Content uploaded by Shlomo Hareli
Author content
All content in this area was uploaded by Shlomo Hareli on Jan 08, 2014
Content may be subject to copyright.
Soc Psychol Educ (2008) 11:259–272
DOI 10.1007/s11218-008-9059-x
When does feedback about success at school hurt?
The role of causal attributions
Shlomo Hareli ·Ursula Hess
Received: 14 September 2007 / Accepted: 8 April 2008 / Published online: 29 April 2008
© Springer Science+Business Media B.V. 2008
Abstract The present research provides evidence that attribution theory can serve
as an explanatory framework not only to explain achievers’ reactions to their achieve-
ment based on their self generated understandings of what brought these achievements
about but also when such information is provided by others. Thus, when we succeed
at school, others may comment on the likely reasons for this success. The present
research addressed the question what it is that makes certain types of feedback on
the reasons for success at school hurtful. The results of two studies conducted in
the context of a school setting demonstrated that the causal structure implied by an
explanation for success explains why some explanations are perceived as hurtful and
elicit anger, shame, and guilt rather than pleasure or pride. Interestingly, the perceived
validity of the explanation is of relatively less importance for the elicitation of hurt
feelings and anger than the content of the explanation. Overall, these results provide
further evidence for the importance of attributional information for social emotions in
educational settings.
Keywords Attribution theory ·Hurt feeling ·Anger, shame, guilt ·Success at
school ·Peer feedback
1 Introduction
Attempts of explaining achievers’ reactions to their outcomes in educational settings
often use attribution theory as a theoretical framework (see e.g., Bar-Tal 1978,Forsyth
S. Hareli (B
)
Graduate School of Management, University of Haifa, Haifa 31905, Israel
e-mail: shareli@gsb.haifa.ac.il
U. Hess
Department of Psychology, University of Quebec at Montreal Canada, CP 8888, station ‘centre-ville’,
Montreal, QC, Canada H3C 3P8
123
260 S. Hareli, U. Hess
1986,Weiner 1976). However, most of the discussions connecting attribution theory
and educational achievements focused on the perspective of the achiever. That is, the
focus is on questions such as the way the achiever understands why she failed or
succeeded at school and the emotional and behavioral consequences of this under-
standing (Weiner 1976,Bar-Tal 1978). Nevertheless, more recently developments in
the literature discussing attributions made in educational settings, note that people’s
achievements’ do not occur in a vacuum but rather in a social context (Hareli and
Hess in press,Hareli and Weiner 2002,Hareli et al. 2006). Indeed achievements are
important events for individuals and hence achievers are motivated to understand their
causes (Weiner 1986) and occasionally the source of this understanding may be others
such as teachers or peers (Hareli and Weiner 2002). Such external feedback not only
increases the achievers’ understanding of the achievement situation but also affects
their response to the achievement. As suggested by Hareli and Weiner (2002) attribu-
tion theory can serve as a theoretical framework that explains the reaction of achievers
to feedback. Employing this line of reasoning, the goal of the present research is
to examine how the emotions that achievers experience following success at school
are shaped by observers’ feedback attempting to explain what brought about one’s
success.
When achievers receive feedback from others about the reasons for their success,
the achiever may feel pleasure or pride because his/her achievement did not go unno-
ticed and because the feedback may suggest that the achiever has superior qualities
(Hareli and Weiner 2002,Weiner 1995). However, others’ performance feedback can
also entrain negative social emotions such as hurt feelings as well as anger, shame
and guilt if it reflects negatively on the achiever. In line with this claim, Hareli and
Hess (in press) recently showed that certain types of observer feedback that include
explanations for the causes for a person’s failure leads to such emotions. In particular,
their research showed that the attributional information underlying observers’ expla-
nations determines if and to what extent an explanation will induce hurt feelings, anger,
guilt and shame in the achiever. The present research moves further by investigating
whether a similar process may also render success feedback hurtful. We first explain
the rationale underlying our research question.
1.1 Communicated explanations for success, causal attribution and emotions
In general, messages are emotionally hurtful if they imply a relational devaluation
(Leary et al. 1998,Leary and Springer 2001,Vangelisti 1994). Such messages
also elicit other emotions such as anger, guilt, anxiety, shame and embarrassment
(Gabriel 1998,Leary et al. 1998). In the context of feedback aimed at explaining a
person’s achievements, Hareli and Hess (in press) based on Weiner’s (1986) attribu-
tion theory of emotion and motivation have shown that an explanation for failure is
hurtful if it implies a devaluation of the person’s worth (see also Hareli and Weiner
2002).
Similarly, certain explanations for a person’s success may also devalue both the
success and the achiever’s worth (Hareli and Weiner 2002). In the following we dis-
cuss how attribution theory of emotion and motivation explains the link between hurt
123
Casual attribution and hurt feelings 261
feelings, anger, guilt and shame and the attributional information regarding the cause
of a success contained in messages explaining success at school.
Weiner (1985,1986) describes three dimensions of perceived causes: causes can
be considered internal or external to the actor (locus), controllable or uncontrollable
by the actor or others (controllability), and varying or unvarying over time (stability).
