Ò
PAIN 155 (2014) 2656–2661
www.elsevier.com/locate/pain
Impact of pain behaviors on evaluations of warmth and competence Claire E. Ashton-James a,b,⇑, Daniel C. Richardson c, Amanda C. de C. Williams d, Nadia Bianchi-Berthouze e, Peter H. Dekker a a
Department of Social and Organizational Psychology, VU University Amsterdam, The Netherlands Department of Oral and Maxillofacial Surgery/Oral Pathology, VU University Medical Center/Academic Centre for Dentistry Amsterdam (ACTA), Amsterdam, The Netherlands c Department of Experimental Psychology, University College London, London, UK d Research Department of Clinical, Educational & Health Psychology, University College London, London, UK e UCL Interaction Center, University College London, London, UK b
Sponsorships or competing interests that may be relevant to content are disclosed at the end of this article.
a r t i c l e
i n f o
Article history: Received 7 July 2014 Received in revised form 21 September 2014 Accepted 23 September 2014
Keywords: Observers’ social judgments Pain estimates Perceived mood Body movement
a b s t r a c t This study investigated the social judgments that are made about people who appear to be in pain. Fiftysix participants viewed 2 video clips of human figures exercising. The videos were created by a motion tracking system, and showed dots that had been placed at various points on the body, so that body motion was the only visible cue. One of the figures displayed pain behaviors (eg, rubbing, holding, hesitating), while the other did not. Without any other information about the person in each video, participants evaluated each person on a variety of attributes associated with interpersonal warmth, competence, mood, and physical fitness. As well as judging them to be in more pain, participants evaluated the person who displayed pain behavior as less warm and less competent than the person who did not display pain behavior. In addition, the person who displayed pain behavior was perceived to be in a more negative mood and to have poorer physical fitness than the person who did not, and these perceptions contributed to the impact of pain behaviors on evaluations of warmth and competence, respectively. The implications of these negative social evaluations for social relationships, well-being, and pain assessment in persons in chronic pain are discussed. Ó 2014 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved.
1. Introduction Difficulties with social relationships are a major problem in chronic pain [21,30,55]. Children, adolescents, and adults with chronic pain report feeling misunderstood, stigmatized, and excluded by others because of their pain [29,44,52]. Scientific explanations of this experience include social withdrawal due to embarrassment about pain [55,56], reduced ability to participate in work, school, and social activities [4,37], and reduced sociability because of pain-induced negative mood [16,25,36,43]. These explanations suggest that the source of relationship dysfunction is the person with pain. Empirically, the role of the interaction partner has received less attention. People with pain consistently report that others evaluate them negatively [39,40,45], and some studies of doctor–patient and family interactions support that view [11,13,36]. However, because
chronic pain is also associated with a negative social information processing bias [10,22], the extent of actual stigmatization remains unclear. We aimed to clarify the impact of pain behavior1 on interpersonal evaluations that are critical to social relationships—namely, evaluations of warmth (trustworthiness, friendliness, or sincerity) and competence (intellect, skill, or capability) [20]. These core interpersonal judgments are made spontaneously on the basis of very little information [61], and predict social motivations to approach or to avoid others [1,19,20,32,38,49,62]. Judgments of high warmth and high competence elicit uniformly positive, affiliative behaviors (eg, admiration, respect, helping, cooperation), whereas judgments of low warmth and low competence are associated with uniformly negative, disaffiliating social responses (eg, contempt, disgust, harm, neglect) [8,9]. Understanding the impact of pain behaviors on warmth and competence evaluations may
⇑ Corresponding author at: Department of Social and Organizational Psychology, VU University Amsterdam, van der Boechorststraat 1, 1081BT Amsterdam, The Netherlands. Tel.: +31 639781224. E-mail address: [email protected] (C.E. Ashton-James).
1 We make no distinction here between protective (such as guarding) or communicative (such as facial expression) pain behavior [57], since protective pain behavior is also communicative [7] and, from an evolutionary perspective, communicative pain behavior is also protective by eliciting help from others [59].
http://dx.doi.org/10.1016/j.pain.2014.09.031 0304-3959/Ó 2014 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved.
Ò
C.E. Ashton-James et al. / PAIN 155 (2014) 2656–2661
provide insight into the social relationship difficulties experienced by people with pain. A person’s warmth is often inferred from that individual’s perceived mood. Positive mood is used as a cue for warmth, prompting a desire for social connection, whereas the perception or anticipation of another’s negative mood motivates avoidance of social contact [8]. Pain is associated with the experience and expression of negative mood [3,16,23,24,42,53]. To the extent that people perceive pain behavior to be indicative of negative mood, therefore, we predict that pain behavior may cue judgments of lower warmth. Competence is inferred from evolutionarily profitable traits such as cleverness, physical strength, and social status, whereas low competence is inferred from signs of physical, mental, social, or psychological vulnerability [9,46,47]. Because pain behaviors can signal physical vulnerability [41], we predict that pain behaviors will elicit less favorable judgments of competence than pain-free behavior, and that this relationship will be mediated by perceptions of physical fitness. We tested these predictions in a within-subjects experiment in which participants evaluated a person who displayed pain behavior and another who displayed none. Participants’ evaluations were based on body motion alone, avoiding the potential confounds inherent in social interaction and controlling for perceptual cues such as facial expressions, gender, and appearance.
