Content uploaded by Nickola C Overall
Author content
All content in this area was uploaded by Nickola C Overall on Jul 28, 2014
Content may be subject to copyright.
Experiences and Interpersonal Consequences of Hurt Feelings and Anger
Edward P. Lemay, Jr.
University of New Hampshire Nickola C. Overall
University of Auckland
Margaret S. Clark
Yale University
This research compared the experiences and consequences of hurt feelings and anger in 3 retrospective
studies (Studies 1a, 1b, and 2), a dyadic daily diary study (Study 3), and a dyadic behavioral observation
study (Study 4). Although victims felt both hurt and angry in response to perpetrators’ behaviors that
signaled relational devaluation (Studies 1–4), hurt and anger differed in terms of victims’ subjective
experiences and behaviors, perpetrators’ responses, and relationship consequences. Hurt was character-
ized by the experience of commitment, dependence, and vulnerability; goals to restore the perpetrator’s
acceptance; and constructive behavior. Moreover, victims’ hurt was associated with perpetrators evalu-
ating victims and victims’ commitment more positively, with perpetrators’ feelings of guilt and empathy
and with perpetrators’ constructive responses. Hurt also had positive consequences for relationships. In
contrast, victims’ anger was generally independent of commitment and characterized by the experience
of control, invulnerability, and low dependence; goals to change perpetrators’ behavior; and victims’
destructive behavior. Furthermore, victims’ anger was associated with perpetrators perceiving victims to
be less committed and elicited reciprocated anger and destructive behaviors from perpetrators. These
findings suggest that despite relational devaluation being a cause of both hurt and anger, these feelings
have distinct social functions. Hurt reflects a desire to maintain interpersonal connection and repair
relationships, which will often successfully elicit repair attempts by perpetrators, whereas anger reflects
a desire to control others via antagonistic destructive behaviors, which exacerbate interpersonal
difficulties.
Keywords: hurt feelings, anger, emotion, relational devaluation
The emotional experience colloquially referred to as “hurt feel-
ings” is a form of social pain that is frequently experienced when
people perceive rejection or devaluation by someone (Feeney,
2004,2005;Leary, Springer, Negel, Ansell, & Evans, 1998;Mac-
Donald & Leary, 2005;Vangelisti, Young, Carpenter-Theune, &
Alexander, 2005).This pain is associated with general emotional
distress (Leary et al., 1998), activates the same brain regions
involved in physical pain (Eisenberger, Lieberman, & Williams,
2003), and rivals physical pain in its unpleasantness (Chen, Wil-
liams, Fitness, & Newton, 2008). In the current research, we
examined the social functions and consequences of hurt feelings.
Consistent with the view that negative emotions can have positive
functions and consequences (e.g., Algoe, Haidt, & Gable, 2008;
Bartlett & DeSteno, 2006;Fischer & Manstead, 2008;Graham,
Huang, Clark, & Helgeson, 2008;Keltner & Haidt, 1999;Salovey,
1991;Van Kleef, 2010), we predicted that hurt would often result
in relationship repair following these experiences of devaluation.
We contrasted the social functions and consequences of hurt
with those of anger. Anger also is elicited by relational devaluation
(see Leary, Twenge, & Quinlivan, 2006) and often blends or
co-occurs with hurt. Yet, we posited that anger is a fundamentally
different and typically more interpersonally destructive experi-
ence. Hurt and anger may have been conflated in prior research
and theorizing, and there is little empirical evidence on the differ-
ences between them. The present research was an attempt to fill
this void.
Our objective was to test whether hurt and anger have different
social functions and consequences. At the broadest level, we
predicted that hurt functions to repair relationships following in-
terpersonal threats, whereas anger exacerbates relationship diffi-
culties. Table 1 details the more specific underlying experiences
and consequences that we predicted would be associated with hurt
feelings and anger. In four studies, involving retrospective ques-
tionnaires, daily reports by both members of a couple, and dyadic
behavioral observation, we tested whether the experience of hurt
and anger differed in appraisals of dependence and vulnerability
versus personal control, goals to restore acceptance versus control
perpetrator behavior, and subsequent constructive versus destruc-
tive behaviors. We also tested how perpetrators respond to victims’
hurt versus anger, including their relational inferences, emotions,
and behaviors. These dimensions are important criteria for estab-
This article was published Online First September 17, 2012.
Edward P. Lemay, Jr., Department of Psychology, University of New
Hampshire; Nickola C. Overall, Department of Psychology, University of
Auckland, Auckland, New Zealand; Margaret S. Clark, Department of
Psychology, Yale University.
This research was supported by National Science Foundation research
Grant BCS 1145349.
Correspondence concerning this article should be addressed to Edward
P. Lemay, Jr., Department of Psychology, Conant Hall, University of New
Hampshire, Durham, NH 03824. E-mail: edward.lemay@unh.edu
Journal of Personality and Social Psychology © 2012 American Psychological Association
2012, Vol. 103, No. 6, 982–1006 0022-3514/12/$12.00 DOI: 10.1037/a0030064
982
lishing that hurt and anger have different social functions (see
Clark, Fitness, & Brissette, 2001;Fischer & Manstead, 2008;
Frijda, Kuipers, & ter Schure, 1989;Frijda & Parrott, 2011;
Roseman, Wiest, & Swartz, 1994;Van Kleef, 2010).
The Experience and Social Functions of Hurt Feelings
Victims’ Experiences and Responses
The top left side of Table 1 outlines our predictions regarding
victims’ experiences and responses when feeling hurt in response
to relational devaluation. The expectation is that hurt involves
subjective appraisals of dependence and vulnerability. The capac-
ity to feel hurt implies the victim is vulnerable to psychological
pain at the hands of the perpetrator, and this vulnerability likely
comes about because the victim desires or needs a relationship
with the perpetrator. Indeed, several scholars have proposed that
the victim’s dependence on a perpetrator for support, security, and
protection creates the potential to feel hurt by relational devalua-
tion (see Leary et al., 1998;Sanford & Rowatt, 2004;Shaver,
Mikulincer, Lavy, & Cassidy, 2009;Vangelisti & Young, 2000).
Desiring or needing a close relationship creates dependence and
vulnerability because the fulfillment of this desire or need depends
on the continued investment and relational valuing of the partner
(Kelley et al., 2003;Kelley & Thibaut, 1978). When valued
partners are rejecting or hostile or betray the self, they signal that
they do not value the relationship as highly, which should heighten
feelings of vulnerability and pain. Themes of personal injury in
reports of hurtful events, and associated feelings such as damaged,
heartbroken, and shattered (Feeney, 2005), indicate that feelings
of vulnerability are poignant when feeling hurt.
Moreover, if dependence is a central component of feeling hurt
in response to relational devaluation, victims’ who highly value
and are committed to their relationships should be especially likely
to feel hurt in these contexts. Relationship commitment summa-
rizes multiple sources of the subjective want and need that should
create a vulnerability to hurt feelings, including feelings of satis-
faction, the belief that the relationship is more desirable than
alternative relationships and activities, and the investment of re-
sources that would render relationship dissolution especially dif-
ficult (see Rusbult & Buunk, 1993). Consistent with this predic-
tion, some prior research has found associations between greater
relationship “quality” (an average of satisfaction and commitment)
and hurt feelings in retrospective accounts of hurtful events (Bach-
man & Guerrero, 2006; see also Leary & Leder, 2009).
The subjective sense of dependence that we predicted would be
associated with hurt feelings should motivate victims to relieve the
psychological pain of hurt feelings by restoring a sense of being
valued by the perpetrator. Thus, we predicted that hurt feelings
would trigger the goal to restore acceptance (see row 1, column 2,
Table 1) and associated relationship-constructive and repairing
behaviors (see row 1, column 3). Although relational devaluation
often triggers aggressive and destructive behavior (Leary et al.,
2006), feeling hurt may mitigate such destructive behavior, as this
behavior would only frustrate the goal to restore perpetrators’
acceptance and would put victims at risk for more severe and
hurtful rejection. Given their sense of vulnerability and depen-
dence, hurt victims should see this as an undesirable strategy.
Moreover, hurt victims appear to lack the sense of certainty and
control (see Fitness & Warburton, 2009) that seems to promote
aggressive responses (see Lerner & Tiedens, 2006). Instead, hurt
victims should be more likely to respond in constructive ways that
could restore the perpetrators’ valuing of the relationship, partic-
ularly if they perceive such action could adequately restore the
perpetrators’ valuing of the relationship and will not be met with
additional rejection (see also Murray, Holmes, & Collins, 2006).
Some findings do suggest that hurt victims respond in a construc-
tive manner (Bachman & Guerrero, 2006;Vangelisti & Crumley,
1998), and other research suggests that the desire and subjective
need for a relationship, which we have argued characterizes hurt,
motivates constructive responses to interpersonal threats (Rusbult,
Verette, Whitney, Slovik, & Lipkus, 1991). However, other find-
ings suggest that hurt motivates destructive responses (e.g., Bach-
man & Guerrero, 2006;Caughlin, Scott, & Miller, 2009;Leary et
al., 1998;MacDonald & Leary, 2005;Vangelisti & Young, 2000).
We believe there are methodological factors that may explain these
findings, and we review them in this article.
Perpetrators’ Responses
The top right side of Table 1 outlines our predictions regarding
perpetrators’ responses to victims’ hurt. We predicted that when
perceiving victims’ hurt, perpetrators would: (a) more positively
evaluate victims and victims’ commitment to their relationship, (b)
feel empathic concern and guilt, and (c) behave in constructive
ways to alleviate victims’ pain. First, hurt may signal victims’
commitment to perpetrators. If the capacity to feel hurt arises
because victims value relationships with perpetrators, then perpe-
trators’ lay understanding of hurt should lead them to infer from
victims’ hurt that victims are committed to the relationship. This
communication of commitment may motivate perpetrators to re-
pair the relationship, consistent with findings that people enact
pro-relationship behaviors when they believe they are valued (Le-
Table 1
Theorized Experiences and Consequences of Hurt Versus Anger
Victim’s
emotion
Victim’s experiences and responses Perpetrator’s responses
Appraisals Goals Behaviors Cognitions Emotions Behaviors
Hurt Dependence and
vulnerability
Restore perpetrator’s
acceptance
Constructive Positive evaluations of victim
and victim’s commitment
Empathy and
guilt
Constructive
Anger Control and power Change/control perpetrator’s
behavior
Destructive Negative evaluations of
victim and victim’s
commitment
Anger Destructive
983
HURT FEELINGS AND ANGER
may & Clark, 2008;Murray et al., 2006;Reis, Clark, & Holmes,
2004;Wieselquist, Rusbult, Foster, & Agnew, 1999).
Second, because hurt conveys the hurt person’s distress and
need, it should also elicit the perpetrator’s empathic concern,
especially in communal relationships, and such empathic concern
should motivate helping behavior (Batson, Eklund, Chermok,
Hoyt, & Ortiz, 2007). Moreover, because perpetrators have caused
victims’ pain, perpetrators should feel guilt—a constructive emo-
tion that motivates relationship repair and usually arises when
people believe they have harmed a communal relationship partner
(Baumeister, Stillwell, & Heatherton, 1994;Tangney, Wagner,
Fletcher, & Gramzow, 1992). These emotional responses may
repair the relationship and alleviate victims’ hurt (Feeney, 2004;
Leary et al., 1998;Ohbuchi, Kameda, & Agarie, 1989;Vangelisti
& Sprague, 1998). Finally, victims’ hurt may serve as an implicit
request for reassurance and support (Sanford & Rowatt, 2004).
Hence, when perpetrators believe they have hurt victims, they
should feel motivated to alleviate their pain through expressing
care for them and the relationship. This should be especially
common when perpetrators and victims share communal relation-
ships in which members care for the other’s welfare (Clark &
Lemay, 2010;Clark, Ouellette, Powell, & Milberg, 1987). Indeed,
most hurtful events do appear to involve such relationships (Leary
et al., 1998).
Consistent with these predictions, people tend to believe that
sharing hurt feelings is a more intimacy-enhancing conflict-
resolution strategy than either calm, rational discussion or expres-
sions of anger (Frey, Holley, & L’Abate, 1979). In addition,
laboratory studies in which victims’ physical pain cues have been
manipulated have revealed that people become less aggressive
toward victims when victims seem to be suffering (Baron, 1971a,
1971b;Buss, 1966), perhaps because the pain cues elicit empathy
or guilt. We theorized that victims’ expressions of social pain
would similarly elicit perpetrators’ constructive behavior. How-
ever, other findings are inconsistent with these predictions. For
example, in a study of autobiographical narratives of hurtful inci-
dents (Leary et al., 1998), the frequency of narratives featuring a
perpetrator apologizing or asking for forgiveness was quite low,
and narratives featuring perpetrators’ destructive responses to vic-
tims’ hurt, such as doing nothing or acting as if they did not care,
were just as common. Moreover, in contrast to our broader view
that hurt facilitates relationship repair, narratives about hurtful
events have frequently featured descriptions of negative relation-
ship consequences (Leary et al., 1998). Again, we will consider
methodological factors that may explain why these prior findings
do not fit with our prediction that people usually respond in a more
constructive manner to hurting others.
The Experience and Social Functions of Anger
Victims’ Experiences and Responses
The bottom left side of Table 1 outlines our predictions regard-
ing victims’ experiences and responses when experiencing anger
in response to relational devaluation. Although, like hurt feelings,
anger is often elicited by relational devaluation, we believe that
anger is distinct from hurt on all of the other dimensions discussed
previously and outlined in Table 1. Rather than the dependence
and vulnerability that we predicted would be associated with hurt
feelings, anger is characterized by appraisals of certainty and
control (or power; Frijda et al., 1989;Mikulincer, 1994;Roseman,
Antoniou, & Jose, 1996;Smith & Ellsworth, 1985), and effects of
anger on risk perceptions and risky decision making suggest that
anger is associated with feeling invulnerable (Lerner, 2001; for a
review, see Lerner & Tiedens, 2006). Anger also encourages
reliance on relatively automatic, superficial, and heuristic thought
processes (Bodenhausen, Sheppard, & Kramer, 1994;Lerner,
Goldberg, & Tetlock, 1998;Tiedens & Linton, 2001), directs
attention to new angering events, and motivates quick action (for
a review, see Lerner & Tiedens, 2006). Hence, angry victims are
unlikely to keep in mind their own dependence on perpetrators or
to carefully consider the effects of their actions on the welfare of
the relationship.
Rather than restoring acceptance, the principal goal associated
with anger is to change the perpetrator’s behavior through coercion
or punishment (Averill, 1982;Fischer & Roseman, 2007). Such
change is unlikely to be done in a diplomatic manner that restores
the perpetrator’s acceptance. Anger is associated with a state of
action readiness that Frijda et al. (1989) described as moving
against. That is, anger is characterized by urges to change the
situation by fighting, harming, or conquering an opponent (see also
Roseman et al., 1994). In addition, the cognitive appraisals and
shifts in information processing accompanying anger may under-
mine the victim’s ability to attend to broader relationship concerns,
such as the perpetrator’s acceptance, and may instead cultivate a
myopic focus on fulfilling antagonistic urges. As a result of the
attendant antagonistic action tendencies, appraisals, and shifts in
cognitive processing, we expected that the experience of anger in
response to relational devaluation usually would motivate victims
to behave destructively, including blaming or criticizing the per-
petrator and engaging in other actions that harm the perpetrator.
Indeed, links between anger and such behavior are well docu-
mented (e.g., Averill, 1982;deRivera & Grinkis, 1986;Fincham &
Bradbury, 1992;Fischer & Roseman, 2007;Fitness & Fletcher,
1993;Frijda et al., 1989;Lerner & Tiedens, 2006;Roseman et al.,
1994;Shaver, Schwartz, Kirson, & O’Connor, 1987;Smith &
Lazarus, 1993), and anger mediates effects of social exclusion on
destructive behavior (Chow, Tiedens, & Govan, 2008). However,
prior research has not examined whether these experiences differ-
entially characterize hurt and anger.
Perpetrators’ Responses
Given the victims’ experiences and responses we have de-
scribed, in contrast to hurt feelings, victims’ anger should elimi-
nate or hinder perpetrators’ reparative responses. Rather than sig-
naling dependence and commitment, angry victims’ antagonistic
behaviors likely convey to perpetrators that victims do not care
about them or the relationship. Indeed, victims’ anger may warn
perpetrators that their welfare is in jeopardy. Such messages may
be conveyed even in the absence of aggressive behavior, as per-
petrators likely infer victims’ disapproval and aggressive motives
from their expressions of anger (Averill, 1982;Knutson, 1996;
Tiedens, 2001;Van Kleef et al., 2009), which should motivate
them to distance themselves from angry victims (Dunham, 2011;
Miller, Maner, & Becker, 2010). The felt threat and desire for
distance may limit perpetrators’ capacity to focus on angry vic-
tims’ welfare, reduce their empathic concern (see Mikulincer et al.,
984 LEMAY, OVERALL, AND CLARK
2001) and guilt (see Leith & Baumeister, 1998), and inhibit their
efforts to make amends. Because people infer high control and low
dependence from others’ anger (Knutson, 1996;Tiedens, 2001),
perpetrators are unlikely to perceive angry victims to be in need of
help. Instead, perpetrators may respond to angry victims’ destruc-
tive behavior, efforts to control, and apparent lack of concern for
affiliation by defensively reciprocating their negativity (see Aver-
ill, 1982;Baumeister, Stillwell, & Wotman, 1990;Levenson &
Gottman, 1983). Hence, we expected that perpetrators would re-
spond to victims’ anger in a less constructive and more destructive
way relative to their response to victims’ hurt. Prior research has
not examined this issue.
Why Our Predictions Are Inconsistent With Some
Prior Research on Hurt Feelings
Although we noted some findings that are consistent with our
view that hurt facilitates relationship repair and is not a destructive
emotional experience, we also noted findings that are inconsistent
with this view. Specifically, some findings indicate that hurt is
associated with victims’ destructive responses, that perpetrators
are not especially constructive in response to victims whom they
have hurt, and that hurtful events have negative relationship con-
sequences. Several factors may have contributed to this apparent
discrepancy.
First, many apparently negative effects of hurt feelings may be
negative effects of anger. Although we suggest they are qualita-
tively different experiences, victims who feel hurt also tend to
report anger (Fehr, Baldwin, Collins, Patterson, & Benditt, 1999;
Fine & Olson, 1997;Leary & Leder, 2009;MacDonald & Leary,
2005;Sanford & Rowatt, 2004). Both emotions are negative, and
both are elicited by others’ aversive behavior that signals relational
devaluation. These common features may contribute to a positive
correlation between hurt and anger. Moreover, in some situations,
victims may experience some features we predicted of hurt and
other features we predicted of anger. For example, when victims
strongly want to be accepted by the perpetrator and have the sense
of control to coerce change in the perpetrator’s aversive behaviors,
they may report both hurt and anger (and associated experiences
and behaviors in Table 1). In addition, reports of both hurt and
anger may occur when people shift from one emotion to the other
in a short time span. When they report globally on the event, they
may report both emotions. For instance, when the restoration of
acceptance does not seem forthcoming, hurt victims may protect
themselves from the pain of rejection by devaluing the relationship
with the perpetrator (see Murray et al., 2006). Accordingly, they
may give up on their goal to restore acceptance and instead pursue
the goal of punishing the perpetrator’s aversive behavior. We
expected that such a transition in motivational state would coincide
with a transition from hurt to anger. Alternatively, victims may
shift from anger to hurt after considering the deeper relational
meaning underlying an offensive remark or after contemplating
their own dependence on the perpetrator. Such a shift would be
consistent with findings suggesting that people often exhibit im-
mediate impulses to reciprocate destructive behavior but undergo
a transformation of motivation (Yovetich & Rusbult, 1994)in
which these impulses are inhibited after considering broader rela-
tionship concerns. Finally, although we believe most people use
hurt and angry labels to describe qualitatively different experi-
ences, some individuals may not understand the difference be-
tween these terms and may use them interchangeably (see Salovey
& Mayer, 1990).
