6
Personality influences on interpretations of aggressive behavior: The role of provocation sensitivity and trait aggression Claire Lawrence * , Ella Hodgkins Risk Analysis and Social Processes and Health (RASPH) Group, School of Psychology, University of Nottingham, Nottingham NG7 2RD, UK article info Article history: Received 10 June 2008 Received in revised form 10 October 2008 Accepted 20 October 2008 Available online 3 December 2008 Keywords: Trait aggression Sensitivity to aggressive triggers Group processes Interpretations of provocation abstract Some individuals are more prone to interpret the behaviors of others as aggressive, which can increase the likelihood of aggressive behavior being performed. However, work in this field has typically focused on the influence of general trait aggression on the perception of text, and static images. These studies assess the extent to which both global trait aggression and provocation sensitivity influence interpreta- tions of real-life footage of aggressive behavior. In Study 1, participants viewed closed circuit television (CCTV) footage depicting a female being assaulted by a male after she physically provoked him. In Study 2, participants viewed two clips of provocation followed by violence, one between two females and one between two males. Provocation sensitivity influenced evaluations of provoking individuals across all three clips above and beyond the influence of trait aggression, and mood in Study 2. These findings sug- gest the influence of individual differences in provocation sensitivity on interpretations of actual aggres- sive behavior beyond overall trait aggression. Possible implications of individual differences in sensitivities to aggressive triggers on the witnessing of aggressive events are discussed. Ó 2008 Elsevier Ltd. All rights reserved. 1. Introduction In contemporary models of aggressive behavior, such as the General Aggression Model (GAM: Anderson and Bushman (2002)), the role of cognitions about others’ behavior is central. The GAM posits that these cognitions derive from the interaction of person (e.g., trait) and situation (e.g., provocation) factors and influence individuals’ responses to the prevailing situation. A growing body of research demonstrates that aggressive individu- als interpret others’ behavior as more hostile and malevolent than non-aggressive individuals (e.g., Dodge & Crick, 1990). Such work typically finds that individuals scoring high in general trait aggression attend more to aggressive photographs and words, process aggressive text faster and are more disrupted by aggres- sive words on emotional Stroop tasks (e.g., Smith & Waterman, 2005; Wingrove & Bond, 2005). Aggressive individuals also inter- pret behaviors performed by others to be less aggressive than non-aggressive controls, arguably because they are less sensitized to aggressive behavior (Smith, Waterman, & Ward, 2006). Tiedens (2001, p. 248) suggested that this propensity for aggressive indi- viduals to perceive aggressive stimuli more readily results in a ‘‘vicious cycle” of aggression. However, aggressive individuals do not act aggressively across all triggering situations. Rather there is evidence for individual differences in the triggers that individ- uals respond aggressively to (Lawrence, 2006; Marshall & Brown, 2006). Accordingly, the literature on aggression and individual differ- ences raises two key questions: (1) do trait sensitivities to aggres- sive triggers influence perceptions of aggressive stimuli over and above general trait aggression? (2) do these aggression-related traits influence the interpretation of real-life footage of aggressive behavior. 1.1. Individual differences in sensitivity to provocations and frustrations Lawrence (2006) recently demonstrated that individuals vary in the extent to which they are sensitive to frustrating or provoking situations, measured by the Situational Triggers of Aggressive Re- sponses (STAR) scale. Sensitivity to provocations is associated with increased physical and verbal aggression while sensitivity to frus- trations is associated with increased trait hostility and anger (Law- rence, 2006: Study 3). Therefore provocation sensitivity may be an important influence on expressed aggression. What is not known, however, is whether those sensitive to provocations not only re- spond to provocations more, but also interpret the provocative behavior of others as being more provoking. Accordingly, due to its proximal relevance, and previous work demonstrating the supe- riority of specific rather than general traits (e.g., Chapman & Hay- slip, 2005), it is predicted that provocation sensitivity rather than frustration sensitivity will predict aggressive interpretations of provoking individuals above and beyond trait aggression. 0191-8869/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.paid.2008.10.022 * Corresponding author. Tel.: +44 0115 951 5326; fax: +44 0115 951 5324. E-mail address: [email protected] (C. Lawrence). Personality and Individual Differences 46 (2009) 319–324 Contents lists available at ScienceDirect Personality and Individual Differences journal homepage: www.elsevier.com/locate/paid

Personality influences on interpretations of aggressive behavior: The role of provocation sensitivity and trait aggression

Embed Size (px)

Citation preview

Page 1: Personality influences on interpretations of aggressive behavior: The role of provocation sensitivity and trait aggression

Personality and Individual Differences 46 (2009) 319–324

Contents lists available at ScienceDirect

Personality and Individual Differences

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

Personality influences on interpretations of aggressive behavior: The roleof provocation sensitivity and trait aggression

Claire Lawrence *, Ella HodgkinsRisk Analysis and Social Processes and Health (RASPH) Group, School of Psychology, University of Nottingham, Nottingham NG7 2RD, UK