2657
a forward bend (right and left) side bends, a back bend, and sitting. Reflective markers were placed on the joints and limbs of the actor and they were filmed by 6 high-speed infrared cameras. The locations of the markers in 3D space were reconstructed by the Vicon Nexus motion tracking system (see Appendix for video stills). The marker positions were then displayed as dots onscreen, and the recordings were edited to produce short video clips of the actor. Hence, the figures that participants viewed were composed only of point-light displays [34], with no facial features or expression visible, nor other individuating features that may influence social evaluations such as skin tone or clothing. The pain and no-pain videos were designed to be as similar as possible. The same actor performed the same actions in the same order, with indistinguishable levels of expressiveness or animation. Hence, the only observable difference between the figures in each video was the presence or absence of pain behaviors. Specifically, the pain behaviors were as follows: hesitation on all movements, and consequently the figure in the pain video took longer to perform the movements, but the difference was small; the actor performed a variety of guarding actions, for instance, when moving to sit, stand, and bend; using support when sitting from standing, and bracing of the back when standing from sitting, and in general, movements were performed more stiffly. The actor, who had experienced chronic pain herself in the past, was trained in pain behavior by a physiotherapist experienced in treating pain for maximum realism in both conditions.2
2. Methods 2.1. Participants and design A total of 56 university students (44 female) participated in a within-subjects experiment for university course credit or payment. Participants watched 2 video clips of an unidentifiable actor performing a series of exercises. In 1 of the 2 video clips, the person showed pain behaviors (eg, rubbing, guarding, hesitating) while completing the exercises, and in the other the exercises were performed normally. The order of presentation was counterbalanced. 2.2. Procedure Participants were recruited on a university campus by flyers advertising the opportunity to participate in research. After signing a consent form to participate in a study of ‘‘person judgments,’’ they were led to private cubicles to complete a computer-based study. Their first task was to watch two 2-minute videos of human figures performing a series of exercises (eg, side stretch, leaning forward, sitting down, standing up). Immediately after watching each video, participants were asked to evaluate the person in the video on a variety of different attributes related to warmth, competence, mood, and physical condition. The task instructions explained to participants, ‘‘We are interested in the impressions that people form of others based only on their bodily movements. For example, when you meet someone for the very first time, you often make a quick evaluation of their personality as you see them walking towards you.’’ After watching each video and completing each set of evaluative judgments, participants reported their perception of the level of pain experienced by the person in each video and reported their age and gender. Before participants were compensated and debriefed, they completed a variety of additional surveys as part of an ongoing research.
2.3.2. Evaluative judgments After watching each video, participants were given the following instructions: ‘‘Please give your impressions of the person that you saw in the video using the scales provided. We understand that you have very little information about this person, but we are interested in how much you can pick up based only on the information you received.’’ Participants then used a 7-point scale (1 = strongly disagree, 7 = strongly agree), to indicate their impression of the person in the video on a variety of attributes related to warmth (warm, friendly, cooperative, good natured, trustworthy, sincere, tolerant, honest, easy going, carefree; reverse-score items: mean, irritable, impatient), competence (confident, skillful, intelligent, competent, capable, prestigious, ambitious, lazy [reversed], economically successful, well educated, hardworking, persistent), mood (happy, depressed [reversed], relaxed, anxious [reversed]), and physical fitness (healthy, unfit [reversed], energetic, athletic). The order in which these attributes were presented was randomized to control for potential question order effects. The traits that together captured impressions of warmth and competence were selected based on their demonstrated validity and reliability in previous research (reviewed by Cuddy et al. [9]). Finally, participants reported how much pain they thought that the person in each video was experiencing (pain intensity) on an 11-point scale (0 = no pain at all, 10 = worst pain imaginable). 2.4. Data reduction and analytic approach
2.3. Materials and measures
Mean evaluations of warmth, competence, mood, and physical fitness were computed for evaluations of the person displaying pain behavior and the person showing no pain behavior. Reliabilities were high with 1 exception: warmth: pain behavior a = 0.82, no pain behavior a = 0.80; competence: pain behavior a = 0.83, no pain behavior a = 0.88; mood: pain behavior a = 0.80, no pain behavior a = 0.64; physical fitness: pain behavior a = 0.86, no pain behavior a = 0.89. We expect that there is greater heterogeneity in
2.3.1. Video stimuli Each video portrayed a human figure executing 5 simple exercises (3 times each) at a slow, steady pace: (right and left) leg lifts,