Because causal dimensions describe the nature of a cause, causal dimensions that
underlie causal success feedback also describe the link between the specific outcome
and the person attaining it. The locus of the causality determines whether something
about the person or something outside the person brought about the success. In addi-
tion, causal stability defines how permanent or transitory the cause is, and, for internal
causes, the dimension of controllability describes the extent to which the outcome is
the result of the person’s own actions. For example, both effort and luck are unstable,
yet effort is internal and under the person’s own control whereas luck is external and
uncontrollable. Therefore, when effort is described as the cause of an achievement
this implies that the person is responsible for it. However, when the outcome was
brought about by luck, this leads to the appraisal that the person is not responsible for
the outcome. When these causes are presented as reasons for a success, they convey
information on the role that the individual has played in bringing the success about
(Hareli and Hess in press,Hareli and Weiner 2000,2002,Weiner 1985,1986).
Given that communications that devalue a person’s worth tend to be hurtful (Leary
et al. 1998,Vangelisti and Young 2000,Young 2004), explanations that suggest that
the individual is not responsible for the success (i.e., the cause of success is external
to the individual) are likely to be hurtful. This expectation fits the fact that locus is
known to affect self-esteem and related emotions (Weiner 1986) as self esteem can be
considered to be a subjective estimate of relational evaluation (Buckley et al. 2004,
Leary and Downs 1995,Leary et al. 1995). Consequently, all else being equal, feed-
back from another person that explains a success by referring to luck or help from
others devalues the achiever’s worth and hence should elicit more hurt feelings than
feedback that explains the same outcome by referring to ability or effort.
The causal dimensions of stability and controllability are also expected to play a role
in inducing hurt feelings. In general, causal stability is associated with expectations
for the future (Weiner 1986). Accordingly, if an explanation for a success suggests
that it was caused by a factor that is unstable, and thus not characteristic of the self,
this implies that further success is not necessarily to be expected as the current one
might have been the result of chance. Such explanations are expected to be particularly
hurtful because they extend the devaluation of the person into the future.
As regards the controllability of the cause conveyed in the explanation, ascription
of a cause controllable by another individual should entrain more intense hurt feelings.
This, because controllability in this context implies that one’s fate is in the hands of
others. If, in addition, such a cause is stable, i.e., success will always be achieved this
way, this presents a particularly strong devaluation of the person. Thus, the ascription
of success to an external, controllable and stable cause should be most hurtful, as
it suggests not only lack of worth (due to dependence on others) in the present but
predicts that future success will also be the fruit of a similar cause.
In sum, both stability and controllability are expected to affect hurt feelings in
interaction with locus. Furthermore, based on the considerations outlined above, it is
123
262 S. Hareli, U. Hess
expected that the most hurtful message would be a message ascribing the success to
an external, stable and controllable cause.
As mentioned above, hurt feelings are usually accompanied by other emotions
such as anger, shame and guilt. We predict that the extent to which these emotions
are elicited by the success feedback is also determined by the implied attributional
information. Anger is an emotion that is determined, among other things, by the degree
of damage inflicted by another person (Averill 1982,Frijda 1986,Ortony et al. 1988).
Hence it is expected that the more hurtful the message, the more anger will be elicited
as well. That is, the same underlying causal structure that leads to more intense hurt
will also lead to increased levels of anger.
Unlike anger, which is an emotion directed at another person, shame and guilt are
both self-directed negative emotions (Frijda 1986). It has been contended that shame
is elicited by a public characteristic of the self that is not under volitional control. In
contrast, guilt follows from a violation of a norm and is evoked by a self-judgment
of responsibility (see Izard 1977,Smith et al. 2002,Wicker et al. 1983). That is, as
Roseman et al. (1996) noted, both shame and guilt result from a self-caused outcome;
however, shame follows when the cause is one of character, whereas for guilt, the cause
is behavioral. Accordingly, it is expected that success feedback that implies external
and stable causes will lead to higher levels of shame—as it points to a lack of ability in
the person, whereas success feedback that implies personal responsibility and control
will elicit more guilt if it also implies that the success could have been achieved in
a way more flattering to the achiever had s/he wished it. Study 1 was planned to test
these predictions in the context of explanations given by a peer for one’s success at
school.
For this a vignette paradigm was chosen. Vignettes have been criticized due to their
low ecological validity (see for example, Parkinson and Manstead 1993, for a discus-
sion of this issue). As such, vignettes are not well suited to assess how people would
actually react in a given social situation. However, for the same reasons vignettes are an
excellent tool to assess the symbolic knowledge that people apply when judging social
interactions. Specifically, we wanted to assess the impact of the causal information
implied in success feedback. Yet any given actual interpersonal situation varies with
regard to a number of ancillary aspects and other social norms and rules may interact
with that information. For example, feedback from a peer perceived as envious of the
achiever (and who may be perceived as competitive as a result) may well be perceived
as less acceptable and less hurtful than feedback from a peer who seems caring (for
example, because s/he is considered as a good friend and a reliable person). However,
to a large extent vignettes allow for an abstraction from the specific characteristics of
the feedback provider as well as other situational factors and are hence the appropriate
tool for the assessment of the question at hand.