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

Article history:Received 10 June 2008Received in revised form 10 October 2008Accepted 20 October 2008Available online 3 December 2008

Keywords:Trait aggressionSensitivity to aggressive triggersGroup processesInterpretations of provocation

0191-8869/$ - see front matter � 2008 Elsevier Ltd. Adoi:10.1016/j.paid.2008.10.022

* Corresponding author. Tel.: +44 0115 951 5326; fE-mail address: [email protected] (

Some individuals are more prone to interpret the behaviors of others as aggressive, which can increasethe likelihood of aggressive behavior being performed. However, work in this field has typically focusedon the influence of general trait aggression on the perception of text, and static images. These studiesassess the extent to which both global trait aggression and provocation sensitivity influence interpreta-tions of real-life footage of aggressive behavior. In Study 1, participants viewed closed circuit television(CCTV) footage depicting a female being assaulted by a male after she physically provoked him. In Study2, participants viewed two clips of provocation followed by violence, one between two females and onebetween two males. Provocation sensitivity influenced evaluations of provoking individuals across allthree clips above and beyond the influence of trait aggression, and mood in Study 2. These findings sug-gest the influence of individual differences in provocation sensitivity on interpretations of actual aggres-sive behavior beyond overall trait aggression. Possible implications of individual differences insensitivities to aggressive triggers on the witnessing of aggressive events are discussed.

� 2008 Elsevier Ltd. All rights reserved.

1. Introduction

In contemporary models of aggressive behavior, such as theGeneral Aggression Model (GAM: Anderson and Bushman(2002)), the role of cognitions about others’ behavior is central.The GAM posits that these cognitions derive from the interactionof person (e.g., trait) and situation (e.g., provocation) factors andinfluence individuals’ responses to the prevailing situation. Agrowing body of research demonstrates that aggressive individu-als interpret others’ behavior as more hostile and malevolent thannon-aggressive individuals (e.g., Dodge & Crick, 1990). Such worktypically finds that individuals scoring high in general traitaggression attend more to aggressive photographs and words,process aggressive text faster and are more disrupted by aggres-sive words on emotional Stroop tasks (e.g., Smith & Waterman,2005; Wingrove & Bond, 2005). Aggressive individuals also inter-pret behaviors performed by others to be less aggressive thannon-aggressive controls, arguably because they are less sensitizedto aggressive behavior (Smith, Waterman, & Ward, 2006). Tiedens(2001, p. 248) suggested that this propensity for aggressive indi-viduals to perceive aggressive stimuli more readily results in a‘‘vicious cycle” of aggression. However, aggressive individuals donot act aggressively across all triggering situations. Rather thereis evidence for individual differences in the triggers that individ-

ll rights reserved.

ax: +44 0115 951 5324.C. Lawrence).

uals respond aggressively to (Lawrence, 2006; Marshall & Brown,2006).

Accordingly, the literature on aggression and individual differ-ences raises two key questions: (1) do trait sensitivities to aggres-sive triggers influence perceptions of aggressive stimuli over andabove general trait aggression? (2) do these aggression-relatedtraits influence the interpretation of real-life footage of aggressivebehavior.

1.1. Individual differences in sensitivity to provocations andfrustrations

Lawrence (2006) recently demonstrated that individuals vary inthe extent to which they are sensitive to frustrating or provokingsituations, measured by the Situational Triggers of Aggressive Re-sponses (STAR) scale. Sensitivity to provocations is associated withincreased physical and verbal aggression while sensitivity to frus-trations is associated with increased trait hostility and anger (Law-rence, 2006: Study 3). Therefore provocation sensitivity may be animportant influence on expressed aggression. What is not known,however, is whether those sensitive to provocations not only re-spond to provocations more, but also interpret the provocativebehavior of others as being more provoking. Accordingly, due toits proximal relevance, and previous work demonstrating the supe-riority of specific rather than general traits (e.g., Chapman & Hay-slip, 2005), it is predicted that provocation sensitivity rather thanfrustration sensitivity will predict aggressive interpretations ofprovoking individuals above and beyond trait aggression.

Page 2: Personality influences on interpretations of aggressive behavior: The role of provocation sensitivity and trait aggression

320 C. Lawrence, E. Hodgkins / Personality and Individual Differences 46 (2009) 319–324

1.2. Perceptions of real aggressive events

Despite frequent calls for more ecologically valid examinationsof behavior (e.g., Sims, Noel, & Maisto, 2007), research examininginfluences on the perception of aggression typically provides par-ticipants with either (i) written vignettes (O’Connor, Archer &Wu, 2001; Smith et al., 2006) or (ii) aggressive vs. neutral wordsor photographs in dot-probe or Stroop-type paradigms (e.g., Smith& Waterman, 2005). Very little work has examined individuals’interpretations of real-world behavior (although, see Dill, Ander-son, Anderson, and Deuser (1997)). Vignette-based approachescan be useful in allowing a wide variety of situations to be exam-ined, however, the situations used are typically heterogeneous,meaning that the nature and effects of individuals’ trigger sensitiv-ity (e.g., frustration, provocation) cannot be assessed. Further, thenature of textual vignettes means that we cannot know whetherindividuals are imagining the same scenario. Consequently, wecannot be sure whether differences between individuals stem fromthe way a scenario is imagined, or the way behavior is interpreted.There is a need, therefore, to understand perceptions of aggressivebehavior using the presentation of authentic rather than artificialstimuli.

