the self-concept and
mood, Sedikides (1992) found that, among participants who were
given a self-focused attention manipulation involving writing
about themselves while facing the reflective side of a mirror, those
with more positive self-conceptions made higher happy mood rat-
ings relative to those with more negative self-conceptions. On the
other hand, among those in the other-focused control condition
involving writing about an acquaintance while facing the backside
of a mirror, no differences were found in mood ratings between

0092-6566/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jrp.2009.11.001
 N.P. Field et al. / Journal of Research in Personality 44 (2010) 70–77
positive and negative self-conception participants. These results
thus suggest that the impact of self-focused attention on mood is
moderated by the evaluative content of participants’ self-
conceptions.
In a similar vein, Phillips and Silvia (2005) examined self-dis-
crepancy (i.e., negative self-conception) as a moderator of the link-
age between self-focused attention and mood. Participants were
assigned to either a high self-awareness condition involving facing
the reflective side of a large mirror or a low self-awareness condi-
tion in which they faced the backside of the mirror. They were then
asked to complete a measure assessing the extent to which their
actual self was discrepant with their ideal self and measures of cur-
rent mood state. As predicted, a significant interaction effect was
found for self-awareness and self-discrepancy on mood. In effect,
the self-focus manipulation served to increase participants’ aware-
ness of discrepancies between their self-conceptions and ideal self-
standards, leading to more negative mood among participants with
higher self-discrepancies relative to those with lower self-discrep-
ancies, whereas no such differences were found in the low self-
aware condition.
1.2. Present study
The present study extends previous work on self-focused atten-
tion and mood through the use of experience sampling methodol-
ogy. This method offers a more ecologically valid means for
investigating the moderating effect of the evaluative tone of the
self-concept on the impact of self-focused attention on mood than
studies relying on an experimentally induced self-awareness
manipulation. Equally if not more important, intensive longitudi-
nal data gathering made possible through the use of experience
sampling methodology provides a means for examining within-
person relationships between focus of attention and mood. Exist-
ing studies typically rely on between-person or -group analyzes
in which one group is administered a self-focus manipulation
while participants in the control condition are given instructions
promoting attentional focus away from the self. These study de-
signs are limited in not being able to draw conclusions regarding
within-person relationships. This is a serious problem in that the
findings from between-person designs have been known to differ
substantively from those obtained using within-person designs,
such that the findings obtained using the former may not apply
at the level of the individual (see Tennen, Affleck, Armeli, & Carney,
2000). To our knowledge, a study conducted by Csikszentmihalyi
and Figurski (1982) is the only other study using experience sam-
pling methodology to examine the effect of self-focused attention
on mood. Their study did not address individual differences in
self-concept valence on the relationship between self-focused
attention and mood, however. Furthermore, in simply aggregating
within-person focus of attention and mood scores in their analysis,
instead of using multilevel statistical procedures that could pro-
vide a means for examining within-person relationships between
these variables as well as between-person factors that might mod-
erate these within-person relationships, it was not possible to an-
swer within-person questions regarding the association between
self-awareness and mood.
This study also has theoretical implications in addressing the
trait–state interface in terms of the interaction between time-
invariant dispositional influences, involving the evaluative tone
of the self-concept, and time-varying situational influences, involv-
ing self-focused versus non-self-focused attention, on mood state.
Through repeated measurements of attentional focus and mood
state throughout the course of a week via experience sampling, it
was possible to obtain a sufficient within-person sampling of these
mental states to examine systematically such trait–state interac-
tion effects in daily life.
71
In line with the results of Sedikides (1992) and Phillips and Sil-
via (2005), we hypothesized that the effect of the evaluative con-
tent of the self-concept on mood would be more pronounced
under conditions of self-focused attention. Self-esteem and neurot-
icism measures were used to assess self-concept valence. Self-es-
teem provides a direct measure of the evaluative tone of the self-
concept (Rosenberg, 1965). Because self-esteem and neuroticism
are known to be strongly correlated (Robins, Tracy, & Trezesniew-
ski, 2001; Watson, Suls, & Haig, 2002), reflecting their shared focus
on negative self experience and dissatisfaction, neuroticism pro-
vides an additional albeit indirect measure of the evaluative con-
tent of the self-concept. We predicted that both measures would
be significant moderators of the impact of self-focused attention
on mood, such that higher scores in neuroticism and lower scores
in self-esteem would be linked to a greater shift toward negative
mood under naturally occurring states of self-focused attention
relative to naturally occurring states in which attention was direc-
ted away from the self.
Finally, stemming from the assumption that the impact of self-
esteem and neuroticism on mood lies through their effect on the
evaluative tone of the self-concept activated during states of self-
focused attention, it follows that the latter should mediate the ef-
fects of these personality variables on mood. We therefore ex-
pected that on experience samplings involving self-focused
attention, participants with lower self-esteem and higher neuroti-
cism would have more negative thoughts about themselves rela-
tive to those with higher self-esteem and lower neuroticism, and
that this would serve as a mechanism by which these personality
variables affected mood as shown in a mediational analysis.