2 Study 1
The main goal of Study 1 was to test the predictions outlined above by examining the
role that the causal information underlying communicated explanations for success
play in inducing hurt feelings and the emotions of anger, guilt and shame.
123
Casual attribution and hurt feelings 263
Tab le 1 The causal structure of the explanations used in Study 1 and Study 2
Locus Stability Controllability Explanation
Internal Stable Uncontrollable High ability
Internal Stable Controllable Like always invested high effort
Internal Unstable Uncontrollable Sudden insight of an idea for the work
Internal Unstable Controllable Invested exceptional effort that time
External Stable Uncontrollable Always lucky in getting good grades with demanding lecturers
External Stable Controllable Like in all other cases, helped by T.A.
External Unstable Uncontrollable Luck in topic choice
External Unstable Controllable Helped by T.A.
2.1 Method
2.1.1 Participants
Participants were 116 (92 female, 24 male) undergraduate students enrolled in psy-
chology courses at different departments in the University of Haifa, and the Iezreel
Valley College, who participated in the experiment voluntarily during regular class
sessions.
2.1.2 Materials and procedure
Participants read one of eight variations of a scenario that described a situation in
which they were asked to imagine receiving a very high grade (90 out of 100) for a
paper submitted for an important course. The instructions further asked participants
to imagine that on the following day, with no third party present, they meet another
student from the same course in the corridor of the department and a discussion about
the paper ensues. The participant was asked to imagine telling the other student about
the grade and expressing bewilderment about how this grade came about. In response,
the peer offers one out of eight different explanations for the success. Finally, partici-
pants were instructed to assume that they felt that the provided explanation reflected
the true reason for the failure. Each explanation reflected a specific combination of one
level of each of the causal dimensions of locus, stability and controllability. Table 1
describes each explanation and the underlying causal dimensions. This resulted in a
between-subjects factorial design with the factors Locus (internal vs. external), Sta-
bility (stable vs. unstable) and Controllability (controllable vs. uncontrollable).
The following is an example of one variation on the scenario. The parts that var-
ied between experimental conditions to achieve the desired manipulations of type of
explanation appear in italics.
“You took a seminar that requires students to submit a paper to fulfill its require-
ments. Your paper received a very good grade (90). The next day, you happen
to meet in the corridor of the department a fellow student who also took this
course. During your conversation with the other student, you tell him/her about
your grade and that you wonder why you received such a good grade. In response,
the other student answers that you managed to prepare a good paper because you
invested exceptional effort that time.”
123
264 S. Hareli, U. Hess
Upon reading the description of the scenario, participants answered questions in-
tended to measure their reactions to the feedback by the other student. All ratings were
made on seven-point scales anchored with 1=“not at all” to 7=“to a large extent.”
2.1.3 Emotion ratings
In addition to providing ratings of the degree that their feelings would be hurt and they
would feel insulted, they rated expected shame and guilt, as well as levels of anger
towards the other student. Level of insult was assessed because in Hebrew the notion
of insult typically refers to the more general situation of hurt feelings.
2.1.4 Causal attributions
Questions on causal attributions served as manipulation checks. Participants were
asked, “To what extent is the cause the other student mentioned as the reason for your
grade … (a) located within you or within the environment?” (ranging from 1 =“located
within you” to 7=“located within the environment”—locus); (b) something that is
permanent or temporary?” (ranging from 1=“temporary” to 7=“permanent”—sta-
bility); (c) controllable or uncontrollable by you (control by self) and (d) controllable
or uncontrollable by another person (ranging from 1=“controllable” to 7=“not-con-
trollable”).
2.2 Results and discussion
2.2.1 Manipulation checks
To test the effectiveness of the causal dimension manipulations, a 2 (Locus)×2
(Stability)×2 (Controllability) Analysis of Variance was conducted. As expected, a
significant main effect emerged for each of the causal dimensions—locus, F(1,108)=
36.19, p<.001, partial η2=.25; stability, F(1,108)=71.28, p<.001, partial
η2=.40; and controllability, F(1,106)=6.76, p<.05, partial η2=.06. Overall,
external explanations (M=4.41, SD =1.66) were perceived to be more located
in the environment than internal ones (M=2.36, SD =1.46), stable explanations
were seen as more stable (M=4.92, SD =1.75) than unstable ones ( M=2.73,
SD =1.23), and controllable explanations (the sum rating of the person’s own control-
lability and other’s controllability) were perceived as more controllable (M=7.53,
SD =2.22) than uncontrollable explanations (M=8.75, SD =2.96). In addition,
a significant interaction between stability and controllability emerged for perceived
locus, F(1,108)=4.71, p<.05, partial η2=.04; however, post-hoc tests ( p<.05)
failed to find any significant differences between conditions.