Yet, even if hurt and anger experiences are often blended, they
are considered by many researchers, ourselves included, to be
unique (see Feeney, 2005;Fitness & Warburton, 2009;Leary &
Leder, 2009;Shaver et al., 2009;Vangelisti, 2009). That is, some
features are thought to be more characteristic of one emotion
relative to the other. Understanding these differences requires
control of their positive association. Otherwise, characteristics that
are highly associated with one experience may be falsely attributed
to the other only because these emotions are sometimes blended
(i.e., a confound effect). Research by Tangney and colleagues
(Tangney et al., 1996;Tangney, Wagner, Fletcher, et al., 1992;
Tangney, Wagner, & Gramzow, 1992) on shame and guilt illus-
trates a useful approach for grappling with the covariance between
emotions. Like hurt and anger, reports of shame and guilt are
strongly positively associated, but these studies suggest very dif-
ferent cognitive and behavioral signatures of shame and guilt once
their direct effects are examined through statistical partialing (see
also Sanford & Rowatt, 2004). We used a similar approach to
understand hurt and anger. This approach can reveal which aspects
of experience are due to hurt feelings and which are due to anger,
even when hurt and anger are sometimes blended.
Other methodological factors also may have contributed to
the view of hurt as a destructive experience. Some of the most
influential studies of hurt feelings are studies of retrospective
accounts of hurtful events (Feeney, 2005;Leary et al., 1998;
Vangelisti & Young, 2000). These studies required participants
to identify and then describe any hurtful event they experi-
enced. Participants in these studies tend to select events that
occurred in the distant past, despite the fact that most partici-
pants had been hurt more recently (Feeney, 2004;Leary &
Springer, 2001). This may indicate that participants selected
events that were especially damaging or unresolved. The viv-
idness of these events, as well as a Zeigarnik effect—the
tendency to exhibit heightened memory for unfinished tasks
(Zeigarnik, 1938)—may explain why participants would select
them. Indeed, people may forget typical hurtful events that have
been met with apologies, repair, and resolution. If these spec-
ulations are correct, then these studies may have overestimated
the negative impact of more typical hurtful episodes.
In addition, some influential studies characterized as being
about hurt feelings have not clearly demonstrated that the
experience of hurt per se predicts negative outcomes. For ex-
ample, findings suggesting that destructive responses and neg-
ative relationship outcomes are frequently described features of
hurtful events (Leary et al., 1998) do not necessarily imply that
the presence or intensity of hurt was responsible for negative
outcomes. Other aspects of these events could cause negative
responses. For example, given that most of these events have
themes of relational devaluation, devaluation may have been a
direct cause, consistent with findings suggesting that devalua-
tion can predict aggressive behavior and reduced constructive
behavior without being mediated by negative emotion (Twenge,
Baumeister, DeWall, Ciarocco, & Bartels, 2007;Twenge,
Baumeister, Tice, & Stucke, 2001), findings illustrating the
power of reciprocity in interpersonal attraction (e.g., Kenny &
la Voie, 1982), and findings suggesting that people are reluctant
985
HURT FEELINGS AND ANGER
to invest in relationships when they doubt acceptance (Murray
et al., 2006). Likewise, some influential studies of hurt feelings
have examined other variables, such as perceived intentionality
(Vangelisti & Young, 2000), perceived rejection (Feeney,
2004), the perpetrator’s behaviors (Feeney, 2004), and other
negative affective responses (Feeney, 2004). Of course, these
are not the same as hurt feelings.
Related to the previous point, reviews of consequences of hurt
feelings (e.g., Caughlin et al., 2009;Leary & Springer, 2001;
MacDonald & Leary, 2005) have described antisocial responses to
relational devaluation as if they were responses to hurt. According
to the implicit logic present in these reviews, if relational devalu-
ation causes hurt feelings and causes negative outcomes, then hurt
feelings must cause negative outcomes. However, as we have
described, hurt feelings may not be the culprit. Relational deval-
uation elicits a variety of emotional experiences that could mediate
its effects on negative outcomes (Leary, 1990;Leary et al., 2006),
and devaluation may elicit negative outcomes directly (Twenge et
al., 2001;2007).
Current Research
Despite agreement among many scholars that hurt and anger
are distinct experiences, prior research has not adequately dif-
ferentiated them. As outlined in Table 1, we discriminated hurt
and anger according to victims’ appraisals, goals, and behaviors
and perpetrators’ cognitive, emotional, and behavioral re-
sponses. To summarize, our prototype of hurt is as follows:
Hurt occurs when the victim feels devalued by a perpetrator
who is highly valued by the victim. Hurt is thus characterized
by a feeling of dependence and vulnerability to harm, and by the
goal of restoring a sense of acceptance and being valued. As
such, hurt usually occurs when victims feel committed to per-
petrators, it does not motivate victims’ destructive behavior,
and it may motivate constructive responses to repair the rela-
tionship. Victims’ hurt conveys their commitment to perpetra-
tors, elicits perpetrators’ empathy and guilt, and motivates
perpetrators’ provision of reassurance. These responses should
contribute to relationship repair. In contrast, our prototype of
anger is as follows: Anger arises when a victim is dissatisfied
with behavior enacted by a perpetrator, is characterized by
appraisals of low dependence and high power, and involves the
goal of coercing change in that behavior, usually through an-
tagonistic, punishing behaviors. When anger is perceived by
perpetrators, it conveys the victims’ dissatisfaction, lack of felt
dependence, and antagonistic urges, and thus often elicits the
perpetrators’ continued enactment of destructive behaviors.
These responses likely exacerbate and perpetuate interpersonal
problems.
Across four studies, we tested these distinctions between hurt
and anger and addressed the previously described methodolog-
ical issues in several ways. First, we examined direct effects of
victims’ hurt and anger intensity, and our analyses ruled out the
alternative explanation that these effects are confounds of re-
lational devaluation. Second, we tested our predictions in a
methodologically pluralistic manner. We conducted retrospec-
tive studies of subjective experiences and subsequent behaviors
associated with hurt and anger (Studies 1a, 1b, and 2), perpe-
trators’ responses (Study 2), and relationship consequences
(Study 2). However, we also used other methodological ap-
proaches. Study 3 is a daily diary study in which we examined
perpetrators’ responses to romantic partners’ hurt and anger.
Study 4 is a behavioral observation study in which we examined
behavioral responses to victims’ hurt and anger during ongoing
interactions. In the latter two studies, we also tested the pre-
diction that victims’ commitment creates a vulnerability to hurt
feelings.
Studies 1a and 1b
In our first two studies, we attempted to replicate prior
findings indicating that relational devaluation elicits both hurt
feelings and anger, and tested predictions regarding the subjec-
tive experiences and subsequent behaviors characterizing hurt
and anger. We predicted that hurt would be associated with
victims’ felt vulnerability and dependence and with their goals
to restore acceptance (see left-hand section of Table 1). In
contrast, we predicted that anger would be associated with
victims’ subjective sense of power, and desires to coerce
change in the perpetrator’s behavior. We also predicted that
anger, but not hurt, would be associated with victims’ destruc-
tive responses, and we speculated that hurt may instead be
associated with victims’ constructive responses.
In Study 1a, participants recalled an incident in which they
felt hurt or angry and then completed measures assessing the
constructs described and listed in Table 1. This procedure
allowed us to examine whether our predictions could be sup-
ported while randomly assigning participants to hurt or angry
conditions, which would suggest differences in hurt and angry
states rather than individual difference phenomena (i.e., quali-
ties of hurt-prone or anger-prone people). However, the explicit
instructions to report on a hurtful or angry event could have
activated general knowledge about hurt and anger, which could
have biased participants’ responses (see Robinson & Clore,
2002). We addressed this issue in Study 1b: participants re-
called a time in which they were mistreated by someone and
then completed the same measures. Unlike participants in Study
1a, the concepts of hurt and anger were not made salient. In
fact, hurt and anger were mentioned only in the two questions
assessing intensity of these experiences. Hence, consistency in
findings across these studies would suggest that they cannot be
explained by activating general knowledge of hurt and anger.
Method
Participants. Study 1a included 422 participants (Mage ⫽38
years; 189 men and 229 women). Study 1b included 218 partici-
pants (Mage ⫽33 years; 50 men and 168 women). Participants in
both studies completed a questionnaire posted on the Internet and
were recruited using two methods. Advertisements were posted on
local Internet bulletin boards across the continental United States
inviting participants to complete the questionnaire in exchange for
entry in cash raffles. In addition, participants were recruited
through Amazon Mechanical Turk, an open website that offers
“workers” the ability to complete brief tasks over the Internet in
exchange for a small payment. Prior research suggests that samples
collected via Amazon Mechanical Turk are more demographically
diverse than typical online and college student samples and that
986 LEMAY, OVERALL, AND CLARK
data quality is comparable (Buhrmester, Kwang, & Gosling,
2011). In Study 1a, the racial distribution was as follows: 79.1%
White, 5% African American, 9% Asian, and 7% other. In Study
1b, the racial distribution was as follows: 78.7% White, 8.2%
African American, 8.2% Asian, and 4.6% other.
1
Measures and procedure. In Study 1a, participants were
assigned randomly to recall an incident in which someone hurt
their feelings or made them angry. In Study 1b, participants
recalled an incident in which they were the target of negative
treatment from another person. Participants in both studies then
described the incident and completed the rating scale questions
described in the following text. We refer to participants as victims
and the individuals reported by participants as enacting the focal
negative behaviors as perpetrators.
Perpetrators’ devaluation behaviors. To assess the degree to
which perpetrators engaged in behaviors that could communicate
relational devaluation, we asked participants to complete eight
questions that assessed perpetrators’ behaviors, including (a) ex-
plicit statements of relational devaluation (e.g., “Said that he/she
does not desire a relationship with you”; “Said that he/she has
negative feelings about you”), (b) betrayal (e.g., “Lied to you”;
“Betrayed your trust”), and (c) social exclusion and ostracism
(e.g., “Excluded you from some activity”; “Ignored you”). Items
were completed on 5-point response scales (1 ⫽not at all char-
acteristic of [perpetrator name]’s behavior;5⫽extremely char-
acteristic of [perpetrator name]’s behavior; Study 1a Cronbach’s
␣⫽.86; Study 1b Cronbach’s ␣⫽.85).
2
Victims’ appraisals. Participants completed questions assess-
ing their subjective experiences after perpetrators enacted the
initial behavior to test the key distinctions outlined in Table 1. Four
items assessed subjective dependence on the relationship (e.g., “I
felt that I needed my relationship with [perpetrator name]”; “I felt
dependent on [perpetrator name]”; Study 1a Cronbach’s ␣⫽.90;
Study 1b Cronbach’s ␣⫽.86), three items assessed vulnerability
(e.g., “I felt vulnerable to harm”; “I felt cautious”; Study 1a
Cronbach’s ␣⫽.71; Study 1b Cronbach’s ␣⫽.72), and four
items assessed subjective control and power (e.g., “I felt power-
ful”; “I felt strong enough to overpower [perpetrator name] if
needed”; “I felt independent”; Study 1a Cronbach’s ␣⫽.75; Study
1b Cronbach’s ␣⫽.77). Items were completed on 6-point scales
(1 ⫽strongly disagree,6⫽strongly agree).
Victims’ goals. Participants also completed questions assess-
ing the relevant goals presented in Table 1. Five items assessed
goals to restore the perpetrator’s acceptance (e.g., “I felt a desire to
be loved by [perpetrator name]”; “I wanted [perpetrator name] to
care about me more than he/she seemed to”; “I wanted [perpetrator
name] to express positive views of me”; Study 1a Cronbach’s ␣⫽
.90; Study 1b Cronbach’s ␣⫽.88). Four items assessed goals to
change perpetrators’ behavior through punishment or coercion
(e.g., “I wanted to force [perpetrator name] to change his/her
behavior”; “I wanted to punish [perpetrator name] for his/her
wrongdoing”; “I wanted to teach [perpetrator name] a lesson”;
Study 1a Cronbach’s ␣⫽.85; Study 1b Cronbach’s ␣⫽.84).
Victims’ subsequent behavior. Participants completed items
assessing their behavioral reactions after perpetrators enacted the
initial behavior. Four items assessed destructive responses (e.g.,
“Blamed, criticized, or insulted [perpetrator name]”; “Was sarcastic or
patronizing to [perpetrator name]”; “Yelled at [perpetrator name]”;
Study 1a Cronbach’s ␣⫽.78; Study 1b Cronbach’s ␣⫽.80), and
seven items assessed constructive responses (e.g., “Calmly expressed
your feelings or point of view to [perpetrator name] ”; “Expressed
positive feelings about [perpetrator name] ”; “Expressed a willingness
to compromise to [perpetrator name]”; Study 1a Cronbach’s ␣⫽.88;
Study 1b Cronbach’s ␣⫽.88). Items were completed on 5-point
scales (1 ⫽not at all characteristic of my behavior,5⫽extremely
characteristic of my behavior).
Participants’ felt intensity of hurt and anger. Participants also
completed one item assessing the extent to which they felt angry
during the incident and one item assessing the extent to which they
had hurt feelings during the incident. Items were completed on 5-point
response scales (1 ⫽not at all,5⫽extremely).
Results and Discussion
We first attempt to replicate prior findings suggesting that perpe-
trators’ devaluation behaviors elicit victims’ hurt feelings and anger.
We then examine associations of victims’ hurt and anger with their
subjective experiences and subsequent behaviors.
Effects of relational devaluation on hurt feelings and anger.
We regressed victims’ ratings of hurt feelings and anger on the
index of perpetrators’ devaluation behaviors. Perpetrators’ deval-
uation behaviors predicted victims’ hurt feelings, Study 1a: ⫽
.35, t⫽7.29, p⬍.001; Study 1b: ⫽.35, t⫽5.24, p⬍.001, and
victims’ anger, Study 1a: ⫽.23, t⫽4.66, p⬍.001; Study 1b:
1
The challenges of conducting research over the Internet are well docu-
mented (see Kraut et al., 2004). These challenges include multiple submissions
by the same individual; inability to control the testing environment, which
introduces noise; high dropout rates; and reductions of participants’ investment
of time and energy into the research tasks. We took a number of steps to
address these issues. First, we tracked Internet protocol (IP) addresses and
eliminated submissions that were identical to a prior submission with regard to
both the IP address and the participant’s age, which should have addressed the
problem of repeat responders. Using these same criteria, we eliminated data
from the two participants who appeared to provide responses in both Study 1a
and Study 1b from Study 1a. Second, to address issues of reduced investment
of time and energy, we tracked questionnaire completion times and eliminated
all responses provided by participants who completed the study in 5 min or
less. This criterion was established before we conducted our data analysis, and
implementing this rule had a negligible effect on the obtained results. The
pattern of statistical significance remained the same, and coefficients changed
in miniscule ways. Implementing this rule reduced the sample size in the
primary analyses by approximately five participants in Study 1a and by
approximately 10 participants in Study 1b. Many other participants who did
not meet this 5-min criterion would have been excluded from analyses anyway
because they discontinued their participation very early in the questionnaire,
leaving missing responses on all of the primary variables. This dropout issue
is common in Internet-based research (e.g., Birnbaum, 2004;Williams,
Cheung, & Choi, 2000). All analyses and descriptive statistics are from
analyses conducted after implementing this rule. Third, to mitigate the issue of
increased noise, we collected large samples.
2
A complete list of scale items is available upon request.
3
An alternative view is that one of these emotions is secondary, only a
result of feeling the other. That is, perhaps victims felt angry by devaluation
only because they initially felt hurt, or they felt hurt only after they were
initially made angry. These data did not support this view. Even after control-
ling for victims’ anger, perpetrators’ devaluation behaviors predicted victims’
hurt feelings, Study 1a: ⫽.37, t⫽7.63, p⬍.001; Study 1b: ⫽.31, t⫽
4.55, p⬍.001. Likewise, even after controlling for victims’ hurt feelings,
perpetrators’ devaluation behaviors predicted victims’ anger, Study 1a: ⫽
.26, t⫽4.91, p⬍.001; Study 1b: ⫽.16, t⫽2.18, p⬍.05.
987
HURT FEELINGS AND ANGER
⫽.23, t⫽3.3, p⬍.001. These results replicate prior findings
suggesting that devaluation elicits both hurt feelings and anger.
3
Victims’ appraisals, goals, and behaviors associated with
hurt and anger.
Mean comparisons. We expected that the experience of hurt
would be associated with perceptions of dependence on a relation-
ship with the perpetrator, a sense of vulnerability to harm, goals to
restore the perpetrator’s acceptance, and subsequent constructive
behavior. In contrast, we expected that the experience of anger
would be associated with a subjective sense of control, goals to
control the perpetrator’s behavior, and subsequent destructive be-
havior. Using data from Study 1a, we compared victims’ subjec-
tive experiences and subsequent behaviors across the hurt and
anger conditions with independent samples ttests (see Table 2).
Relative to participants who were assigned to report on an expe-
rience in which they were made angry, participants who were
assigned to report on a hurtful experience reported greater depen-
dence, more vulnerability, and stronger acceptance goals. Relative
to participants who were assigned to report on a hurtful experience,
participants who were assigned to report on an angry experience
reported higher subjective control, stronger behavior control goals,
and more destructive behavior. Constructive behavioral responses
did not vary across conditions. With this one exception, these
results provide strong support for our predictions.
These findings demonstrate that random assignment to reporting on
hurtful or angry experiences produces differences in reports of sub-
jective experiences and subsequent behaviors that are consistent with
our predictions. This method is useful because the results suggest
differences between the states of hurt and anger rather than effects of
individual differences. However, the method is limited because the
comparisons are constrained. Although participants assigned to report
a hurtful incident reported significantly more hurt feelings during the
incident (M⫽4.39) than participants assigned to report an anger
incident (M⫽3.64), t(385) ⫽6.58, p⬍.001, it is clear that hurt was
elevated across both conditions. Likewise, participants assigned to
report an angry experience reported more anger (M⫽4.25) than
participants assigned to report a hurtful experience (M⫽3.59),
t(385) ⫽⫺6.40, p⬍.001, but again it is clear that anger was elevated
across both conditions. The condition comparisons may have under-
estimated differences between hurt and anger because these condi-
tions do not vary widely on hurt and anger. Moreover, the condition
comparisons do not indicate whether differences are due to effects of
hurt, effects of anger, or both. To address these limitations, we used
data from both Studies 1a and 1b to conduct analyses involving the
continuous ratings of hurt and anger intensity.
Analyses using continuous ratings of hurt and anger intensity.
First we examined effects of hurt and anger on cognitive appraisals
(i.e., perceived dependence, vulnerability, and control). We used data
from both Study 1a and Study 1b and regressed each of these vari-
ables on victims’ hurt and anger intensity. Given that both hurt and
anger were associated with perpetrators’ devaluation behaviors, we
controlled for perpetrators’ devaluation behaviors to be sure that
support for our predictions was not due to the association of hurt or
anger with perpetrators’ devaluation behaviors. In fact, results sup-
ported our predictions regardless of whether we included this control.
Results from both Study1a and Study1b were consistent with our
predictions (see Table 3). Victims’ hurt was positively associated with
subjective dependence and vulnerability and was inversely associated
with subjective control. Victims’ anger was inversely associated with
dependence and (in Study 1b only) vulnerability and was positively
associated with subjective control.