2. Method
2.1. Participants
One hundred and seven undergraduate psychology students (59
males and 48 females) were recruited for a study of personality
and experience sampling. Participants were paid $120 for their
involvement in the study.
2.2. Procedure
The measures used in this study were taken from a larger re-
search project on personality and ongoing experience. For previous
published research using this database, see Hart, Field, Garfinkle,
and Singer (1997) and Oosterwegel, Field, Hart, and Anderson
(2001). Although data from the experience sampling measures
was used in these past studies, the present study involved the first
attempt to examine within-person relationships on this database.
The Hart et al. study examined proximities between representa-
tions of self and others in their relation to personality measures
of adjustment and experience sampling aggregate mood scores.
The Oosterwegal et al. study examined the relationship between
self-esteem variation, as assessed in experience sampling, and per-
sonality measures of psychological adjustment.
Those who agreed to participate were given a packet of ques-
tionnaires to complete at home consisting of personality and mood
measures that included the Rosenberg Self-Esteem Inventory
(Rosenberg, 1965) NEO-FFI Neuroticism (Costa & McCrae, 1992),
and the Self-Consciousness Scale (Fenigstein, Scheier, & Buss,
1979). After completing and returning the packet of question-
naires, participants were then given a booklet of experience sam-
pling forms and a preprogrammed electronic organizer with an
alarm function capable of being programmed to sound at a random
time within a fixed interval. Participants were electronically sig-
naled seven times a day between 8 A.M. and 10 P.M. at random
72 N.P. Field et al. / Journal of Research in Personality 44 (2010) 70–77
times within each two-hour block interval over the course of a
week. Upon hearing the signal, they completed an experience sam-
pling form that included mood, focus of attention, and self-evalua-
tive thoughts measures. Participants completed on average 37 of
49 experience sampling forms (73%; SD = 9.7).
2.3. Measures
2.3.1. Self-esteem
The Rosenberg Self-Esteem Scale (SES; Rosenberg, 1965) was
developed as a unidimensional measure of global self-esteem. It
consists of 10 statements addressing overall feelings of self-worth
and self-acceptance. The items are answered on a four-point scale
ranging from strongly agree to strongly disagree. The SES has dem-
onstrated acceptable reliability and validity (Hagborg, 1993). The
internal consistency of the SES in the present study sample was
satisfactory (Cronbach alpha = .88).
2.3.2. Neuroticism
As part of the 60-item NEO Five-Factor Inventory (NEO-FFI), this
neuroticism measure was developed to provide a concise means
for assessing neuroticism (Costa & McCrae, 1992). NEO-FFI-Neurot-
icism consists of a 12 item scale involving a Likert response format.
It has been shown to possess good internal consistency (Costa &
McCrae, 1992) as well as good validity (Pytlik Zillig, Hemenover,
& Dienstbier, 2002). The internal consistency of the NEO-FFI in
the present study sample was satisfactory (Cronbach alpha = .90).
2.3.3. Self-Consciousness Scale
The Fenigstein, Scheier, and Buss (1975) Self-Consciousness
Scale is a widely used trait measure of self-awareness. It consists
of 23 items assessing three different facets of self-consciousness:
(1) private self-consciousness addressing one’s tendency to attend
to one’s inner thoughts and feelings (10 items; alpha = .75 in the
present study sample); (2) public self-consciousness assessing
one’s tendency toward focusing on oneself as a social object (7
items; alpha = .85 in the present study sample) and (3) social anx-
iety assessing the degree of discomfort experienced in the presence
of other people (6 items; alpha = .76 in the present study sample).
It was included in the present study as a validation check for the
experience sampling self-awareness measure developed for the
present study described below.
2.3.4. Experience sampling form
The experience sampling form contained items addressing var-
ious aspects of participants’ experiences and behaviors at the time
of being signaled. Items relevant to the present study were an-
swered in the following order: mood, focus of attention, and eval-
uative thoughts about the self.
2.3.4.1. Experience sampling mood measure. The form included an 8-
item mood measure consisting of the following set of bipolar
items: depressed-bright, irritable-cheerful, closed-open,
ashamed-proud, lonely-sociable, confused-clear, sad-happy, com-
petitive-cooperative (alpha = .91), answered on a scale ranging
from 1 to 5. An average score was calculated for each sampling
within each participant. For more detailed psychometric informa-
tion on the mood measures, see Hart et al. (1997).
2.3.4.2. Experience sampling self-focused attention measure. After
completing the experience sampling mood items, participants an-
swered the following experience sampling item assessing self-fo-
cused attention: ‘‘At the time you were paged, the focus of your
thoughts was: (a) on other people rather than on your self; (b)
on objects, ideas, or activities; (c) on your own thoughts, feelings,
and obligations and (d) on the way you imagine other people think
about you”. Participants were instructed that they could endorse
more than one alternative. Aggregating across samplings, partici-
pants on average endorsed an other person focus ‘‘a” on 27.3%
(SD = 13.5%) of the samplings, a non-social focus ‘‘b” on 53.6%
(SD = 17.2%) of the samplings, private self-focus ‘‘c” on 24.5%
(SD = 14.8%) of the samplings, and a public self-focus ‘‘d” on 9.6%
(SD = 1.0%) of the samplings.