For perceived stability, a main effect for locus, F(1,108)=14.85, p<.001,
partial η2=.12, emerged such that internal causes (M=4.41, SD =1.87) were
perceived as somewhat more stable than external ones (M=3.35, SD =1.74).
In addition a significant interaction between locus and stability, F(1,108)=4.90,
p<.05, partial η2=.04, suggested that overall, the internal and stable causes
123
Casual attribution and hurt feelings 265
(M=5.76, SD =1.18) were perceived as most stable, followed by external, stable
causes (M=4,13, SD =1.84). As intended, the two unstable causes were perceived
as significantly less stable than the two stable causes, and did not differ from one
another (M=2.52, SD =1.15; and, M=2.96, SD =1.29, for external, unstable
and internal, unstable, respectively). Thus, as expected stable causes were indeed per-
ceived as more stable than unstable causes. In sum, these data generally confirm the
validity of the experimental manipulations.
2.2.2 Hypotheses testing
As the two scales asking about level of hurt and level of feeling insulted correlated
very highly (r=.93) they were combined into one variable, which we will refer to as
hurt feelings. A 2 (Locus)×2 (Stability) by 2 (Controllability) analysis of Variance
was conducted to assess the relation between the causal dimensions and hurt feel-
ings. As predicted, a significant main effect of Locus, F(1,107)=30.34, p<.001,
η2=.22, emerged, such that explanations that included an internal cause for success
were seen as less hurtful (M=2.71, SD =1.74) than explanations including an
external cause (M=4.22, SD =1.53). A main effect of Stability also emerged,
F(1,107)=19.16, p<.001, η2=.15, such that explanations suggesting stable
causes were perceived as less hurtful (M=2.97, SD =1.90) than those suggesting
unstable ones (M=4.04, SD =1.51). In addition, a two-way interaction between
locus and stability, F(1,107)=13.18, p<.001, partial η2=.11, as well as
between stability and controllability, F(1,107)=13.09, p<.001, partial η2=.11
emerged. Post-hoc tests indicated that explanations suggesting internal and stable
causes (M=1.74; SD =1.15) were significantly less hurtful than all other expla-
nations, which did not differ significantly from one another (M=3.76; SD =1.67;
M=4.15; SD =1.73, and, M=4.29; SD =1.32; for internal unstable, exter-
nal stable and external unstable, respectively). Further, uncontrollable unstable causes
(M=4.53; SD =1.45) were significantly more hurtful than all other explana-
tions, which did not differ significantly from one another (M=3.36; SD =1.87;
M=3.50; SD =1.51, and, M=2.46; SD =1.85; for controllable stable, control-
lable unstable and uncontrollable stable, respectively).
In sum, the present findings suggest that an explanation that ascribes success to
something that is either external to the achiever or internal but uncharacteristic of him
or her (i.e., help or having exceptionally invested hard work) rather than to the person’s
own stable characteristics (ability or being a hard worker), are perceived as hurtful.
Explanations that describe success as stemming from causes that seem random (luck
or sudden insight) are even more hurtful. This latter finding may be due to the fact that
such a success appears to be very detached from the achiever, that is, not at all linked
to his or her abilities and actions.
Thus, contrary to our expectations these findings suggest that it is not necessary
for the communicated causes for success to describe an external cause for success—
indicating that the achiever did not contribute significantly to the attainment of the
achievement—and thus to hurt his or her feelings. Uncontrollable and unstable causes
seem to have a similar impact on such feelings regardless of their locus.
123
266 S. Hareli, U. Hess
Tab le 2 F,p,andη2for the causal dimensions underlying observer’s communicated explanations for
success for anger, guilt and shame: Study 1
Causal dimension Anger F(1,108)η
2Guilt F(1,108)η
2Shame F(1,108)η
2
Locus Internal M2.63 15.80∗∗ .13 1.66 8.89∗.08 1.91 3.69 .03
SD 1.77 1.10 1.42 (p=.057)
External M3.78 2.38 2.40
SD 1.64 1.42 1.46
Stability Stable M2.78 12.08∗.10 2.05 .002 .0 2.07 1.06 .01
SD 1.81 1.42 1.44
Unstable M3.70 2.02 2.27
SD 1.66 1.21 1.48
Controllability Controllable M3.00 2.33 .02 2.07 .06 .00 2.20 .08 .004
SD 1.70 1.39 1 1.36
Uncontrollable M3.46 2.00 2.13
SD 1.88 1.25 1.56
Note.N=116; Higher values indicate a greater degree of that variable. ∗p<.01; ∗∗ p<.001
As mentioned above, causal attributions for success should also impact on other
emotions. As Table 2indicates, a significant main effect of locus emerged for anger,
guilt (marginally) for shame, with higher levels of each of these emotions for exter-
nal causes. Further, a main effect of stability emerged for anger, such that unstable
causes led to higher levels of anger. These main effects were qualified by significant
interactions.
