ORIGINAL ARTICLE
Sugaring the Pill: Assessing RhetoricalStrategies Designed to Minimize DefensiveReactions to Group Criticism
Matthew J. Hornsey1, Erin Robson1, Joanne Smith1,Sarah Esposo1, & Robbie M. Sutton2
1 School of Psychology, University of Queensland, St. Lucia, Queensland 4072, Australia
2 School of Psychology, University of Kent, Keynes College, Canterbury CT27NP, UK
People are considerably more defensive in the face of group criticism when the criticism
comes from an out-group rather than an in-group member (the intergroup sensitivity
effect). We tested three strategies that out-group critics can use to reduce this height-
ened defensiveness. In all studies, Australians received criticism of their country either
from another Australian or from a foreigner. In Experiment 1, critics who attached
praise to the criticism were liked more and agreed with more than were those who did
not. In Experiment 2, out-group critics were liked more and aroused less negativity
when they acknowledged that the problems they identified in the target group were
shared also by their own in-group. In both experiments, the ameliorative effects of
praise and acknowledgment were fully mediated by attributions of constructiveness.
Experiment 3 tested the strategy of spotlighting; that is, of putting on the record that
you intend your comments to apply to just a portion of the group rather than to the
whole group. This strategy—which did not directly address the attributional issues that
are presumed to underpin the intergroup sensitivity effect—proved ineffective. Practical
and theoretical implications for intergroup communication are discussed.
doi:10.1111/j.1468-2958.2007.00314.x
In literature on interpersonal and small group communication, researchers hold an
ambivalent approach toward criticism. On one hand, there is an acknowledgmentthat criticism can be destructive. Poor use of criticism, for example, is perceived to be
a major cause of conflict in organizations (Baron, 1988) and perceptions that one’sspouse is critical of you is a major predictor of relapse among depressives (Hooley &
Teasdale, 1989). Personal criticism is seen as a dysfunctional strategy for dealing withconflict (e.g., Hocker & Wilmot, 1991), whereas absence of criticism is seen to bea relational maintenance strategy (Canary & Stafford, 1992). In his work on marital
Corresponding author: Matthew J. Hornsey; e-mail: [emailprotected] article was accepted under the editorship of Jim Dillard.
Human Communication Research ISSN 0360-3989
70 Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association
relationships, Gottman (1994; Gottman & Levenson, 1992) showed that criticism ofone’s partner contributes to cascades of isolation and withdrawal and is a key factor
that characterizes distressed couples. Gottman concluded that to achieve maritalstability, five positive behaviors should be communicated for every one negative
behavior.Despite this, it is also clear that people occasionally need criticism to grow and
to develop. As Winston Churchill said: ‘‘Criticism may not be agreeable, but it is
necessary. It fulfills the same function as pain in the human body. It calls attention toan unhealthy state of things.’’ Criticism helps signpost weaknesses and potentially
provides a map for how to correct them. Providing negative feedback about past andcurrent behavior is a cornerstone of many performance interventions; for example,
performance appraisals in the workplace, feedback on assignments, and reviews ofscholarly articles. Even in intimate relationships, it might on occasion be functional
to highlight aspects of a partner’s behavior that need addressing. Indeed, whereasconflict engagement negatively predicts marital happiness when measured concur-rently, there is evidence that conflict engagement predicts improvement in marital
satisfaction in the long run (Gottman & Krokoff, 1989).Criticism can be directed at people as individuals (e.g., ‘‘you are racist’’) or it can
be directed at people’s groups (e.g., ‘‘Australians are racist’’). In the current paper, weexamine the second form of criticism. Specifically, we examine how people respond
when, for example, they receive negative comments about their country, or theirprofession, or their religion. Like criticism of individuals, criticism of groups can be
a prerequisite for reform. Where a nation, for example, is out of touch with inter-national norms of environmental responsibility, or military conduct, or respect for
human rights, then criticism from within and outside the country can be a catalystfor positive change. If received defensively, however, criticism can lead to dishar-mony, schism, conflict, and denial.
This paper examines some of the strategies that can be used to reduce resistanceto criticism of groups. In examining this question, we first review the limited
research that has been conducted to date on when and why people respond ina defensive versus an open-minded fashion to group criticism. We then describe
and test three strategies for reducing resistance. Although all three are intuitivelyappealing strategies, we argue (and show) that only two of these strategies help
overcome the active ingredient that promotes defensiveness: suspicious attributionsabout motive. Throughout, we acknowledge points of contact between the researchon group criticism and the research on interpersonal criticism but maintain that
insights that can be drawn from one arena cannot be translated unproblematicallyinto the other. One phenomenon that helps illustrate the unique nature of inter-
group criticism is the intergroup sensitivity effect.
Intergroup sensitivity effect
It is only very recently that researchers have systematically examined the psychology
of responses to group-directed criticism. One factor that has been shown to have
M. J. Hornsey et al. Strategies to Soften Criticism
Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association 71
a profound effect on how people respond to group criticism is the group member-ship of the critic. Put simply, when criticism of a group is delivered by outsiders
(otherwise known as out-group members), the criticisms arouse more defensivenessthan when the very same comments are delivered by insiders (or in-group members).
This phenomenon is labeled the intergroup sensitivity effect. This effect appears to berobust, having now been demonstrated across at least eight different intergroupcontexts, each using different criticisms, and using participants from both collectivist
and individualist cultures (see Hornsey, 2005, for a review). Furthermore, the effectis relatively large: Criticisms that might be cautiously accepted when delivered by an
in-group member are aggressively denied if delivered by an out-group member(Hornsey & Imani, 2004; Hornsey, Oppes, & Svensson, 2002; Sutton, Elder, &
Douglas, 2006).With its focus on group identities, influence, and threat, the intergroup sensi-
tivity effect can be located within the language and meta-theoretical umbrella of thesocial identity approach (Tajfel & Turner, 1979; Turner, 1991). Self-categorizationtheory (Turner, Hogg, Oakes, Reicher, & Wetherell, 1987), for example, argues that
the more strongly people identify with a salient group, the more they shift their self-definition from the personal to the collective level. This depersonalization around
the group prototype is accompanied by a perceptual enhancement of intragroupsimilarities and intergroup differences, a process that is assumed to underpin con-
formity and group-mediated attitude change (Turner, 1991). One consequence ofthis is that people are more likely to embrace messages from in-group members than
from out-group members.There are two reasons, however, to believe that the intergroup sensitivity effect is
not merely a specific example of this tendency for people to be cognitively predis-posed to embrace in-group messages more readily than out-group messages. First,proponents of the social identity approach would not argue that members would
assimilate around the attitudes of any group member; indeed, researchers in this fieldtypically argue that in-group deviants and dissenters face exaggerated levels of hos-
tility as the group tries to preserve its positive distinctiveness (Abrams, Hogg, &Marques, 2005). Second, the tendency to accept in-group comments more than out-
group comments does not apply when people make positive comments about thegroup. When participants read praise of their university (Hornsey et al., 2002, Study
1b) or of their country (Hornsey & Imani, 2004; Hornsey et al., 2002, Study 1a), theydo not rate the speaker or the comments differently depending on whether they areattributed to an in-group or an out-group member. This suggests that the intergroup
sensitivity effect is specific to criticisms and not a more generalized tendency tocognitively assimilate to in-group members (see Hornsey, 2006, for a deeper discus-
sion of the points of similarity and dissimilarity between the intergroup sensitivityeffect and the social identity perspective).
Heightened defensiveness in the face of criticism from outsiders is a concernbecause there are times when it is important that groups listen to criticism. If a group
is not criticized, it can become complacent and stagnant as maladaptive, corrupt, or
Strategies to Soften Criticism M. J. Hornsey et al.
72 Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association
inefficient practices continue unchallenged. Indeed, a lack of dissent and criticismhas been shown to lead to suboptimal decision making (Postmes, Spears, & Cihangir,
2001), the consequences of which can be disastrous (Janis, 1982). Furthermore, thereare times when negative feedback needs to come from outside the group because in-
group members are unwilling or unable to recognize the problems within their ownculture. In short, groups occasionally need a nudge along from outsiders to helpthem pick up their game and reform their culture. And yet research on the inter-
group sensitivity effect suggests that such pleas for change face heightened resistance.One intuitive way that outsiders could guard themselves against defensiveness is
to engage in credentialing; that is, to equip themselves with a great deal of knowledgeand experience of the target group (see Cupach & Metts, 1994, for a discussion of this
strategy in the context of interpersonal conflict). Interestingly, however, this strategydoes not appear to work. Hornsey and Imani (2004) designed studies in which
experience and group membership could be manipulated independently of eachother. In these studies, Australians received an extract from an interview with a per-son who criticized Australians for being uncultured and racist. These comments were
attributed either to another Australian, a foreigner who had spent many years livingin Australia, or a foreigner with no experience of Australia. The consistent finding
was that in-group critics aroused less defensiveness than outsiders and that experi-ence did not help the outsiders. In other words, critics who had spent large chunks of
their life in Australia were treated no differently from critics who had never set footin the country. This suggests that perceptions of epistemic authority did not under-
pin the intergroup sensitivity effect and that outsiders cannot reduce defensivenessmerely by boosting and communicating their credentials as experienced judges.
Using language to overcome the intergroup sensitivity effect
The failure of credentialing to overcome the intergroup sensitivity effect poses the
question of what outsiders can do to neutralize or overcome defensiveness in the faceof group criticism. One possibility is to adjust the language they use. Before engaging
in (interpersonal) communications that might be considered argumentative oraggressive, people edit their arguments before uttering them (Hample & Dallinger,
1988). Arguments are tailored and so is the language used to express them. Com-municators use disclaimers (Hewitt & Stokes, 1975), equivocations (Bavelas, Black,
Bryson, & Mullett, 1988), and ingratiation tactics (Berscheid & Walster, 1978) tomaximize the effectiveness of their message and to conform to rules about politenessand maintenance of face (Brown & Levinson, 1987; Cupach & Carson, 2002; Cupach
& Metts, 1990, 1994; Oetzel & Ting-Toomey, 2003; Rogan & Hammer, 1994; Ting-Toomey, 1988). Like any communication, criticisms can be delivered well and they
can be delivered poorly, depending on the sensitivity and skill with which criticstailor their language to deliver the message (Baron, 1988; Tracy, van Dusen, &
Robinson, 1987). Although research on this question is surprisingly scarce, certainprinciples can be intuited: Criticism should be specific, should be oriented toward
the future rather than the past, should avoid attributing poor performance to
M. J. Hornsey et al. Strategies to Soften Criticism
Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association 73
internal causes, should be clear, and should avoid being ‘‘biting’’ or sarcastic (Baron,1988; Ogilvie & Haslett, 1985).