To assess the convergent and discriminant validity of the self-
focused attention measure, a simple correlational analysis was
conducted involving the aggregate form of this attentional focus
item with the Fenigstein et al. (1975) self-consciousness measure.
Some support for the convergent validity was shown in a signifi-
cant, albeit weak, positive correlation between the proportion of
samplings for which private self-focus was endorsed and the pri-
vate self-consciousness subscale of the Fenigstein et al. measure
(r = .20, p < .05). No support was shown for its discriminant valid-
ity, however, in that private self-focus was also significantly posi-
tively correlated with the public self-consciousness (r = .23,
p < .05) and social anxiety (r = .28, p < .01) subscales of the Fenig-
stein et al. measure. It should also be noted that the proportion
of samplings for which public self-focus was endorsed was not sig-
nificantly correlated with the public self-consciousness subscale
(r = .08, ns), nor with the private self-consciousness (r = .17, ns)
and social anxiety (r = .15, ns) subscales.
Although, at best, only limited support was shown for the valid-
ity of the private self-focus state measure when comparing it with
a dispositional measure of self-focused attention, it is not alto-
gether clear how strong a relationship should be expected here, gi-
ven that the former was a psychological state measure while the
latter constituted a trait measure. Because only the private self-fo-
cus item was significantly correlated with the Fenigstein et al.
measure, this provided some justification for selecting it as the
independent variable self-focus measure in the present study. This
is also consistent with the extent literature on self-focused atten-
tion and mood that typically addresses private self-focus, as re-
viewed in the Introduction. Finally, it should be noted that the
results of the main set of analyzes remain essentially unchanged
if using a self-focused attention measure based on combining the
private and public self-focused attention scores.
2.3.4.3. Experience sampling self-evaluation item. In samplings for
which participants responded with ‘‘c”, indicating private self-fo-
cus, they were then asked to judge how negatively versus posi-
tively they were thinking about themselves on a 5-point scale,
ranging from very negative (1) to very positive (5). This provided
a measure of the evaluative tone of the self state activated on a gi-
ven self-focused attention sampling. Support for the convergent
validity of this item was shown in the significant correlation be-
tween its mean score, aggregated across samplings, and both the
self-esteem (r = .35, p < .001) and neuroticism (r = .38, p < .001)
measures.
3. Results
Given the hierarchical nature of the data, wherein level-1 with-
in-person experience sampling repeated measurements of mood
and self-focused attention are nested within level-2 between-per-
son measurements (i.e., self-esteem and neuroticism), hierarchical
linear modeling was selected as the appropriate method for this
analysis. This analysis permits a simultaneous analysis of level-1
and level-2 variation. For the level-1 self-focused attention vari-
able, non-self-focused attention samplings were coded as ‘‘0” and
self-focused attention samplings as ‘‘1” in the analyzes. Self-fo-
cused attention was person-centered in all the analyzes. Further-
more, analyzes involving the level-2 between-person variables
 N.P. Field et al. / Journal of Research in Personality 44 (2010) 70–77 73

(self-esteem and neuroticism) were group-centered. All analyzes Table 1b

were conducted using HLM6 statistical software (Raudenbush, Random effects model for self-focused attention predicting mood.

Bryk, Cheong, Congdon, & du Toit, 2004), unless otherwise stated. Random effects Standard deviation Variance component Chi- v2
Intercept (u0j) 0.279 0.078 1036.381***
3.1. Unconditional intercept and slope model Slope (r1j) 0.172 0.029 166.716**
Error (rij) 0.556 0.309
Prior to addressing the study hypotheses regarding the moder- **
p < 0.01.
ating role of individual differences in self-esteem and neuroticism ***
p < 0.001.
on the relationship between self-focused attention and mood, an
initial unconditional intercept and slope model analysis was con-
ducted focusing exclusively on the experience sampling re- attentional states condition relative to non-self-focus attentional
peated-measures self-focused attention and mood measures. The states.
level-1 within-person unconditional intercept and slope model is The Table 1b results summarize the random effects model
shown in addressing whether significant between-person differences exist
in the intercept and slope coefficients. The significant variance esti-
Moodij ¼ b0j þ b1j Self -focusij þ rij: ð1Þ
mate 0.078 (p < .001) for the intercept and variance estimate 0.029
This equation specifies a unique relationship between self-focused (p < .001) for the slope indicated that participants differed from
attention and mood for each person. The criterion variable Moodij one another in their average mood ratings and in their change
represents the mood rating for person j on experience sampling oc- scores from the non-self-focused to the self-focused states condi-
casion i. Self-focusij is person j’s self-focus rating on occasion i tions respectively.