For anger, the expected two-way interaction between locus and stability,
F(1,108)=8.67, p<.01, partial η2=.07, emerged, such that internal and stable
causes (M=1.79; SD =1.21) induced less anger than any of the remaining condi-
tions, which did not differ significantly (M=3.52; SD =1.89; M=3.71 SD =
1.81, and, M=3.86; SD =1.46; for internal, unstable, external stable and external
unstable, respectively). A two-way interaction was also found between stability and
control, F(1,108)=9.34, p<.01, partial η2=.08; uncontrollable and unstable
causes (M=4.31; SD =1.39) led to the significantly more anger than any of the
other causes, which did not differ significantly (M=2.56; SD =1.93; M=2.97;
SD =1.72, and, M=3.04; SD =1.70; for uncontrollable, stable, controllable,
stable and controllable, unstable, respectively). This effect was further qualified by a
significant three-way interaction, F(1,108)=6.46, p<.05, η2=.06. As is appar-
ent from Table 3, an explanation ascribing success to a sudden insight or luck led to
the highest levels of anger. Temporary effort led to an intermediate level of anger that
was no different than that of internal, stable causes (permanent effort and ability), but
also no different than the anger arising in the context of external causes. This again
indicates that success ascribed to factors that seem random is hurtful. Overall, these
findings replicate the picture that emerged in the context of hurt feelings. Thus, as
found by Hareli and Hess (in press) for failure events, hurt feelings are closely and
positively associated with anger for success feedback as well.
Another significant three-way interaction emerged for shame, F(1,108)=7.16,
p<.01, η2=.06. As is apparent from Table 3, an explanation ascribing success to
a sudden insight led to higher degrees of shame than did an explanation that ascribed
success to ability. There were no other differences between explanations in terms of
123
Casual attribution and hurt feelings 267
Tab le 3 Means and standard deviations for anger and shame ratings as a function of the causal dimensions
underlying the explanation: Study 1
Variable Causal dimension Explanation
Locus Int. Int. Int. Int. Ext. Ext. Ext. Ext.
Stability Unsta. Unsta. Sta. Sta. Unsta. Unsta. Sta. Sta.
Controllability Uncont. Cont. Uncont. Cont. Uncont. Cont. Uncont. Cont.
Anger M4.77c2.36ab 1.38a2.13a3.94c3.77bc 3.64bc 3.76bc
SD 1.30 1.55 .96 1.31 1.39 1.59 1.99 1.72
Shame M2.92b1.57ab 1.15a2.00ab 2.00ab 2.69ab 2.43ab 2.53ab
SD 1.85 .85 .38 1.55 1.46 1.38 1.70 1.36
Note.N=116; Int.=Internal, Ext.=External, Unsta. =Unstable, Sta. =Stable, Uncont.=Uncontrolla-
ble, Cont.=Controllable; Subscripts based on post-hoc tests at p<.05. Higher numbers represent greater
level of hurt feelings. Numbers with different subscripts differ at p<.05. For anger, the internal, unstable,
uncontrollable explanation is marginally significantly different than all other explanations (p=.54)
level of shame that they induced. Again, this finding implies the undesirability of
explaining success as resulting from a random factor such as sudden insight. How-
ever, overall, shame did not vary a lot across conditions indicating that in the context
of positive achievement, shame is of low likelihood even when the achiever did not
contribute much to bring about the success.
In this study, participants were instructed to assume that the explanation provided
by the other student was valid. However, people do also provide explanations that
convey invalid information and this may even lead to higher levels of hurt feelings,
especially if the perceiver is aware that the communicator purposefully presents a
wrong explanation (Hareli and Weiner 2002). Under such circumstances the achiever
is likely to assume that the main intention of the speaker was in fact to hurt his or her
feelings and intentional hurtful acts lead to higher degrees of hurtful feelings (Leary
et al. 1998;Vangelisti and Young 2000). A similar effect should be obtained for anger.
However, in the context of guilt and shame, the opposite effect is expected, that is, true
causes should elicit higher levels of shame and guilt as these emotions are closer to the
person’s character and actions (Lewis 1971) and less to their relationships with others.
However, Hareli and Hess (in press) found that the validity of the communication had
a very limited effect on the emotional reaction of the achiever. The goal of Study 2
was to asses the influence of the validity of the explanation with regard to hurt feelings
and related emotions.
3 Study 2
3.1 Method
3.1.1 Participants
Participants were 214 (141 female, 73 male) undergraduate students enrolled in
psychology courses at different departments in the University of Haifa who partici-
pated in the experiment voluntarily during regular class sessions.
123
268 S. Hareli, U. Hess
3.1.2 Materials and procedure
To simplify the design and given that the more clearly hurtful effects emerged in
the context of external explanations, only the four scenarios used in Study 1 that
described an external explanation for the success (four last explanations appearing in
the forth column of Table 1) were used. There were two versions for each explana-
tion. In one version the participants were instructed to imagine that the explanation
described the real reason for the success, whereas the other version mentioned that the
participant is aware to the fact that the friend knows for sure that the real reason for
success was high ability. High ability was chosen as the real reason because it poten-
tially allows for the greatest discrepancy between the true reason for the success and
the communicated reason in terms of distancing the person from the cause of the
achievement when the explanation is false. This formed a between-subjects
factorial design combining the validity of the explanation and the causal dimensions
of stability and controllability in a 2 (Validity: valid vs. invalid)×2 (Stability: stable
vs. unstable)×2 (Controllability: controllable vs. uncontrollable) between subjects
design. The same emotion ratings scales were used as in Study 1. Because the vignettes
were the ones used in Study 1, manipulation checks were not included in this study.