The research question examined in this paper is: What strategies are availableto the outsider who wishes to promote change in another group? To answer this
question, we sought a theoretical base that went beyond intuitive notions of respect-fulness, specificity, and so forth, and could grapple with the emergent processes thatgovern intergroup as opposed to interpersonal criticism. To achieve this, it is nec-
essary to engage with the deeper question of what it is about a critic’s out-groupstatus that arouses so much defensiveness in the first place. Only by coming to
a theoretical understanding of what drives the intergroup sensitivity effect, can wegenerate predictions about which strategies should work and which should not.
The literature to date has shown quite consistently that at the heart of theintergroup sensitivity effect is an attributional bias. Attributions play a central role
in communication theories, having been found to influence responses to excuses(Weiner, Amirkhan, Folkes, & Verette, 1987), resistance to compliance requests(Wilson, Cruz, Marshall, & Rao, 1993), relational disengagement (Cody, Kersten,
Braaten, & Dickson, 1992), rejection (e.g., Folkes, 1982), bad news (e.g., Bies &Sitkin, 1992), and conflict (e.g., Bradbury & Fincham, 1990). With respect to
group-directed criticism, Hornsey (2005) argues that people look past the contentof the words and make judgments about the integrity of the critic’s motives; in other
words they ask themselves the question: ‘‘Why would they say that?’’ If they perceivethat the critic has relatively sinister or destructive motives, then this provides an
opportunity to dismiss the message, and heightens negativity toward the speaker andhis or her comments. However, if they can see no reason to assume that the speaker
has destructive motives, they are free to assess the content of the message on itsmerits. When making this judgment about motive, receivers of criticism factor ina number of considerations, not least of which is the group membership of the critic.
In short, when in-group members criticize the group, people are more likely toassume that they are motivated by constructive reasons than when the same com-
ments are delivered by an outsider, and these differing perceptions of motive drivethe effect. This is consistent with much evidence showing that people expect in-
group members to look after them and to show reciprocity of favors (e.g., Brewer,1981; Tanis & Postmes, 2005; Yamagishi & Kiyonari, 2000), whereas they expect out-
group members to have hostile and competitive intentions (e.g., Judd, Park, Yzerbyt,Gordijn, & Muller, 2005; Vivian & Berkowitz, 1993).
Evidence for this attributional explanation includes mediational analyses dem-
onstrating that the intergroup sensitivity effect tends to disappear when attributionsof constructiveness are controlled for (e.g., Hornsey & Imani, 2004). There is also
experimental evidence reinforcing the notion that it is attributions of constructive-ness that are the most proximal driver of the effect. For example, if participants are
led to believe that the in-group critic is not committed to the group—either becausethey are a low identifier (Hornsey, Trembath, & Gunthorpe, 2004) or because they
are a newcomer to the group (Hornsey, Grice, Jetten, Paulsen, & Callan, 2007)—they
Strategies to Soften Criticism M. J. Hornsey et al.
74 Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association
arouse just as much defensiveness as outsiders. Figure 1 provides a summary of theattributional model of responses to group criticism.
In devising and testing strategies for reducing resistance to group criticism, wewere guided by this attributional account of the intergroup sensitivity effect. If it is
true that heightened levels of defensiveness in the face of out-group criticisms aredriven by suspicions about the motives of the out-group speaker, then any strategyused by out-group speakers to minimize distrust with regard to their motives should
be particularly useful in reducing defensiveness. One such strategy might be topreface criticism with positive feedback (‘‘sweetening’’). An alternative strategy avail-
able to out-group critics might be to acknowledge the failings of their own group as
CRITICISM
DATA (examples …)
Group membership (in-group or out-group?) Tenure of critic (newcomer or oldtimer?) Strategy (e.g., praise, acknowledgment)
DECISION
If answer to original hypotheses is YES:
Relatively low levels of emotionalnegativity to criticism Relatively generous appraisal of thecritic Relative high acceptance of comments
HYPOTHESIS TESTING
Does the critic care about us?
DECISION
If answer to original hypotheses is NO:
High levels of emotional negativity tocriticismHarsh appraisal of the criticRejection of comments as untrue
Is the critic trying to be constructive?
Figure 1 Responses to group criticism: An attributional model.
M. J. Hornsey et al. Strategies to Soften Criticism
Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association 75
well as the target group (‘‘sharing’’; e.g., ‘‘we also are racist’’). In each case, the out-group members are presenting the criticisms in such a way as to disarm people’s
expectations that they are making the criticism as part of a wider intergroup com-petition for prestige and status. With these suspicions assuaged, in-group members
might be better positioned psychologically to focus on the content of the message ina balanced and nondefensive way. These two strategies are tested in Experiments 1and 2, respectively. Experiment 3 tests an alternative strategy that has powerful
intuitive appeal but does not address the attributional concerns that have beenargued to be the primary driver of defensiveness: namely, putting on the record that
you intend your comments to apply to just a portion of the group rather than to thewhole group (‘‘spotlighting’’). If attributions of constructiveness really do underpin
the intergroup sensitivity effect, we should see the first two strategies work, whereasthe third strategy should not.
Experiment 1
It is commonly assumed that the strategic use of praise (‘‘sweetening’’) can helpsoften responses to negative feedback; for example, in our department, tutors are
advised to sandwich their negative feedback between pieces of praise when markingassignments. This common wisdom has filtered into textbooks and manuals, but
there have been surprisingly few controlled empirical tests of whether the strategyactually works.
Descriptive evidence for the power of praise emerged in a study by Tracy et al.(1987), in which participants were asked to recall examples of ‘‘good’’ and ‘‘bad’’
criticism they had received in the past. In 19% of cases of well-delivered criticism,the specific negative comments were framed in a broader positive context. Of theexamples of poorly delivered criticism, only 2% did so. To our knowledge, experi-
mental tests of the use of praise as a buffer to criticism are rare. One exception isa study by Cohen, Steele, and Ross (1999), who exposed Black and White students
to negative evaluations of a written task from a White teacher. In one condition,the criticism was unbuffered; in another condition, it was buffered by praise.
Evidence for the buffering effect of praise was mixed. Black (but not White)students were marginally more task motivated when the criticism was buffered
by praise than when it was not. On a measure of whether participants identifiedwith their writing task (e.g., ‘‘How would you rate your overall competence asa writer?’’), praise had no effect for either Black or White participants. One final
experiment of note was conducted by Davies and Jacobs (1985), who tested thesandwiching approach described earlier using 28 attendants at a 2-hour self-
development workshop. Participants who received the ‘‘sandwich’’ rated the feed-back as more desirable and credible (but no more emotionally positive) than those
who received other permutations of positive and negative feedback. However, itwas not possible to compare the effectiveness of praise per se because praise was
present in all conditions.
Strategies to Soften Criticism M. J. Hornsey et al.
76 Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association
In summary, there has been surprisingly little empirical research examiningwhether praise can buffer the effects of criticism (and certainly none from an inter-
group perspective). Of the research that has been conducted, the evidence for praiseis mixed. Where effects have been found, there has been no examination of what
might mediate the effects. Experiment 1 was designed to respond to these limitations.From the perspective of research on the intergroup sensitivity effect, it is reason-
able to expect that praise might help reduce defensiveness toward group criticism.
Our argument is that, when responding to criticism of their groups, people drawhypotheses about what is going on in the heart and mind of the critic: Do they care
about us? Are they trying to be constructive? If praise is attached to the criticism, itmight be that recipients would be more likely to answer these questions in the
affirmative. The more generous attributions associated with praise should then flowon to reduced levels of defensiveness.
One potential problem with sweetening, however, is that the strategy might beseen by a skeptical in-group audience as an overly transparent attempt to softencriticism. This could result in the praise simply being dismissed, or even worse, it
could result in group members feeling as though the speaker is being overaccom-modating or patronizing. Even in the absence of skepticism about the motives for the
praise, it could be that the praise would not be processed or remembered, given thatit is potentially overshadowed by the more threatening and ego-involving criticism.
Indeed, social theorists have long recognized the tendency for some people to selec-tively attend to negative experiences and to discount positive experiences when re-
ceiving feedback from the social world, a phenomenon sometimes referred to as thenegativity effect (Fiske, 1980), negativity bias (Baumeister, Bratslavsky, Finkenauer,
& Vohs, 2001; Rozin & Royzman, 2001), or mnemic neglect (Green, Pinter, &Sedikides, 2005). Consistent with this, an analysis of student reactions to teacherfeedback revealed that bad feedback had a much stronger effect on perceptions than
good feedback (Coleman, Jussim, & Abraham, 1987). Although good feedback wasseen to be more credible, bad feedback was seen to be more diagnostic of what the
teacher really thought. Thus, despite the assumed consensus that praise is helpful insoftening the blow when delivering negative feedback, there are theoretical reasons
to test this assumption empirically.In Experiment 1, Australian participants read criticisms of their group that were
attributed either to another Australian (in-group critic) or to a non-Australian(out-group critic). Depending on condition, the criticism was either contextualizedwith praise or it was not. Consistent with the intergroup sensitivity effect, we
predicted that out-group critics would be liked less than in-group critics and thattheir comments would arouse more negativity and less agreement than the in-group
critic’s comments. We also predicted that, when praise was attached to the criticism,levels of interpersonal liking and agreement would increase and levels of negativity
would decrease relative to when praise was not used. If we were to obtain effectsof praise, it was expected that these effects would be mediated by attributions
of constructiveness.
M. J. Hornsey et al. Strategies to Soften Criticism
Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association 77
Method
Participants and design
Participants were 107 undergraduate psychology students (79 females, 28 males,M = 21.61 years) who participated in the study in return for course credit. Assuming
a moderate effect size of .30, we needed only 90 participants to achieve acceptablepower (80%); thus, the current sample size was considered sufficiently powerful.
Participants were randomly allocated to the four conditions of a 2 (speaker type:in-group vs. out-group) 3 2 (strategy: praise vs. no praise) between-groups
design. Participants were only included if they nominated Australia as their primarynational identity.