(coded 0 for non-self-focused attention and 1 for self-focused atten-
tion). The intercept coefficient B0 represents person j’s mood, cen- 3.2. Intercepts and slopes as – an outcome model for level-2 predictors
tered on his or her average self-focus score. B1 is person j’s slope
coefficient for the self-focus variable, while rij is the random com- This model involves introducing between-person predictors in
ponent for person j’s mood on occasion i. the level-2 model. In light of the significant between-person vari-
Eqs. (2) and (3) show the specification of the level-2 uncondi- ation in the intercept and slope coefficient scores shown in the
tional intercept and slope model addressing between-person dif- above random effects model results (Table 1b), the final set of ana-
ferences in the estimation of the intercept b0j and slope b1j lyzes involved identifying between-person predictor variables that
coefficients respectively: might explain this variation. In the present study context, this in-
b0j ¼ y00 þ u0j ð2Þ volved introducing the self-esteem and neuroticism measures as
level-2 predictors, as specified in the level-2
b1j ¼ y10 þ u1j ð3Þ

Here, y00 and y10 represent the estimated intercept and slope coef-
b0j ¼ y00 þ y01  ðself-esteem=neuroticismÞ þ uij ð4Þ
ficients respectively for the study sample as a whole averaged b0j1 ¼ y10 þ y11  ðself-esteem=neuroticismÞ þ u0j ð5Þ
across participants, while u0j and u1j represent the random residual
Prior to conducting the final set of analyzes addressing the main
component identifying variation between individuals in their
study hypotheses regarding the moderating effect of each these le-
respective intercept and slope coefficient scores.
vel-2 predictors on the relationship between self-focused attention
A preliminary analysis was conducted to determine if there
and mood, it was first necessary to perform an analysis including
were changes in mood ratings as a function of length of time in
gender and number of samplings as level-2 between-person pre-
the study. This analysis was performed to rule out the possibility
dictors. This was required in order to determine whether it would
that participating in experience sampling over time had an impact
be necessary to statistically control for these variables in the final
on mood ratings. This involved entering the level-1 study sampling
set of analyzes. The results indicated that gender was not predic-
variable as a predictor of mood. The non-significant results for the
tive of the intercept [y01 (gender) = .087, t = 1.549, ns] or slope
estimated average slope coefficient for time (y10 = .002, t = 1.602,
[y11 (gender) = .032, t = 0.585, ns]; nor were number of samplings
ns) indicated that mood ratings did not differ as a function of time.
predictive of the intercept [y01 (samplings) = .002, t = 0.361, ns] or
Therefore, it was not necessary to control for this variable in sub-
slope [y11 (samplings) = .006, t = .981, ns].
sequent analyzes.
A similar preliminary analysis was conducted for mood varia-
The results for the unconditional intercept and slope analyzes
tion in the experience sampling. In previous research, individual
involving self-focused attention as a predictor of mood are shown
differences in mood variation has been shown to be associated
in Tables 1a and 1b. Table 1a summarizes the fixed effects model
with personality measures related to psychological adjustment
results wherein the intercept coefficient 3.227 (p < .001) represents
such as neuroticism (Larsen & Cutler, 1996). In the prior study by
the grande mean y00 average mood rating for the sample as a
Oosterwegel et al. (2001) on a subset of the participants and expe-
whole while the slope coefficient 0.214 (p < .001) represents the
rience samplings from the present study database, those with
grande mean average mood change score as a function of non-
greater mood variation in their experience samplings had higher
self-focused wherein versus self-focused attention for the sample
scores in neuroticism. If indeed neuroticism and self-esteem are re-
as a whole. The significant negative slope coefficient here indicated
lated to mood variation then this may confound the results of the
that participants reported less positive mood under self-focused
main study hypotheses; specifically, those with higher neuroticism
and lower self-esteem may be found to have a stronger relation-
Table 1a ship between focus of attention and mood simply as a result of
Fixed effects model for self-focused attention predicting mood. greater mood variation in their experience samplings – knowing
Fixed effects Coefficient T-ratio
that the strength of correlations can be influenced by extent of var-
iation in the distribution of scores. It was therefore important to
Intercept (y00) 3.227 111.967***
determine whether individual differences in mood variation mod-
Slope (y10) 0.214 7.440***
erated the strength of the within-person correlations between
***
p < 0.001. attentional focus and mood. This analysis required first calculating
74 N.P. Field et al. / Journal of Research in Personality 44 (2010) 70–77

the standard deviation for the experiencing sampling mood scores

on each participant as a measure of within-person mood variabil-
ity. This variable was then entered as a level-2 predictor in the
regression analysis. The results indicated that although mood var-
iation was not predictive of the intercept [y01 (mood variabil-
ity) = .165, t = 0.291, ns], it was predictive of the slope [y11 (mood
variability) = .533, t = 2.680, p < .01]. In effect, those with greater
mood variation indicated a stronger negative relationship between
self-focused attention and mood relative to those with less mood
variation. Consequently, it was important to statistically control
for this.