3.2 Results and discussion
As in Study 1, perceptions of insult and hurt feelings correlated highly (r=.83) and
were combined into one measure of hurt feelings. A 2 (Validity) ×2 (Stability) ×(Con-
trollability) analysis of variance was conducted to assess the effect of the validity of the
explanation and the causal dimensions of stability and controllability on hurt feelings.
For validity, only a marginally significant main effect emerged, F(1,203)=2.82,
p=.10, η2=.01, such that true explanations were somewhat more hurtful than
false explanations (M=4.08, SD =1.79 vs. M=3.66, SD =1.78). A three-way
interaction between validity, stability and controllability, F(1,203)=4.36, p<.05,
partial η2=.02 also emerged, however, post-hoc tests failed to find any significant
differences between explanations. Thus, the validity of the explanation did not play
a significant role in affecting the extent to which the feelings of the achiever were
expected to be hurt. This replicates previous findings by Hareli and colleagues (Hareli
and Hess in press,Hareli et al. 2006) in the context of feedback for school achieve-
ments. For anger as well no significant effects emerged.
In contrast to the minor role that validity played for hurt feelings and anger, for
guilt, F(1,204)=9.97, p<.01, partial η2=.05, and shame, F(1,203)=25.23,
p<.001, partial η2=.11, main effects for validity emerged such that, as pre-
dicted, a valid explanation was expected to elicit higher levels of both emotions
(guilt: M=2.54, SD =1.77; shame: M=2.88, SD =1.93) than a false one
(guilt: M=1.89, SD =1.83, shame: M=1.78, SD =1.26). In addition, a
main effect for stability emerged for shame, F(1,214)=4.86, p<.05, partial
η2=.02, such that stable explanations increased shame ratings in comparison to
unstable ones (M=2.56, SD =1.74 vs. M=2.08, SD =1.65). This replicates
the findings of Study 1 and is congruent with the view that shame is related to an
123
Casual attribution and hurt feelings 269
undesirable characteristic of the individual (Lewis 1971). No other significant effects
emerged.
4 General discussion
Attribution theory is an important theoretical framework allowing to explain how
achievers react to their educational achievements as a function of how they explain
to themselves the reasons for their outcomes. However, attribution theory can also
explain achievers’ reactions in the context of their achievements when this attribu-
tional information is provided by others (Hareli and Hess in press,Hareli and Weiner
2002). The goal of the present study was to use attribution theory as an explanatory
tool for the way achievers react emotionally to peers’ explanations concerning the
reasons for their success. In particular, we were interested to examine the conditions
under which achievers may experience hurt feelings, anger, guilt and shame due to a
relational devaluation implied in attributions for success provided by peers. Guided by
Weiner’s attribution theory (1985,1986) and following recent research by Hareli and
Hess (in press) we predicted that the specific pattern of causal properties underlying
such attributions together with their validity are important determinants of these reac-
tions. As regards hurt feelings and anger we predicted that explanations that distance
the person from the cause of the success, that is, external causes and causes con-
trollable by other, will be most hurtful and angering. Overall these predictions were
confirmed. In particular, not only external causes but also explanations that ascribe
success to random factors (such as sudden insight or luck)—regardless of the locus of
these causes—increased hurt feelings and anger as these later causes also undermine
the person’s potential contribution to the success. Thus, as for hurt feelings and anger
arising in response to explanations for failure (Hareli and Hess in press), in the context
of success too, the causal structure of the communicated cause explains the extent to
which the communication elicited hurt feelings and anger.
Further, in both studies, the pattern of anger and hurt feelings was similar. Interest-
ingly, as found by Hareli and Hess (in press) in the context of failure, the level of anger
and hurt feelings was not strongly determined by the validity of the explanation—and
this despite the fact that providing knowingly an invalid external explanation such as
luck for a success which is actually due to high ability may be considered an act of
ill-will. Rather, it seems that both for hurt feelings and anger what matters more is
what others say rather than whether the messages convey true or false information.
This fits findings in other domains that involve reactions to social information, for
example, social judgments and social emotions, and that indicate that such effects are
often relatively insensitive to validity of the information. For example, inferences of
arrogance are determined mostly by the way people explain their achievements with
little regard to whether the achievement was in fact major (Hareli and Weiner 2000)
or whether the explanations given for the achievement reflect the true reason for the
success (Hareli et al. 2006).
In addition to hurt feelings and anger, guilt and shame were also assessed. We pre-
dicted that both emotions would be determined by the causal dimensions underlying
communicated explanations for success as well as by the validity of the information.