Materials and procedure
To reduce the transparency of the experiment, participants were told that the study
was interested in how people make personality evaluations on the basis of minimalinformation. They were told that they would be given an extract from an interview
with someone who would talk a little about themselves and a little about what theythink of Australia. They would be asked to read these extracts and to evaluate both
the speaker and his or her comments. Before starting the questionnaire, participantsrecorded the extent to which they agreed with the following statements: ‘‘Being anAustralian is an important part of my self-image,’’ ‘‘I have a lot in common with
other Australians,’’ and ‘‘I identify as an Australian.’’ These items were combined toform a single measure of national identification (a = .87).
Manipulation of speaker
This section of the questionnaire consisted of a biography of the speaker and was
intended to manipulate speaker type. A summary of the speaker’s demographicinformation was provided, including name (J. Benson), country of birth (Australiaor United States), current residence (Brisbane or Washington), and age (22). In
response to the question, ‘‘Tell us a little bit about yourself,’’ the speaker answered:‘‘My parents own their own landscaping business. I’ve got two older brothers and an
older sister. I enjoy hanging out with my friends and reading.’’ In response to thequestion, ‘‘What country are you from?’’ the speaker said ‘‘I was born in Australia (or
America, depending on condition) and I’ve lived here all my life.’’ To accentuate thiscondition, the national flag of Australia or the United States (depending on condi-
tion) was pasted next to the biographical information.Participants then rated the extent to which they thought the speaker was intel-
ligent, trustworthy, friendly, open-minded, likeable, respected, interesting, andnice (Time 1 trait evaluations; a = .90). Trait ratings of the speaker were measuredbefore the criticisms were read by the participants in order to assess preexisting
differences in trait evaluations of in-group and out-group members. These mea-sures can be used as a baseline against which postcriticism trait evaluations can
be compared.
Strategies to Soften Criticism M. J. Hornsey et al.
78 Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association
Criticisms
The scripts used to present the criticisms were similar to those used by Hornsey and
colleagues (Hornsey & Imani, 2004; Hornsey et al., 2002, 2004). In response to thequestion, ‘‘What do you think of Australians?’’ the critic said:
I think of them as being fairly racist. They’re racist towards Aborigines, andthey’re intolerant of Asians. Also, my understanding of Australia is that, on the
whole, they’re not as cultured as most societies.
The script was identical in both in-group and out-group conditions, except in the
in-group conditions they used inclusive pronouns (i.e., ‘‘us’’ and ‘‘we’’ rather than‘‘them’’ and ‘‘they’’). This is consistent with all previous research on the intergroup
sensitivity effect.In the no-praise condition, there was no further text. In the praise condition,
the negative comments were contextualized with three positive comments:
Having said that, I think Australians are generally fairly friendly and warm
people. I also think they’re a very educated society. And they seem to havea good sense of humor. They can see the funny side to things which is good.
The comments used in the scripts were based on a pilot study in which Australianswere asked to rate the extent to which various negative and positive statements
applied to Australians. The traits used in the scripts were used because we foundsome consensus that Australians themselves believed them to be true (see Hornsey
et al., 2002, for details). The order of the praise relative to the criticism was counter-balanced, such that half of the time the praise came before the criticism and half the
time the praise came after the criticism.
Dependent measures
All subsequent items were measured on a 7-point scale from 1 (not at all) to 7(very much). The first part of the questionnaire was designed to test the extent to
which the positive and negative comments were salient to participants. Participantswere reminded of each of the three criticisms and were asked after each to indicate
the extent to which, when reading the script, this particular comment had ‘‘stuck intheir mind’’ and the extent to which it had ‘‘stood out from the rest in their mind.’’
For participants in the praise condition, they repeated the exercise in relation to eachof the pieces of praise. The responses for the pieces of criticism (a = .70) and for
the pieces of praise (a = .88) were combined into scales of salience.The rest of the questionnaire measured the key dependent variables. Construc-
tiveness was measured by asking the extent to which the participants believed the
speaker’s comments were intended to be constructive, were intended to be destruc-tive (reverse scored), and were made with Australia’s best interests at heart (a = .80).
Negativity toward the criticisms was measured by asking the extent to which theparticipants felt the criticisms of Australia were disappointing, irritating, offensive,
judgmental, and insulting (a = .93). Following this, participants completed the
M. J. Hornsey et al. Strategies to Soften Criticism
Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association 79
same trait evaluation scale used at Time 1 (Time 2 trait evaluations; a = .92).Agreement with the criticisms was measured by asking the extent to which the
participants agreed with the speaker’s criticisms about Australia, and the extent towhich they felt the criticisms were true (r = .93).
To determine whether constructiveness, negativity, agreement, and likeabilityrepresent discrete scales, a factor analysis was conducted on these items using prin-cipal components extraction with oblimin rotation. A three-factor solution emerged,
with the first factor comprising both the likeability items (Time 2) and the construc-tiveness items (eigenvalue = 10.05, 55.81% of variance explained). The second factor
comprised the negativity items (eigenvalue = 2.45, 13.61% of variance explained),and the third factor comprised the agreement items (eigenvalue = 1.15, 6.40%
of variance explained). Although the likeability and constructiveness items areclearly highly correlated, on face value, we were confident that these constructs were
separate, given that the former represents a set of trait evaluations of the critic andthe latter represents attributions of motive. It should also be noted that in Experi-ments 2 and 3 (reported later), factor analysis of the same items reveals a perfect
four-factor solution.
Results
Preliminary analyses
Preliminary analyses showed that there were no significant main effects of sex on any
measures and that sex did not interact with speaker type or strategy. Analysis of theidentification measure showed that participants identified strongly with their
national identity overall (M = 5.66) and that levels of identification were statisticallyequivalent across all conditions. For participants in the praise condition, indepen-dent t tests were conducted to see if the order of the praise relative to the criticism
had any effect on the dependent measures. None of the t tests were significant(all ps . .32), meaning that it did not matter whether praise was delivered before
or after the criticism. Consequently, the data for the praise conditions were collapsedacross order.
To test how salient the praise was relative to the criticisms, we conducted a2 (speaker type) 3 2 (praise vs. criticism) mixed model analysis of variance
(ANOVA) on the salience scales. The only significant effect was that criticisms onthe whole were more salient to participants (M = 5.34) than the praise (M = 4.26),F(1, 53) = 31.75, p, .001, h2 = .38. Speaker type had no effect on salience, either as
a main effect (p = .91) or as an interaction (p = .42).
Main analyses
Results for constructiveness, negativity, and agreement were analyzed using a series
of 2 (speaker type: in-group vs. out-group) 3 2 (strategy: praise vs. no praise)between-groups ANOVAs. Results for Time 2 likeability were analyzed using
a 2 3 2 analysis of covariance (ANCOVA) with Time 1 likeability scores entered
Strategies to Soften Criticism M. J. Hornsey et al.
80 Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association
as a covariate. All effect sizes reported here and in subsequent studies are partial h2.Means and standard deviations are summarized in Table 1.
As predicted, main effects of speaker type emerged on constructiveness, F(1, 103)= 30.00, p , .001, h2 = .23, negativity, F(1, 103) = 18.01, p , .001, h2 = .15, Time 2
likeability, F(1, 102) = 6.49, p = .012, h2 = .06, and agreement, F(1, 103) = 7.98,p = .006, h2 = .07. Consistent with the intergroup sensitivity effect, in-group criticswere liked more (M = 4.04), were attributed more constructive motives (M = 4.28),
and were agreed with more (M = 4.08) than were out-group critics (Ms = 3.49, 3.06,and 3.13, respectively). In-group critics also aroused less negativity (M = 3.97) than
did out-group critics (M = 5.18).Main effects of praise also emerged on constructiveness, F(1, 103) = 27.70,
p , .001, h2 = .21, Time 2 likeability, F(1, 102) = 13.26, p , .001, h2 = .12,and agreement, F(1, 103) = 4.74, p = .032, h2 = .04. Critics who used praise were
liked more (M = 4.16), were attributed more constructive motives (M = 4.24), andwere agreed with more (M = 3.95) than critics who did not praise (Ms = 3.37, 3.06,and 3.22, respectively). The tendency for critics who praised (M = 4.34) to arouse less
negativity than those who did not praise (M = 4.83) was nonsignificant, F(1, 102) =2.55, p = .11, h2 = .02. On none of the measures did strategy significantly inter-
act with speaker type (all ps . .64).
Why does praise help? The role of constructiveness
In summary, attaching praise to criticism helped critics overcome defensiveness on
measures of agreement and likeability. It was predicted that the reason for this wouldbe that critics who praised would be attributed more constructive motives for their
criticism than those who did not praise. To test for mediation, a series of regressionswas conducted with strategy dummy-coded such that 0 = no praise and 1 = praise.
Consistent with the strategy for conducting the ANOVAs, when conductingregressions on Time 2 likeability, Time 1 likeability scores were controlled for atthe first stage.
When included in a regression along with strategy, constructiveness significantlypredicted both agreement (b = .59, p, .001) and likeability (b = .68, p, .001), thus
Table 1 Effects of Sweetening on Responses to Criticism: Experiment 1
Measure
Australian
(In-Group Critic)
Non-Australian
(Out-Group Critic)
Praise No Praise Praise No Praise
Constructiveness 4.87c (1.28) 3.63b (1.18) 3.58b (1.02) 2.54a (0.96)
Time 2 likeability 4.40c (1.30) 3.68b (1.08) 3.92bc (1.21) 3.06a (1.22)
Negativity 3.81a (1.65) 4.14ab (1.64) 4.89bc (1.35) 5.47c (1.16)
Agreement 4.50c (1.85) 3.60ac (1.87) 3.39ab (1.52) 2.87a (1.46)
Note: Numbers in parentheses are standard deviations. Means that do not share a subscript
are significantly different according to Duncan posthoc test (p , .05). Means for Time 2
likeability are adjusted after having covaried out Time 1 likeability.
M. J. Hornsey et al. Strategies to Soften Criticism
Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association 81
satisfying a key condition for mediation. Furthermore, after including constructive-ness in the model, significant effects of strategy on agreement (b = .21, p = .031) and
likeability (b = .31, p, .001) were reduced to nonsignificance (agreement: b = 2.04,p = .64; likeability: b = .03, p = .65). In each case, this represented a significant drop
in variance according to the Sobel test (agreement: z = 3.89, p , .001; likeability:z = 3.81, p , .001). In sum, there was evidence that attributions of constructivenessfully mediated the effects of praise on agreement and likeability.