The first analysis addressing the main study hypotheses in-
volved entering self-esteem and mood variation in the level-2
model. The results are shown in Table 2. The significant self-es-
teem intercept coefficient of 0.016 (t = 3.348, p < .01) indicated that
a one-unit change in the self-esteem score was associated with a
0.016 unit change in the average mood score. The significant
self-esteem slope coefficient of 0.009 (t = 2.250, p < .05) indicated Fig. 1. Interaction of mood and non-self-focused versus self-focused attention
that a one-unit change in the self-esteem score was associated across level of self-esteem. SD = regression line at one standard deviation below
the mean for self-esteem; SD+ = regression line at one standard deviation above the
with a 0.009 change in a slope value for the effect of self-focused
mean for self-esteem. Positive mood value of three represents the midpoint
attention on mood. Thus, as hypothesized, self-esteem was a sig- between positive and negative mood.
nificant moderator of the effect of self-focused attention on mood,
such that participants with higher self-esteem showed less of a
shift toward negative mood under self-focused attention condi- Table 3
tions relative to those with lower self-esteem. This significant Self-focused attention predicting mood as moderated by mood variation and
neuroticism.
interaction effect is illustrated in Fig. 1 in which regression lines
based on estimated plus and minus one standard deviation scores Fixed effects Coefficient T-ratio
for self-esteem are identified. Noteworthy here is that those who Intercept (y00) 3.228 127.741***
were low in self-esteem (1 SD) showed a steeper negative slope Mood variation (y01) 0.282 2.249*
toward less positive mood as a function of non-self-focused versus Neuroticism (y02) 0.014 4.721***
Slope (y10) 0.203 7.761***
self-focused attention relative to those high in self-esteem (+1 SD).
Mood variation (y11) 0.490 2.529*
A similar analysis was conducted in which neuroticism and Neuroticism (y12) 0.007 2.643**
mood variation were entered as a level-2 predictors. The results *
p < 0.05.
are shown in Table 3. The significant neuroticism intercept coeffi- **
p < 0.01.
cient of .014 (t = 4.721, p < .001) indicated that a one-unit in- ***
p < 0.001.
crease in neuroticism was associated with a .014 unit shift
toward negative mood. As hypothesized, the significant slope coef-
ficient .007 (t = 2.643, p < .01) for neuroticism indicated that
those who were higher in neuroticism showed a greater shift to-
ward negative mood under self-focused conditions relative to
those lower in neuroticism. The results are illustrated in Fig. 2.
Similar to that shown for self-esteem, those high in neuroticism
(+1 SD) showed a steeper negative slope toward less positive mood
as a function of non-self-focused versus self-focused attention rel-
ative to those low in neuroticism (1 SD).
To highlight the value of a hierarchical linear modeling analysis
as used here to identify within-person relationships and between-
person variables that moderate these within-person relationships,
an exclusively between-persons analysis was performed as a point
of comparison. This analysis required first aggregating the within-
person self-focused attention and mood scores; the aggregate fo-
cus of attention score represented the proportion of self-focused

Table 2
Self-focused attention predicting mood as moderated by mood variation and self- Fig. 2. Interaction of mood and non-self-focused versus self-focused attention
esteem. across level of neuroticism. SD = regression line at one standard deviation below
the mean for neuroticism; SD+ = regression line at one standard deviation above the
Fixed effects Coefficient T-ratio mean for neuroticism. Positive mood value of three represents the midpoint
Intercept (y00) 3.228 118.664*** between positive and negative mood.
Mood variation (y01) 0.128 0.942
Self-esteem (y02) 0.016 3.348**
Slope (y10) 0.204 7.840*** attention samplings while the aggregate mood score represented
Mood variation (y11) 0.556 2.836** the average mood score. Regression analyzes were then performed
Self-esteem (y12) 0.009 2.250* using these aggregated scores in conjunction with self-esteem and
*
p < 0.05. neuroticism respectively. The predictors were centered prior to the
**
p < 0.01. analyzes and the interaction terms were computed based on these
***
p < 0.001. centered scores.
 N.P. Field et al. / Journal of Research in Personality 44 (2010) 70–77
The first regression analysis involved entering self-focus and
self-esteem along with their interaction as predictors of mood.
The results indicated that, although self-esteem was significantly
associated with mood [B = .085, t = 3.006, p < .01], no significant ef-
fect was shown for self-focused attention on mood [B = .038,
(t = 1.323, ns) or for the self-focused attention by self-esteem inter-
action on mood [B = .026, (t = .836, ns)]. Similar results were found
in the regression involving self-focused attention and neuroticism
along with their interaction as predictors of mood. Again, although
neuroticism was significantly associated with mood [B = .135,
t = 4.797, p < .001] no significant effect was shown for self-fo-
cused attention on mood [B = .005, (t = 197, ns) or for the self-fo-
cused attention by neuroticism interaction on mood [B = .021,
(t = .681, ns). This exclusively between-persons analysis therefore
failed to identify relationships between attentional focus and mood
as well as failing to identify the moderating effect of self-esteem
and neuroticism.