Next we used the same approach to examine effects of hurt and
anger on victims’ goals (see Table 4). Results again supported our
predictions across both Study 1a and Study 1b. Victims’ hurt was
positively associated with acceptance goals and (Study 1a only)
inversely associated with behavior control goals, whereas victims’
anger was positively associated with behavior control goals and
inversely associated with acceptance goals.
We conducted analogous regression models to examine the asso-
ciations of hurt feelings and anger with victims’ constructive and
destructive behavior following the event (see Table 5). In both
Studies 1a and 1b, victims’ anger predicted increased destruc-
tive behavior and decreased constructive behavior. In contrast,
victims’ hurt did not predict destructive behavior and predicted
increased constructive behavior, although this effect was found
only in Study 1a.
Summary. Results of both studies are consistent with the con-
ceptual distinctions outlined in Table 1. Hurt was associated with
feelings of dependence on the relationship with the perpetrator, vul-
nerability to harm, and reduced control, and with goals to restore
acceptance. Anger, in contrast, was associated with feelings of greater
subjective control, reduced dependence, and reduced vulnerability,
and with goals to coerce change in perpetrators’ behavior. In addition,
anger, but not hurt, was associated with victims’ increased destructive
behavior and reduced constructive behavior.
4
4
In Study 1a, hurt and anger reports were not significantly correlated,
r(380) ⫽⫺.02, p⬎.73. This null association appeared to be due to the fact
that assignments to conditions resulted in one group of participants report-
ing more hurt and less anger than another group, contributing to a negative
correlation between hurt and anger. Indeed, the partial correlation between
ratings of hurt and anger after condition was controlled was marginal and
positive, r(378) ⫽.10, p⫽.06. In Study 1b, hurt and anger were only
moderately positively correlated, r(196) ⫽.26, p⬍.001. Given that the
associations between hurt and anger were only moderate in these studies,
regression analyses that excluded one of these as a predictor produced a
nearly identical pattern of results. We added interactions with gender to all
regression models to examine whether effects were moderated by gender.
In 28 tests of gender interactions across Studies 1a and 1b, none of the
moderating effects reached conventional significance levels, ps⬎.07.
Moreover, the one marginal interaction in Study 1a was not marginal in
Study 1b, and the two marginal interactions in Study 1b were not marginal
in Study 1a, suggesting that gender effects were not replicable.
Table 2
Comparisons Across Hurt and Anger Accounts (Study 1a)
Victims’ experiences
and responses Hurt condition Anger condition t
Dependence 3.29 (1.52) 2.69 (1.56) 3.88
ⴱⴱⴱ
Vulnerability 3.93 (1.34) 3.27 (1.47) 4.66
ⴱⴱⴱ
Subjective control 2.52 (1.10) 3.00 (1.22) ⫺4.03
ⴱⴱⴱ
Acceptance goals 3.73 (1.4) 3.04 (1.59) 4.57
ⴱⴱⴱ
Behavior control goals 3.45 (1.45) 4.17 (1.42) ⫺4.93
ⴱⴱⴱ
Constructive behavior 2.40 (1.05) 2.38 (1.06) 0.25
Destructive behavior 2.08 (1.01) 2.42 (1.06) ⫺3.18
ⴱⴱ
Note. Means appear outside parentheses, and standard deviations appear
inside parentheses. Due to missing values, degrees of freedom for the ttests
ranged from 379 to 400.
ⴱⴱ
p⬍.01.
ⴱⴱⴱ
p⬍.001.
988 LEMAY, OVERALL, AND CLARK
Study 2
In our second study, we attempted to replicate some of the
theoretically central findings reported in Studies 1a and 1b, includ-
ing findings regarding victims’ goals and subsequent behaviors. In
addition, and consistent with the distinctions outlined in Table 1,
we tested the additional predictions that victims’ hurt feelings
signal high levels of commitment to perpetrators, whereas victims’
anger signals low commitment; and victims’ hurt feelings elicit
perpetrators’ constructive responses, whereas victims’ anger elicits
perpetrators’ destructive responses.
In addition, we examined relationship consequences of hurtful
and angry events. Prior research suggests the hurtful events have
generally negative consequences (Leary et al., 1998). However, we
expected that as a result of constructive responses to victims’ hurt,
the experience of hurt may have positive relationship conse-
quences. Indeed, given destructive responses to anger, anger
should have more negative consequences than hurt. As stated
earlier, one reason that prior retrospective studies suggest negative
consequences of hurtful events relates to the methodology used to
sample events. Participants in prior retrospective studies were
permitted to select any hurtful event. Events with especially de-
structive consequences (e.g., relationship termination), with severe
transgressions (e.g., infidelity), or that are unfinished (e.g., the
perpetrator has yet to make amends) may have been especially
memorable and therefore overrepresented. As a result, these stud-
ies may have produced an overly negative portrayal of hurtful
events. To address this issue, we asked participants to describe
recent events, with the expectation that this would pull for more
typical episodes of hurt. Given the systematic biases that arise as
a function of recalling events from victim and perpetrator perspec-
tives (i.e., perpetrators portray themselves as less culpable and
destructive than victims portray perpetrators; Baumeister et al.,
1990), we solicited reports of events from both perspectives.
Method
Participants. Undergraduate college students (Mage ⫽19.72
years; 115 men and 148 women) were recruited from a participant
pool and received partial credit in their psychology courses in
exchange for participation. The majority (94%) of participants
were White.
Table 3
Effects of Perpetrators’ Devaluation Behaviors, Victims’ Hurt Feelings, and Victims’ Anger on
Victims’ Cognitive Appraisals (Studies 1a and 1b)
Predictor Dependence Vulnerability Subjective control
Study 1a
Perpetrators’ devaluation behaviors .20 (4)
ⴱⴱⴱ
.16 (3.2)
ⴱⴱ
.06 (1.1)
Victims’ hurt feelings intensity .28 (5.68)
ⴱⴱⴱ
.38 (7.69)
ⴱⴱⴱ
⫺.33 (⫺6.51)
ⴱⴱⴱ
Victims’ anger intensity ⫺.21 (⫺4.35)
ⴱⴱⴱ
.02 (.37) .17 (3.37)
ⴱⴱ
Study 1b
Perpetrators’ devaluation behaviors .25 (3.59)
ⴱⴱⴱ
.22 (3.11)
ⴱⴱ
.12 (1.63)
Victims’ hurt feelings intensity .36 (5.21)
ⴱⴱⴱ
.38 (5.4)
ⴱⴱⴱ
⫺.25 (⫺3.24)
ⴱⴱ
Victims’ anger intensity ⫺.22 (3.38)
ⴱⴱ
⫺.14 (⫺2.1)
ⴱ
.16 (2.18)
ⴱ
Note. Standardized coefficients appear outside parentheses, and tvalues appear inside parentheses. In Study 1a,
residual degrees of freedom ranged from 377 to 378. In Study 1b, residual degrees of freedom ranged from 189
to 190.
ⴱ
p⬍.05.
ⴱⴱ
p⬍.01.
ⴱⴱⴱ
p⬍.001.
Table 4
Effects of Perpetrators’ Devaluation Behaviors, Victims’ Hurt
Feelings, and Victims’ Anger on Victims’ Goals
(Studies 1a and 1b)
Predictor
Goals
Acceptance Behavior
control
Study 1a
Perpetrators’ devaluation behaviors .24 (5.12)
ⴱⴱⴱ
.13 (2.71)
ⴱⴱ
Victims’ hurt feelings intensity .39 (8.51)
ⴱⴱⴱ
⫺.15 (⫺3.13)
ⴱⴱ
Victims’ anger intensity ⫺.19 (⫺4.35)
ⴱⴱⴱ
.45 (9.75)
ⴱⴱⴱ
Study 1b
Perpetrators’ devaluation behaviors .23 (3.47)
ⴱⴱ
.14 (2.05)
ⴱ
Victims’ hurt feelings intensity .44 (6.68)
ⴱⴱⴱ
⫺.07 (⫺1.02)
Victims’ anger intensity ⫺.30 (3.47)
ⴱⴱ
.45 (6.74)
ⴱⴱ
Note. Standardized coefficients appear outside parentheses, and tvalues
appear inside parentheses. In Study 1a, residual degrees of freedom ranged
from 377 to 378. In Study 1b, residual degrees of freedom ranged from 189
to 190.
ⴱ
p⬍.05.
ⴱⴱ
p⬍.01.
ⴱⴱⴱ
p⬍.001.
Table 5
Effects of Perpetrators’ Devaluation Behaviors, Victims’ Hurt
Feelings, and Victims’ Anger on Victims’ Constructive and
Destructive Behaviors (Studies 1a and 1b)
Predictor
Behaviors
Constructive Destructive
Study 1a
Perpetrators’ devaluation behaviors .21 (3.96)
ⴱⴱⴱ
.14 (2.69)
ⴱⴱ
Victims’ hurt feelings intensity .17 (3.16)
ⴱⴱ
⫺.08 (⫺1.53)
Victims’ anger intensity ⫺.19 (⫺3.69)
ⴱⴱⴱ
.37 (7.56)
ⴱⴱⴱ
Study 1b
Perpetrators’ devaluation behaviors .34 (4.66)
ⴱⴱⴱ
.23 (3.15)
ⴱⴱ
Victims’ hurt feelings intensity .03 (0.41) ⫺.11 (⫺1.56)
Victims’ anger intensity ⫺.28 (⫺3.94)
ⴱⴱⴱ
.32 (4.61)
ⴱⴱⴱ
Note. Standardized coefficients appear outside parentheses, and tvalues
appear inside parentheses. In Study 1a, residual degrees of freedom were
367. In Study 1b, residual degrees of freedom were 186.
ⴱⴱ
p⬍.01.
ⴱⴱⴱ
p⬍.001.
989
HURT FEELINGS AND ANGER
Procedure and measures. Participants completed a question-
naire on which they described and rated two recent incidents
(“within the past month or so”) that occurred with a relationship
partner, one relevant to hurt feelings and one relevant to anger.
They were instructed to describe incidents involving the same
romantic partner (n⫽112) or the same friend if they were not
romantically involved (n⫽142).
5
Participants who were assigned
to the perpetrator perspective condition (perpetrator reporters;
n⫽123) described one incident in which they hurt this person’s
(the victim’s) feelings and one incident in which they made this
person angry. Participants assigned to the victim perspective con-
dition (victim reporters;n⫽141) described incidents in which
they felt hurt by this person (the perpetrator) and angry with this
person. Order was counterbalanced. After describing the incident,
participants completed the measures described in the following
text. Except for 10 cases (2% of all requested events), participants
were able to identify relevant events.
Perpetrators’ devaluation behaviors. Perpetrator reporters re-
sponded to five questions assessing their enactment of behaviors
that instigated the situation and made the victim upset. These
questions assessed common ways of communicating relational
devaluation to victims, including betrayal (“I lied to or betrayed
him/her”), criticism (“I criticized him/her”), cold behavior (“I was
cold toward him/her”), selfishness (“I was being selfish”), and
unfair treatment (“I treated him/her unfairly”). Victim reporters
completed analogous measures of the perpetrator’s behavior. Items
were completed on 9-point response scales (1 ⫽extremely dis-
agree,9⫽extremely agree; Cronbach’s ␣s⫽.71 and .73, respec-
tively).
Victims’ hurt feelings and anger intensity. Participants re-
porting on their experience as victims completed two items assess-
ing the intensity of their hurt feelings and anger during each
incident (i.e., “My feelings were hurt”; “I was angry”). Perpetrator
reporters completed two analogous items assessing their percep-
tions of the victim’s emotions (e.g., “His/her feelings were hurt”).
Items were completed on 5-point scales (1 ⫽very slightly or not
at all,5⫽extremely).
Perpetrators’ perception of victims’ state commitment.
Perpetrator reporters completed a three-item measure of their per-
ceptions of the victim’s commitment during the incident (e.g.,
“This person felt very attached to me”; “This person felt commit-
ted to maintaining our relationship”; Cronbach’s ␣⫽.54). Items
were completed on 9-point response scales (1 ⫽extremely dis-
agree,9⫽extremely agree).
Victims’ goals. Victim reporters completed a two-item mea-
sure assessing their goals to restore the perpetrator’s relational
valuing during the incident (i.e., “I wanted this person to express
more positive feelings about me”; “I wanted this person to care
more about our relationship”; Cronbach’s ␣⫽.80) and a four-item
measure assessing goals to change the perpetrator’s behavior (e.g.,
“I wanted this person to do things my way”; “I wanted to change
this person’s behavior”; Cronbach’s ␣⫽.74). Items were com-
pleted on 9-point response scales (1 ⫽extremely disagree,9⫽
extremely agree). Participants assigned to the perpetrator perspec-
tive condition completed analogous items assessing perceptions of
the victim’s goals to restore the perpetrator’s relational valuing and
to change the perpetrator’s behavior, using identical response
scales (e.g., “He/she wanted me to care more about our relation-
ship”; Cronbach’s ␣s⫽.77 and .78, respectively).
Victims’ behavioral responses. Victim reporters completed a
four-item measure of constructive behaviors they enacted after
they became hurt or angered by the perpetrator (e.g., “I expressed
positive feelings about him/her and our relationship”; “I asked
him/her for reassurance regarding how he/she felt about me”;
Cronbach’s ␣⫽.79) and a five-item measure of destructive
behaviors (e.g., “I yelled at him/her”; “I said something nasty or
critical to him/her”; Cronbach’s ␣⫽.81). Perpetrator reporters
completed analogous measures assessing perceptions of the vic-
tim’s constructive or destructive behavior after the victim became
hurt or angry (i.e., “He/she yelled at me”; Cronbach’s ␣s⫽.74 and
.82, respectively). Items were completed using 5-point response
scales (1 ⫽not at all characteristic of how I [he/she] behaved;5⫽
extremely characteristic of how I [he/she] behaved).
Perpetrators’ responses. Using the same 5-point response
scales, perpetrator reporters completed measures of the behaviors
they enacted after victims became hurt or angry, including a
five-item measure of constructive behavior (e.g., “I apologized”; “I
was affectionate toward him/her”; Cronbach’s ␣⫽.81) and a
four-item measure of destructive behavior (e.g., “I criticized him/
her”; “I blamed him/her”; Cronbach’s ␣⫽.85). Victim reporters
completed analogous items assessing their perceptions of perpe-
trators’ behavior (Cronbach’s ␣s⫽.86 and .91, respectively).
Using 5-point response scales (1 ⫽very slightly or not at all,5⫽
extremely), perpetrator reporters also completed measures of their
emotions at the end of the event, after the victim became hurt or
angry, including a four-item measure of guilt (e.g., “Guilty”;
“Sorry”; Cronbach’s ␣⫽.90); a four-item measure of empathic
concern, which was adapted from prior research (Coke, Batson, &
McDavis, 1978; e.g., “Warm”; “Compassionate”; Cronbach’s ␣⫽
.83); and a four-item measure of anger or hostility (e.g., “Anger”;
“Hostile”; Cronbach’s ␣⫽.89). Victim reporters completed anal-
ogous items assessing perceptions of the perpetrator’s emotions
(Cronbach’s ␣s⫽90, .91, and .91, respectively). Given that
reports of perpetrators’ constructive behavior, empathy, and guilt
were highly correlated (rs ranged from .60 to .67; Cronbach’s ␣⫽
.84) and that analyses using the individual variables produced
parallel results, we averaged these three variables to create a
composite measure of perpetrators’ constructive responses for use
in the primary analyses. Likewise, reports of perpetrators’ destruc-
tive behavior and anger or hostility were highly correlated (r⫽
.75; Cronbach’s ␣⫽.84), and analyses using these individual
variables produced parallel results. Hence, we averaged them to
create an index of perpetrators’ destructive responses.
Relationship consequences. Victim reporters completed a
seven-item measure assessing positive consequences of the event
(e.g., “As a result of this event, I care more about our relationship”;
“As a result of this event, this person treats me better”; “As a result
of this event, overall our relationship is stronger”; Cronbach’s ␣⫽
.94). They also completed a four-item measure assessing analo-
gous negative consequences (e.g., “As a result of this event, I care
less about our relationship”; “As a result of this event, our rela-
tionship is damaged”; Cronbach’s ␣⫽.92). Items were completed
5
The pattern of results was similar across participants reporting on
romantic and nonromantic interactions. Ten participants did not follow
instructions and described incidents involving family members or former
romantic partners. We retained their responses in the statistical analyses.
990 LEMAY, OVERALL, AND CLARK
on 9-point response scales (1 ⫽extremely disagree,9⫽extremely
agree). Participants assigned to the perpetrator condition com-
pleted analogous measures using the same response scales (Cron-
bach’s ␣s⫽.89 and .87, respectively).
Results and Discussion
Each participant provided data relevant to two incidents. We
tested our predictions using two-level regression models that mod-
eled these two incidents as nested within each participant. A
compound-symmetry error structure modeled the covariance of
the criterion variable across the two incidents, which accounts
for the nesting of the data. Our primary analyses examined effects
of the continuous ratings of victims’ hurt and anger intensity.
6
We
centered all continuous predictor variables on their sample means.
Effects of relational devaluation on hurt feelings and anger.
Using data provided by participants taking the perspective of
victim, we regressed victims’ hurt feelings and anger on the index
of perpetrators’ devaluation behaviors. Perpetrators’ devaluation
behaviors predicted victims’ self-reports of their own hurt feelings,
b⫽0.34, t⫽7.42, p⬍.001, and anger, b⫽0.40, t⫽9.19, p⬍
.001. Likewise, using data provided by participants reporting from
the perpetrator perspective, we found that perpetrators’ devalua-
tion behaviors predicted their perceptions of victims’ hurt feelings,
b⫽0.29, t⫽5.97, p⬍.001, and anger, b⫽0.21, t⫽4.06, p⬍
.001. These results replicate prior findings suggesting that deval-
uation elicits hurt and anger.
7
Victims’ goals associated with hurt feelings and anger. We
expected that victims’ hurt feelings would be associated with their
goal to restore acceptance from the perpetrator, whereas victims’
anger would be associated with the goal to change the perpetrator’s
behavior (see Table 1). To test this prediction, we regressed
victims’ acceptance and behavior control goals on victims’ hurt
and anger. We controlled for perpetrators’ devaluation behaviors
to be sure that support for our predictions could not be explained
by the association of victims’ hurt or anger with devaluation
behaviors, although results supported our predictions regardless of
whether we included this control. Results for participants reporting
from the victim perspective appear in the upper portion of Table 6.
Victims’ hurt was associated with their goals to restore acceptance,
whereas victims’ anger was associated with their goals to control
the perpetrator’s behavior.
We sought to replicate these findings using data provided by
participants reporting from the perspective of perpetrator. That is,
do perpetrators understand that hurt and angry victims have dif-
ferent goals? Results of analogous models are displayed in the
lower section of Table 6. Perpetrators’ perceptions of victims’ hurt
predicted their inferences of victims’ acceptance goals, whereas
perpetrators’ perceptions of victims’ anger predicted their infer-
ences of victims’ behavior control goals. These results support our
prediction that hurt and anger involve different goals, and they
were consistent across participants taking victim and perpetrator
perspectives.
Victims’ hurt feelings and anger as commitment signals.