123
270 S. Hareli, U. Hess
As expected, explanations describing external causes led to slightly higher levels of
shame than did explanations describing internal causes. Also, stable causes increased
shame (Study 2 only) as well as guilt feelings. Further, valid explanations increased
the experience of guilt and shame, yet there were no interactions between the validity
of the explanation and its underlying causal dimensions for both emotions.
Overall, these studies provide reasonable evidence for the applicability of Weiner’s
attribution theory of emotion and motivation (1985,1986) to explain the link between
communicated explanations for success, hurt feelings and related emotions. Previ-
ous research emphasized the important role of relational devaluation for hurt feelings
as well as which other emotional reactions are likely to arise in such contexts (Leary
et al. 1998,Leary and Springer 2001,Vangelisti 1994), the present research extends
this prior work and, importantly, specifies the type of information that will lead a per-
son to feel devalued. Most importantly, this research provides support for the idea that
attribution theory is useful in explaining not only how achievers react to their achieve-
ments as a function of their understanding of the reasons for their success, but also the
way observers’ feedback shapes such reactions. This is in particular important when
considering achievement motivation as shaped not only by ones’ self generated cogni-
tions but also by the information that others (e.g., peers, teachers and parents) provide
(Hareli and Weiner 2002). School settings provide ample opportunities for students
to put their abilities and skills to test and understand their role in their achievements.
Often this information is provided by others and when this occurs achievers respond
to such feedback in a way that may further determine how they will behave (Hareli
and Weiner 2002). The present research exemplifies that attribution theory is a useful
tool for understanding this understudied aspect of achievement motivation.
Although the present research gathered reasonable evidence for the important role
that causal dimensions underlying communicated explanations for success has for
inducing hurt feelings and related emotions, further research is required to better assess
the generality of these findings and the limiting factors. For example, in the present
study the importance of the achievement was controlled, yet level of importance may
be a moderating factor in so far as emotions arise in the context of personally significant
events and that their intensity is positively correlated with the subjective significance
of these events (Frijda 1986,1988). Likewise the quality of the relationship with the
other person may also be pertinent. People tend to interpret criticism voiced by others
as more constructive and hence as less hurtful if their relations with the others are
satisfying (Bradbury and Fincham 1990). However, if such explanations are known
to be false and there are reasons to suspect that the other is aware of this, then the
quality of relationships is expected to have the reverse effect. That is, the closer the
other person is the more hurtful such an invalid explanation should be. In such a case,
the message is expected to be more hurtful as the importance of the achievement for
the individual increases. Finally, the presence of a third person can further affect that
achiever’s reaction to the explanation. Yet, all of these factors are expected to work in
interaction with the causal dimensions underlying the communicated explanation.
The present research used a vignette approach. By using vignettes and asking par-
ticipants to imagine the relevant emotional situation, one cannot be assured that the
results represent people’s actual emotional reactions. Recent discussions exploring
the question of the extent to which self reports made in the context of hypothetical
123
Casual attribution and hurt feelings 271
scenarios reflect actual experience suggest that such reports are most likely to represent
participants’ beliefs about such situations rather than actual experiences (Robinson and
Clore 2002). Yet, under many circumstances the discrepancy between actual experi-
ence and prospective self report is a matter of degree rather than of the pattern of
response. That is, prospective self-reports such as those used here, are exaggerations
or underestimations of the actual reaction (Gilbert et al. 1998,Robinson and Clore
2002,Wilson et al. 2000). In addition, the beliefs that people have about hurtful events
are highly pertinent in their own right as hurtful events have long-term consequences
(Leary and Springer 2001) and the long-term consequences of emotional events tend
to be more strongly influenced by the beliefs that people have about these events than
by the actual emotions experienced in response to them (Robinson and Clore 2002).
Finally, the present research centered on communicated explanations for a specific
event, yet, hurtful comments can be made without reference to specific events. We
suggest that such comments will be perceived as hurtful for the same reasons that
comments in the present research were thus perceived, that is, because they distance
the person from achievement or conversely associate the person with failure (Hareli
and Hess in press). Overall, we believe that the present research provides a valuable
framework for the study of the determinants of the emotional consequences of hurtful
communication.
References
Averill, J. R. (1982). Anger and aggression: An essay on emotion. New York: Springer-Verlag.
Bar-Tal, D. (1978). Attributional analysis of achievement-related behavior. Review of Educational
Research, 48, 259–271.
Bradbury, T. N., & Fincham, F. D. (1990). Attributions in marriage: Review and critique. Psychological
Bulletin, 107, 3–23.
Buckley, K. E., Winkel, R. E., & Leary, M. R. (2004). Reactions to acceptance and rejection: Effects of
level and sequence of relational evaluation. Journal of Experimental Social Psychology, 40, 14–28.
Forsyth, D. R. (1986). An attributional analysis of students’ reactions to success and failure. In: R. S.
Feldman (Ed.), The social psychology of education: Current research and theory. New York: Cambridge
University Press.
Frijda, N. H. (1986). The emotions. Cambridge, UK: Cambridge University Press.
Frijda, N. H. (1988). The laws of emotion. American Psychologist, 43, 349–358.