3.3. Self-evaluation as a mediator of the impact of self-esteem and
neuroticism on mood
A final analysis was conducted to examine more directly
whether the significant findings for the moderating effects of
self-esteem and neuroticism on mood were attributable to the im-
pact of these personality variables on the evaluative tone of
thoughts about the self that are assumed to become activated un-
der self-focused attention conditions. This involved a mediational
analysis in which the effect of self-esteem and neuroticism on
mood during states of self-focused attention would be mediated
by the evaluative thoughts about the self that become accessible
during these states. It was possible to address because participants’
evaluative ratings of their thoughts about themselves were avail-
able for experience samplings in which self-focused attention
was endorsed.
A variable is regarded as a mediator insofar as it carries the
influence of a given predictor variable on a given criterion variable.
Mediation can be said to occur when the following conditions are
met: (1) the predictor variable affects the mediator; (2) the predic-
tor variable affects the criterion variable in the absence of the
mediator and (3) the influence of the predictor variable on the cri-
terion variable is significantly reduced when the mediator is en-
tered into the model, while the mediator is shown to exert a
significant unique effect on the criterion variable (Baron & Kenny,
1986). This multilevel mediational analysis was conducted using
SPSS.17 mixed regression analysis, adapted from Krull and MacK-
innon (2001) for repeated-measures level-1 data.
The first set of analyzes involved determining whether self-
evaluative thoughts mediated the relationship between self-es-
teem and mood. An intercept as outcome model analysis, involving
self-esteem as a level-2 between-person predictor of participants’
estimated mean scores for experience sampling self-evaluative
thoughts and mood respectively, verified conditions 1 and 2 above
that self-esteem was predictive of self-evaluative thoughts
[B = .030, t = 4.000, p < .001], as well as mood [B = .014, t = 2.522,
p < .01]. In support of a mediational model as outlined in condition
3 above, when both self-esteem and self-evaluative thoughts were
entered as predictors of mood in a main effects model, self-esteem
was no longer predictive of mood [B = .003, t = .841, ns] whereas
self-evaluative thoughts was shown to be predictive of mood
[B = .353, t = 24.388, p < .001]. A Sobel test confirmed that the indi-
rect link between self-esteem and mood, as mediated by self-eval-
uative thoughts, was significant (Z = 4.362, p < .001).
A comparable set of analyzes were performed to determine
whether self-evaluative thoughts mediated the relationship be-
tween self-esteem and mood. In support of conditions 1 and 2
above respectively, neuroticism was shown to be predictive of
75
self-evaluative thoughts [B = .024, t = 5.176, p < .001] and mood
[B = .021, t = 6.236, p < .001]. Consistent with condition 3 above,
when both neuroticism and self-evaluative thoughts were entered
as predictors, support for a partial mediational model was shown
in the reduced influence of neuroticism on mood [B = .009,
t = 3.635, p < .001] in conjunction with a significant effect for
self-evaluative thoughts on mood [B = .352, t = 22.500, p < .001].
A Sobel test confirmed that the indirect link between trait neurot-
icism and mood, as mediated by self-evaluative thoughts, was sig-
nificant (Z = 4.690, p < .001).
4. Discussion
This study addressed individual differences in the evaluative
content of the self-concept as a significant moderator of the with-
in-person relationship between self-focused attention and mood. It
was hypothesized that those with a more negatively valenced self-
concept, as assessed using measures of self-esteem and neuroti-
cism, would show a greater shift toward negative mood under
self-focused attention versus non-self-focused attention conditions
relative to those with a less negative self-concept. As predicted,
both self-esteem and neuroticism were shown to be significant
moderators of the relationship between self-focused attention
and mood in a hierarchical linear modeling analysis, such that
higher self-esteem was associated with less of a within-person
shift toward more negative mood while higher neuroticism was
associated with a more pronounced shift toward more negative
mood under self-focused attention conditions.
These results highlight the importance of individual differences
in the evaluative content of the self-concept underlying the rela-
tionship between self-focused attention and mood under naturally
occurring changes in self-focused attention and mood respectively
as assessed through the use of experience sampling methodology.
It thus extends previous findings where self-focused attention was
experimentally manipulated (e.g., Sedikides, 1992; Phillips & Silvia,
2005). The present study also extends previous research in
addressing within-person relationships between self-focused
attention and mood; through the use of experience sampling in
conjunction with hierarchical linear modeling statistical proce-
dures, it provides a basis for being able to draw conclusions that
apply to the individual in a way that is not possible in the afore-
mentioned experimental studies on self-focused attention and
mood that rely exclusively on between-person designs.
The value of a hierarchical linear modeling approach was high-
lighted further in comparing the results using this analysis with
those obtained when performing an exclusively between-persons
regression analysis on the same data. In this followup regression
analysis using aggregate self-focused attention and mood scores,
a significant relationship was no longer found for self-focus and
mood or for the moderating role of self-esteem and neuroticism
on the relationship between self-focus and mood. Thus, different
conclusions would have been drawn based on these findings rela-
tive to those addressing within-person relationships.