We expected that victims’ hurt feelings would convey high levels
of commitment to perpetrators, whereas victims’ anger would
convey low levels of commitment (see perpetrator responses in
Table 1). Using data reported by participants taking the perspective
of perpetrator, we regressed perpetrators’ perceptions of victims’
commitment during the incident on perpetrators’ perceptions of
victims’ hurt feelings and anger. Again, we controlled for perpe-
trators’ relational devaluation behaviors to be sure that results
could not be explained by associations of hurt or anger with these
devaluation behaviors, although results supported our prediction
regardless of whether we included this control. Consistent with our
predictions, perpetrators’ perceptions of victims’ hurt predicted
perpetrators’ greater inferences of victims’ commitment, b⫽0.19,
t⫽2.29, p⬍.05, whereas perpetrators’ perceptions of victims’
anger predicted perpetrators’ inferences of reduced commitment,
b⫽⫺0.27, t⫽⫺3.3, p⬍.01. Perpetrators’ devaluation behaviors
did not predict perceptions of commitment, p⫽.56. Thus, hurt and
anger appeared to convey different messages to perpetrators re-
garding victims’ commitment.
Destructive and constructive responses following hurt feel-
ings and anger. Next we conducted parallel analyses to test
predictions that victims’ anger would elicit victims’ and perpetra-
tors’ destructive behavior, whereas victims’ hurt would elicit con-
structive responses (see Table 1). Results using data provided by
participants taking the perspective of victim and perpetrator appear
6
Our use of continuous ratings of hurt and anger affords stronger tests
of our predictions relative to comparing the hurt accounts with the angry
accounts because both types of accounts tended to be characterized by both
hurt and anger (Mhurt ⫽3.04 in the hurt accounts; Mhurt ⫽2.64 in the
angry accounts; Manger ⫽2.73 in the hurt accounts; Manger ⫽3.10 in
the angry accounts). As a result, condition comparisons are constrained.
They do not reflect a comparison of hurt (without anger) versus anger
(without hurt).
7
As with the prior studies, we conducted additional analyses to examine
whether hurt or anger mediated effects of perpetrators’ devaluation behav-
iors on the other emotion. For example, do victims only feel angry
following devaluation because devaluation hurt their feelings? When we
controlled for victims’ hurt, perpetrators’ devaluation behaviors continued
to predict victims’ anger: victim perspective b⫽0.29, t⫽6.49, p⬍.001;
perpetrator perspective b⫽0.14, t⫽2.66, p⬍.01. Likewise, when we
controlled for victims’ anger, perpetrators’ devaluation behaviors contin-
ued to predict victims’ hurt: victim perspective b⫽0.19, t⫽4.03, p⬍
.001; perpetrator perspective b⫽0.23, t⫽4.67, p⬍.001. Hence,
devaluation appeared to have independent pathways to victims’ hurt and
anger.
Table 6
Effects of Perpetrators’ Devaluation Behaviors, Victims’ Hurt
Feelings, and Victims’ Anger on Victims’ Goals (Study 2)
Predictor
Victims’ goals
Acceptance Behavior control
Victim perspective participants
Perpetrators’ devaluation behaviors .46 (5.8)
ⴱⴱⴱ
.40 (6.80)
ⴱⴱⴱ
Victims’ hurt feelings .60 (6.18)
ⴱⴱⴱ
⫺.05 (⫺0.71)
Victims’ anger ⫺.14 (⫺1.41) .40 (5.39)
ⴱⴱⴱ
Perpetrator perspective participants
Perpetrators’ devaluation behaviors .47 (5.61)
ⴱⴱⴱ
.33 (5.01)
ⴱⴱⴱ
Perpetrators’ perception of victims’
hurt feelings .53 (4.89)
ⴱⴱⴱ
0(⫺0.05)
Perpetrators’ perception of victims’
anger .03 (0.25) .46 (5.63)
ⴱⴱⴱ
Note. Unstandardized coefficients appear outside parentheses, and tval-
ues appear inside parentheses. Degrees of freedom ranged from 225 to 256.
ⴱⴱⴱ
p⬍.001.
991
HURT FEELINGS AND ANGER
in the upper and lower sections of Table 7, respectively. Victims’
hurt, but not their anger, predicted increases in perpetrators’ con-
structive responses, and this effect was replicated across both
victim and perpetrator perspectives. In contrast, victims’ anger, but
not their hurt, predicted increases in their own and perpetrators’
destructive responses, and these effects also were replicated across
both victim and perpetrator perspectives. Hence, these results
support our predictions regarding responses to victims’ hurt and
anger using data from participants reporting from both perspec-
tives.
One prediction that was not supported was a positive association
between victims’ hurt feelings and their own constructive behav-
ior. The effects were in the predicted direction, but they did not
reach conventional significance levels. Perhaps greater statistical
power would have enabled us to detect the effect. In the analyses
previously reported, we reduced our statistical power by conduct-
ing separate analyses for victim and perpetrator reporters. When
we combined these participants into a single analysis and tested the
model again, we found that victims’ hurt (self-reports of their own
hurt or perpetrators’ perceptions of victims’ hurt) was marginally
associated with their own greater constructive behavior, b⫽0.07,
t⫽1.79, p⫽.07, whereas victims’ anger was associated with their
reduced constructive behavior, b⫽⫺0.10, t⫽⫺2.78, p⬍.01.
Relationship consequences of victims’ hurt feelings and
anger. Finally, we tested analogous models to examine the re-
lationship consequences associated with victims’ hurt and anger.
The interpersonal consequences of hurt versus anger that we have
shown and outlined in Table 1 should mean that hurt leads to more
positive relationship consequences, whereas anger creates more
negative relationship consequences. Results of models constructed
with data provided by participants reporting from the victim per-
spective appear in the upper section of Table 8. Victims’ anger
predicted their increased reports of negative relationship conse-
quences, whereas victims’ hurt predicted their increased reports of
positive relationship consequences. Results of models constructed
with data provided by perpetrator perspective participants appear
in the lower section of Table 8. Perpetrators’ perceptions of vic-
tims’ hurt predicted perpetrators’ increased reports of positive
relationship consequences. These results provide support for the
prediction that victims’ hurt feelings would be associated with
positive relationship consequences in data provided from both
victim and perpetrator perspectives. Support for the prediction that
victims’ anger would be associated with negative relationship
consequences was found in reports from the victim perspective.
Why should hurt and anger differ in their relationship conse-
quences? As we stated earlier, anger should tend to produce
negative relationship consequences due to the destructive re-
sponses that typically arise from anger, whereas the tendency for
hurt to produce positive relationship consequences should occur
due to the constructive responses that often follow hurt. In addi-
tional models, we added perpetrators’ and victims’ responses as
mediators to explicitly test these assumptions. After controlling for
victims’ and perpetrators’ destructive responses, we found that the
effect of victims’ anger on negative relationship consequences
(which was previously significant for victim perspective partici-
pants) was no longer significant, p⫽.78. Perpetrators’ destructive
responses predicted negative relationship consequences in this
model, b⫽0.66, t⫽4.91, p⬍.001. After controlling for victims’
and perpetrators’ constructive responses, we found that the effect
of victims’ hurt on positive relationship consequences (which was
previously significant for victim and perpetrator perspective par-
ticipants) was no longer significant, ps⬎.12. Instead, for victim
perspective participants, perpetrators’ constructive responses pre-
dicted positive relationship consequences, b⫽0.79, t⫽8.08, p⬍
.001, and for perpetrator perspective participants, both perpetra-
tors’ constructive responses and victims’ constructive responses
Table 7
Effects of Perpetrators’ Devaluation Behaviors, Victims’ Hurt Feelings, and Victims’ Anger on Victims’ and Perpetrators’ Responses
(Study 2)
Predictor
Victims’ responses Perpetrators’ responses
Constructive Destructive Constructive Destructive
Victim perspective participants
Perpetrators’ devaluation behaviors 0 (⫺0.02) .14 (4.35)
ⴱⴱⴱ
⫺.12 (⫺2.93)
ⴱⴱ
.19 (5.88)
ⴱⴱⴱ
Victims’ hurt feelings .07 (1.39) .02 (0.48) .14 (2.69)
ⴱⴱ
⫺.05 (⫺1.29)
Victims’ anger ⫺.13 (⫺2.57)
ⴱ
.33 (8.32)
ⴱⴱⴱ
⫺.03 (⫺0.58) .22 (5.55)
ⴱⴱⴱ
Perpetrator perspective participants
Perpetrators’ devaluation behaviors .03 (0.74) .04 (1.16) .03 (0.85) .09 (3.03)
ⴱⴱ
Perpetrators’ perception of victims’ hurt feelings .08 (1.47) .05 (1.28) .23 (4.85)
ⴱⴱⴱ
.05 (1.44)
Perpetrators’ perception of victims’ anger ⫺.07 (⫺1.27) .40 (10.14)
ⴱⴱⴱ
⫺.02 (⫺0.41) .19 (5.41)
ⴱⴱⴱ
Note. Unstandardized coefficients appear outside parentheses, and tvalues appear inside parentheses. Degrees of freedom ranged from 204 to 254.
ⴱ
p⬍.05.
ⴱⴱ
p⬍.01.
ⴱⴱⴱ
p⬍.001.
Table 8
Effects of Victims’ Hurt Feelings and Anger on Relationship
Consequences (Study 2)
Predictor
Relationship consequences
Positive Negative
Victim perspective participants
Perpetrators’ devaluation behaviors ⫺.03 (⫺0.57) .38 (5.69)
ⴱⴱⴱ
Victims’ hurt feelings .21 (2.63)
ⴱⴱ
⫺.08 (⫺0.96)
Victims’ anger .02 (.25) .16 (2.05)
ⴱ
Perpetrator perspective participants
Perpetrators’ devaluation behaviors .23 (3.77)
ⴱⴱⴱ
.10 (1.89)
†
Victims’ hurt feelings .17 (2.26)
ⴱ
0 (0.96)
Victims’ anger ⫺.10 (⫺1.33) .10 (1.48)
Note. Unstandardized coefficients appear outside parentheses, and tval-
ues appear inside parentheses. Degrees of freedom ranged from 184 to 238.
†
p⬍.10.
ⴱ
p⬍.05.
ⴱⴱ
p⬍.01.
ⴱⴱⴱ
p⬍.001.
992 LEMAY, OVERALL, AND CLARK
predicted positive relationship consequences, b⫽0.39, t⫽3.53,
p⬍.01, and b⫽0.34, t⫽3.58, p⬍.001, respectively. These
results confirm the hypothesis that constructive responses to hurt
and destructive responses to anger explain their divergent relation-
ship consequences.
8
Summary. Victims’ hurt feelings and anger were both pre-
dicted by perpetrators’ relational devaluation behaviors, but they
were associated with different goals and behavioral responses, and
these differences were consistent with the distinctions outlined in
Table 1. Replicating findings from Study 1, victims’ hurt was
associated with goals to restore the perpetrator’s acceptance,
whereas victims’ anger was associated with goals to change the
perpetrator’s behavior. Also replicating findings from Study 1,
victims’ anger, but not hurt, was associated with victims’ destruc-
tive behavior and reductions in constructive behavior. In fact,
victims’ hurt was associated with increased constructive behavior,
although the effect was small.
New to Study 2 were findings regarding the commitment that
hurt signals to perpetrators, perpetrators’ responses to victims’
hurt and anger, and the associated relationship consequences.
Just as hurt and anger had countervailing effects on feelings of
dependence (Studies 1a and 1b) and acceptance goals (Studies
1a, 1b, and 2), perpetrators in Study 2 inferred higher levels of
commitment when victims were more hurt and perceived lower
levels of commitment when victims were more angry. In addi-
tion, victims’ hurt predicted perpetrators’ constructive re-
sponses and positive relationship consequences, whereas vic-
tims’ anger predicted perpetrators’ destructive responses and
negative relationship consequences. The links with relationship
consequences also indicated that hurt had positive conse-
quences for the relationship because hurt elicited victims’ and
perpetrators’ constructive responses, whereas anger had nega-
tive consequences because anger elicited perpetrators’ destruc-
tive responses.
9
Most of our findings were replicated across
victim and perpetrator perspectives, suggesting that they cannot
be explained by biases that often arise as a result of taking one
of these perspectives.
Study 3
In Study 3, we examined hurt feelings and anger using a daily
report study of romantically involved couples. New to this study
was our examination of relationship commitment as an antecedent
of hurt feelings. As described previously, victims who are highly
committed to a relationship with the perpetrator should be more
likely to experience hurt feelings, and this would provide addi-
tional evidence, beyond the dependence and acceptance goals and
behaviors examined in the prior studies, suggesting that hurt in-
volves relationship maintenance concerns.
A second major contribution of Study 3 is the examination of
dyadic and prospective processes. Our predictions regarding per-
petrators’ constructive responses to victims’ hurt and destructive
responses to victims’ anger imply interpersonal processes in which
one person’s emotions elicit responses in his or her relationship
partner. By incorporating both members of relationships in Study
3, we could more directly test these dyadic processes. Moreover,
this study allowed us to test our predictions prospectively (e.g.,
effects of victims’ daily emotions on perpetrators’ responses the
following day). We used this approach to test our predictions that
victims’ hurt and perpetrators’ perceptions of this hurt convey high
commitment to perpetrators and elicit perpetrators’ constructive
responses, whereas victims’ anger and perpetrators’ perceptions of
this anger convey low commitment and elicit destructive re-
sponses.
A third contribution of Study 3 is our attempt to replicate
patterns across two levels of analysis. In addition to studying
specific emotional states, emotion researchers have studied emo-
tions in terms of trait tendencies to experience emotions (e.g.,
Watson & Clark, 1991;1992), and it is likely that people have
relatively stable tendencies to experience both hurt and anger. We
examined whether the predictions we made for states of hurt and
anger—in terms of their similar associations with relational deval-
uation, their distinct relations with victims’ commitment, and the
different messages they convey to perpetrators regarding victims’
commitment—could be replicated at the trait level of analysis.
Method
Participants. A sample of 105 dating or married couples was
recruited via advertisements on local bulletin boards, in newspa-
pers, and on a psychology department subject pool website.
Twenty-three couples were married or engaged or had a civil
8
We examined whether participants’ gender moderated the effects of
hurt and anger. Of 32 interactions examined, only one gender interaction
was marginally significant; all other ps⬎.10. The one marginal interaction
suggested that victims’ anger predicted their behavioral control goals more
strongly for female victims, ⫽.56, t⫽3.85, p⬍.001, than for male
victims, ⫽.27, t⫽2.8, p⬍.01. However, this interaction was not
significant in Studies 1a or 1b.
Our theorizing suggests the importance of modeling both hurt and anger
effects because these variables are correlated. Indeed, victims’ hurt and
anger were positively associated, anger ¡hurt b⫽0.48, t⫽8.97, p⬍
.001, as were perpetrators’ perceptions of victims’ hurt and anger, b⫽
0.35, t⫽5.77, p⬍.001. We briefly summarize the notable changes that
occurred when we excluded either hurt or anger, and, hence, we did not
account for their association. Some previously nonsignificant effects of
victims’ hurt became marginal or significant when anger was excluded
from the model, including an effect on victims’ destructive behavior—
victim perspective: b⫽0.13, p⬍.001; perpetrator perspective: b⫽0.16,
p⬍.001—and an effect on perpetrators’ destructive behavior—perpetrator
perspective only: b⫽0.10, p⬍.01. In addition, for perpetrator perspective
participants, victims’ anger no longer predicted negative relationship con-
sequences when hurt was not controlled, p⬎.10. Hence, victims’ hurt
seemed more destructive when victims’ anger was excluded from the
model, and victims’ anger seemed less destructive when the positive effects
of hurt were not controlled. However, all of the other findings reported in
the text remained significant when one of the emotions was excluded.
9
As we stated earlier, a possible explanation of why hurt seemed to have
less negative relationship consequences in our research relative to the
research by Leary et al. (1998) is that our strategy of asking participants to
select only recent events created a sample of events that were, on average,
less severe than the sample obtained by Leary and colleagues. We used a
5-point response scale to measure hurt feelings that was comparable to the
scale used by Leary and colleagues, allowing us to compare average hurt
feelings across our studies. Indeed, victims in our study (those reporting on
a hurtful event) reported less hurt (M⫽2.98; SD ⫽1.23) than victims in
the study by Leary and colleagues (M⫽4.0, SD ⫽.81), t(130) ⫽9.51, p⬍
.001. In addition, perpetrators in our study (those reporting on a hurtful
event) saw the victim as significantly less hurt, on average (M⫽3.12,
SD ⫽1.21), than did perpetrators in the study by Leary and colleagues
(M⫽3.9), t(117) ⫽⫺6.99, p⬍.001. Hence, events in our sample were
reported as less severe, on average, which may explain the difference in
findings. Of course, we cannot be certain of the cause of the difference in
findings given the many variables that likely vary across these two studies.
993
HURT FEELINGS AND ANGER
union. The remaining couples were dating. Three were same-sex
female couples. The remaining 102 couples were heterosexual.
Couples were predominately White and, on average, 24 years old.
They received a payment of $50 or research participation credit in
a psychology class and entry in a $100 lottery.
Procedure. The study was advertised as a study on daily events
in relationships. Participants arrived to an initial laboratory session
with their romantic partners to complete the baseline measures de-
scribed in the following section (among others). Participants were
instructed to start the daily report component the day following the
laboratory session and to complete a daily report between 7 p.m. and
midnight for a period of 7 consecutive days. Due to missing reports
and the loss of data when we examined lagged effects, the number of
daily observations included in the analyses ranged from 1,058 to
1,130. Daily items described in the following text were embedded in
a larger questionnaire. Hence, participants likely did not perceive a
focus on hurt feelings and anger.
Baseline measures.
Chronic relationship commitment. During the baseline ses-
sion, participants completed six items assessing relationship commit-
ment, which were adapted from the Investment Model Scale (Rusbult,
Martz, & Agnew, 1998; e.g., “I want our relationship to last for a very
long time”‘; “I am committed to maintaining my relationship with this
person”). Items were completed on 9-point scales (1 ⫽extremely
disagree,9⫽extremely agree; Cronbach’s ␣⫽.82).
General proneness to anger and hurt. Participants completed
four items assessing proneness to anger, which were adapted from
the Multidimensional Anger Inventory (Siegel, 1986). Items as-
sessed the frequency, duration, and extremity of anger toward the
partner (e.g., “I frequently get angry with this person”; “When I get
angry at this person, I stay angry for hours”; Cronbach’s ␣⫽.88).
They also completed four analogous items assessing frequency,
duration, and extremity of hurt feelings (e.g., “It is easy for this
person to hurt my feelings”; Cronbach’s ␣⫽.85). Items were
completed using the 9-point response scales described previously.
Participants also completed analogous measures of their percep-
tions of their partner’s proneness to anger (e.g., “This person
frequently gets angry with me”; Cronbach’s ␣⫽.91) and prone-
ness to hurt (e.g., “This person frequently feels hurt by me”;
Cronbach’s ␣⫽.84).
Chronic perception of the partner’s valuing. Participants
completed a four-item measure of their perceptions of their partner’s
regard (e.g., “This person thinks I have a number of good character-
istics”; Cronbach’s ␣⫽.80); a six-item measure of their perceptions
of their partner’s commitment, which was adapted from the Invest-
ment Model Scale (Rusbult et al., 1998; e.g., “This person wants our
relationship to last for a very long time”; Cronbach’s ␣⫽.91); and a
10-item measure of their perception of their partner’s care for their
welfare, which was adapted from the Communal Strength Scale
(Mills, Clark, Ford, & Johnson, 2004; e.g., “This person would go out
of his/her way to help me”; Cronbach’s ␣⫽.91). Items were com-
pleted using the 9-point response scales described earlier. Perceived
regard, commitment, and care scales were strongly correlated (rs
ranged from .54 to .77). Hence, they were averaged to create a
composite index of perceived partner valuing of the relationship
(Cronbach’s ␣⫽.85).
Daily measures.