Discussion
There has been limited empirical research examining whether negative feedback is
more likely to be absorbed if the criticisms are contextualized by praise, either at theinterpersonal level or the intergroup level. Although such a strategy is intuitively
appealing, there are theoretical reasons to challenge the assumed link between praiseand responses to criticism that make empirical scrutiny of the relationship important.If praise is recognized as simply a strategy to reduce defensiveness toward the
criticisms, then people might dismiss or gloss over the positive feedback and focusexclusively on the more diagnostic, threatening, and ego-involving criticisms, ren-
dering the praise ineffectual. Indeed, in our sample, we found that people who wereexposed to both praise and criticism of Australia were far more likely to cognitively
focus on the negative than the positive feedback.Despite this, there was some evidence that praise could help reduce defensiveness
toward group criticism. The effects of praise were most marked on ratings of howlikeable the critic was seen to be: Critics who used praise were seen to be much more
likeable than those who did not. A weaker but still reliable effect emerged on agree-ment: When praise was used participants agreed with the criticisms more than whenpraise was not used. On negativity, however, praise had no reliable effect at all.
Despite the mixed picture, the overall conclusion is that it is better to praise thanto not praise, particularly if being liked is important to you.
In addition to testing the effects of praise on reducing defensiveness, another aimof the current study is to examine why praise might help reduce defensiveness. Our
mediation analysis showed evidence consistent with our argument that attributionsunderpin responses to group-directed criticism. Specifically, critics who used praise
were assumed to be motivated by more constructive reasons than did those who didnot praise, and this led to higher ratings on likeability and agreement. Thus, thecurrent study does not just provide an early attempt to empirically assess the merits
of praise, it also provides clues as to the psychological mechanisms through whichpraise might work.
It should be noted that in all cases, the effects of praise need to be interpretedagainst the backdrop of a robust intergroup sensitivity effect. In all cases, out-group
critics aroused more defensiveness than did in-group critics. The effects of praiseworked equally for in-group and out-group critics, so this strategy should be seen as
a way of reducing defensiveness per se rather than a way of reducing the intergroup
Strategies to Soften Criticism M. J. Hornsey et al.
82 Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association
sensitivity effect. Inspection of effect sizes shows that the effect of group membershipwas greater than the effect of praise on ratings of agreement and negativity. Indeed,
across all the measures, an out-group critic who used praise aroused just as muchdefensiveness as an in-group member who did not use this strategy. This helps rein-
force the uphill battle that people face when directing negative feedback at out-groups.
Experiment 2
We have seen in Experiment 1 that praise can help outsiders reduce defensiveness
toward criticism, at least on dimensions of agreement and likeability. An alternativestrategy is for out-group critics to acknowledge the failings of their own group as well
as the in-group with regard to the criticism (sharing). As described earlier, the socialidentity perspective suggests that people gain and maintain self-esteem through
membership in groups that offer positive distinctiveness. Thus, when an out-groupmember criticizes a group to which we belong, we might be quick to assume thattheir comments are simply part of this ongoing struggle for intergroup supremacy.
If, however, an out-group critic also articulates the shortcomings of his or her owngroup, it would help defuse the notion that the out-group critic is motivated by
a need to demonstrate his or her group’s superiority. In the absence of this expla-nation for the criticisms, the in-group members might be more likely to rely on an
alternative explanation: that the person is making these comments with the bestinterests of the group in mind. These attributions of constructiveness might then
flow on to lower levels of negativity, more positive evaluations of the critic, anda greater willingness to embrace the truth within the criticisms. To our knowledge,
there have been no empirical tests of whether such a strategy would work, either inthe context of interpersonal or intergroup criticism.
It is important, however, not to automatically assume that sharing would work
solely because it promotes more positive attributions about motive. An alternativeand equally plausible explanation is that critics who acknowledge problems within
their own group are seen to be less hypocritical than those who do not, which leadsto a more positive attitude toward the comments. Indeed, qualitative research on
persuasion has shown that perceptions of hypocrisy in others are a key reason forresisting messages stemming from schools and wider society (e.g., Booth-Butterfield,
Anderson, & Williams, 2000; Flacks, 1967). It seems reasonable, then, to test whetherany effects of acknowledgment are driven by constructiveness over and aboveperceptions of hypocrisy.
In Experiment 2, Australian participants read what they believed to be an extractfrom an interview in which a person criticized Australians for being racist toward
indigenous people and Asians. In the in-group condition, the criticisms wereattributed to another Australian. In the out-group conditions, the criticisms were
attributed to a foreigner. Three versions of the out-group criticism were included inthe study. In the acknowledgment condition, the out-group critic acknowledged the
failings of his or her own country in regard to racism and in the no-acknowledgment
M. J. Hornsey et al. Strategies to Soften Criticism
Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association 83
condition they did not. We also included a control condition in which the out-groupcritic acknowledged racism existed in other countries but did not mention his or her
own (other-group acknowledgment). By including this condition, we were able todisentangle the unique effects of own-group acknowledgment from the effects of
merely diffusing criticism across multiple groups.In line with the intergroup sensitivity effect, it was expected that the in-group
critic would arouse less defensiveness than the out-group critics who did not
acknowledge similar problems within their own country (the no-acknowledgmentand the other-group acknowledgment conditions). It was expected that the out-
group critic who acknowledged the failings of his or her own group would arouseless defensiveness than the other out-group critics, resulting in attenuation or even
elimination of the intergroup sensitivity effect. Furthermore, it was expected that theeffects of acknowledgment would be mediated by constructiveness, over and above
perceptions of hypocrisy. Specifically, it was expected that the out-group critic whoacknowledged his or her own group’s failings would be attributed more constructivemotives than would the out-group critic who did not use acknowledgment and that
this difference in attribution would underpin the effects of acknowledgment onlikeability, agreement, and negativity.
Method
Participants and design
Participants were 160 undergraduate student volunteers (118 females, 42 males,M = 21.26 years) who nominated Australia as their primary national identity.
Assuming a moderate effect size of .30, we needed only 128 participants to achieveacceptable power (80%); thus, the current sample size was considered sufficientlypowerful. Participants were randomly allocated to one of the four between-groups
conditions: in-group criticism, out-group criticism with no acknowledgment, out-group criticism with acknowledgment of another group’s flaws, and out-group
criticism with acknowledgment of own group’s flaws.
Procedure
The cover story and procedure were similar in most respects to that used in Experi-
ment 1. After completing demographic details and the three-item measure ofnational identification (a = .86), participants read details about the target speaker.The demographic descriptions of the speaker were the same as those used in Experi-
ment 1, except this time three out-group nationalities were used: speakers weredescribed as being either citizens of the United States, New Zealand, or Canada
(out-group nationality was counterbalanced within each out-group condition).As in Experiment 1, the speakers gave a small amount of information about their
hobbies, their families, and the fact that they had lived in their country of birthall their life. After receiving this information, participants completed the Time 1
likeability items described earlier (a = .91).
Strategies to Soften Criticism M. J. Hornsey et al.
84 Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association
Participants then read the speaker’s response to the question ‘‘What do you thinkof Australians?’’ In the in-group condition and the out-group/no-acknowledgment
condition the response read: ‘‘When I think of Australians I think of us (them) asfairly racist. We’re (They’re) racist toward Aborigines and intolerant of Asians.’’ In
the condition where the out-group speaker acknowledged similar flaws in othercountries, the following sentences were added: ‘‘But I think this is something othercountries struggle with too. For instance, (members of X country) are racist towards
their indigenous people and towards Asian people.’’ In this condition, the speakeralways referred to members of another out-group; for example, a New Zealander
might say that Canadians are also racist, or a Canadian might say that U.S. citizensare also racist. In the condition where the out-group speaker acknowledged similar
flaws in his or her own country, these sentences were replaced with the following:‘‘But I think this is something we struggle with too. For instance, I also think we are
racist towards our indigenous people and towards Asian people.’’To account for ratings of hypocrisy, participants rated the extent to which they
found the criticisms to be ‘‘hypocritical.’’ They also completed the same measures of
constructiveness (a = .85), Time 2 likeability (a = .95), negativity (a = .92), andagreement (r = .90) used in Experiment 1. This time, a factor analysis with oblimin
rotation revealed a four-factor solution corresponding perfectly to the four key-dependent measures.
Results
Preliminary analyses
Preliminary analyses showed that there were no significant main or interactive effectsof sex on any measures. Again, participants identified strongly with their national
identity (M = 5.28), and an ANOVA revealed that levels of identification wereequivalent across all conditions (p = .22). A series of ANOVAs was run among the
participants in the out-group conditions to check whether ratings of negativity,likeability, agreement, and constructiveness differed depending on which out-groupparticipants were exposed to. Results showed that ratings were equivalent regardless
of whether the out-group critic was a New Zealander, a Canadian, or a U.S. citizen(all ps . .19). Thus, we felt comfortable collapsing these groups together.
Main analyses
Analyses were conducted using one-way between-groups ANOVAs with four levels.When analyzing Time 2 likeability, Time 1 likeability was entered as a covariate.
Significant main effects were followed up with Duncan’s posthoc tests. Means andstandard deviations are summarized in Table 2.
Unexpectedly, no effects of condition emerged on ratings of agreement, F(3, 156)= 1.11, p = .35, h2 = .02. However, as predicted, main effects of condition emergedon constructiveness, F(3, 156) = 15.73, p , .001, h2 = .23, likeability, F(3, 155) =
6.59, p , .001, h2 = .11, and negativity, F(3, 156) = 10.11, p , .001, h2 = .16. For
M. J. Hornsey et al. Strategies to Soften Criticism
Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association 85
ratings of constructiveness, posthoc analysis revealed that the highest ratings of
constructiveness occurred when the critic was an in-group member and when anout-group critic acknowledged similar problems within his or her own country(these conditions did not differ from each other). Ratings of constructiveness in
these conditions were significantly higher than when an out-group critic acknowl-edged similar problems in another group, which in turn were higher than when out-
group critics did not use acknowledgment at all.For ratings of negativity a similar pattern emerged: Least negativity emerged
when the critic was an in-group member and when an out-group critic acknowl-edged similar problems within his or her own country (these conditions did not
differ from each other). Ratings of negativity were significantly lower in the in-groupcondition than in the condition where out-group critics acknowledged similar prob-lems in other groups. Negativity in the out-group/no-acknowledgment condition
was significantly higher than in all other conditions.Finally, posthoc analyses on the Time 2 likeability scores showed that the highest
likeability ratings occurred when the critic was an in-group member and when anout-group critic acknowledged similar problems within his or her own country
(these conditions did not differ from each other). Ratings of likeability were signif-icantly higher in these conditions than when an out-group critic acknowledged
similar problems in another group, or did not use acknowledgment at all (these lasttwo conditions also did not differ from each other).