The mediational analysis provided more direct support for the
impact of individual differences in self-esteem and neuroticism
on mood as due to the effect of these personality variables on
the evaluative tone of the kinds of thoughts about the self that be-
came active during self-focused states. Here, significant indirect ef-
fects were found for both self-esteem and neuroticism on mood, as
mediated by the valence of thoughts about the self aroused by self-
focused attention. These findings are consistent with a social cog-
nitive perspective on the self in which the self-concept as a cogni-
tive structure exerts its influence on experience and behavior in
situations that serve to activate this structure, such as those induc-
ing self-focused attention. In effect, the valence of the working self-
76 N.P. Field et al. / Journal of Research in Personality 44 (2010) 70–77
concept activated during such states is a mechanism by which the
evaluative content of the chronic self-concept as a cognitive struc-
ture affects mood.
It should be noted that although support for a full mediational
model was found for self-esteem, this was not so for neuroticism;
neuroticism had both indirect and direct links with mood. An
explanation for the direct link between neuroticism and mood
may be due to the fact that, unlike the self-esteem measure which
directly assesses the evaluative content of the self-concept, neurot-
icism is an indirect measure of this. As a measure of dispositional
negative affectivity, neuroticism may therefore affect mood
through additional mechanisms beyond its linkage with the evalu-
ative content of the self-concept.
The results for the unconditional intercept and slope model
showing a significant relationship between self-focused attention
and shift toward more negative mood for the study sample as a
whole is consistent with the literature on self-focused attention
and negative mood (e.g., Ingram et al., 1987; Pyszczynski et al.,
1987). It is noteworthy that although self-esteem and neuroticism
moderated this relationship, such that those with higher self-es-
teem or lower neuroticism showed less of a shift in mood as a func-
tion of focus of attention, a strong significant main effect
nevertheless remained for self-focus and mood. This suggests that
the relationship between self-focus and mood is not fully ex-
plained by the evaluative content of the dispositional self-concept.
An explanation for this finding may be that there is a normative
tendency for negative mood states, whatever it is that causes them,
to elicit self-focused attention. These results appear to run counter
to other research showing that self-focused attention is also linked
with positive mood (e.g., Ickes et al., 1973; Salovey, 1992). How-
ever, because the present study used bipolar mood rating scales,
it was not possible to examine the impact of self-focused attention
on negative and positive mood separately.
The results for self-focused attention and mood do not clarify
the direction of causality between these two measures. Although
the results of the mediational analysis suggest that the evaluative
tone of thoughts arising during states of self-focused attention af-
fect mood, this does not provide a basis for concluding that self-fo-
cused attention causes mood in a unidirectional manner. The
mediational analysis results simply imply that when in a state of
self-focused attention, whichever way it is elicited, the valence of
the thoughts that are activated have an effect on mood. Thus, It
is possible that a bi-directional relationship exists between focus
of attention and mood, such that negative mood may lead people
to focus their attention on themselves while self-focus also incites
negative mood. Nevertheless, the significant finding for the moder-
ating effect of self-concept valence on the relationship between
self-focused attention and mood suggests that it plays a causal role
in the strength of the self-focused attention and mood linkage.
This study has some limitations. First, although self-focused
attention was found to be a significant predictor of mood, it was
not possible to determine the direction of causality because the hier-
archical linear modeling analyzes focused exclusively on the syn-
chronous relationship between these two variables. Lagged
analyzes would help to clarify this, but it was not possible to conduct
such analyzes given the random interval timing between samplings.
Second, because participants completed the experience sam-
pling in booklet form, they may have answered the forms in
batches rather than completing them at the time of being signaled.
Since it is unlikely that participants would be able to guess the
study hypotheses and fabricate their ratings accordingly, complet-
ing the measures in batches, if anything, would be likely to result
in an increase in random error and thereby reduce the possibility
of identifying significant relationships among variables addressing
the main study hypotheses. One might argue, therefore, that the
results represent a conservative estimate of these relationships.
Third, the self-focused attention measure used in this study has
not been previously validated. Fourth, as previously mentioned,
the use of the bipolar mood ratings precludes the ability to deter-
mine conclusively whether self-focused attention differentially im-
pacts negative and positive mood. Finally, one should be careful in
generalizing these findings beyond a university student popula-
tion, such as clinical populations.
There are a number of considerations for future research. Sepa-
rate mood measures for negative and positive moods should be
considered in future research. The use of palm pilot technology
also needs to be considered in future studies. This would allow par-
ticipants to answer directly on the hand-held computer, thus mak-
ing it possible to verify that the measures were answered at the
time of being signaled. Finally, future research could involve
extending this work to clinical populations such as individuals
with major depression. For example, knowing that ruminative
self-focus is an important type of thinking underlying depression
(Nolen-Hoeksema, 2000), it might be useful to distinguish rumina-
tive self-focus from more adaptive types of cognitions activated
during self-focused attention – such as appraised ability to regu-
late negative mood. Personality measures such as ruminative style
(Nolen-Hoeksema, 2000) and trait measures of ability to regulate
negative mood (Catanzaro & Mearns, 1990) might also be consid-
ered as potential moderators here.