Therefore, these studies are the first to investigate the relativeinfluence of individual differences in sensitivity to provocationover and above general trait aggression on interpretations of real,dynamic aggressive behavior.

Table 1Pearson’s correlations between trait aggression, sensitivity to provocations andfrustrations and interpretations of the protagonists’ behavior.

1 2 3 4 a Mean (SD)

SP (1) .77 3.02 (.56)SF (2) .63** .78 2.66 (.65)GTA (3) .29** .13 .85 2.71 (.52)Male Responder (4) �.02 �.21* �.22* .68 4.65 (.36)Female Provoker (5) .31** .15 .12 �.03 .70 4.07 (.58)

Key: SP, sensitivity to provocations; SF, sensitivity to frustration; and GTA, traitaggression.* p < .05.** p < .01.

2. Study 1

This study examines the influence of general trait aggression(GTA) and, particularly, sensitivity to provocation and frustrationon interpretations of protagonists’ behavior during CCTV footageof a provoking female being physically assaulted by a male. Thosesensitive to provocation are expected to interpret the behavior ofthe provoking protagonist as more aggressive, beyond the influ-ence of GTA. Following Smith et al. (2006), it is expected thatthose high in GTA will interpret the aggressively respondingmale’s behavior as being less aggressive. As the footage depictedinvolves a male assaulting a provoking female, it is also possiblethat the sex of participants will influence interpretations ofbehavior. Previous work has demonstrated that when presentedinformation about a male physically or sexually assaulting a fe-male, males are less negative in their evaluations of the maleaggressor (e.g., Whatley, 2005). Therefore the main and interac-tive effects of participant sex on interpretations of both protago-nists’ behavior are assessed.

2.1. Method

2.1.1. ParticipantsOne hundred undergraduates from a large UK University partic-

ipated in this study (50% were males) with a mean age of 21(SD = 2.34). Participants were recruited and tested individually.

2.1.2. Materials2.1.2.1. CCTV footage. A 2 min segment of CCTV footage was ex-tracted from a series of CCTV segments broadcast on British televi-sion showing street-violence in UK cities. Overall, 48 segmentswere examined. Each were assessed for inclusion if they: (i) de-picted an incident involving two key individuals in order to iden-tify protagonists easily (ii) involved clear provocation by oneprotagonist and an aggressive response by the other, and (iii) pro-gressed at a pace which allowed viewers to see the unfolding inter-actions between the protagonists. One segment fulfilled all threerequirements. It depicted a female pointing and shouting at a male,

who responded by pushing her down roughly and repeatedly, face-first onto the bonnet (hood) of a parked car. The footage did notcontain audio stimuli.

To ensure that the male was perceived to be aggressive (and notreacting due to fear) and that the female was perceived to be pro-voking, a pilot study was conducted. Eighty-one undergraduatestudents (70 were female, mean age = 21.0, SD = 0.32) watchedthe CCTV footage. Participants rated the extent to which the femaleprotagonist’s behavior was provoking and how aggressive theyperceived both protagonists to be on a Likert-scale ranging from1 (strongly disagree) to 5 (strongly agree). Participants judgedthe male responder (mean = 4.88; SD = 0.46) to be more aggressivethan the female provoker (mean = 3.90; SD = 1.11) (t = 8.29,p < .001). The female was judged to be moderately provoking(mean = 3.57, SD = 1.02) and therefore suitable for an examinationof the effect of sensitivity to provocation on interpretation of herbehavior. Participants did not judge the male to have acted dueto fear, with participants assessing his behavior as significantlymore aggressive (mean = 4.88; SD = 0.46) than fear-invoked(mean = 1.60; SD = 0.8; t = 29.0, p < .001).

2.1.2.2. CCTV interpretation questionnaire (CCTV-IQ). This comprisedeight statements. The following six statements were presentedseparately: ‘‘The female/male acted aggressively”; ‘‘The female/male acted violently”; ‘‘The female/male’s behavior was justified”(reversed). An additional item was generated for each protagonistto ask about the specifics of their behavior and avoid responseacquiescence. These were: ‘‘The male attempted to gain powerover the female”; ‘‘The female aimed to humiliate the male”. Par-ticipants indicated how much they agreed with each item on afive-point Likert-scale (1 = ‘‘Strongly disagree” and 5 = ‘‘Stronglyagree”). Internal reliabilities were acceptable (see Table 1).