Why does acknowledgment help? The role of constructiveness
By acknowledging similar problems within one’s own country, out-group critics
could reduce defensiveness on likeability and negativity to the point where theintergroup sensitivity effect was eliminated. Consistent with the attributional argu-
ments presented here, it was predicted that these effects would be mediated byattributions of constructiveness. To test this, we selected only participants in the
out-group/no-acknowledgment and the out-group/own-group acknowledgmentconditions. This way we were able to reduce our four-level IV into a dichotomous
Table 2 Effects of Sharing on Responses to Criticism: Experiment 2
In-Group Critic
Out-Group Critic
Acknowledgment of Group Failings
None Other Group Own Group
Constructiveness 4.32c (1.30) 2.69a (1.04) 3.31b (1.28) 3.98c (0.94)
Time 2 likeability 4.18b (0.90) 3.46a (1.06) 3.60a (1.08) 4.13b (0.86)
Negativity 3.71a (1.45) 5.31d (1.12) 4.69bc (1.49) 4.18ab (1.34)
Agreement 4.25 (1.62) 3.78 (1.47) 3.66 (1.54) 3.97 (1.47)
Note: Numbers in parentheses are standard deviations. Means that do not share a subscript
are significantly different according to Duncan posthoc test (p , .05). Means for Time 2
likeability are adjusted after having covaried out Time 1 likeability.
Strategies to Soften Criticism M. J. Hornsey et al.
86 Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association
variable that could be used in regression. This variable was dummy coded such that theno-acknowledgment condition was coded 0 and the own-group acknowledgment
condition was coded 1. When conducting the regressions on Time 2 likeability, ratingsof likeability at Time 1 were included at the first step as a control.
When included in a regression along with condition, constructiveness signifi-cantly predicted both negativity (b = 2.41, p , .001) and likeability (b = .51,p , .001). After including constructiveness into the model, significant effects of
condition on negativity (b = 2.42, p , .001) and likeability (b = .32, p , .001)were reduced to being either marginally significant (negativity: b = 2.19, p = .090)
or nonsignificant (likeability: b = .04, p = .63). In each case, this change in varianceexplained was significant according to the Sobel test (negativity: z = 23.08, p = .002;
likeability: z = 4.06, p , .001). This suggests that acknowledgment helped the out-group critic because acknowledgment increased attributions of constructiveness.
Why does acknowledgment help? The role of hypocrisy
Despite the mediation analyses reported above, it is important also to test whether
perceptions of hypocrisy play a mediating role. ANOVA on the hypocrisy ratingsconfirmed that out-group critics who did not acknowledge (M = 4.68) or who
acknowledged the failings of groups other than their own (M = 5.14) were seen tobe more hypocritical than those who acknowledged failings in their own group
(M = 3.80), who in turn were seen to be more hypocritical than in-group speakers(M = 3.02), F(3, 156) = 15.19, p , .001, h2 = .23. To test whether acknowledgment
aroused less defensiveness in part because it reduced perceptions of hypocrisy, werepeated the regressions reported above, but this time with hypocrisy included
alongside constructiveness as a mediator. When entered together with strategy, bothconstructiveness and hypocrisy proved to be significant predictors of negativity andlikeability (all ps , .001). Subsequent regressions were performed without construc-
tiveness in the model to see if perceptions of hypocrisy on their own could mediatethe effects. In each case, however, the effects of strategy remained highly significant
after controlling for hypocrisy (negativity: b = 2.25, p = .003; likeability: b = .22,p = .011). Thus, there is little evidence that hypocrisy is playing a unique role in
explaining the effects of acknowledgment.
Discussion
Overall, the data suggest that acknowledgment (or sharing) is an effective strategy for
reducing defensiveness. When claiming that Australians were racist, out-group crit-ics were liked more and aroused less negativity when they acknowledged that their
own country had problems with racism than when they did not. The positive effect ofacknowledgment was strong enough to eradicate the intergroup sensitivity effect on
these measures. Furthermore, we were able to gain insights into the psychologicalunderpinnings of the effects. Consistent with our theoretical focus on attributions,
out-group critics who acknowledged problems within their own group were seen to
M. J. Hornsey et al. Strategies to Soften Criticism
Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association 87
be motivated by more constructive reasons than were those who did not acknowl-edge, and it was this shift in attributions that explained the effect. Although acknowl-
edgment also helped reduce perceptions that the comments were hypocritical, therewas no evidence that perceptions of hypocrisy were playing a mediating role.
Although the data broadly conformed to expectations, there were two unex-pected results. First, although the means trended in a direction that mirrored theother dependent measures, the manipulations had no significant effect on agree-
ment. This includes a failure to find the intergroup sensitivity effect on agreement;that is, there is no significant difference between the in-group and the out-group/no-
acknowledgment condition. This noneffect is particularly surprising given that theintergroup sensitivity effect has proven very reliable in the past and that sizeable
intergroup sensitivity effects emerged on the other dependent measures of the cur-rent experiment. Furthermore, an intergroup sensitivity effect emerged on this mea-
sure in Experiment 1 and (as will be seen later) in Experiment 3, which used almostidentical methods to the current experiment. Regardless of the reason for the non-effect, it should be noted that no conclusions can be drawn with regard to the power
of acknowledgment to increase levels of agreement with critical messages.A second unexpected effect was that acknowledging failings in groups other than
your own succeeded in reducing defensiveness (relative to a no-acknowledgmentcondition) on some measures. The effect of other-group acknowledgment was not as
strong as own-group acknowledgment; for example, it did not eliminate the inter-group sensitivity effect on any of the measures and it had no reliable effect at all on
ratings of likeability. But it does seem to be the case that when critics diffused thenegative feedback across multiple out-groups, participants were more likely to see
the critic as being driven by constructive motives and felt less negatively toward thecriticisms in general. In other words, acknowledging that similar problems exist ingroups other than the target group is more effective than not acknowledging at all.
This raises the question of whether in-group critics can benefit from such a strategy.For example, if in-group critics criticized their country, but contextualized the
criticisms by saying that the problems were shared by other groups as well, it mightbe that this would also help reduce defensiveness. Further research is needed to
clarify some of the nuances associated with the power of other-group acknowledg-ment to reduce defensiveness.
Experiment 3
In Experiment 3, we assess a third strategy for reducing defensiveness that possessesa similar intuitive appeal to the first two but would not be likely to change attri-
butions of constructiveness. This is the strategy of spotlighting; that is, of putting onthe record that you intend your comments to apply to just a portion of the group
rather than to the whole group. If our attributional explanation of defensiveness iscorrect, then we should find that this strategy has little or no effect on responses
to criticism.
Strategies to Soften Criticism M. J. Hornsey et al.
88 Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association
Despite the fact that spotlighting does not directly address the attributional biasthat drives the intergroup sensitivity effect, there are at least three reasons why it
might be otherwise expected to work. First, spotlighting seems consistent with theprinciple of specificity that researchers on interpersonal criticism consider important
when presenting feedback (Baron, 1988). Second, research suggests that peopledisapprove of individuals who make generalizations about groups, even if the gen-eralizations apply to ostensibly positive traits (Mae & Carlston, 2005). Thus, critics
who are careful not to make their generalizations about the group overly sweeping(i.e., by spotlighting) might be seen as less objectionable and their comments might
subsequently seem more palatable.A third possible reason that spotlighting might work is that it helps reduce any
threat to personal identity associated with group criticism. When somebody criticizesa group to which we belong, this undoubtedly represents a challenge to our social or
group identity. However, it is ambiguous whether this also represents a criticism of ourvalues as an individual. Imagine if your country was criticized for being racist using thekind of language used in the scripts of the current experiments; it is unclear in this
situation whether the critic sees these values in all, most, or some of your compatriots.But our assumptions about this might have an impact on the degree of threat we feel.
If we assume that the critic only intended the criticism to apply to some of ourcompatriots, we might feel relaxed in the knowledge that the critic would not automat-
ically assume that we were racist individuals, and that he or she would probablyexonerate us from this allegation if the critic were to meet us. In this case, the criticism
represents a threat to our social identity without representing a threat to our personal orindividual identity. But if we assume that the critic were to believe that all or most of our
compatriots share this quality, then this potentially adds another stratum of threat to thesituation; not only is this a criticism of the values of our country but also is a criticism ofour own personal values. This might result in a heightened degree of defensiveness.
There is reason to believe that any anxieties associated with feeling ‘‘lumpedtogether’’ in the face of group criticism might be particularly strong when being
criticized by an out-group member. Research shows that people typically have a morehom*ogenized view of out-groups than they do of their own groups (Ostrom &
Sedikides, 1992), partly because they have more experience with their own group(Linville, Fischer, & Salovey, 1989) and partly because they have more concrete
representations of in-group exemplars (Park, Judd, & Ryan, 1991). If people havenaı̈ve theories in line with this out-group hom*ogeneity effect, then they wouldprobably assume that criticisms from out-group members were intended to apply
to the majority of the group. In contrast, when we hear criticisms from in groupmembers, we would assume that the in-group critic would be aware of the high
degree of variation that exists within the group in terms of attitudes and behaviors.In short, it could be that criticisms from outsiders are seen to be more overly
generalized than criticisms from insiders, and that this might help contribute tothe intergroup sensitivity effect independent of any attributions about whether or
not the critic intended the comments to be constructive. One way for an out-group
M. J. Hornsey et al. Strategies to Soften Criticism
Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association 89
critic to overcome this perception is to make it clear that the criticisms do not applyto everyone, but rather to a portion of the group population.
In Experiment 3, Australian participants received criticism of their country fromeither an in-group or an out-group member. The critic either used spotlighting or did
not. We predicted an intergroup sensitivity effect, such that out-group memberswould elicit more defensiveness than in-group members. Given the lack of previousresearch on generalizations with regard to criticism, it was not possible to make strong
predictions with regard to the effects of spotlighting. On the basis of the logic outlinedabove, we might expect that spotlighting would result in less defensiveness than
a control condition and that the ameliorative effects of spotlighting might be particu-larly noticeable for out-group criticisms. However, if we are to take seriously the
notion that attributions of constructiveness play a central role in driving defensiveness,we might expect that spotlighting will have no reliable effect on responses to criticism.
Method
Participants and design
Eighty-two undergraduate psychology students (62 females, 20 males, M = 20.87
years) participated in the study in return for course credit. Participants were ran-domly allocated to the four conditions of a 2 (speaker type: in-group vs. out-group)
3 2 (strategy: spotlighting vs. no spotlighting) between-group design. Participantswere only included if they nominated Australia as their primary national identity.