References
Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in
social psychological research: Conceptual, strategic, and statistical
considerations. Journal of Personality and Social Psychology, 51, 1173–1182.
Carver, C. S. (1979). A cybernetic model of self-attention processes. Journal of
Personality and Social Psychology, 37, 1251–1281.
Carver, C. S., & Scheier, M. (1998). On the self-regulation of behavior. New York:
Cambridge University Press.
Catanzaro, S. J., & Mearns, J. (1990). Measuring general expectancies for negative
mood regulation: Initial scale development and implications. Journal of
Personality Assessment, 54, 546–563.
Costa, P. T., Jr., & McCrae, R. R. (1992). Revised NEO Personality Inventory (NEO-PI-R)
and NEO Five-Factor Inventory (NEO-FFI) professional manual. Odessa, FL:
Psychological Assessment Resources.
Csikszentmihalyi, M., & Figurski, T. J. (1982). Self-awareness and aversive
experience in everyday life. Journal of Personality, 50, 15–28.
Duval, T. S., & Wicklund, R. A. (1972). A theory of objective self-awareness. New York:
Academic.
Fenigstein, A., Scheier, M. F., & Buss, A. H. (1975). Public and private self-
consciousness: Assessment and theory. Journal of Consulting and Clinical
Psychology, 43, 522–527.
Hagborg, W. J. (1993). The Rosenberg Self-Esteem Scale and Harter’s Self-Perception
Profile for Adolescents: A concurrent validity study. Psychology in the Schools,
30, 132–136.
Hart, D., Field, N. P., Garfinkle, J. R., & Singer, J. (1997). Representations of self and
other: A semantic space model. Journal of Personality, 65(1), 77.
Ickes, W. J., Wicklund, R. A., & Ferris, C. B. (1973). Objective self awareness and self
esteem. Journal of Experimental Social Psychology, 9, 202–219.
Ingram, R. E., Lumry, A. E., Cruet, D., & Sieber, W. (1987). Attentional processes in
depressive disorders. Cognitive Therapy and Research, 11, 351–360.
Krull, J. L., & MacKinnon, D. P. (2001). Multilevel modeling of individual and group
level mediated effects. Multivariate Behavioral Research, 36, 249–277.
Larsen, R. J., & Cutler, S. E. (1996). The complexity of individual emotional lives: A
within-subject analysis of affect structure. Journal of Social and Clinical
Psychology, 15, 206–230.
Nolen-Hoeksema, S. (2000). The role of rumination in depressive disorders and
mixed anxiety/depressive symptoms. Journal of Abnormal Psychology, 109,
504–511.
Oosterwegel, A., Field, N. P., Hart, D., & Anderson, K. (2001). The relation of self-
esteem variability to emotion variability, mood, personality traits, and
depressive tendencies. Journal of Personality, 69, 689–708.
Phillips, A. G., & Silvia, P. J. (2005). Self-awareness and the emotional consequences
of self-discrepancies. Personality and Social Psychology Bulletin, 31(5), 703–713.
Pyszczynski, T., Holt, K., & Greenberg, J. (1987). Depression, self-focused attention,
and expectancies for positive and negative future life events for self and others.
Journal of Personality and Social Psychology, 52(5), 994–1001.
Pytlik Zillig, L. M., Hemenover, S. H., & Dienstbier, R. A. (2002). What do we assess
when we assess a Big 5 trait? A content analysis of the affective, behavioral, and
cognitive porcesses represented in Big 5 personality inventories. Personality and
Social Psychology Bulletin, 28, 847–858.
 N.P. Field et al. / Journal of Research in Personality 44 (2010) 70–77
Raudenbush, S., Bryk, A., Cheong, Y. F., Congdon, R., & du Toit, M. (2004). HLM6:
Hierarchical linear and nonlinear modeling. Lincolnwood, IL: Scientific Software
International, Inc.
Robins, W. R., Tracy, J. L., & Trezesniewski (2001). Personality correlates of self-
esteem. Journal of Research in Personality, 35, 463–482.
Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton
University Press.
Salovey, P. (1992). Mood induced self-focused attention. Journal of Personality and
Social Psychology, 62(4), 699–707.
77
Sedikides, C. (1992). Attentional effects on mood are moderated by chronic self-
conception valence. Personality and Social Psychology Bulletin, 18, 580–584.
SPSS advanced statistics 17. SPSS incorporated. Chicago, IL.
Tennen, H., Affleck, G., Armeli, S., & Carney, M. A. (2000). A daily process approach to
coping. American Psychologist, 55, 626–636.
Watson, D., Suls, J., & Haig, J. (2002). Global self-esteem in relation to structural
models of personality and affectivity. Journal of Personality and Social Psychology,
83(1), 185–197.