Daily hurt feelings and anger. Each day, participants com-
pleted two items assessing their hurt feelings and anger caused by
their partner (i.e., “To what extent did this person hurt your
feelings today?”; “How angry did you feel toward this person
today?”). Items were completed on 9-point response scales (1 ⫽
not at all,9⫽extremely).
Daily perception of partner’s hurt feelings and anger.
Participants completed two analogous items assessing perceptions
of their partner’s hurt feelings and anger using the same 9-point
response scales (e.g., “To what extent were this person’s feelings
hurt by you today?”).
Daily perception of partner’s commitment. Participants com-
pleted a single item assessing perceptions of their partner’s com-
mitment (e.g., “For how much longer does your partner want your
relationship to last today?”) using a 9-point response scale (1 ⫽
end now,9⫽last forever).
Daily valuing of partners and perceived partner valuing.
Participants completed four items assessing their valuing of the
relationship with their partner, including commitment (i.e., “For
how much longer do you want your relationship to last with this
person today”; 1 ⫽I want it to end now,9⫽forever), regard for
the partner (i.e., “Overall, how do you view this person today?”
1⫽extremely negatively,9⫽extremely positively), and care (i.e.,
“How concerned for this person’s well-being were you today?”;
“How motivated were you to attend to this person’s needs today?”;
1⫽not at all,9⫽extremely; Cronbach’s ␣⫽.74). They also
completed analogous items assessing their perceptions of their
partner’s valuing using the same 9-point scales (e.g., “How did this
person view you today”; Cronbach’s ␣⫽.77).
Daily guilt. Participants completed a single item assessing
daily guilt (i.e., “Today, how much did you feel guilty or sorry
about how you have treated your study partner?”), which was
completed on 9-point response scales (1 ⫽not at all guilty,9⫽
extremely guilty).
Results and Discussion
Analysis strategy. Some of our predictions involved only the
measures administered during the baseline session. To test these
predictions, we used multilevel regression models (using the SAS
MIXED procedure) that treated the two dyad members as nested
within dyads and specified a compound-symmetry error structure
to estimate the covariance between the two dyad members on the
criterion variable, which accounts for potential dyadic interdepen-
dence (Kenny, Kashy, & Cook, 2006). Due to limited degrees of
freedom in these types of designs, we modeled slopes as fixed. We
refer to this type of model as a person-level model.
Other predictions involved variables measured at the daily level.
We tested these hypotheses using multilevel regression models
that modeled days and people as nested within dyad. The models
estimated person-level intercepts, which represent participants’
average score on the criterion variable across days, as well as the
correlation of these intercepts across the two dyad members. In
addition, the models estimated correlations of the day-specific
residuals across the two dyad members (Kenny et al., 2006). The
daily predictor variables were centered on each person’s mean.
Hence, they are orthogonal to individual difference variables and
their effects represent within-person effects. Again, we continued
to model the slopes as fixed, given the limited degrees of freedom
and to address model convergence problems. We refer to this type
of model as a day-level model.
994 LEMAY, OVERALL, AND CLARK
To clarify presentation of results, we use the term perpetrators
to refer to participants who were responding to their partner’s hurt
or anger, and we use the term victims to refer to the partner who
may have been hurt or angry, although it is important to note that
all participants provided data relevant to both roles. That is, in
some observations, Partner A provided data relevant to the victim
role and Partner B provided data relevant to the perpetrator role. In
other observations, Partner A provided data relevant to the perpe-
trator role and Partner B provided data relevant to the victim role.
Effects of relational devaluation on hurt feelings and anger.
We expected to replicate prior findings suggesting that relational
devaluation predicts both hurt feelings and anger. We tested this
prediction in “chronic” and “state” forms. With regard to evidence
for chronic processes, using the person-level model, we regressed
victims’ hurt and anger proneness on their chronic perceptions of
relational valuing (all administered during the baseline session).
Consistent with our prediction and prior findings, victims’ percep-
tions of perpetrators’ relational valuing was inversely associated
with victims’ reports of their hurt proneness, b⫽⫺0.90, t⫽
⫺8.77, p⬍.001, and anger proneness, b⫽⫺0.91, t⫽⫺9.43, p⬍
.001.
To test the same prediction in terms of state feelings of hurt and
anger, we used the day-level model and regressed victims’ daily
hurt and anger on their daily perceptions of the perpetrator’s
valuing. Daily perceived valuing was inversely associated with
victims’ daily hurt, b⫽⫺0.56, t⫽⫺17.34, p⬍.001, and daily
anger, b⫽⫺0.53, t⫽⫺15.64, p⬍.001. Hence, at both the
chronic and daily levels of analysis, relational devaluation pre-
dicted both hurt and anger.
10
Commitment as a vulnerability to hurt feelings. We ex-
pected that highly committed victims would be especially likely to
feel hurt in response to perceived devaluation by perpetrators.
Again, we tested this prediction with regard to chronic and state
responses. Using the person-level model, we regressed victims’
general hurt proneness on their chronic relationship commitment,
their chronic perceptions of the perpetrator’s relational valuing (all
measured during the baseline session), and a product term repre-
senting their interaction. Our prediction regarding commitment
and hurt was with regard to the unique component of hurt—the
variance of hurt that does not overlap with anger, as we did not
expect that commitment to be associated with a vulnerability to
anger (or general emotional volatility). Hence, to conduct a more
focused test of our prediction and to be sure that findings could not
be explained by anger, we controlled for victims’ general anger
proneness (for a similar approach, see Tangney et al., 1996;
Tangney, Wagner, Fletcher, et al., 1992). The Chronic Commit-
ment ⫻Chronic Perceptions of the Partner’s Valuing interaction
predicted hurt proneness, b⫽⫺0.17, t⫽⫺1.95, p⫽.05.
Predicted values are presented in Figure 1. We examined condi-
tional effects of chronic perceived partner valuing at low (1 SD
below the mean) and high (maximum score, which was slightly
lower than 1 SD above the mean) levels of commitment. Consis-
tent with our prediction, the link between victims’ perceptions of
the perpetrator’s valuing and victims’ hurt proneness was stronger
for highly committed victims, b⫽⫺0.72, t⫽⫺5.25, p⬍.001,
than for victims who were low in commitment, b⫽⫺0.36, t⫽
⫺2.48, p⬍.05.
Does commitment create the same vulnerability to anger? When
we regressed victims’ general anger proneness on an analogous set
of predictors (victims’ chronic relationship commitment, chronic
perception of the perpetrator’s valuing, their interaction, and vic-
tims’ hurt proneness), we did not find the same Commitment ⫻
Perceived Valuing interaction, p⬎.54. Hence, commitment was
not associated with vulnerability to anger.
We used the day-level model and daily measures of hurt feel-
ings, anger, and perceived partner valuing to test this same pre-
diction in terms of state feelings of hurt and anger. We expected
that chronically committed victims would be especially likely to
experience hurt feelings on days they felt devalued by perpetrators.
We regressed victims’ daily hurt feelings on their daily perceptions
of the perpetrator’s valuing, their chronic commitment to the
perpetrator, and a product term representing their interaction.
Again, we controlled for victims’ daily anger to examine the aspect
of hurt that is not correlated with anger and to be sure that daily
anger did not explain the effect. The Commitment ⫻Daily Per-
ceived Valuing interaction was a significant predictor of victims’
daily hurt feelings, b⫽⫺0.04, t⫽⫺1.98, p⬍.05, and is plotted
in Figure 2. Consistent with our prediction, highly committed
victims (1 SD above the mean) exhibited a greater inverse relation
between daily perceived partner valuing and daily hurt feelings,
b⫽⫺0.28, t⫽⫺7.72, p⬍.001, relative to victims low in
commitment, b⫽⫺0.20, t⫽⫺6.13, p⬍.001. A parallel model
examining daily feelings of anger did not reveal a Commitment ⫻
Daily Perceived Partner Valuing interaction, p⬎.12. Hence,
chronic relationship commitment moderated the effect of daily
perceived devaluation on daily hurt feelings and not daily anger.
10
As in our prior studies, we conducted additional analyses to discern
whether relational devaluation was uniquely associated with hurt and anger
proneness or whether one of these emotional experiences was primary and
mediated effects of relational devaluation on the other. Perceived valuing
by the perpetrator continued to predict victims’ anger proneness while their
hurt proneness was controlled, b⫽⫺0.43, t⫽⫺4.97, p⬍.001, and it
continued to predict victims’ hurt proneness while anger proneness was
controlled, b⫽⫺0.32, t⫽⫺3.25, p⬍.01. Similarly, daily perceived
valuing predicted victims’ daily hurt feelings while their daily anger was
controlled, b⫽⫺0.23, t⫽⫺8.42, p⬍.001, and it predicted victims’ daily
anger while their daily hurt was controlled, b⫽⫺0.20, t⫽⫺6.84, p⬍
.001. Hence, neither emotional experience appeared to mediate effects of
devaluation on the other.
Figure 1. General hurt proneness as a function of relationship commit-
ment and chronic perceptions of the partner’s valuing of the relationship
(Study 3).
995
HURT FEELINGS AND ANGER
Victims’ hurt feelings as a commitment signal. Next we
tested the prediction that victims’ hurt feelings function to signal
high levels of commitment to perpetrators, whereas victims’ anger
signals low commitment. Again, we tested this prediction in both
chronic and state forms. With regard to chronic processes, we
regressed perpetrators’ chronic perceptions of victims’ commit-
ment on victims’ self-reported general proneness to hurt feelings
and anger (all administered during the baseline session) using the
person-level modeling strategy described earlier. Victims’ self-
reported proneness to hurt feelings was positively associated with
perpetrators’ perceptions of victims’ commitment, b⫽0.19, t⫽
2.76, p⬍.01, whereas victims’ self-reported proneness to anger
was inversely associated with perpetrators’ perceptions of victims’
commitment, although this effect did not reach conventional sig-
nificance levels, b⫽⫺0.10, t⫽⫺1.51, p⫽.13.
A limitation of this analysis is that it assumes that perpetrators
correctly understood victims’ emotional tendencies, but errors in
understanding may have weakened the links between victims’
emotional tendencies and perpetrators’ perceptions of commit-
ment. Perpetrators’ perceptions of victims’ emotions may be more
proximally related to their responses (i.e., perpetrators must de-
code victims’ hurt as hurt before they infer high commitment).
Hence, in a second analysis, we replaced victims’ baseline self-
reports of their own hurt and anger proneness with perpetrators’
baseline perceptions of victims’ hurt and anger proneness. Con-
sistent with our predictions, perpetrators’ perception of victims’
hurt proneness was positively associated with perpetrators’ per-
ception of victims’ commitment, b⫽0.31, t⫽4.69, p⬍.001,
whereas perpetrators’ perception of victims’ anger proneness was
inversely associated with perpetrators’ perception of victims’ com-
mitment, b⫽⫺0.36, t⫽⫺6.13, p⬍.001. These results support
our view that hurt and anger communicate divergent messages
about victims’ commitment using trait measures of hurt and anger
proneness.
We tested these predictions in terms of states by using the
day-level model described earlier. The daily component permitted
us to examine these effects in a prospective manner that estab-
lished temporal precedence. We examined the effects of victims’
daily hurt feelings and anger on a given day (i.e., on day d), on
perpetrators’ subsequent (i.e., on day d⫹1) perceptions of vic-
tims’ commitment. To model change in this outcome, we con-
trolled for perpetrators’ initial (i.e., on day d) perceptions of
victims’ commitment. In addition, to rule out the alternative ex-
planation that these findings are explained by victims’ feeling
devalued by perpetrators (we reported associations of victims’
daily hurt feelings and anger with their daily perceptions of de-
valuation in preceding text), we also controlled for victims’ daily
perceptions of devaluation (i.e., on day d), although results sup-
ported our predictions regardless of whether we included this
control.
11
In the model using victims’ self-reports of their daily
emotion, victims’ anger tended to predict temporal declines in
perpetrators’ perceptions of victims’ commitment, b⫽⫺0.06, t⫽
⫺1.74, p⫽.08. In contrast, male victims’ hurt feelings tended to
predict temporal increases in female perpetrators’ perceptions of
victims’ commitment, b⫽0.09, t⫽1.93, p⫽.05. (The effect of
female victims’ hurt feelings was not significant, p⫽.18.)
Again, to address potential issues with perpetrators not detecting
victims’ emotions, we tested a second model in which we replaced
victims’ self-reports of their own emotions with perpetrators’
perceptions of victims’ emotions. Male perpetrators’ perceptions
of female victims’ anger predicted temporal decreases in their
perceptions of these victims’ commitment, b⫽⫺0.16, t⫽⫺4.51,
p⬍.001. (The effect was not significant for female perpetrators,
p⬎.86.) In addition, male and female perpetrators’ perceptions of
victims’ hurt feelings tended to predict temporal increases in their
perceptions of victims’ commitment, b⫽0.06, t⫽1.71, p⬍.09.
These results support our predictions at a state level—victims’
daily hurt feelings appeared to signal high commitment to perpe-
trators, whereas victims’ daily anger appeared to signal low com-
mitment.
Perpetrators’ responses to victims’ daily hurt feelings.
Next, using the same time-lagged “day-level” analysis strategy
described earlier, we examined perpetrators’ subsequent responses
to victims’ daily hurt feelings and anger (and perpetrators’ per-
ceptions of victims’ emotions). Whereas perpetrators likely be-
haved in ways that communicated devaluation on days they elic-
ited victims’ hurt feelings and anger, they are likely to respond
quite differently the following day as a function of whether victims
felt hurt or anger. We used three outcome measures to index
perpetrators’ responses, including their daily valuing of the victim
(i.e., care, commitment, and regard), their daily guilt, and their
daily anger. We examined the effects of victims’ daily hurt feel-
ings and anger on the perpetrator’s responses the following day.
We included the same controls described earlier. We controlled
for the lagged assessment of the criterion to model change. To
be sure that effects of hurt and anger were not explained by the
association with victims’ perceived devaluation, we controlled
for victims’ daily perceived devaluation, although, again, sup-
port for our predictions did not depend on including this con-
trol. These control variables were measured on the same day as
the emotion predictors.
11
We did not control for victims’ perceptions of devaluation in the trait
analyses described because that model used only concurrent measures. As
a result, including victims’ perceived devaluation as a control would render
the model nonrecursive (i.e., perceived partner commitment serving as both
a covariate, as part of the perceived devaluation composite, and an out-
come).
Figure 2. Daily hurt feelings as a function of relationship commitment
and daily perceptions of the partner’s valuing of the relationship (Study 3).
996 LEMAY, OVERALL, AND CLARK
Results of analyses of victims’ self-reported daily hurt feelings
and anger appear in the upper section of Table 9. Victims’ self-
reported daily hurt feelings predicted increases in perpetrators’
valuing of victims and decreases in perpetrators’ anger toward
victims the following day. In contrast, victims’ self-reported daily
anger predicted decreases in perpetrators’ valuing of victims and
increases in perpetrators’ anger toward victims the following day.
Again, we tested analogous models in which we replaced vic-
tims’ daily self-reports of emotion with perpetrators’ daily percep-
tions of victims’ emotion. Results are presented in the lower
section of Table 9. Perpetrators’ perceptions of victims’ hurt
predicted increases in their valuing of victims, increases in their
guilt, and decreases in their anger the following day. In contrast,
perpetrators’ perceptions of victims’ anger predicted decreases in
perpetrators’ valuing of victims (female victims only) and in-
creases in their anger toward victims the following day. These
results are consistent with our predictions. Victims’ hurt (and
perpetrators’ perceptions of this hurt) predicted perpetrators’ con-
structive responses the following day and reduced anger, whereas
victims’ anger (and perpetrators’ perceptions of this anger) had
opposite effects.
12
Summary. This study provided support for several of the
distinctions between hurt and anger outlined in Table 1. Victims’
perceived devaluation predicted hurt and anger, and victims’
chronic relationship commitment moderated the effect on hurt
feelings but not on anger, which suggests that vulnerability to hurt
reflects felt desire or need for a relationship with the perpetrator.
Consistent with this logic, victims’ hurt appeared to signal high
commitment to perpetrators, whereas victims’ anger appeared to
convey low commitment. Support for all of these predictions was
found using both trait measures of proneness to hurt and anger and
daily measures of hurt and anger. Lagged daily analyses also
suggested that perpetrators tried to make amends on days after they
hurt victims’ feelings (i.e., increased care for the partner, increased
commitment to the relationship, increased guilt, and decreased
anger), whereas perpetrators did not appear to make amends, and
experienced more anger, on days after they made victims’ angry.
These results are consistent with our predictions that victims’ hurt
feelings and anger would elicit divergent responses from perpe-
trators.
Study 4
In our final study, we examined the behavioral responses that
follow from hurt and anger within ongoing interactions using a
behavioral observation measure. Members of romantic dyads were
asked to discuss ways they wanted their partner to change, which
could elicit the partner’s hurt feelings or anger. These interactions
were recorded, and we coded relationship-constructive and
relationship-destructive behavioral responses. We tested predic-
tions regarding the associations of victims’ hurt feelings and anger
with victims’ and perpetrators’ behavior. In addition, we attempted
to replicate the finding reported previously suggesting that com-
mitment involves a vulnerability to hurt feelings but not anger.
Method
Participants. One hundred and eighty heterosexual couples
responded to article and electronic announcements posted across a
New Zealand university and associated student-based organiza-
tions. All notices informed potential participants that couples had
to have been involved for at least 1 year. Sixty-one percent of the
sample were living together or married, and the mean length of
relationships was 2.95 years (SD ⫽2.26). Participants ranged from
18 to 45 years of age (M⫽23.07, SD ⫽4.18). Couples were paid
NZ$70 for participation in a 3-hr session.
Procedure. Participants completed scales assessing relation-
ship commitment and perceptions of the partner’s commitment and
then identified and ranked in order of importance three aspects of
their partner that they wanted to improve and which they under-
stood would be discussed in the next phase of the study. The most
important ranked feature from each dyad member was selected for
discussion. Couples then had two 7-min discussions that were
unobtrusively recorded. In one discussion, couples discussed the
feature the man wanted to change about his female partner, and in
the other, couples discussed the feature the woman wanted to
change about her male partner. Order of discussion was counter-
balanced across couples. After completing both discussions, part-
ners were directed to separate rooms where they reviewed their
discussions and reported on their hurt and anger during the dis-
cussion (among other questions). This type of review procedure
provides a sensitive measure of people’s subjective emotions dur-
ing their discussions (see Welsh & Dickson, 2005).
Measures.
Relationship commitment. Responses to five items from the
Investment Model Scale (Rusbult et al., 1998) were averaged to
index participants’ commitment to their relationship (e.g., “I am
committed to maintaining my relationship with my partner”).
Items were completed using 7-point response scales (1 ⫽not at
all,7⫽extremely; Cronbach’s ␣⫽.88 and .86 for women and
men, respectively).