2.1.2.3. Individual differences in frustration and provocation sensitiv-ity. To measure sensitivity to provocations and frustrations, partic-ipants completed the 22-item Situational Triggers of AggressiveResponse scale (Lawrence, 2006) comprising two scales: sensitivityto frustration (SF: 10 items, e.g., ‘‘I feel aggressive when someoneignores me”) and sensitivity to provocations (SP: 12 items e.g., ‘‘Ifeel aggressive when someone makes offensive remarks to me”).Both scales typically show good internal reliability (SP a = .82; SFa = .80, Lawrence (2006)).

2.1.2.4. Individual differences in trait aggression. Trait aggressionwas measured using Buss and Perry’s (1992) 29-item aggressionquestionnaire (BPAQ). This scale comprises four sub-scales: Physi-cal aggression (9 items, e.g., ‘‘Occasionally, I can’t control the urgeto strike someone”); Verbal aggression (5 items, e.g., ‘‘I have threa-tened people I know”); Anger (7 items, e.g., ‘‘I sometimes fly off thehandle for no good reason”); and Hostility (8 items, e.g., ‘‘I amsometimes eaten up with jealousy”). Participants’ general trait

Page 3: Personality influences on interpretations of aggressive behavior: The role of provocation sensitivity and trait aggression

C. Lawrence, E. Hodgkins / Personality and Individual Differences 46 (2009) 319–324 321

aggression (GTA) scores were calculated by averaging their scoresacross all items. The overall scale is typically reliable (a varies be-tween .85 and .95).

2.1.3. ProcedureParticipants were recruited individually. Each participant

watched the footage and then completed the CCTV-IQ followedby the BPAQ and STAR scales.

2.2. Results

2.2.1. Descriptive statisticsMean scores for items pertaining to the male and female pro-

tagonists’ aggressiveness, SP, SF and GTA were calculated for eachparticipant. Table 1 shows the internal reliabilities, means, stan-dard deviations and Pearson’s correlations. Higher scores indi-cated greater judgments of protagonists’ aggression, higher SP,SF and GTA.

As predicted, SP, but not SF, was related to interpretations of theprovoking female’s aggression, with those sensitive to provoca-tions interpreting the female’s behavior as more aggressive. Alsoas predicted, those higher in GTA judged the aggressive male tobe less aggressive. Finally, participants high in SF also viewed theaggressive male to be less aggressive.

2.2.2. Influence of trait aggression and sensitivity to provocations andfrustrations on interpretations of aggressive behavior

Two hierarchical regressions were conducted with interpreta-tions of the protagonists’ behavior as the DVs. Participants’ sexwas entered at step one, GTA scores at step two and SP andSF scores at step three. To assess the moderating influence ofparticipant sex on the relationships between trait measuresand interpretation, the interaction terms were entered at stepfour.

As predicted, SP influenced participants’ interpretations of theprovoking female’s behavior beyond the influence of GTA and par-ticipants’ sex (see Table 2). Also as predicted, males and those highin GTA judged the aggressive male to have behaved less aggres-sively. However, SF also predicted interpretations of the aggressivemale’s behavior beyond both participant sex and GTA. None of theinteractions between participant sex and trait measures signifi-cantly influenced interpretations of protagonists’ behavior, there-fore these are not reported further.

2.3. Discussion

The influence of individual sensitivity to aggressive triggerson interpretations of aggressive behavior was found, with inter-

Table 2Influence of participants’ sex, trait aggression, and sensitivity to provocation andfrustration on judgments of the protagonists.

Regression Male responder Female provoker

DR2 b DR2 b

Step 1 .05* .02Sex �.22* �.12

Step 2 .036* .02GTA �.19* .14

Step 3 .09** .07*

SP .24 �.31*

SF �.38** �.05

Key: SP, sensitivity to provocations; SF, sensitivity to frustration; and GTA, traitaggression.* p < .05.** p < .01.

pretations of the provoking female’s behavior solely influencedby participants’ provocation sensitivity. In addition, frustrationsensitivity influenced judgments of the aggressive male beyondthe influence of general trait aggression and participants’ sex.

The expected relationship between participant sex and inter-pretation of the protagonists’ behavior was found for males only.Male participants judged the male protagonist to have acted lessaggressively than female participants. However, this effect waseclipsed by the influence of sensitivities to frustration. Finally,those high in general trait aggression judged the aggressive maleto be less aggressive. This supports previous findings (e.g., Smithet al., 2006) showing that aggressive individuals generally viewaggressive events less negatively.