Prospective power analyses revealed that this sample gave us a 76.5% chance ofdetecting a moderate effect size.
Materials and procedure
The cover story, materials, and questionnaire were similar to those used in Experi-
ment 2. In the no-spotlighting condition, the speaker said:
When I think of Australians I think of them (us) as being fairly racist. They’re(We’re) racist towards Aborigines and intolerant of Asians.
In the spotlighting condition the speaker said:
When I think of Australians I think some of them (us) are fairly racist. Some of
them (us) are racist towards Aborigines and intolerant of Asians. Of course notall Australians are like that, but many are.
These comments were either attributed to an in-group member (an Australian) or anout-group member (a U.S. citizen).
Immediately after the scripts, participants were asked: ‘‘How many Australiansdo you think J. Benson believes his comments are true of?’’ Participants could circle
one option from an 11-point scale, which began with 0% and increased by 10%integers until reaching the maximum (100%). This measure was included as a check
on the manipulation of strategy. Scales of likeability (Time 1 a = .94; Time 2 a = .96),
Strategies to Soften Criticism M. J. Hornsey et al.
90 Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association
constructiveness (a = .84), negativity (a = .93), and agreement (r = .92) were thesame as those used in Experiments 1 and 2. Furthermore, factor analysis of all the
items using oblimin rotation revealed a four-factor solution corresponding perfectlyto the constructs described above.
Results
Preliminary analyses
Because some of the conditions did not include more than one male, it was not
possible to run the full design crossed with participant sex. However, a series of 2(participant sex) 3 2 (strategy) ANOVAs revealed no significant main or interactive
effects of sex on any measures. Given that sex in the previous studies was shown notto interact with speaker type, we felt comfortable collapsing the data across sex.Given the small representation of male participants, it should be noted that caution
should be exercised before generalizing the current results to males.As in previous studies, our sample reported strong identification with Australia
(M = 5.59). Unexpectedly, a 2 (speaker type) 3 2 (strategy) ANOVA on identifica-tion revealed a main effect of strategy, F(1, 78) = 4.50, p = .037, h2 = .05, and
a significant two-way interaction, F(1, 78) = 4.21, p = .044, h2 = .05. Follow-upanalyses revealed unusually low levels of identification in the in-group/spotlighting
condition (M = 5.00) relative to the in-group/no-spotlighting condition (M = 6.00).Because identification scores were measured prior to the manipulation of both
independent variables, this effect can only be attributed to chance variation inparticipant assignment to conditions. To ensure that the effects of the independentvariables are not confounded by identification, all analyses below were conducted
using 2 (speaker type) 3 2 (strategy) ANCOVAs, with identification included asa covariate. It should be noted that the pattern of results remained the same even
when identification was not controlled for. Consistent with the policy used in Experi-ments 1 and 2, Time 1 likeability was included as an additional covariate when
conducting analyses on Time 2 likeability. Adjusted means and standard deviationsfor all measures are reported in Table 3.
Manipulation check
A 2 3 2 ANCOVA on the manipulation check revealed only a main effect of strategy,
F(1, 77) = 4.11, p = .046, h2 = .05. Participants in the no-spotlighting conditionbelieved that the critic intended their comments to apply to more Australians
(M = 60.49) than did participants in the spotlighting condition (M = 50.73). Thus,the manipulation was successful.
Main analyses
Consistent with the intergroup sensitivity effect, main effects of speaker type
emerged on constructiveness, F(1, 77) = 63.90, p , .001, h2 = .45, negativity,F(1, 77) = 12.75, p , .001, h2 = .14, Time 2 likeability, F(1, 76) = 17.92, p ,
.001, h2 = .19, and agreement, F(1, 76) = 11.77, p , .001, h2 = .13. In-group critics
M. J. Hornsey et al. Strategies to Soften Criticism
Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association 91
were liked more (M = 4.18), were attributed more constructive motives (M = 4.64),and were agreed with more (M = 4.14) than were out-group critics (Ms = 3.21, 2.76,
and 2.88, respectively). In-group critics also aroused less negativity (M = 3.86) thandid out-group critics (M = 5.01).
In contrast, strategy had no significant main or interactive effects on any of themeasures. There was a marginal main effect of strategy such that speakers whospotlighted tended to be attributed more constructive motives for their comments
(M = 3.93) than those who did not spotlight (M = 3.47), F(1, 77) = 3.71, p = .058,h2 = .05, but all other effects of strategy were emphatically nonsignificant (ps . .26,
h2 range from .00 to .02). Clearly, on the constructiveness measure, it is possible thatmore participants could have pushed the effect past the conventional level of signifi-
cance. But it should be noted that the effect size (h2 = .05) compares unfavorably tothe double-digit effect sizes found with respect to sweetening and sharing. Further-
more, power analyses revealed that at least 165 participants would have to be testedto achieve an 80% likelihood of detecting a significant effect on the basis of this effectsize. On negativity, the effect size (partial h2) for the main effect and the interaction
was .004. Posthoc power analyses revealed that at least 1,961 participants would haveto be tested to achieve an 80% likelihood of detecting a significant effect on the basis
of this effect size. On measures of Time 2 likeability and agreement, the meanstrended such that participants reported the critic to be less likeable and agreed with
the critic less in the spotlighting condition than in the no-spotlighting condition. Onthe basis of the direction of the means, it is clear that extra statistical power would
not uncover a beneficial effect of spotlighting on these measures.
Discussion
Consistent with the intergroup sensitivity effect, people’s responses to criticism were
not just influenced by the content of the message but by the source of the message:In-group critics were liked more, were attributed more constructive motives, were
agreed with more, and aroused less negativity than out-group critics. In this case,
Table 3 Effects of Spotlighting on Responses to Criticism: Experiment 3
Australian (In-Group) Non-Australian (Out-Group)
Spotlighting No Spotlighting Spotlighting No Spotlighting
Constructiveness 4.85c (1.03) 4.42c (1.28) 3.01b (0.90) 2.51a (0.96)
Time 2 likeability 3.86b (1.10) 4.50c (1.19) 3.27a (1.33) 3.15a (1.13)
Negativity 3.78a (1.43) 3.95a (1.35) 4.92b (1.53) 5.11b (1.56)
Agreement 4.07b (1.68) 4.21b (1.65) 2.90a (1.56) 2.86a (1.71)
Note: Numbers in parentheses are standard deviations. Means with different subscripts are
significantly different according to Duncan posthoc test (p, .05). Means for all measures are
adjusted after having covaried out national identification. Means for Time 2 likeability are
adjusted after also having covaried out Time 1 likeability.
Strategies to Soften Criticism M. J. Hornsey et al.
92 Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association
however, the strategy did not have any reliable effects on responses. Spotlighting hadno significant effect on how people felt about the critic or the message.
Naturally, caution needs to be exercised when interpreting nonsignificant effects,but it is difficult in this case to attribute the null result to anything other than the fact
that the strategy was ineffective. It cannot be attributed to a failed manipulationbecause the manipulation check worked: Those who spotlighted were assumed to begeneralizing their comments to a smaller percentage of group members than those
who did not spotlight. It cannot be attributed to unresponsive or insensitive meas-ures because the same measures have yielded sizeable effects of group membership.
Neither is it likely to be a question of statistical power, given that the effect sizes onnegativity, likeability, and agreement were negligible. Furthermore, on likeability and
agreement, the means trended nonsignificantly in the opposite direction to whatwould be expected if spotlighting were to work. It seems unlikely that increases in
sample size would have changed the broad pattern of results.In sum, it appears that spotlighting simply does not help. On one hand, this
might seem surprising because spotlighting is an intuitive and widely used strategy
(‘‘I’m not saying you’re all like this .’’). From a theoretical point of view, the reasoncould be that the strategy does not significantly alter the attributions that people
make about the motives for the criticism. Qualifying the comments does not sub-stantially change whether the comments are motivated by constructive or destructive
reasons and so does not help ease the suspicion that triggers defensiveness.
General discussion
There is a generalized tendency for people to feel more defensive when they receivegroup-directed criticism from an outsider than an insider. The three studies herereinforce this point: Out-group critics were liked less, agreed with less (In Experi-
ments 1 and 3 only), and aroused more negativity than when the same commentswere made by an in-group member. The reason for this is that, in the absence of any
other information, outsiders are attributed less constructive motives than insiders,which then promotes high levels of defensiveness. This finding on the one hand is
relatively pessimistic because it suggests that outsiders face defensiveness because ofsomething they cannot readily change—their group membership. But it also offers
a signpost as to how to overcome the problem posed by the intergroup sensitivityeffect. If outsiders can defuse suspicion about the integrity of their motives, thentheoretically defensiveness should be reduced.
The studies reported here investigated three strategies that, on the basis of bothintuition and theory, might be expected to be effective in reducing defensiveness
in the face of group criticism. In Experiments 1 and 2, some encouraging signsemerged. If out-group critics contextualize their criticisms with praise, or if they
acknowledge that similar problems exist in their own group, then defensiveness isreduced, at least on some measures. This suggests that, although out-group member-
ship is an impediment to acceptance when it comes to criticism, this impediment can
M. J. Hornsey et al. Strategies to Soften Criticism
Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association 93
be overcome through the strategic use of language. This is not a trivial point.Identifying strategies to overcome the intergroup sensitivity effect is something that
is potentially of applied relevance to members of oppressed minorities, human rightsgroups, environmental groups, and other groups for whom collective action is a
matter of trying to win over change from the outside. In this sense, reducing defen-siveness in the face of out-group criticisms seems an important precondition forreform, reconciliation, and reinvention, and identifying strategies that can assist out-
group members to be heard is an important applied research question.The news was less positive, however, with regard to the third strategy tested here.
Spotlighting—in the sense of putting on the record that the criticisms were notintended to apply to everyone—had no reliable effect on responses, either positive
or negative. Although there are inevitable interpretational difficulties associated withdrawing meaning out of a null result, from an attributional perspective it can be
argued that spotlighting failed because it did not address the suspicion about motivethat has previously been shown to work against outsiders who criticize.