Perception of the partner’s commitment. Participants com-
pleted an analogous five items to assess perceptions of their
partner’s commitment (e.g., “My partner is committed to main-
taining his/her relationship with me”). Items were completed using
12
Hurt and anger were positively associated at the daily level, anger ¡
hurt b⫽0.68, t⫽34.69, p⬍.001, as were perpetrators’ perceptions of
victims’ hurt and anger, perceived partner anger ¡perceived partner hurt
b⫽0.53, t⫽26.83, p⬍.001. We conducted additional analyses of
perpetrators’ responses using only hurt or anger to examine the effects of
ignoring these associations. In models examining effects of victims’ self-
reported emotions, effects of victims’ hurt and anger on perpetrators’
subsequent valuing were no longer significant, ps⬎.26, the effect of
victims’ anger on perpetrators’ subsequent anger was no longer significant,
p⬎.47, and the effect of victims’ hurt on perpetrators’ subsequent anger
became marginal, b⫽⫺0.06, p⬍.10. In models examining perpetrators’
perceptions of victims’ emotions, the effect of perceived hurt on perpetra-
tors’ subsequent valuing was no longer significant, p⬎.13, the effect of
perceived hurt on perpetrators’ subsequent guilt remained significant, b⫽
0.11, p⬍.05, the effect of perceived hurt on perpetrators’ subsequent
anger was marginal, b⫽⫺0.06, p⬍.10, the effect of perceived anger on
male perpetrators’ subsequent valuing remained significant, b⫽⫺0.08,
p⬍.05, and the effect of perceived anger on perpetrators’ anger was not
significant, p⬎.30. Hence, about half of the effects were no longer
observed. For these, victims’ hurt appeared less constructive and victims’
anger appeared less destructive when the direct effects of each emotion
were not distinguished from their indirect effects via the association
between the two emotions. These results underscore the importance of
examining the independent effects of hurt and anger.
997
HURT FEELINGS AND ANGER
the same 7-point response scales (Cronbach’s ␣⫽.89 and .78 for
women and men, respectively).
Assessing hurt and anger within the discussion. Participants
watched both of their discussions in 30-s segments. At the end of
each 30-s section of the discussion, the recording was stopped, and
using the same 7-point response scales, participants rated the
degree to which they felt “angry” and “hurt” during that segment
of the discussion.
Coding behavior. Six trained coders who were unaware of
participants’ responses to self-report measures independently rated
the extent to which each partner exhibited destructive and con-
structive communication behavior. The specific behaviors were
selected for their consistency across major coding systems of
relationship conflict behavior and have been shown to predict
important relationship outcomes, such as problem resolution (see
Overall, Fletcher, Simpson, & Sibley, 2009). Ratings of destruc-
tive communication captured harsh and aggressive strategies, such
as derogating and blaming the partner, being hostile and demand-
ing, and invalidating or rejecting the partner. Ratings of construc-
tive communication indexed the degree to which intimates at-
tempted to soften conflict and maintain positivity via affection and
positive affect (e.g., affiliative humor), accommodation, and vali-
dation.
Coders were given a detailed description of destructive and
constructive communication, including a list of associated tac-
tics, and were instructed to take into account the frequency,
intensity, and duration of each set of behaviors for each 30-s
segment of the discussion (1–2 ⫽low, 3–5 ⫽moderate,6–7⫽
high). The behaviors exhibited by men and women were coded
in separate viewings. For half of the discussions, men were coded
first; for the other half, women were coded first. Between two and
four coders from the team of six rated each participant, and ratings
were averaged across coders to index the amount of destructive
(intraclass correlation coefficient [ICC] ⫽.91) and constructive
(ICC ⫽.86) behavior exhibited in each 30-s segment of the
discussion.
Results
Analysis strategy. We used the individual discussion seg-
ments (14 per discussion) as the units of analysis. Our multilevel
analysis strategy was similar to the day-level strategy we used in
Study 3, except that discussion segments, rather than days, were
the repeated measurements at the lowest level of analysis. The
models estimated an intercept for each discussion and variances of
these intercepts across dyads. These variance estimates account for
the interdependence due to repeated segments belonging to the
same discussion. In addition, the models estimated the covariance
between these two intercepts within the same dyad, which models
the interdependence due to having each dyad participate in two
discussions. A repeated statement allowed for separate residual
error variances (at the level of the discussion segment) for men and
women and estimated the covariance between these errors. To
address model convergence problems, we modeled the slopes as
fixed.
Gender interactions were examined for all effects. When they
were significant, separate conditional effects for men and women
are reported. Again, we used the terms perpetrators and victims to
clarify presentation of results, although all participants provided
data relevant to both roles. Perpetrators were participants who
were discussing ways they would like their partner to change.
Victims were their romantic partners, whose hurt feelings and
anger were of interest.
Commitment as a hurt vulnerability. We expected that vic-
tims’ commitment would amplify effects of perpetrators’ destruc-
tive behavior on victims’ hurt feelings. To test this prediction, we
regressed victims’ hurt feelings on perpetrators’ destructive behav-
ior during the same interaction segment, victims’ chronic relation-
ship commitment, and a product term representing their interac-
tion. Consistent with the model in Study 3, we controlled for
victims’ anger to examine the unique aspect of hurt and to be sure
that findings were not explained by anger. Predictors were grand-
mean centered. The Perpetrators’ Destructive Behavior ⫻Victims’
Commitment interaction predicted victims’ hurt feelings, b⫽0.03,
Table 9
Effects of Victims’ Hurt and Anger (Upper Section) and Perpetrators’ Perceptions of Victims’ Hurt and Anger (Lower Section) on
Changes in Perpetrators’ Valuing of Victims, Guilt, and Anger (Study 3)
Predictor
Perpetrators’ response
Valuing partner Guilt Anger
Models examining victims’ self-reported emotion
Lagged criterion ⫺.07 (⫺1.98) ⫺.12 (⫺3.21)
ⴱⴱ
⫺.11 (⫺3.47)
ⴱⴱⴱ
Victims’ perceived valuation .13
M
(2.57)
ⴱ
;⫺.04
W
(⫺0.72) ⫺.09 (⫺1.72)
†
⫺.01 (⫺0.21)
Victims’ self-reported hurt .09 (2.36)
ⴱ
.03 (0.49) ⫺.14 (⫺3)
ⴱⴱ
Victims’ self-reported anger ⫺.08 (⫺2.28)
ⴱ
.05 (0.98) .12 (2.63)
ⴱⴱ
Models examining perpetrators’ perceptions of victims’ emotion
Lagged criterion ⫺.07 (⫺1.82)
†
⫺.15 (⫺3.89)
ⴱⴱⴱ
⫺.14 (⫺4.05)
ⴱⴱⴱ
Victims’ perceived valuation .16
M
(3.47)
ⴱⴱⴱ
;⫺.04
W
(⫺0.93) ⫺.10 (⫺2.10)
ⴱ
⫺.01 (⫺0.33)
Perpetrators’ perception of victims’ hurt .07 (2.11)
ⴱ
.12 (2.08)
ⴱ
⫺.10 (⫺2.3)
ⴱ
Perpetrators’ perception of victims’ anger .05
M
(1.18); ⫺.11
W
(⫺2.97)
ⴱ
.0 (⫺0.05) .07 (1.97)
ⴱ
Note. Unstandardized coefficients appear outside parentheses, and tvalues appear inside parentheses. Coefficients with the subscript “M” are coefficients
for male victims. Coefficients with the subscript “W” are coefficients for female victims. Coefficients without subscripts did not significantly vary across
gender. Degrees of freedom ranged from 672 to 923.
†
p⬍.10.
ⴱ
p⬍.05.
ⴱⴱ
p⬍.01.
ⴱⴱⴱ
p⬍.001.
998 LEMAY, OVERALL, AND CLARK
t⫽1.95, p⫽.05. Predicted values are presented in Figure 3.
Conditional effects of perpetrators’ destructive behavior were ex-
amined at low (1 SD below the mean) and high (1 SD above the
mean) levels of victims’ relationship commitment. Perpetrators’
destructive behavior was positively associated with victims’ hurt
feelings for high-commitment victims, b⫽0.04, t⫽1.85, p⬍.07,
but not for low-commitment victims, p⬎.72. This pattern sup-
ports our prediction.
Did commitment also involve a vulnerability to anger? We
tested a similar model in which we predicted victims’ anger from
perpetrators’ destructive behavior and victims’ commitment while
controlling for victims’ hurt. The Destructive Behavior ⫻Com-
mitment interaction also was significant in this model, b⫽⫺0.03,
t⫽⫺2.31, p⬍.05. Predicted values are plotted in Figure 4.Ina
pattern opposite to that reported previously, perpetrators’ destruc-
tive behavior was more strongly predictive of victims’ increased
anger for low-commitment victims, b⫽0.10, t⫽6.33, p⬍.001,
relative to high-commitment victims, b⫽0.05, t⫽2.65, p⬍.01.
Hence, commitment was associated with more hurt reactivity but
less anger reactivity.
Behavioral responses to hurt and anger. Next we tested
prospective effects of victims’ hurt and anger on perpetrators’ and
victims’ behavioral responses. These models examined effects of
victims’ hurt feelings and anger in a given discussion segment
(segment t) on victims’ and perpetrators’ constructive and destruc-
tive behavior in the following segment (segment t⫹1). To model
residualized change, we controlled for the assessment of the be-
havioral criterion occurring during the same segment as the vic-
tims’ emotion predictors (during segment t). Consistent with the
day-level analyses in Study 3, we person-centered the predictors to
isolate within-person effects.
We expected that victims’ anger, but not hurt, would predict
their own subsequent destructive responses, and we speculated that
victims’ hurt might predict their own subsequent constructive
responses. As shown in the first two columns of Table 10, victims’
anger predicted decreases in victims’ constructive behavior and
increases in their destructive behavior during the subsequent dis-
cussion segment. In contrast, victims’ hurt did not predict victims’
constructive behavior, and for male victims, their hurt predicted
decreases in their destructive behavior. We also expected that
victims’ anger would predict perpetrators’ destructive responses,
whereas victims’ hurt would predict perpetrators’ constructive
responses. As shown in the last two columns of Table 10, these
predictions were supported.
Summary. This study provided additional support for our
predictions using behavioral observation measures. Victims’ rela-
tionship commitment predicted increased hurt reactivity and de-
creased anger reactivity in response to perpetrators’ destructive
behavior, indicating that hurt and anger vary in the extent to which
they reflect dependence and relationship maintenance concerns. In
addition, victims’ hurt predicted increases in perpetrators’ con-
structive reactions, whereas victims’ anger predicted increases in
their own and perpetrators’ destructive reactions.
General Discussion
The experiences, consequences, and functions of hurt feelings
and anger have not been clearly differentiated in prior research.
Consistent with our predictions outlined in Table 1, our findings
suggest that the experience of hurt feelings involves perceived
devaluation by a needed partner and a desire to restore connection
and security in the partner’s care for the self. Expressing hurt
conveys victims’ relationship dependence and need for support to
perpetrators, and then perpetrators tend to enact constructive re-
sponses that contribute to relationship repair. Hence, an important
social function of hurt appears to be the expression and mainte-
nance of interpersonal connection. Anger, in contrast, involves a
sense of independence and control, is associated with the motiva-
tion to coerce perpetrators to change their behavior and seems to
further exacerbate relationship difficulties over and above perpe-
trators’ initial behavior. In following text, we review the support-
ive evidence from our four studies, which provide strong and
consistent support for the distinctions between hurt and anger
outlined in Table 1.
The Victim’s Experience
The results across four studies supported our view that the
capacity to feel hurt reflects victims’ vulnerability to psychological
pain at the hands of the perpetrator, and this vulnerability is the
result of wanting or needing a relationship with the perpetrator.
Figure 4. Victims’ anger as a function of victims’ relationship commit-
ment and perpetrators’ destructive behavior (Study 4).
Figure 3. Victims’ hurt feelings as a function of victims’ relationship
commitment and perpetrators’ destructive behavior (Study 4).
999
HURT FEELINGS AND ANGER
Further, we proposed that an experience of hurt occurs when this
vulnerability is met with perceived devaluation by the perpetrator.
Accordingly, we expected that the experience of hurt would be
characterized by a subjective sense of vulnerability to pain and
dependence on a relationship with the perpetrator. Moreover, hurt
should be associated with reduced perceptions of control. Given
their dependence on acceptance, which is currently in question,
hurt individuals should feel powerless. In addition, given the sense
of wanting and needing a relationship with the perpetrator, we
expected that the primary goal characterizing the experience of
hurt would involve restoring the perpetrator’s positive valuing of
the relationship. As such, hurt victims should be unlikely to engage
in destructive behavior that would further jeopardize the relation-
ship and may even engage in constructive behavior. In contrast,
many prior studies have suggested that anger is characterized by
perceptions of independence, invulnerability, and control (for a
review, see Lerner & Tiedens, 2006) and that anger involves the
goal of changing the perpetrator’s aversive behavior through co-
ercion or punishment (Fischer & Roseman, 2007). Consequently,
anger should predict victims’ destructive behavior.
Our findings support our view that hurt and anger differ in these
ways. In Studies 1a and 1b, the intensity of victims’ hurt feelings
in prior events was positively associated with their subjective
sense of vulnerability and dependence during those events, and it
was negatively associated with perceptions of control and power.
Moreover, victims who were high in commitment to perpetrators
exhibited greater proclivity for hurt feelings relative to victims
who were low in commitment, suggesting that hurt arises out of
victims’ desires to maintain relationships. This pattern was found
using self-reported general hurt proneness in response to global
perceptions of devaluation (Study 3), daily hurt feelings in re-
sponse to daily perceptions of devaluation (Study 3), and momen-
tary hurt in response to destructive behavior within ongoing inter-
actions (Study 4). Findings regarding goals also were consistent
with predictions. The intensity of hurt was associated with goals to
restore perpetrators’ acceptance and valuing of the relationship
(Studies 1a, 1b, and 2). In accordance with such goals, the intensity
of victims’ hurt did not predict victims’ destructive responses in
any of our studies (Studies 1a, 1b, 2, or 4). In fact, in two studies
(Studies 1a and 2), victims’ hurt predicted greater constructive
behaviors. Nonetheless, effects of hurt on constructive behavior
were small and inconsistent, which may be a result of the uncer-
tainty involved with experiences of hurt. On the one hand, hurt
victims may value the relationship and wish to enact behaviors that
could repair it. On the other, hurt victims often may be uncertain
of whether repair attempts would be welcomed or effective, and so
they may often restrict their constructive responses until the per-
petrator reaffirms interest in the relationship (see also MacDonald
& Leary, 2005). Indeed, hurt is characterized by subjective uncer-
tainty (Fitness & Warburton, 2009).
In contrast to hurt, the intensity of victims’ anger was associated
with greater control and power and reduced vulnerability and
dependence (Studies 1a and 1b). Anger was not associated with
victims’ commitment in one study (Study 3) and was inversely
associated with commitment in the other (Study 4). Hence, anger
did not involve the same sense of wanting or needing a relationship
with the perpetrator. Anger also was characterized by goals to
coerce perpetrators to change their behavior (Studies 1a, 1b, and
2), and seemed to reflect lack of concern for the perpetrator’s
sentiments, as victims’ intensity of anger was inversely associated
with their goals to improve the perpetrator’s valuing of the rela-
tionship (Studies 1a and 1b). These findings are also consistent
with our argument that anger involves a focus on immediate
behavior and a myopic oversight of broader relationship concerns.
In accordance with the goal to coerce behavioral change and with
their disregard for restoring acceptance, victims’ anger intensity
predicted their destructive behavior (Studies 1a, 1b, 2, and 4), a
pattern that is well established in prior research (e.g., Averill,
1982;deRivera & Grinkis, 1986;Fischer & Roseman, 2007;Frijda
et al., 1989;Roseman et al., 1994).
Hence, hurt and anger differed in appraisals, goals, and behav-
iors, which are indicators that they serve different social functions
(Fischer & Manstead, 2008;Fischer & Roseman, 2007;Frijda et
al., 1989;Frijda & Parrott, 2011;Lazarus, 1991b;Roseman et al.,
1994;Simon, 1967;Tooby & Cosmides, 1990). These findings
support our arguments that the social function of hurt is to restore
acceptance by valued relationship partners, whereas the social
function of anger is the control of others’ behavior to render the
environment more consistent with one’s needs or desires, which do
not appear to involve acceptance by others.
Perpetrators’ Responses and Interpersonal
Consequences
In many cases, the social functions of emotions involve influ-
encing and conveying messages to others (Clark et al., 1987;
Keltner & Haidt, 1999;Van Kleef, 2010). As outlined in Table 1,
we proposed that the consequences and functions of hurt and anger
involve such dyadic processes. Specifically, we predicted that
victims’ hurt would signal to perpetrators that victims valued the
Table 10
Effects of Victims’ Emotions on Victims’ Subsequent Constructive and Destructive Behavior (Study 4)
Predictor
Victims’ subsequent behaviors Perpetrators’ subsequent behaviors
Constructive Destructive Constructive Destructive
Lagged criterion .18 (11.92)
ⴱⴱⴱ
.22 (14.48)
ⴱⴱⴱ
.17 (10.64)
ⴱⴱⴱ
.24 (15.86)
ⴱⴱⴱ
Perpetrators’ destructive behaviors ⫺.02 (⫺1.65)
†
.03 (2.31)
ⴱ
⫺.05 (⫺3.17)
ⴱⴱ
—
Victims’ hurt 0 (⫺.27) ⫺.06
M
(⫺2.89)
ⴱⴱⴱ
;0
W
(⫺0.20) .03 (2.45)
ⴱ
⫺.02 (⫺1.34)
Victims’ anger ⫺.03 (⫺1.87)
†
.07 (5.05)
ⴱⴱⴱ
0(⫺0.08) .04 (2.72)
ⴱⴱ
Note. Unstandardized coefficients appear outside parentheses, and tvalues appear inside parentheses. Coefficients with the subscript “M” are coefficients
for men. Coefficients with the subscript “W” are coefficients for women. Coefficients without subscripts did not significantly vary across gender.
†
p⬍.10.
ⴱ
p⬍.05.
ⴱⴱ
p⬍.01.
ⴱⴱⴱ
p⬍.001.
1000 LEMAY, OVERALL, AND CLARK
relationship and would elicit perpetrators’ constructive responses,
including feelings of empathic concern and guilt and actions to
make amends. In contrast, because anger communicates indepen-
dence, a motivation to control others, and antagonistic urges, we
expected that victims’ anger would signal to perpetrators that
victims do not value the relationship or care for perpetrators’
welfare, and elicit perpetrators’ reciprocated anger and defensive,
destructive behavior.
Our studies provided strong support for these predictions. Vic-
tims’ hurt predicted constructive responses by perpetrators, includ-
ing empathy, guilt, and constructive behavior, in retrospective
accounts (Study 2). In our daily report study (Study 3), perpetra-
tors were more caring, admiring, committed, and guilty on days
following victims’ hurt feelings or perpetrators’ perceptions of
victims’ hurt. In our behavioral observation study, perpetrators
enacted more constructive behavior during ongoing interactions
with victims after victims felt hurt (Study 4). In contrast, victims’
anger predicted perpetrators’ reduced constructive responses
(Studies 3 and 4) and increased destructive responses (Studies 2, 3,
and 4).
Moreover, our results supported the prediction that victims’ hurt
and anger would convey different messages regarding victims’
relationship concerns. Victims’ hurt predicted perpetrators seeing
victims as desirous of their acceptance and more committed in
retrospective accounts (Study 2). Similarly, victims’ chronic
proneness to hurt feelings was associated with perpetrators’
chronic perceptions of victims’ commitment in our dyadic study
(Study 3), and victims’ daily hurt feelings predicted perpetrators
perceiving victims as more committed to the relationship the
following day in that same study (Study 3). These findings suggest
that based on their own conceptual knowledge of hurt, perpetrators
intuitively know that the propensity to feel hurt signals that the
victim values the relationship. So when victims express hurt, they
communicate commitment to the perpetrator. In contrast, victims’
anger intensity predicted perpetrators seeing victims as less com-
mitted and wanting to control them (Studies 2 and 3). These
findings strongly suggest that hurt conveys dependence and moti-
vates perpetrators’ constructive responses, whereas anger signals
threat (i.e., conveys destructive urges and withdrawal from recon-
nection) and motivates perpetrators’ defensive retaliation.
These diverging responses from perpetrators likely influence the
impact of hurtful and angry events on interpersonal relationships.