In sum, the results from Study 1 indicate that sensitivity toaggressive triggers influence interpretations of aggressive behav-ior beyond trait aggression. Those sensitive to provocation inter-preted the behavior of a provoking individual to have been moreaggressive. This finding is potentially important, as although theliterature typically explains the propensity for aggressive individ-uals to interpret stimuli as aggressive in general terms, thesefindings show the value in examining the specific nature of indi-viduals’ trigger sensitivity. However, to test the generalizabilityof these findings, further examples of aggressive behavior involv-ing provocation need to be examined. This is particularly perti-nent because the footage used in Study 1 depicted a maleassault on a female, and as such, the influence of trait or sex-based influences on interpretations of behavior may be limitedto this situation. In addition, the mood-congruency effect (Bower,1991) proposes that interpretations are typically made in thedirection of prevailing mood (Niedenthal & Setterlund, 1994;Rusting, 1999). Therefore as well as examining trait influenceson the interpretation of stimuli, Study 2 assesses the additionalinfluence of participants’ mood on interpretations of aggressiveand provoking behavior.

3. Study 2

Study 2 extends Study 1 in two ways. Firstly, two footage seg-ments are presented: the first depicts a female provoker being as-saulted by another female, and the second depicts a male provokerbeing assaulted by another male. Secondly, this study also exam-ines the impact of perceivers’ mood on interpretations of the pro-tagonists. Previous work has typically shown mood-congruenteffects, with those in a negative mood making more negative inter-pretations (Forgas, 1994) and those in a positive mood spontane-ously thinking of more positive constructs when cued (Mayer &Hanson, 1995). However, more trait-congruent effects are foundwhen natural (i.e. not induced) mood is measured (Gomez & Go-mez, 2002). Indeed Gomez and Gomez (2002) argue that measur-ing natural mood is a more valid method of assessing mood-congruency effects on the interpretation of stimuli. Accordingly,mood will not be induced in this study; rather mood is measuredprior to presentation of footage. It is predicted that provocationsensitivity will influence interpretations of the provoking protago-nists beyond the impact of trait aggression and perceiver sex. Inaddition, mood-congruent effects on interpretations of protago-nists are predicted, but are expected to diminish in the light ofthe influence of provocation sensitivity.

3.1. Method

3.1.1. ParticipantsOne hundred undergraduates participated (61 females: 39

males; mean age = 20, SD = 0.98). Participants were recruited andtested individually.

Page 4: Personality influences on interpretations of aggressive behavior: The role of provocation sensitivity and trait aggression

322 C. Lawrence, E. Hodgkins / Personality and Individual Differences 46 (2009) 319–324

3.1.2. Materials3.1.2.1. Mood. Mood was measured using the 20-adjective Positiveand Negative Affectivity Scale (PANAS) (Watson, Clark, & Tellegen,1988). This comprises ten positive (e.g., proud, enthusiastic) andten negative adjectives (e.g., tense, distressed). Participants wereasked to rate how they felt ‘‘right now” from 1 (not at all or onlyslightly) to 5 (extremely). Positive affectivity (PA) scores were cal-culated by averaging participants’ scores across all positive adjec-tives, and negative affectivity (NA) scores were calculated byaveraging their scores across all negative adjectives. The PANASis a reliable measure of mood (PA a typically varies between .87and .90 and NA a typically varies between .84 and .87).

3.1.2.2. Trait aggression and individual differences in sensitivity toaggressive triggers. Individual sensitivities to provocations andfrustrations were measured using the 22-item STAR scale (Law-rence 2006), and general trait aggression was measured using theBuss and Perry (1992) Aggression Questionnaire (BPAQ) as de-scribed in Study 1.

3.1.2.3. Footage of violent episodes. Internet and television searchesfor real footage of aggressive behavior were undertaken. Overall,50+ segments were examined. They were included in the studyif they: (i) depicted an aggressive incident involving two malesor two females; (ii) depicted an incident involving two key indi-viduals in order to identify protagonists easily; (iii) showed aclearly provoking protagonist and (iv) involved real, i.e. not dra-matized violence. The search generated four segments: twodepicting female-on-female violence and two depicting male-on-male violence. All four segments were presented to 20 under-graduates taken from the same population as the main study andrated in terms of how provoking and aggressive each protago-nist’s behavior had been on five-point Likert-scales (1 = not atall to 5 = extremely). These participants were then excluded fromthe main study to avoid carry-over effects to the main study. Fe-male-on-female violence footage ‘A’ (‘A’ provoking mean = 3.3,SD = 1.44; aggressive mean = 3.5, SD = 1.7: ‘B’ provokingmean = 2.7, SD = 0.87; aggressive mean = 3.2, SD = 1.1), andmale-on-male violence footage ‘A’ (‘A’ provoking mean = 4.1,SD = 0.72; aggressive mean = 4.0, SD = 0.85: ‘B’ provokingmean = 2.83, SD = 1.34; aggressive mean = 3.1, SD = 1.7) were se-lected. In the chosen female footage one female gestures to andgoads another female in a ‘‘reality television” context. The footageends with the provoking female being physically assaulted by thesecond female. In the chosen male footage, one male provokes theother verbally and takes steps towards the other aggressively un-til he is pushing against him and is subsequently physically as-saulted during a TV studio interview.