Gabriel, Y. (1998). An introduction to the social psychology of insults in organizations. Human Relations,
51, 1329–1354.
Gilbert, D. T., Pinel, E. C., Wilson, T. D., Blumberg, S. J., & Wheatley, T. P. (1998). Immune neglect:
A source of durability bias in affective forecasting. Journal of Personality and Social Psychology, 75,
617–638.
Hareli, S., & Hess, U. (in press). The role of causal attribution in hurt feelings and related social emotions
elicited in reaction to other’s feedback about failure. Cognition and Emotion.
Hareli, S., & Weiner, B. (2000). Accounts for success as determinants of perceived arrogance and modesty.
Motivation and Emotion, 17, 215–236.
Hareli, S., & Weiner, B. (2002). Social emotions and personality inferences: A scaffold for a new research
direction in the study of achievement motivation. Educational Psychologist, 37, 183–193.
Hareli, S., Weiner, B., & Yee, J. (2006). Honesty does not always pay – the role of honesty of accounts for
success in inferences of modesty and arrogance. Social Psychology of Education, 9, 119–138.
Izard, C. E. (1977). Human emotions. New York: Plenum Press.
Leary, M. R., & Downs, D. L. (1995). Interpersonal functions of the self-esteem motive: The self-esteem
system as a sociometer. In M. Kernis (Ed.), Efficacy, agency, and self-esteem (pp. 123–144). New York:
Plenum.
123
272 S. Hareli, U. Hess
Leary, M. R., & Springer, C. (2001). Hurt feelings: The neglected emotion. In R. M. Kowalski (Ed.), Aversive
behaviors and relational transgressions. Washington, DC: American Psychological Association.
Leary, M. R., Springer, C., Negel, L., Ansell, E., & Evans, K. (1998). The causes, phenomenology, and
consequences of hurt feelings. Journal of Personality and Social Psychology, 74, 1225–1237.
Leary, M. R., Tambor, E. S., Terdal, S. K., & Downs, D. L. (1995). Self-esteem as an interpersonal monitor:
The sociometer hypothesis. Journal of Personality and Social Psychology, 68, 518–530.
Lewis, H. B. (1971). Shame and guilt in neurosis. New York: International Universities Press.
Ortony, A., Clore, G. L., & Collins, A. (1988). The cognitive structure of emotions. New-York: Cambridge
University Press.
Parkinson, B., & Manstead, A. S. R. (1993). Making sense of emotion in stories and social life. Cognition
and Emotion, 7, 295–323.
Robinson, M. D., & Clore, G. L. (2002). Belief and feeling: An accessibility model of emotional self-report.
Psychological Bulletin, 128, 934–960.
Roseman, I. J., Antoniou, A. A., & Jose, P. E. (1996). Appraisal determinants of emotions: Constructing a
more accurate and comprehensive theory. Cognition and Emotion, 10, 241–277.
Smith, R. H. Webster, J. M. Parrott, W. G., & Eyre, H. L. (2002). The role of public exposure in moral and
nonmoral shame and guilt. Journal of Personality and Social Psychology, 83, 138–159.
Vangelisti, A. L. (1994). Messages that hurt. In W. R. Cupach & B. Spitzberg (Eds.), Thedarksideof
interpersonal communication (pp. 53–82). Hillsdale, NJ: Erlbaum.
Vangelisti, A. L., & Young, S. L. (2000). When words hurt: The effects of perceived intentionality on
interpersonal relationships. Journal of Social and Personal Relationships, 17, 393–424.
Weiner, B. (1976). An attributional approach for educational psychology. Review of Research in
Education, 4, 179–209.
Weiner, B. (1985). Spontaneous causal thinking. Psychological Bulletin, 97, 74–84.
Weiner, B. (1986). An attributional theory of motivation and emotion. New York: Springer-Verlag.
Weiner, B. (1995). Judgments of responsibility: A foundation for a theory of social conduct.NewYork:
Guilford.
Wicker, F. W., Payne, G. C., & Morgan, R. D. (1983). Participant description of guilt and shame. Motivation
and Emotion, 7, 25–39.
Wilson, T. D., Wheatley, T. P., Meyers, J. M., Gilbert, D. T., & Axsom, D. (2000). Focalism: A source of
durability bias in affective forecasting. Journal of Personality and Social Psychology, 78, 821–836.
Young, S. L. (2004). Factors that influence recipients’ appraisals of hurtful communication. Journal of
Social and Personal Relationships, 21, 291–303.
Author Biographies
Shlomo Hareli is a Lecturer of Psychology at the Graduate School of Management, University of Haifa.
His areas of interest include psychology of emotions, attribution processes, social emotions and social
judgments as these are related to social settings and organizations.
Ursula Hess is a Professor of Psychology at the University of Quebec at Montreal. She received her Ph.D.
in Psychology from Dartmouth College and she was a Fellow at the Association for Psychological Science.
Her current research is in the area of emotion psychology—in particular, the communication of affect with
an emphasis on two main lines of research: psychophysiological measures and the role of social influences
in the encoding and decoding of emotional expressions.
123