In Study 2, we found that victims’ hurt intensity predicted more
positive perceptions of relationship consequences (e.g., more trust
in partners, more care for the relationship), which was explained
by perpetrators’ and victims’ constructive responses to hurt,
whereas victims’ anger predicted negative consequences, which
was explained by perpetrators’ destructive responses to anger.
These findings contrast with prior findings suggesting that hurtful
events generally have negative interpersonal consequences (Leary
et al., 1998). Methodological differences may have contributed to
these effects. Our studies examined associations with intensity of
both hurt and anger. Moreover, in Study 2, which assessed rela-
tionship consequences, we restricted participants’ selection of
hurtful events to relatively recent events, which may have resulted
in a sample of events that were more typical and less destructive.
These findings highlight the relationship-protective function of
hurt. When people strongly value a relationship, they have much to
lose from its termination, including the resources and security they
need and desire (see Baumeister & Leary, 1995;Murray et al.,
2006;Rusbult, 1980). In contrast, the termination of undesired or
unnecessary relationships may be of little consequence. Hence,
mechanisms that protect and maintain relationships should occur
primarily when those relationships are valuable (Baumeister &
Leary, 1995;Keltner, Haidt, & Shiota, 2006;Leary & Baumeister,
2000). Our findings regarding victims’ subjective experiences,
perpetrators’ responses, and relationship consequences suggest
that hurt feelings, but not anger, may be one such mechanism (see
also MacDonald & Leary, 2005;Shaver et al., 2009). Anger
instead appears to involve a mechanism of controlling others’
behavior to make the social environment more conducive to angry
victims’ needs or desires. However, given that the focus of anger
is on coercing behavioral change and not on restoring acceptance,
anger appears to motivate unresponsive behavior that further
threatens relationships.
Some theoretical perspectives suggest that expressing negative
emotions builds intimate relationships (see Clark et al., 2001;
Graham et al., 2008). The reasoning behind these views is that this
expression conveys needs and a willingness to reveal vulnerabil-
ities. We are in general agreement with this view. However, the
current findings suggest that specific emotions also convey one’s
stance with regard to the relationship. Frameworks regarding the
interpersonal consequences of emotions could benefit from a con-
sideration of both the needs and relationship motives that are
communicated by the experience and expression of specific emo-
tions.
Boundary Conditions, Strengths and Limitations, and
Future Directions
The characteristics we described for hurt and anger should be
viewed as probabilistic rather than absolute. Many events will not
fit the prototypes we described (see Russell & Fehr, 1994). Situ-
ational, relationship, and dispositional factors undoubtedly impact
their experiences and consequences. For example, people may
modulate expression of emotion as a result of contextualized
predictions of the consequences of expression. People may sup-
press feelings of hurt to avoid expressing vulnerabilities that could
be exploited by others, suppress feelings of anger to avoid pun-
ishment by powerful perpetrators, express hurt to manipulate oth-
ers’ sympathy, and express anger to gain control (see Gross, 1998;
Tiedens, 2001). Subjective goals associated with hurt and anger
also may vary as a function of context. For example, hurt victims
may usually not retaliate because they feel dependent on the
perpetrator’s affections and want to restore acceptance. In some
cases, however, a hurtful act may leave victims relatively certain
that they are not valued and the relationship is permanently dam-
aged. Under these circumstances, hurt victims may distance them-
selves from perpetrators (see Feeney, 2004;Vangelisti & Young,
2000). These events should feature especially high-intensity hurt
feelings, although we would guess that distancing would be a
result of the certainty of loss rather than hurt per se. Individual
differences also may contribute to these perceptions. People who
are chronically insecure about acceptance, such as people with low
self-esteem or high attachment anxiety, may be more prone to
interpret an act as unequivocal evidence that they are devalued
(Murray et al., 2006).
1001
HURT FEELINGS AND ANGER
Responses of perpetrators also may vary. For example, although
perpetrators of hurt usually value relationships with victims, and
their hurtful acts often occur by accident, thoughtlessness, or
insensitivity (Leary & Springer, 2001), sometimes perpetrators
engage in hurtful acts because they truly do not value a relation-
ship with victims. In such cases, victims may be better off sup-
pressing hurt, as perpetrators may not exhibit constructive re-
sponses to the request for reassurance implicit in victims’
expression of hurt, which could exacerbate victims’ distress. Prior
research suggests that a desire for a relationship determines how
people respond to others’ distress (Clark et al., 1987). Hence,
examining the role of perpetrators’ relationship motivations in
responses to hurt would be an important avenue for future re-
search.
We examined victims’ commitment to a relationship with the
perpetrator as a moderator variable because we believe that hurt
feelings often involve threat to a desired and needed relationship.
However, sometimes victims feel hurt not because they value a
relationship with the perpetrator, but because the hurtful act threat-
ens victims’ sense of being valued by other relationship partners
(Vangelisti et al., 2005). Indeed, hurt victims often report that the
hurtful events threaten their sense of social desirability (Leary et
al., 1998). This may explain why even low-commitment victims
showed some vulnerability to hurt. In addition, sometimes people
feel hurt by behaviors that seem unrelated to relational devaluation
(e.g., expressions of jealousy), suggesting that hurt may be elicited
by other threats to relationships (Feeney, 2005). Hence, further
delineation of the elicitors of hurt is an important topic for future
research.
Many of our findings suggested divergent experiences and con-
sequences of hurt and anger even at the zero-order level (i.e., when
the other emotion was not controlled). All of the findings from
Studies 1a, 1b, and 4 and many of the findings from Studies 2 and
3 did not vary as a function of whether the other emotion was
controlled. However, in a few cases in Studies 2 and 3, our clearest
findings emerged when we used methods that allowed us to
examine the direct (i.e., unique) effects of hurt and anger, and this
was because reports of hurt and anger were correlated. Examining
direct effects allowed us to understand how hurt and anger are
different—what features of experience and interactions are a result
of hurt and what features are a result of anger—even when some
experiences are blends of hurt and anger or involve transitions
between them. In such episodes, people may walk along a fine line
in response to partners’ misdeeds, hovering between feeling and
maintaining dependence and vulnerability on one side, and striving
for independence and control on the other. The factors that con-
tribute to these blends and how they guide victims’ and perpetra-
tors’ responses are intriguing directions for future research.
Our findings have important practical implications and support
therapeutic interventions that focus on emotions to treat relation-
ship distress. Emotion-focused couple therapy (Johnson, 2004), for
example, involves reducing damaging interaction cycles that ac-
company hostile emotions (such as anger) by helping partners
recognize and express more vulnerable emotions (such as hurt). By
encouraging couples to share their vulnerabilities, emotion-
focused couple therapy aims to help partners become more emo-
tionally engaged, compassionate, and responsive to each other’s
needs. Our results support that identifying and fostering the ex-
pression of more vulnerable emotions, such as hurt, is likely to
have positive consequences for the relationship by eliciting repar-
ative efforts by perpetrators, thereby building care and trust. In
contrast, although anger might help individuals retain personal
control and punish partners’ transgressions, anger is most likely to
have interpersonal costs by escalating destructive interactions.
Future research should test the long-term relationship conse-
quences of these processes.
We consistently referred to “hurt” and “anger” as emotional
experiences. Whereas anger is featured in many taxonomies of
emotion, hurt is understudied in the emotions literature. Our aim in
the current research was not to qualify hurt feelings as an emo-
tional experience in terms of criteria proposed by other emotions
scholars, although hurt may qualify according to some standards.
For example, some have proposed as a definition a variant of the
following: an emotional experience is a valenced state that is (a)
appraised as personally significant, (b) has an object (i.e., the state
is about something), and, hence, (c) motivates a problem-focused
response (e.g., Clore & Ortony, 2000;Frijda, 1986;Lazarus,
1991b). Hurt appears highly negatively valenced, given its strong
association with emotional distress (Leary et al., 1998). Moreover,
the current findings clearly indicate that hurtful experiences are
personally significant (i.e., threat to an important relationship) and
have an object and a problem-focused motivational state (i.e.,
perceiving devaluation by a valuable relationship partner and
wanting to restore acceptance). Hence, hurt would qualify as an
emotion according to these criteria. Of course, other theorists have
different criteria, and this is beyond the scope of the current
research. We expect that, with further studies dedicated to the
issue, hurt feelings will pass the test(s) for emotional experience.
In addition, our aim was not to show how hurt is different from
other emotional experiences beyond anger. Hurt and sadness may
be positively associated in some contexts and have similar inter-
personal consequences (see Sanford, 2007;Sanford & Rowatt,
2004;Shaver et al., 1987). Hurt also may overlap with other
negative emotions such as anxiety and shame (Vangelisti &
Young, 2000). As argued by others (see Feeney, 2005;Fitness &
Warburton, 2009;Leary & Leder, 2009;Shaver et al., 2009;
Vangelisti, 2009); however, we believe that hurt is a unique
emotional experience. Sadness represents a sense of finality that
we do not ascribe to hurt. It is characterized by the appraisal that
events are beyond anyone’s control (Smith & Ellsworth, 1985), a
sense of irrevocable loss (Lazarus, 1991a), and a reduction in
active goal pursuit (Roseman et al., 1994). In contrast, our findings
suggest that hurt involves the wish for pain relief—the hope and
desire to restore feelings of value and connection. Anxiety is
characterized by a desire to avoid an unpleasant outcome (Rose-
man et al., 1994;Smith & Ellsworth, 1985), whereas hurt usually
occurs when an unpleasant outcome has already been experienced.
In addition, hurt and anger seem primarily social—inflicted by
another person—whereas sadness and anxiety can occur without a
human agent (see Roseman, Spindel, & Jose, 1990). A number of
investigations, beyond our own, provide empirical evidence for
viewing hurt as a unique experience. Leary and Leder (2009) have
shown that hurt has unique variance that cannot be explained by
other negative emotions and that the size of this unique variance is
comparable to the unique variance of many other emotions (e.g.,
fear, hostility, sadness, and guilt). Hence, hurt seems to be no more
a “blend” than these others. These authors also found that the
appraisal that one was not valued as much as one had desired
1002 LEMAY, OVERALL, AND CLARK
distinguished hurt from anger, sadness, anxiety, and guilt. More-
over, Feeney (2005) reported evidence suggesting that hurt is
distinct from other emotions in terms of themes regarding personal
injury. Fitness and Warburton (2009) differentiated hurt from
anger and sadness on a number of conventional appraisal dimen-
sions. These findings, along with the current findings, suggest that
hurt is a unique experience.
Finally, although we found consistent evidence suggesting that
relational devaluation elicits both hurt and anger, our research does
not speak to the subtler features of the eliciting situation that may
lead one to feel one of these emotions over the other. Such features
can prime preexisting knowledge about hurt or anger, which may
impact current emotional experience and labeling (Barrett, 2006),
and cause people to appraise the current situation in fundamentally
different and emotionally evocative ways (e.g., Clore & Ortony,
2000). Feeney (2005) has provided some initial support for the
utility of examining finer details of eliciting events. Retrospective
accounts of hurt feelings featured themes of relational devaluation,
but specific forms of relational devaluation were related to reports
of other emotions alongside hurt. Reports of anger were overrep-
resented when perpetrators engaged in infidelity and deception,
and they were underrepresented when perpetrators engaged in
active and passive disassociation (i.e., explicit rejection, abandon-
ment, ignoring the victim). One possibility is that people report
both hurt and anger when the specific form of devaluation involves
behaviors that victims feel they can rightfully punish or otherwise
control (i.e., deception and infidelity), perhaps because these are
seen as unjust or immoral (see Roseman, 1984). Another possibil-
ity is that certain forms of devaluation may be so severe (i.e.,
infidelity) that they reduce victims’ concerns about the relation-
ship, perhaps promoting a shift to anger.
Conclusion
Our findings suggest that angry individuals have the desire and
conviction to control others’ behavior, often through antagonistic
coercion. Although anger might often serve the purpose of con-
trolling others, our results suggest that the antagonistic elements of
anger will decrease perpetrators’ confidence in victims’ commit-
ment and motivate perpetrators to respond in ways that exacerbate
conflict, including reduced constructive responses and increased
aggression. In contrast, hurt does not involve the same reverber-
ation of interpersonal difficulties. Our results suggest that hurt
facilitates repair of needed relationships. Specifically, hurt indi-
viduals want to restore the perpetrator’s valuing, feel dependent on
the perpetrator’s intentions, and feel hurt especially when they
need and desire the relationship. Thus, hurt does not motivate
destructive responses. Furthermore, hurt signals to perpetrators
that victims care about the relationship and elicits perpetrators’
constructive responses, which have benefits that appear to offset
hurt feelings, including increasing care and trust. The repair func-
tion of hurt appears truly interpersonal—it can be seen not only in
the appraisals, goals, and behaviors of hurt victims but also in the
thoughts, feelings, and behaviors of perpetrators.
References
Algoe, S. B., Haidt, J., & Gable, S. L. (2008). Beyond reciprocity: Grati-
tude and relationships in everyday life. Emotion, 8, 425–429. doi:
10.1037/1528-3542.8.3.425
Averill, J. R. (1982). Anger and aggression: An essay on emotion. New
York, NY: Springer–Verlag. doi:10.1007/978-1-4612-5743-1
Bachman, G. F., & Guerrero, L. K. (2006). Relational quality and com-
municative responses following hurtful events in dating relationships:
An expectancy violations analysis. Journal of Social and Personal
Relationships, 23, 943–963. doi:10.1177/0265407506070476
Baron, R. A. (1971a). Aggression as a function of magnitude of victim’s
pain cues, level of prior anger arousal, and aggressor–victim similarity.
Journal of Personality and Social Psychology, 18, 48–54. doi:10.1037/
h0030643
Baron, R. A. (1971b). Magnitude of victim’s pain cues and level of prior
anger arousal as determinants of adult aggressive behavior. Journal of
Personality and Social Psychology, 17, 236–243. doi:10.1037/h0030595
Barrett, L. F. (2006). Solving the emotion paradox: Categorization and the
experience of emotion. Personality and Social Psychology Review, 10,
20–46. doi:10.1207/s15327957pspr1001_2
Bartlett, M. Y., & DeSteno, D. (2006). Gratitude and prosocial behavior:
Helping when it costs you. Psychological Science, 17, 319–325. doi:
10.1111/j.1467-9280.2006.01705.x
Batson, C. D., Eklund, J. H., Chermok, V. L., Hoyt, J. L., & Ortiz, B. G.
(2007). An additional antecedent of empathic concern: Valuing the
welfare of the person in need. Journal of Personality and Social Psy-
chology, 93, 65–74. doi:10.1037/0022-3514.93.1.65
Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire for
interpersonal attachments as a fundamental human motivation. Psycho-
logical Bulletin, 117, 497–529. doi:10.1037/0033-2909.117.3.497
Baumeister, R. F., Stillwell, A. M., & Heatherton, T. F. (1994). Guilt: An
interpersonal approach. Psychological Bulletin, 115, 243–267. doi:
10.1037/0033-2909.115.2.243
Baumeister, R. F., Stillwell, A., & Wotman, S. R. (1990). Victim and
perpetrator accounts of interpersonal conflict: Autobiographical narra-
tives about anger. Journal of Personality and Social Psychology, 59,
994–1005. doi:10.1037/0022-3514.59.5.994
Birnbaum, M. H. (2004). Human research and data collection via the
Internet. Annual Review of Psychology, 55, 803–832. doi:10.1146/
annurev.psych.55.090902.141601
Bodenhausen, G. V., Sheppard, L. A., & Kramer, G. P. (1994). Negative
affect and social judgment: The differential impact of anger and sadness.
European Journal of Social Psychology, 24, 45–62. doi:10.1002/ejsp
.2420240104
Buhrmester, M., Kwang, T., & Gosling, S. D. (2011). Amazon’s Mechan-
ical Turk. Perspectives on Psychological Science, 6, 3–5. doi:10.1177/
1745691610393980
Buss, A. H. (1966). Instrumentality of aggression, feedback, and frustration
as determinants of physical aggression. Journal of Personality and
Social Psychology, 3, 153–162. doi:10.1037/h0022826
Caughlin, J. P., Scott, A. M., & Miller, L. E. (2009). Conflict and hurt in
close relationships. In A. L. Vangelisti (Ed.), Feeling hurt in close
relationships. (pp. 143–166). New York, NY: Cambridge University
Press. doi:10.1017/CBO9780511770548.009
Chen, Z., Williams, K. D., Fitness, J., & Newton, N. C. (2008). When hurt
will not heal: Exploring the capacity to relive social and physical pain.
Psychological Science, 19, 789–795. doi:10.1111/j.1467-9280.2008
.02158.x
Chow, R. M., Tiedens, L. Z., & Govan, C. L. (2008). Excluded emotions:
The role of anger in antisocial responses to ostracism. Journal of
Experimental Social Psychology, 44, 896–903. doi:10.1016/j.jesp.2007
.09.004
Clark, M. S., Fitness, J., & Brissette, I. (2001). Understanding people’s
perceptions of relationships is crucial to understanding their emotional
lives. In G. Fletcher & M. S. Clark (Eds.), Blackwell handbook of social
psychology: Interpersonal processes (pp. 253–278). Oxford, England:
Blackwell.
1003
HURT FEELINGS AND ANGER
Clark, M. S., & Lemay, E. P., Jr. (2010). Close relationships. In S. T. Fiske,
D. T. Gilbert, & G. Lindzey (Eds.), Handbook of social psychology (5th
ed., Vol. 2, pp. 898–940). New York, NY: Wiley.
Clark, M. S., Ouellette, R., Powell, M. C., & Milberg, S. (1987). Recipi-
ent’s mood, relationship type, and helping. Journal of Personality and
Social Psychology, 53, 94–103. doi:10.1037/0022-3514.53.1.94
Clore, G. L., & Ortony, A. (2000). Cognition in emotion: Always, some-
times, or never? In L. Nadel, R. Lane & G. L. Ahern (Eds.), The
cognitive neuroscience of emotion (pp. 24–61). New York, NY: Oxford
University Press.
Coke, J. S., Batson, C. D., & McDavis, K. (1978). Empathic mediation of
helping: A two-stage model. Journal of Personality and Social Psychol-
ogy, 36, 752–766. doi:10.1037/0022-3514.36.7.752
deRivera, J., & Grinkis, C. (1986). Emotions as social relationships.
Motivation and Emotion, 10, 351–369. doi:10.1007/BF00992109
Dunham, Y. (2011). An angry ⫽outgroup effect. Journal of Experimental
Social Psychology, 47, 668–671. doi:10.1016/j.jesp.2011.01.003
Eisenberger, N. I., Lieberman, M. D., & Williams, K. D. (2003, October
10). Does rejection hurt? An fMRI study of social exclusion. Science,
302, 290–292. doi:10.1126/science.1089134
Feeney, J. A. (2004). Hurt feelings in couple relationships: Towards
integrative models of the negative effects of hurtful events. Journal of
Social and Personal Relationships, Vol. 21, 487–508. doi:10.1177/
0265407504044844
Feeney, J. A. (2005). Hurt feelings in couple relationships: Exploring the
role of attachment and perceptions of personal injury. Personal Rela-
tionships, 12, 253–271. doi:10.1111/j.1350-4126.2005.00114.x
Fehr, B., Baldwin, M., Collins, L., Patterson, S., & Benditt, R. (1999).
Anger in close relationships: An interpersonal script analysis. Person-
ality and Social Psychology Bulletin, 25, 299–312. doi:10.1177/
0146167299025003003
Fincham, F. D., & Bradbury, T. N. (1992). Assessing attributions in
marriage: The Relationship Attribution Measure. Journal of Personality
and Social Psychology, 62, 457–468. doi:10.1037/0022-3514.62.3.457
Fine, M. A., & Olson, K. A. (1997). Anger and hurt in response to
provocation: Relationship to psychological adjustment. Journal of So-
cial Behavior & Personality, 12, 325–344.