A second pilot study was conducted to ensure that the presen-tation of footage did not induce change in observers’ mood andthat both segments were broadly equivalent in terms of complex-ity and overall valence. Two additional groups of participants wereshown one of the clips. Eighty-one participants viewed the femalefootage (20 were male; mean age = 20.5, SD = 3.05), and 80 partic-ipants viewed the male footage (25 were male; mean age = 19.78,

Table 3Influence of male and female footage on mood.

Mood pre footage (SD) Mood post

Male footage PA 2.26 (0.70) 2.24 (0.72)NA 1.45 (0.47) 1.38 (0.42)

Female footage PA 2.63 (0.78) 2.58 (0.80)NA 1.34 (0.42) 1.31 (0.35)

Key: PA, positive affect; NA, negative affect.

SD = 2.5). Participants completed the PANAS prior to and afterwatching the footage, and rated the complexity and overall valenceof the footage on 5-point Likert-scales. Higher scores indicated thefootage was more complex and positive. Mood remained un-changed by either male or female versions of the footage (see Table3). Male and female footage did not differ in terms of overall va-lence (male footage mean = 4.21, SD = 0.52; female footagemean = 4.41, SD = 0.77, t = 1.81, p = .08) and perceived complexity(male footage mean = 2.38, SD = 1.02; female footage mean = 2.41,SD = 1.05, t = 0.15, p = .88).

3.1.2.4. Video interpretation questionnaires (VIQs). For each segment,participants responded to three questions about the protagonists’behavior. Protagonists denoted by an ‘A’ were provokers and ‘B’were the responders: ‘‘Female/Male A acted aggressively”; ‘‘Fe-male/Male B acted aggressively”; ‘‘Female A/Male A provoked Fe-male B/Male B”. For each protagonist, two further items wereadded to ask about the specifics of their behavior and avoid re-sponse acquiescence. These were: ‘‘Female A intended to antago-nize Female B”; ‘‘Female A intended to cause the fight”; ‘‘FemaleB was not justified”; ‘‘Female B was friendly” (reversed); ‘‘MaleA’s behavior was intended to provoke Male B”; ‘‘Male A deservedto be hit”; ‘‘Male B knew he had upset Male A”; ‘‘Male B’s behaviorwas intended to hurt Male A”. Participants responded on five-pointLikert-scales, (1 = Strongly disagree to 5 = Strongly agree). Internalreliabilities were acceptable (see Table 4). Interpretations for eachprotagonist’s behavior were derived by calculating the mean acrossrelevant items.

3.1.3. ProcedureParticipants completed the PANAS, and then watched each vi-

deo segments and completed the corresponding questionnaires(the order of segments was counter-balanced). They then com-pleted the BPAQ, and STAR scales.

3.2. Results

3.2.1. Descriptive statisticsTable 4 shows the internal reliabilities, means, standard devia-

tions and Pearson’s correlations. Higher scores indicated greaterjudgments of protagonists’ aggression, and higher SP, SF, GTA, po-sitive mood and negative mood.

3.2.2. Influence of participant sex, trait aggression, SP and SF oninterpretations of aggressive behavior

Four hierarchical regressions were conducted, with interpreta-tions of the two male and two female protagonists’ behavior asthe DVs. Participants’ sex was entered at step one, GTA and moodat step two and SP and SF at step three. To assess the moderatinginfluence of participant sex on the relationships between traitmeasures and interpretation, the interaction terms were enteredat step four.

Again, SP but not SF influenced aggressive interpretations of theprovoking female’s behavior (see Table 5). Similarly, for interpreta-tions of the provoking male, while the DR2 for step three only ap-

footage (SD) Difference Pearson’s r

F1,79 = 0.20; p = .66, partial g2 = .003 .85F1,79 = 2.43; p = .13, partial g2 = .03 .55

F1,80 = 2.03; p = .16, partial g2 = .03 .25F1,80 = 0.95; p = .33, partial g2 = .01 .55

Page 5: Personality influences on interpretations of aggressive behavior: The role of provocation sensitivity and trait aggression

Table 4Pearson’s correlations between trait aggression, sensitivity to provocations and frustrations and interpretations of the four protagonists’ behavior.

1 2 3 4 5 6 7 8 a Mean (SD)

Provoking female (1) .71 3.99 (.68)Responding female (2) �.39** .65 3.67 (.77)Provoking male (3) .17 .06 .65 4.12 (.68)Responding male (4) �.05 .03 �.41** .70 2.87 (.83)PA (5) .20* .14 .03 �.11 .83 3.28 (.67)NA (6) �.02 .07 �.03 �.05 �.08 .81 2.15 (.68)GTA (7) .33** �.07 �.13 .10 .04 .27** .91 2.53 (.61)SP (8) .32** �.08 .04 �.01 �.19 .31** .57** .73 3.29 (.56)SF (9) .11 �.14 �.13 .04 �.24* .38** .45** .56* .83 2.75 (.73)

Key: PA, Positive Affect; NA, Negative Affect; GTA, Trait Aggression; SP, Sensitivity to Provocations; and SF, Sensitivity to Frustration.* p < .05.** p < .01.