The paradigm we used here, in which participants read a handful of sentences
from what are ostensibly extracts from an interview, was used because we felt itmaintained a sensible balance between internal and external validity. Because par-
ticipants believe the criticisms are real comments from a real person, we have suc-ceeded in placing the participants in a potentially threatening and emotion-laden
environment while simultaneously maintaining a level of experimental control thatwould not be possible if participants were to see a speaker or even to hear the
comments by audiotape. Having said that, we acknowledge that it would be idealin the future to examine the strategies in more naturalistic environments and poten-
tially using longer scripts.We also consciously used the same target group (Australians) and the same
criticisms across experiments so as to create a level playing field within which strat-
egies could be meaningfully compared. If we had changed the target group and thecriticisms across experiments, it would not be possible to determine if any differences
observed across studies was due to the different strategies used or the differentintergroup contexts invoked. As mentioned earlier, the intergroup sensitivity effect
and the role of constructiveness in producing the effect has now been demonstratedacross a range of contexts, including criticism of schools, university students, math
science students, social science students, Queenslanders, allied health professionals,and Muslims (see Hornsey, 2005). Given that these effects have proven reliable androbust, it would be surprising if the strategies had dramatically different effects
depending on target group or the content of the criticisms.
Summary and conclusions
Outsiders who criticize groups struggle to get an open-minded audience, but this is
not to say they are doomed to face defensiveness. Criticisms are a communicative actframed in language, and we have shown here that strategic use of rhetoric can be
effective in reducing hostile responses to legitimate group criticism. So it is not
Strategies to Soften Criticism M. J. Hornsey et al.
94 Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association
advised that outsiders simply let the content of their criticisms to speak for them-selves; if they spend more time framing their criticisms with certain rhetorical flour-
ishes, then this can pay off in terms of reducing hostility and winning acceptanceof the message.
Although we have assessed only a limited number of strategies here, a pattern isbeginning to emerge that helps predict which strategies will work and which will not.No matter how intuitively appealing, it seems that for strategies to work they need to
be successful in defusing the suspicion about motive that underpins the intergroupsensitivity effect. This finding has theoretical import, in the sense that it helps lend
support to our attributional model of defensiveness in the face of group criticism.It also serves a practical purpose in terms of providing a rudimentary how-to guide
for people who are motivated or forced to engage in critical communication acrossgroup divides.
Acknowledgments
The research presented in this paper was supported by an Australian ResearchCouncil-Discovery grant.
References
Abrams, D., Hogg, M. A., & Marques, J. (Eds.) (2005). The social psychology of inclusion and
exclusion. New York: Psychology Press.
Baron, R. A. (1988). Negative effects of destructive criticism: Impact on conflict, self-efficacy,
and task performance. Journal of Applied Psychology, 73, 199–207.
Baumeister, R. F., Bratslavsky, E., Finkenauer, C., & Vohs, K. D. (2001). Bad is stronger than
good. Review of General Psychology, 5, 323–370.
Bavelas, J. B., Black, A., Bryson, L., & Mullett, J. (1988). Political equivocation: A situational
explanation. Journal of Language and Social Psychology, 7, 137–145.
Berscheid, E., & Walster, E. (1978). Interpersonal attraction. Reading, MA: Addison-Wesley.
Bies, R. J., & Sitkin, S. B. (1992). Explanation as legitimization: Excuse-making in
organizations. In M. L. McLaughlin, M. J. Cody, & S. J. Read (Eds.), Explaining the self
to others: Reason-giving in a social context (pp. 183–198). Hillsdale, NJ: Erlbaum.
Booth-Butterfield, M., Anderson, R., & Williams, K. (2000). Perceived messages from schools
regarding adolescent tobacco use. Communication Education, 49, 196–205.
Bradbury, T. N., & Fincham, F. D. (1990). Attributions in marriage: Review and critique.
Psychological Bulletin, 107, 3–33.
Brewer, M. B. (1981). Ethnocentrism and its role in interpersonal trust. In M. B. Brewer
& B. E. Collins (Eds.), Scientific inquiry and the social sciences (pp. 345–360).
San Francisco: Jossey-Bass.
Brown, P., & Levinson, S. C. (1987). Politeness: Some universals in language use. Cambridge,
UK: Cambridge University Press.
Canary, D. J., & Stafford, L. (1992). Relational maintenance strategies and equity in marriage.
Communication Monographs, 59, 243–268.
M. J. Hornsey et al. Strategies to Soften Criticism
Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association 95
Cody, M. J., Kersten, L., Braaten, D. O., & Dickson, R. (1992). Coping with relational
dissolutions: Attributions, account credibility, and plans for resolving conflicts.
In J. H. Harvey, T. L. Orbuch, & A. L. Weber (Eds.), Attributions, accounts, and
close relationships (pp. 93–115). New York: Springer-Verlag.
Cohen, G. L., Steele, C. M., & Ross, L. D. (1999). The mentor’s dilemma: Providing
feedback across the racial divide. Personality and Social Psychology Bulletin, 25,
1302–1318.
Coleman, L. M., Jussim, L., & Abraham, J. (1987). Students’ reactions to teacher evaluations:
The unique impact of negative feedback. Journal of Applied Social Psychology, 17,
1051–1070.
Cupach, W. R., & Carson, C. L. (2002). Characteristics and consequences of interpersonal
complaints associated with perceived face threat. Journal of Social and Personal
Relationships, 19, 443–462.
Cupach, W. R., & Metts, S. (1990). Remedial processes in embarrassing predicaments.
In J. A. Anderson (Ed.), Communication yearbook 13 (pp. 323–352). Newbury Park,
CA: Sage.
Cupach, W. R., & Metts, S. (1994). Facework. Newbury Park, CA: Sage.
Davies, D., & Jacobs, A. (1985). ‘‘Sandwiching’’ complex interpersonal feedback. Small Group
Behavior, 16, 387–396.
Fiske, S. T. (1980). Attention and weight in person perception: The impact of negative and
extreme behavior. Journal of Personality and Social Psychology, 38, 889–906.
Flacks, R. (1967). The liberated generation: An exploration of the roots of student protest.
Journal of Social Issues, 23, 52–75.
Folkes, V. S. (1982). Communicating the causes of social rejection. Journal of Experimental
Social Psychology, 18, 235–252.
Gottman, J. M. (1994). What predicts divorce? Hillsdale, NJ: Lawrence Erlbaum.
Gottman, J. M., & Krokoff, L. J. (1989). Marital interaction and satisfaction: A longitudinal
view. Journal of Personality and Social Psychology, 57, 47–52.
Gottman, J. M., & Levenson, R. W. (1992). Marital processes predictive of later dissolution:
Behavior, physiology, and health. Journal of Personality and Social Psychology, 63, 221–233.
Green, J. D., Pinter, B., & Sedikides, C. (2005). Mnemic neglect and self-threat: Trait
modifiability moderates self-protection. European Journal of Social Psychology, 35, 225–235.
Hample, D., & Dallinger, J. M. (1988). Individual differences in cognitive editing standards.
Human Communication Research, 14, 123–144.
Hewitt, J. P., & Stokes, R. (1975). Disclaimers. American Sociological Review, 40, 1–11.
Hocker, J. L., & Wilmot, W. W. (1991). Interpersonal conflict (3rd ed.). Dubuque, IA: Brown.
Hooley, J. M., & Teasdale, J. D. (1989). Predictors of relapse in unipolar depressives:
Expressed emotion, marital distress, and perceived criticism. Journal of Abnormal
Psychology, 98, 229–235.
Hornsey, M. J. (2005). Why being right is not enough: Predicting defensiveness in the face of
group criticism. European Review of Social Psychology, 16, 301–334.
Hornsey, M. J. (2006). Ingroup critics and their influence on groups. In T. Postmes & J. Jetten
(Eds.), Individuality and the group: Advances in social identity (pp. 74–91). London: Sage.
Hornsey, M. J., Grice, T., Jetten, J., Paulsen, N., & Callan, V. (2007). Group-directed
criticisms and recommendations for change: Why newcomers arouse more defensiveness
than old-timers. Personality and Social Psychology Bulletin, 33, 1036–1048.
Strategies to Soften Criticism M. J. Hornsey et al.
96 Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association
Hornsey, M. J., & Imani, A. (2004). Criticizing groups from the inside and the outside: An
identity perspective on the intergroup sensitivity effect. Personality and Social Psychology
Bulletin, 30, 365–383.
Hornsey, M. J., Oppes, T., & Svensson, A. (2002). ‘‘It’s ok if we say it, but you can’t’’:
Responses to intergroup and intragroup criticism. European Journal of Social Psychology,
32, 293–307.
Hornsey, M. J., Trembath, M., & Gunthorpe, S. (2004). ‘‘You can criticize because you care’’:
Identity attachment, constructiveness, and the intergroup sensitivity effect. European
Journal of Social Psychology, 34, 499–518.
Janis, I. L. (1982). Groupthink: Psychological studies of policy decisions and fiascoes. Boston:
Houghton Mifflin.
Judd, C. M., Park, B., Yzerbyt, V., Gordijn, E. H., & Muller, D. (2005). Attributions of
intergroup bias and outgroup hom*ogeneity to ingroup and outgroup others.
European Journal of Social Psychology, 35, 677–704.
Linville, P. W., Fischer, F. W., & Salovey, P. (1989). Perceived distributions of characteristics
of ingroup and outgroup members: Empirical evidence and a computer simulation.
Journal of Personality and Social Psychology, 42, 193–211.
Mae, L., & Carlston, D. E. (2005). Hoist on your own petard: When prejudiced remarks
are recognized and backfire on speakers. Journal of Experimental Social Psychology, 41,
240–255.
Oetzel, J. G., & Ting-Toomey, S. (2003). Face concerns in interpersonal conflict: A
cross-cultural empirical test of the face negotiation theory. Communication Research,
30, 599–624.
Ogilvie, J. R., & Haslett, B. (1985). Communicating peer feedback in a task group. Human
Communication Research, 12, 79–98.
Ostrom, T. M., & Sedikides, C. (1992). Outgroup hom*ogeneity effects in natural and minimal
groups. Psychological Bulletin, 112, 536–552.
Park, B., Judd, C. M., & Ryan, C. S. (1991). Social categorization and the representation of
variability information. In W. Stroebe & M. Hewstone (Eds.), European Review of Social
Psychology (Vol. 2, pp. 211–245). Chichester, UK: Wiley.
Postmes, T., Spears, R., & Cihangir, S. (2001). Quality of decision making and group norms.
Journal of Personality and Social Psychology, 80, 918–930.
Rogan, R. G., & Hammer, M. R. (1994). Crisis negotiations: A preliminary investigation of
facework in naturalistic conflict discourse. Journal of Applied Communication Research,
22, 216–231.