Fischer, A. H., & Manstead, A. S. R. (2008). Social functions of emotion.
Handbook of emotions (3rd ed., pp. 456–468). New York, NY: Guilford
Press.
Fischer, A. H., & Roseman, I. J. (2007). Beat them or ban them: The
characteristics and social functions of anger and contempt. Journal of
Personality and Social Psychology, 93, 103–115. doi:10.1037/0022-
3514.93.1.103
Fitness, J., & Fletcher, G. J. (1993). Love, hate, anger, and jealousy in close
relationships: A prototype and cognitive appraisal analysis. Journal of
Personality and Social Psychology, 65, 942–958. doi:10.1037/0022-
3514.65.5.942
Fitness, J., & Warburton, W. (2009). Thinking the unthinkable: Cognitive
appraisals and hurt feelings. In A. L. Vangelisti (Ed.), Feeling hurt in
close relationships (pp. 34–49). New York, NY: Cambridge University
Press.
Frey, J., III, Holley, J., & L’Abate, L. (1979). Intimacy is sharing hurt
feelings: A comparison of three conflict resolution models. Journal of
Marital and Family Therapy, 5, 35–41. doi:10.1111/j.1752-0606.1979
.tb01265.x
Frijda, N. H. (1986). The emotions. New York, NY: Cambridge University
Press.
Frijda, N. H., Kuipers, P., & ter Schure, E. (1989). Relations among
emotion, appraisal, and emotional action readiness. Journal of Person-
ality and Social Psychology, 57, 212–228. doi:10.1037/0022-3514.57.2
.212
Frijda, N. H., & Parrott, W. G. (2011). Basic emotions or ur-emotions?
Emotion Review, 3, 406–415. doi:10.1177/1754073911410742
Graham, S. M., Huang, J. Y., Clark, M. S., & Helgeson, V. S. (2008). The
positives of negative emotions: Willingness to express negative emo-
tions promotes relationships. Personality and Social Psychology Bulle-
tin, 34, 394–406. doi:10.1177/0146167207311281
Gross, J. J. (1998). The emerging field of emotion regulation: An integra-
tive review. Review of General Psychology, 2, 271–298. doi:10.1037/
1089-2680.2.3.271
Johnson, S. (2004). The practice of emotionally focused marital therapy
(2nd ed.). New York, NY: Brunner/Routledge.
Kelley, H. H., Holmes, J. G., Kerr, N. L., Reis, H. T., Rusbult, C. E., & Van
Lange, P. A. M. (2003). An atlas of interpersonal situations. New York,
NY: Cambridge University Press.
Kelley, H. H., & Thibaut, J. W. (1978). Interpersonal relations: A theory
of interdependence. New York, NY: Wiley.
Keltner, D., & Haidt, J. (1999). Social functions of emotions at four levels
of analysis. Cognition & Emotion, 13, 505–521. doi:10.1080/
026999399379168
Keltner, D., Haidt, J., & Shiota, M. N. (2006). Social functionalism and the
evolution of emotions. In M. Schaller, J. A. Simpson & D. T. Kenrick
(Eds.), Evolution and social psychology (pp. 115–142). New York, NY:
Psychology Press.
Kenny, D. A., Kashy, D. A., & Cook, W. L. (2006). Dyadic data analysis.
New York, NY: Guilford Press.
Kenny, D. A., & la Voie, L. (1982). Reciprocity of interpersonal attraction:
A confirmed hypothesis. Social Psychology Quarterly, 45, 54–58. doi:
10.2307/3033675
Knutson, B. (1996). Facial expressions of emotion influence interpersonal
trait inferences. Journal of Nonverbal Behavior, 20, 165–182. doi:
10.1007/BF02281954
Kraut, R., Olson, J., Banaji, M., Bruckman, A., Cohen, J., & Couper, M.
(2004). Psychological research online: Report of board of scientific
affairs’ advisory group on the conduct of research on the Internet.
American Psychologist, 59, 105–117. doi:10.1037/0003-066X.59.2.105
Lazarus, R. S. (1991a). Cognition and motivation in emotion. American
Psychologist, 46, 352–367. doi:10.1037/0003-066X.46.4.352
Lazarus, R. S. (1991b). Progress on a cognitive–motivational–relational
theory of emotion. American Psychologist, 46, 819–834. doi:10.1037/
0003-066X.46.8.819
Leary, M. R. (1990). Responses to social exclusion: Social anxiety, jeal-
ousy, loneliness, depression, and low self-esteem. Journal of Social and
Clinical Psychology, 9, 221–229. doi:10.1521/jscp.1990.9.2.221
Leary, M. R., & Baumeister, R. F. (2000). The nature and function of
self-esteem: Sociometer theory. In M. P. Zanna (Ed.), Advances in
experimental social psychology (Vol. 32, pp. 1–62). San Diego, CA:
Academic Press.
Leary, M. R., & Leder, S. (2009). The nature of hurt feelings: Emotional
experience and cognitive appraisals. In A. L. Vangelisti (Ed.), Feeling
hurt in close relationships (pp. 15–33). New York, NY: Cambridge
University Press.
Leary, M. R., & Springer, C. (2001). Hurt feelings: The neglected emotion.
In R. M. Kowalski (Ed.), Behaving badly: Aversive behaviors and
relational transgressions (pp. 151–175). Washington, DC: American
Psychological Association. doi:10.1037/10365-006
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. doi:10.1037/0022-
3514.74.5.1225
Leary, M. R., Twenge, J. M., & Quinlivan, E. (2006). Interpersonal
rejection as a determinant of anger and aggression. Personality and
Social Psychology Review, 10, 111–132. doi:10.1207/
s15327957pspr1002_2
Leith, K. P., & Baumeister, R. F. (1998). Empathy, shame, guilt, and
narratives of interpersonal conflicts: Guilt-prone people are better at
1004 LEMAY, OVERALL, AND CLARK
perspective taking. Journal of Personality, 66, 1–37. doi:10.1111/1467-
6494.00001
Lemay, E. P., Jr., & Clark, M. S. (2008). How the head liberates the heart:
Projection of communal responsiveness guides relationship promotion.
Journal of Personality and Social Psychology, 94, 647–671. doi:
10.1037/0022-3514.94.4.647
Lerner, J. S., Goldberg, J. H., & Tetlock, P. E. (1998). Sober second
thought: The effects of accountability, anger, and authoritarianism on
attributions of responsibility. Personality and Social Psychology Bulle-
tin, 24, 563–574. doi:10.1177/0146167298246001
Lerner, J. S., & Keltner, D. (2001). Fear, anger, and risk. Journal of
Personality and Social Psychology, 81, 146–159. doi:10.1037/0022-
3514.81.1.146
Lerner, J. S., & Tiedens, L. Z. (2006). Portrait of the angry decision maker:
How appraisal tendencies shape anger’s influence on cognition. Journal
of Behavioral Decision Making, 19, 115–137. doi:10.1002/bdm.515
Levenson, R. W., & Gottman, J. M. (1983). Marital interaction: Physio-
logical linkage and affective exchange. Journal of Personality and
Social Psychology, 45, 587–597. doi:10.1037/0022-3514.45.3.587
MacDonald, G., & Leary, M. R. (2005). Why does social exclusion hurt?
The relationship between social and physical pain. Psychological Bul-
letin, 131, 202–223. doi:10.1037/0033-2909.131.2.202
Mikulincer, M. (1994). Human learned helplessness: A coping perspective.
New York, NY: Plenum Press.
Mikulincer, M., Gillath, O., Halevy, V., Avihou, N., Avidan, S., & Eshkoli,
N. (2001). Attachment theory and reactions to others’ needs: Evidence
that activation of the sense of attachment security promotes empathic
responses. Journal of Personality and Social Psychology, 81, 1205–
1224. doi:10.1037/0022-3514.81.6.1205
Miller, S. L., Maner, J. K., & Becker, D. (2010). Self-protective biases in
group categorization: Threat cues shape the psychological boundary
between “us” and “them.” Journal of Personality and Social Psychol-
ogy, 99, 62–77. doi:10.1037/a0018086
Mills, J., Clark, M. S., Ford, T. E., & Johnson, M. (2004). Measurement of
communal strength. Personal Relationships, 11, 213–230. doi:10.1111/
j.1475-6811.2004.00079.x
Murray, S. L., Holmes, J. G., & Collins, N. L. (2006). Optimizing assur-
ance: The risk regulation system in relationships. Psychological Bulletin,
132, 641–666. doi:10.1037/0033-2909.132.5.641
Ohbuchi, K.-i., Kameda, M., & Agarie, N. (1989). Apology as aggression
control: Its role in mediating appraisal of and response to harm. Journal
of Personality and Social Psychology, 56, 219–227. doi:10.1037/0022-
3514.56.2.219
Overall, N. C., Fletcher, G. J. O., Simpson, J. A., & Sibley, C. G. (2009).
Regulating partners in intimate relationships: The costs and benefits of
different communication strategies. Journal of Personality and Social
Psychology, 96, 620–639. doi:10.1037/a0012961
Reis, H. T., Clark, M. S., & Holmes, J. G. (2004). Perceived partner
responsiveness as an organizing construct in the study of intimacy and
closeness. In D. J. Mashek & A. P. Aron (Eds.), Handbook of closeness
and intimacy (pp. 201–225). Mahwah, NJ: Erlbaum.
Robinson, M. D., & Clore, G. L. (2002). Episodic and semantic knowledge
in emotional self-report: Evidence for two judgment processes. Journal
of Personality and Social Psychology, 83, 198–215. doi:10.1037/0022-
3514.83.1.198
Roseman, I. J. (1984). Cognitive determinants of emotions: A structural
theory. In P. Shaver (Ed.), Review of personality and social psychology
(Vol. 5, pp. 11–36). Beverly Hills, CA: Sage.
Roseman, I. J., Antoniou, A. A., & Jose, P. E. (1996). Appraisal determi-
nants of emotions: Constructing a more accurate and comprehensive
theory. Cognition & Emotion, 10, 241–278. doi:10.1080/
026999396380240
Roseman, I. J., Spindel, M. S., & Jose, P. E. (1990). Appraisals of
emotion-eliciting events: Testing a theory of discrete emotions. Journal
of Personality and Social Psychology, 59, 899–915. doi:10.1037/0022-
3514.59.5.899
Roseman, I. J., Wiest, C., & Swartz, T. S. (1994). Phenomenology, behav-
iors, and goals differentiate discrete emotions. Journal of Personality
and Social Psychology, 67, 206–221. doi:10.1037/0022-3514.67.2.206
Rusbult, C. E. (1980). Commitment and satisfaction in romantic associa-
tions: A test of the investment model. Journal of Experimental Social
Psychology, 16, 172–186. doi:10.1016/0022-1031(80)90007-4
Rusbult, C. E., & Buunk, B. P. (1993). Commitment processes in close
relationships: An interdependence analysis. Journal of Social and Per-
sonal Relationships, 10, 175–204. doi:10.1177/026540759301000202
Rusbult, C. E., Martz, J. M., & Agnew, C. R. (1998). The Investment
Model Scale: Measuring commitment level, satisfaction level, quality of
alternatives, and investment size. Personal Relationships, 5, 357–387.
doi:10.1111/j.1475-6811.1998.tb00177.x
Rusbult, C. E., Verette, J., Whitney, G. A., Slovik, L. F., & Lipkus, I.
(1991). Accommodation processes in close relationships: Theory and
preliminary empirical evidence. Journal of Personality and Social Psy-
chology, 60, 53–78. doi:10.1037/0022-3514.60.1.53
Russell, J. A., & Fehr, B. (1994). Fuzzy concepts in a fuzzy hierarchy:
Varieties of anger. Journal of Personality and Social Psychology, 67,
186–205. doi:10.1037/0022-3514.67.2.186
Salovey, P. (1991). The psychology of jealousy and envy. New York, NY:
Guilford Press.
Salovey, P., & Mayer, J. D. (1990). Emotional intelligence. Imagination,
Cognition and Personality, 9, 185–211.
Sanford, K. (2007). Hard and soft emotion during conflict: Investigating
married couples and other relationships. Personal Relationships, 14,
65–90. doi:10.1111/j.1475-6811.2006.00142.x
Sanford, K., & Rowatt, W. C. (2004). When is negative emotion positive
for relationships? An investigation of married couples and roommates.
Personal Relationships, 11, 329–354. doi:10.1111/j.1475-6811.2004
.00086.x
Shaver, P. R., Mikulincer, M., Lavy, S., & Cassidy, J. (2009). Understand-
ing and altering hurt feelings: An attachment-theoretical perspective on
the generation and regulation of emotions. In A. Vangelisti (Ed.), Feel-
ing hurt in close relationships (pp. 92–120). New York, NY: Cambridge
University Press. doi:10.1017/CBO9780511770548.007
Shaver, P., Schwartz, J., Kirson, D., & O’Connor, C. (1987). Emotion
knowledge: Further explorations of a prototype approach. Journal of
Personality and Social Psychology, 52, 1061–1086. doi:10.1037/0022-
3514.52.6.1061
Siegel, J. M. (1986). The multidimensional anger inventory. Journal of
Personality and Social Psychology, 51, 191–200. doi:10.1037/0022-
3514.51.1.191
Simon, H. A. (1967). Motivational and emotional controls of cognition.
Psychological Review, 74, 29–39. doi:10.1037/h0024127
Smith, C. A., & Ellsworth, P. C. (1985). Patterns of cognitive appraisal in
emotion. Journal of Personality and Social Psychology, 48, 813–838.
doi:10.1037/0022-3514.48.4.813
Smith, C. A., & Lazarus, R. S. (1993). Appraisal components, core rela-
tional themes, and the emotions. Cognition & Emotion, 7, 233–269.
doi:10.1080/02699939308409189
Tangney, J. P., Hill-Barlow, D., Wagner, P. E., Marschall, D. E., Boren-
stein, J. K., Sanftner, J.,...Granzow, R. (1996). Assessing individual
differences in constructive versus destructive responses to anger across
the lifespan. Journal of Personality and Social Psychology, 70, 780–
796. doi:10.1037/0022-3514.70.4.780
Tangney, J. P., Wagner, P., Fletcher, C., & Gramzow, R. (1992). Shamed
into anger? The relation of shame and guilt to anger and self-reported
aggression. Journal of Personality and Social Psychology, 62, 669–675.
doi:10.1037/0022-3514.62.4.669
1005
HURT FEELINGS AND ANGER
Tangney, J. P., Wagner, P., & Gramzow, R. (1992). Proneness to shame,
proneness to guilt, and psychopathology. Journal of Abnormal Psychol-
ogy, 101, 469–478. doi:10.1037/0021-843X.101.3.469
Tiedens, L. Z. (2001). Anger and advancement versus sadness and subju-
gation: The effect of negative emotion expressions on social status
conferral. Journal of Personality and Social Psychology, 80, 86–94.
doi:10.1037/0022-3514.80.1.86
Tiedens, L. Z., & Linton, S. (2001). Judgment under emotional certainty
and uncertainty: The effects of specific emotions on information pro-
cessing. Journal of Personality and Social Psychology, 81, 973–988.
doi:10.1037/0022-3514.81.6.973
Tooby, J., & Cosmides, L. (1990). The past explains the present: Emotional
adaptations and the structure of ancestral environments. Ethology and
Sociobiology, 11, 375–424. doi:10.1016/0162-3095(90)90017-Z
Twenge, J. M., Baumeister, R. F., DeWall, C., Ciarocco, N. J., & Bartels,
J. (2007). Social exclusion decreases prosocial behavior. Journal of
Personality and Social Psychology, 92, 56–66. doi:10.1037/0022-3514
.92.1.56
Twenge, J. M., Baumeister, R. F., Tice, D. M., & Stucke, T. S. (2001). If
you can’t join them, beat them: Effects of social exclusion on aggressive
behavior. Journal of Personality and Social Psychology, 81, 1058–1069.
doi:10.1037/0022-3514.81.6.1058
Vangelisti, A. L. (2009). Hurt feelings: Distinguishing features, functions,
and overview. In A. L. Vangelisti (Ed.), Feeling hurt in close relation-
ships (pp. 3–11). New York, NY: Cambridge University Press.
Vangelisti, A. L., & Crumley, L. P. (1998). Reactions to messages that
hurt: The influence of relational contexts. Communication Monographs,
65, 173–196. doi:10.1080/03637759809376447
Vangelisti, A. L., & Sprague, R. (1998). Guilt and hurt: Similarities,
distinctions, and conversational strategies. In P. Anderson & L. Guerrero
(Eds.), Handbook of communication and emotion: Research, theory,
application, and contexts (pp. 123–154). New York, NY: Academic
Press.
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. doi:10.1177/
0265407500173005
Vangelisti, A. L., Young, S. L., Carpenter-Theune, K. E., & Alexander,
A. L. (2005). Why does it hurt? The perceived causes of hurt feelings.
Communication Research, 32, 443–477. doi:10.1177/
0093650205277319
Van Kleef, G. A. (2010). The emerging view of emotion as social infor-
mation. Social and Personality Psychology Compass, 4, 331–343. doi:
10.1111/j.1751-9004.2010.00262.x
Van Kleef, G. A., Homan, A. C., Beersma, B., Van Knippenberg, D., Van
Knippenberg, B., & Damen, F. (2009). Searing sentiment or cold cal-
culation? The effects of leader emotional displays on team performance
depend on follower epistemic motivation. Academy of Management
Journal, 52, 562–580. doi:10.5465/AMJ.2009.41331253
Watson, D., & Clark, L. A. (1991). Self- versus peer ratings of specific
emotional traits: Evidence of convergent and discriminant validity. Jour-
nal of Personality and Social Psychology, 60, 927–940. doi:10.1037/
0022-3514.60.6.927
Watson, D., & Clark, L. A. (1992). On traits and temperament: General and
specific factors of emotional experience and their relation to the five-
factor model. Journal of Personality, 60, 441–476. doi:10.1111/j.1467–
6494.1992.tb00980.x
Welsh, D. P., & Dickson, J. W. (2005). Video-recall procedures for
examining observational data and subjective understanding in family
psychology. Journal of Family Psychology, 19, 62–71. doi:10.1037/
0893-3200.19.1.62
Wieselquist, J., Rusbult, C. E., Foster, C. A., & Agnew, C. R. (1999).
Commitment, pro-relationship behavior, and trust in close relationships.
Journal of Personality and Social Psychology, 77, 942–966. doi:
10.1037/0022-3514.77.5.942
Williams, K. D., Cheung, C. K., & Choi, W. (2000). Cyberostracism:
Effects of being ignored over the Internet. Journal of Personality and
Social Psychology, 79, 748–762. doi:10.1037/0022-3514.79.5.748
Yovetich, N. A., & Rusbult, C. E. (1994). Accommodative behavior in
close relationships: Exploring transformation of motivation. Journal of
Experimental Social Psychology, 30, 138–164. doi:10.1006/jesp.1994
.1007
Zeigarnik, B. (1938). On finished and unfinished tasks. In W. D. Ellis
(Ed.), A source book of Gestalt psychology (pp. 300–314). London,
England: Kegan Paul, Trench, Trubner & Co. doi:10.1037/11496-025
Received March 14, 2012
Revision received August 13, 2012
Accepted August 14, 2012 䡲
1006 LEMAY, OVERALL, AND CLARK
A preview of this full-text is provided by American Psychological Association.
Content available from Journal of Personality and Social Psychology
This content is subject to copyright. Terms and conditions apply.