Table 5Influence of participants’ sex, trait aggression, mood and sensitivity to provocation and frustration on judgments of all protagonists.

Provoking female Responding female Provoking male Responding male

DR2 (R2) b DR2 (R2) b DR2 (R2) b DR2 (R2) b

1 .001 (.001) .002 (.002) .01 (.01) .02 (.02)Sex �.02 �.05 .07 �.15

2 .16*** (.16) .04 (.04) .02 (.02) .02Sex .06 �.02 .08 �.12PA �.19 .15 .06 .07NA �.11 .11 .005 �.07GTA .36*** �.10 �.13 .11

3 .06** (.22) .02 (.05) .05* (.07) .01Sex .07 .02 .13 �.16PA .24** .13 .07 .07NA �.13 .15 .02 -.09GTA .22* -.04 -.18 .10SP .33*** .01 .27** �.10SF �.09 �.16 �.22 .14

Key: PA, positive affect; NA, negative affect; SP, sensitivity to provocations; SF, sensitivity to frustration; and GTA, trait aggression.* p 6 .08 > .05.** p < .05.*** p < .01.

C. Lawrence, E. Hodgkins / Personality and Individual Differences 46 (2009) 319–324 323

proached significance (p = .08), SP but not SF influenced aggressiveinterpretations of the provoking male’s behavior.

Those higher in positive mood viewed the provoking female’sbehavior as being more aggressive. The provoking male’s behaviorwas not influenced by any mood or trait measure. Finally, interpre-tations of neither male nor female aggressive responders’ behaviorwere influenced by sex, mood or trait-based variables. None of theinteractions between participant sex and trait measures were sig-nificant for any of the four protagonists, therefore these are not re-ported further (provoking female: DR2 = .05, R2 = .27; respondingfemale: DR2 = .002, R2 = .06; provoking male: DR2 = .05; R2 = .12;responding male: DR2 = .017; R2 = .07).

3.3. Discussion

The findings of Study 2 replicate those from Study 1. Interpreta-tions of the provoking individuals’ behavior were predominantlyinfluenced by participants’ sensitivity to provocations beyond theinfluence of participant sex, mood or general trait aggression. Thisemphasizes the importance of examining the influence of individ-ual differences in sensitivity to provocation rather than traitaggression alone on the perception of actual aggressive behavior.

Interestingly, the influence of mood on judgments of any of theprotagonists’ behavior was minimal. Where it was found, its influ-ence was contrary to the mood-congruence hypothesis. Conse-quently, the effects of provocation sensitivity on interpretationsof aggressive behavior are not necessarily dependent upon therelationship between trait and mood.

4. Conclusion

Both studies showed consistently across three different dy-namic segments of aggressive behavior that trait sensitivity toprovocation influenced the way provoking behavior was inter-preted and understood. These effects operate above and beyondthe established influence of trait aggression, and in response to realaggressive behavior and so offer additional explanatory value. Thenotion of individual sensitivities (trait) to aggressive triggers (situ-ations) may help explain why individuals respond aggressively tosome triggering situations but not others. This can potentially ex-plain the link between purely situation-based and purely person-based antecedents of aggression in contemporary models ofaggression (e.g., Anderson & Bushman, 2002). The findings alsosupport the notion of a ‘‘vicious cycle” of aggressive interpretation,aggressive behavior, aggressive interpretation (Tiedens, 2001). Thestudies reported here suggest that this may be particularly the casefor those sensitive to provocations during aggressive situationsthat are characterized by provocation.

4.1. The role of individual differences when interpreting aggressiveevents

These findings potentially have relevance for the criminal jus-tice system. The testimonies of individuals who have witnessedaggressive crimes involving provocation may be influenced bytheir own trait aggression and, in particular their sensitivities toprovocation. Such sensitivities may result in biases in the criminal

Page 6: Personality influences on interpretations of aggressive behavior: The role of provocation sensitivity and trait aggression

324 C. Lawrence, E. Hodgkins / Personality and Individual Differences 46 (2009) 319–324

justice process from the perceptions of those reporting the behav-ior initially, through police perceptions of the protagonists in-volved and finally perceptions of the judge or jury. Thus futurework examining the impact of individual differences in sensitivi-ties to aggressive triggers on mock jury verdicts would investigateits potential importance in this field.