Rozin, P., & Royzman, E. B. (2001). Negativity bias, negativity dominance, and contagion.
Personality and Social Psychology Review, 5, 296–320.
Sutton, R. M., Elder, T. J., & Douglas, K. M. (2006). Reactions to internal and external
criticism of outgroups: Social convention in the intergroup sensitivity effect. Personality
and Social Psychology Bulletin, 32, 563–575.
Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup conflict. In W. G. Austin
& S. Worchel (Eds.), The social psychology of intergroup relations (pp. 33–47). Monterey,
CA: Brooks/Cole.
Tanis, M., & Postmes, T. (2005). A social identity approach to trust: Interpersonal perception,
group membership and trusting behaviour. European Journal of Social Psychology, 35,
413–424.
M. J. Hornsey et al. Strategies to Soften Criticism
Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association 97
Ting-Toomey, S. (1988). Intercultural conflict styles: A face-negotiation theory. In Y. Y. Kim
& W. B. Gudykunst (Eds.), Theories in intercultural communication (pp. 213–235).
Newbury Park, CA: Sage.
Tracy, K., Van Dusen, D., & Robinson, S. (1987). ‘‘Good’’ and ‘‘bad’’ criticism: A descriptive
analysis. Journal of Communication, 37, 46–59.
Turner, J. C. (1991). Social influence. Buckingham, UK: Open University Press.
Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D., & Wetherell, M. S. (1987).
Rediscovering the social group: A self-categorization theory. New York: Blackwell.
Vivian, J. E., & Berkowitz, N. H. (1993). Anticipated outgroup evaluations and intergroup
bias. European Journal of Social Psychology, 23, 513–524.
Weiner, B., Amirkhan, J., Folkes, V. S., & Verette, J. A. (1987). An attributional analysis of
excuse giving: Studies of a naive theory of emotion. Journal of Personality and Social
Psychology, 52, 316–324.
Wilson, S. R., Cruz, M. G., Marshall, L. J., & Rao, N. (1993). An attributional analysis of
compliance-gaining interactions. Communication Monographs, 60, 352–372.
Yamagishi, T., & Kiyonari, T. (2000). The group as the container of generalized reciprocity.
Social Psychology Quarterly, 63, 116–132.
Strategies to Soften Criticism M. J. Hornsey et al.
98 Human Communication Research 34 (2008) 70–98 ª 2008 International Communication Association
Évaluation des stratégies visant à minimiser les réactions défensives face à
une critique de groupe : Dorer la pilule
Matthew J. Hornsey, Erin Robson, Joanne Smith, and Sarah Esposo
University of Queensland
Robbie M. Sutton
University of Kent
Résumé
Les gens sont considérablement plus défensifs face à une critique de groupe
lorsque cette critique provient d’un membre hors-groupe plutôt que d’un
membre intra-groupe (l’effet de sensibilité inter-groupe). Nous avons testé trois
stratégies que les critiques hors-groupes peuvent utiliser afin de réduire cette
attitude défensive exacerbée. Dans chaque étude, des Australiens ont reçu des
critiques de leur pays formulées par un autre Australien ou par un étranger. Dans
l’expérience 1, les critiques ayant attaché des éloges à la critique furent plus
appréciés et appuyés que ceux qui ne le firent pas. Dans l’expérience 2, les
critiques hors-groupes furent plus appréciés et soulevèrent moins de négativité
lorsqu’ils reconnurent que les problèmes identifiés dans le groupe cible étaient
aussi partagés par leur propre groupe. Dans les deux expériences, les effets
avantageux de l’éloge et de la reconnaissance furent complètement médiatisés
par l’attribution d’une volonté constructive. L’expérience 3 a testé la stratégie de
focalisation, c’est-à-dire d’indiquer que vos commentaires visent à ne
s’appliquer qu’à une portion du groupe plutôt qu’à tout le groupe. Cette stratégie
(qui ne traitait pas directement des questions d’attribution présumées soutenir
l’effet de sensibilité inter-groupes) s’est révélée inefficace. Les implications
pratiques et théoriques pour la communication inter-groupes sont discutées.
Die Bewertung von Strategien zur Minimierung defensiver Reaktionen auf
Gruppenkritik: Wie man eine bittere Pille versüßt.
Menschen reagieren wesentlich defensiver, wenn sie mit der Kritik einer Person
aus der Outgroup verglichen mit einer Person aus der Ingroup konfrontiert
werden (Intergruppensensitivitätseffekt). Wir testeten drei Strategien, die
Outgroup-Kritiker nutzen können, um diese erhöhte Defensivität zu minimieren.
In allen Studien wurden Australier von einem Australier oder einem Ausländer
mit Kritik am eigenen Land konfrontiert. Im ersten Experiment zeigte sich, dass
Kritiker, deren Kritik mit Lob verbunden war, mehr gemocht wurden und
größere Zustimmung erfuhren als solche, die dies nicht taten. Im zweiten
Experiment zeigte sich, dass ein Outgroup-Kritiker mehr gemocht wurde und
weniger Ablehnung erfuhr, wenn er einräumte, dass das Problem, dass er
angesprochen hat auch in seiner Ingroup existiert. Beide Experimente
verdeutlichen den förderlichen Effekt von Lob und Anerkennung, der gänzlich
durch Konstruktivität moderiert wurde. Experiment 3 testete die Scheinwerfer-
Strategie; das heißt, die Strategie, einen Kommentar so zu formulieren, dass er
nur einen Teil der Gruppe betrifft und nicht die gesamte Gruppe. Diese Strategie
– die nicht direkt das Thema angesprochen hat, dass den
Intergruppensensitivitätseffekt befördern sollte – erwies sich allerdings als
ineffektiv. Praktische und theoretische Schlussfolgerungen für Kommunikation
zwischen Gruppen werden diskutiert.
Evaluando las Estrategias Destinadas a Minimizar las Reacciones
Defensivas hacia las Críticas de Grupo: Endulzando la Píldora
Matthew J. Hornsey, Erin Robson, Joanne Smith, and Sarah Esposo
University of Queensland, Robbie M. Sutton, University of Kent
Resumen
La gente es considerablemente más defensiva hacia la crítica de grupo
cuando esta crítica proviene de un grupo al que no se pertenece, que cuando
proviene de un miembro del grupo de pertenencia (el efecto de sensibilidad
intergrupal). Pusimos a prueba 3 estrategias que las críticas de grupos de no
pertenencia pueden usar para reducir este mecanismo de incrementado. En
todos los estudios, unos australianos recibieron la crítica a su propio país de
otro australiano ó de un extranjero. En el experimento 1, aquellos que
hicieron una crítica junto con elogios fueron más preferidos y concordaron
más que aquellos que no hicieron elogios. En el experimento 2, los críticos
del grupo de no pertenencia fueron más preferidos y despertaron menos
negatividad cuando admitieron que los problemas que ellos identificaron en
el grupo meta eran compartidas también por su propio grupo. En ambos
experimentos, los efectos de alivio de los elogios y de reconocimiento
fueron mediados por completo por las atribuciones constructivas. El
experimento 3 puso a prueba la estrategia de llamar la atención; esto es, de
poner como antecedente que la persona intentaba aplicar sus comentarios
solo a una porción del grupo y no al grupo en su totalidad. Esta estrategia –
que no se dirigió directamente a los asuntos de atribuciones supuestos para
mantener el efecto de sensibilidad intergrupal—comprobó ser ineficiente.
Las implicaciones prácticas y teóricas sobre comunicación intergrupal son
también discutidas.
评价旨在降低人们对群体批评防卫性反应的策略：给药丸披上糖衣
Matthew J. Hornsey, Erin Robson, Joanne Smith, and Sarah Esposo
Queensland 大学
Robbie M. Sutton
肯特大学
摘要
由于群体内敏感性的效果，人们对来自群体外的批评比来自群体内的
批评反应更具防卫性。我们检测了群体外批评者可以使用的 3 种降低
高度防卫性反应的策略。在所有的研究中，澳大利亚人或受到另一个
澳大利亚人的批评，或受到一个外国人的评价。在实验 1 中，既批评
又称赞的人比那些只批评的人受到更多的喜爱，其观点更多地被接
受。在实验 2 中，当群体外批评者承认他们所指出的问题在他们自己
的群体内也存在，他们所发出的批评会受到更多的喜欢、激发更少的
负面效果。在这两种实验中，称赞和承认所带来的减缓效果受到建构
性特征的中介。实验 3 检测了聚焦的效果，即批评只针对群体的一部
分人而不是整个群体。由于没有直接涉及群体内敏感性效果所依托的
归属性议题，这个策略没有效果。最后我们讨论了本研究对群体内传
播的实践及理论涵义。
집단 비판주의에 대한 방어적인 반응을 최소화하기 위해 구성된 평가
전략들에 관한 연구
Matthew J. Hornsey, Erin Robson, Joanne Smith, and Sarah Esposo
University of Queensland
Robbie M. Sutton
University of Kent
요약
사람들은 집단 비판에 직면시 그 비판이 집단내 구성원으로부터 나올때보다 집단 밖에서 나올때 더욱 방어적이 된다. 우리는 집단밖으로부터의 비판들이 이러한 방어성을 줄일수 있는지를 알아보기 위해 3가지의 전략들을 연구하였다. 모든 연구에서, 호주인들은 그들 나라에 대한 비판을 다른 호주인이나 외국인으로 부터 듣게했다. 실험1에서는, 비판에 칭찬을 덧붙인 비평가들에 대해서는 그렇지 않은 사람들보다 더욱 좋아하는 사람들이 많았으며 이들 의견에도 동의하는 비율이 높았다. 실험2 에서는 집단밖의 비판가들이 그들이 목표집단의 구성원들 사이에서 인지하고 있는 문제들을 알고 있고, 그들 집단 구성원들에 의해 그 문제점들이 공유될때, 비판가들이 더욱 선호되며 나쁜 감정들을 유발하지 않았다. 이들 두 실험에서, 칭찬과 인지의 개선적 효과들은 건설적인 형태에 의해 완전하게 중재되었다. 실험 3은 강조전력을 사용하였다. 즉, 비판자가 전체집단을 상대로 하는 것이 아니라 집단의 일부에게만 적용되는 비판을 기록하게 하였다. 이 전략—집단내 민감성 효과를 불러일으킬 수 있는 귀속성 이슈들을 직접적으로 표현하지 않는—은 비효과적인 것으로 나타났다. 집단내 커뮤니케이션을 위한 실행적 그리고 이론적 함의들이 논의되었다.