4.2. Limitations and future work

This study has gone beyond the general impact of trait aggres-sion on interpretations of aggression and has begun to unpick thenature of individual variation in the types of situation individualsperceive as aggressive. However, these studies are correlationalin nature. An experimental design, allocating individuals randomlyto footage of differing levels of provocation and frustration wouldresult in a fuller picture of the impact of these sensitivities toaggressive stimuli. In addition, the influence of provocation andfrustration sensitivity on responses to aggressive and ambiguousbehavior during real interactions would allow us to examine therole of these sensitivities on the ‘‘vicious cycle of aggression”,and permit an analysis of verbal and indirect aggression as wellas its more physical forms. Finally, general trait aggression hasbeen associated with biological bases – in particular reduced base-line autonomic and electrocortical arousal and greater reactivity tostressful and provoking situations (Ortiz & Raine, 2004). It remainsto be seen whether this pattern of reactivity is found when exam-ining those sensitive to provocation during provoking situationsspecifically.

Acknowledgements

We would like to thank Professor Eamonn Ferguson, ProfessorClive Hollin and three anonymous reviewers for their insightfulcomments on drafts of this paper.

References

Anderson, C. A., & Bushman, B. J. (2002). Human aggression. Annual Review ofPsychology, 53, 27–51.

Bower, G. H. (1991). How might emotions affect learning? In S. Christianson (Ed.),The handbook of emotion and memory (pp. 3–31). Hillsdale, NJ: Erlbaum.

Buss, A. H., & Perry, M. P. (1992). The Aggression Questionnaire. Journal ofPersonality and Social Psychology, 63, 452–459.

Chapman, B. P., & Hayslip, B. (2005). Incremental validity of a measure of emotionalintelligence. Journal of Personality Assessment, 85, 154–169.

Dill, K. E., Anderson, C. A., Anderson, K. B., & Deuser, W. E. (1997). Effects ofaggressive personality on social expectations and social perceptions. Journal ofResearch in Personality, 31(2), 272–292.

Dodge, K. A., & Crick, N. R. (1990). Social information processing bases of aggressivebehavior in children. Personality and Social Psychology Bulletin, 16, 8–22.

Forgas, J. P. (1994). The role of emotion in social judgments: An introductory reviewand an affect infusion model (AIM). European Journal of Social Psychology, 24,1–24.

Gomez, A., & Gomez, R. (2002). Personality traits of the behavioral approach andinhibition systems: Associations with processing of emotional stimuli.Personality and Individual Differences, 32, 1299–1316.

Lawrence, C. (2006). Measuring individual responses to aggression triggeringevents: Development of the situational triggers of aggressive responses (STAR)scale. Aggressive Behavior, 32, 241–252.

Marshall, M. A., & Brown, J. D. (2006). Trait aggressiveness and situationalprovocation: A test of the traits as situational sensitivities (TASS) model.Personality and Social Psychology Bulletin, 32(8), 1100–1113.

Mayer, J. D., & Hanson, E. (1995). Mood-congruent judgment over time. Personalityand Social Psychology Bulletin, 21, 237–244.

Niedenthal, P. M., & Setterlund, M. B. (1994). Emotion congruence in perception.Personality and Social Psychology Bulletin, 20, 401–411.

O’Connor, D. B., Archer, J., & Wu, F. W. C. (2001). Measuring aggression: Self-reports,partner reports and responses to provoking scenarios. Aggressive Behavior, 27,79–101.

Ortiz, J., & Raine, A. (2004). Heart rate level and antisocial behavior in children andadolescents: A meta-analysis. Journal of American Academic Child and AdolescentPsychiatry, 43, 154–162.

Rusting, C. L. (1999). Interactive effects of personality and mood on emotion-congruent memory and judgment. Journal of Personality and Social Psychology,77(5), 1073–1086.

Sims, C. M., Noel, N. E., & Maisto, S. A. (2007). Rape blame as a function of alcoholpresence and resistance types. Addictive Behaviors, 32, 2766–2775.

Smith, P., & Waterman, M. (2005). Sex differences in processing aggression wordsusing the Emotional Stroop task. Aggressive Behavior, 31(3), 271–282.

Smith, P., Waterman, M., & Ward, N. (2006). Driving aggression in forensic and non-forensic populations: Relationships to self-reported levels of aggression, angerand impulsivity. British Journal of Psychology, 97, 387–403.

Tiedens, L. Z. (2001). The effect of anger on the hostile inferences of aggressive andnonaggressive people: Specific emotions, cognitive processing and chronicaccessibility. Motivation and Emotion, 25, 223–251.

Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of briefmeasures of positive and negative affect: The PANAS scales. Journal ofPersonality and Social Psychology, 54(6), 1063–1070.

Whatley, M. A. (2005). The effect of participant sex, victim dress, and traditionalattitudes on causal judgments for marital rape victims. Journal of FamilyViolence, 20(3), 191–200.

Wingrove, J., & Bond, A. J. (2005). Correlation between trait hostility and fasterreading times for sentences describing angry reactions to ambiguous situations.Cognition and Emotion, 19(3), 463–472.