2 The video materials developed for this study are available upon request from the first author.
2658
Ò
C.E. Ashton-James et al. / PAIN 155 (2014) 2656–2661
negative affect ratings pertaining to the figure who showed no pain behavior because of the absence of affective (ie, pain) cues in this video. After checking for normal distributions, a multivariate repeated-measures analysis of variance (ANOVA) was conducted to examine whether evaluations of warmth, competence, mood, and physical fitness varied as a function of the presence or absence of pain behavior. Effect sizes were calculated using partial etasquared (g2p). Initial analyses included the order of evaluation (whether participants evaluated the person who displayed pain behavior first or second) as a between-subjects factor in the model. This analysis revealed no significant main or interaction effects involving order of evaluation or participant gender, so these factors are omitted from the model tested and presented in Section 3. Difference scores in participants’ evaluations of the warmth, competence, mood, and physical fitness of the person who showed pain versus no pain were calculated. These difference scores were used to investigate the correlations between the impact of pain behavior on evaluations of warmth, competence, mood, and physical fitness. The extent to which the impact of the pain versus no-pain video on perceived mood and physical fitness mediated the impact on perceptions of warmth and competence, respectively, was evaluated using multilevel modeling. Thereafter, the indirect effects were tested using the bootstrapping technique [54].
Table 1 Mean evaluations of warmth, competence, pain intensity, mood, and physical fitness as a function of pain behaviors.
Pain intensitya Warmth Competence Mood Physical fitness
All participants completed the study. No participants were excluded from analyses. One participant did not complete the perceived pain intensity rating scale. Because of a programming error, participants were unable to report their age, and hence this information is not available. 3.1. Impact of pain behavior on perceived pain intensity A repeated-measures analysis of variance (ANOVA) revealed that participants estimated the intensity of pain experienced by the person with pain behaviors as significantly greater (mean = 5.84, standard deviation [SD] = 2.53, range = 1–9) than the pain experienced by the person with no pain behavior (mean = 2.55, SD = 1.86, range = 1–10), F1,54 = 59.20, P < .001, g2p = 0.523. 3.2. Impact of pain behavior on evaluative judgments The omnibus multivariate repeated measures test revealed a significant main effect of the presence versus absence of pain behavior on evaluative judgments (F4,52 = 7.94, P < .001, g2p = 0.38). As presented in Table 1, compared to the person who exhibited no pain, the person who exhibited pain behaviors was evaluated as being significantly less warm (F1,55 = 6.96, P = .01, g2 = 0.11), less competent (F1,55 = 13.48, P = .001, g2 = 0.20), in a less positive mood (F1,55, = 6.19, P = .02, g2 = 0.10), and in worse physical condition (F1,55 = 24.08, P < .001, g2 = 0.31). 3.3. Interrelationships between the effects of pain behaviors on evaluative judgments As shown in Table 2, participants who estimated larger differences between the pain of the 2 video figures tended to perceive smaller differences between figures’ competence, mood, and physical fitness. This direction of difference was also apparent for perceptions of warmth but was not statistically significant. This means that, to the extent that participants interpreted pain behavior as a sign of pain intensity, they did not evaluate the behavior as a sign of competence, mood, or physical fitness.
No pain behavior
t(55)
5.84 4.41 4.18 4.51 3.65
2.55 4.74 4.69 4.96 4.91
7.69*** 2.64* 3.67** 2.49* 4.91***
(2.54) (.71) (.80) (1.11) (1.49)
(1.86) (.61) (.78) (.85) (1.32)
Numbers in parentheses are standard deviations. Higher scores reflect higher pain ratings, more perceived warmth and competence, better (more positive) mood, and more physical fitness. a df = 54. * P < .05. ** P < .01. *** P < .001.
Table 2 Correlations between difference scores reflecting relationships between the effects of pain behavior on evaluative judgments.
Pain intensity Warmth Competence Mood *
3. Results
Pain behavior
**
Warmth
Competence
0.18
*
–
0.29 0.19 –
Mood
Physical fitness **
0.45 0.55** 0.50**
0.50** 0.19 0.65** 0.66**
P < .05 (2-tailed). P < .01 (2-tailed).
The impact of pain behaviors on perceptions of warmth was correlated with the impact of pain behaviors on perceptions of mood but not physical fitness. This means that, to the extent that pain behaviors were used as a cue for evaluating warmth, they were also used to evaluate mood. However, the impact of pain behavior on evaluations of warmth was independent of the impact of pain behavior on evaluations of physical fitness and competence. The impact of pain behavior on perceptions of competence was associated with the impact of pain behavior on both physical fitness and mood. In other words, to the extent that participants interpreted pain behavior as a sign of (poor) physical fitness and (less positive) mood, they also used pain behavior as a cue for evaluating competence. 3.4. Exploring the impact of pain behaviors on judgments of warmth and competence It was hypothesized that the impact of observed pain behaviors on perceptions of a person’s mood and physical fitness would contribute to judgments regarding a person’s warmth and competence. This hypothesis was investigated using multilevel modeling. The repeated measurement character of the design is accounted for by introducing an unstructured (co-)variance matrix of the 2 sets of measurements (pain vs no-pain video). The impact of the pain versus no-pain video on variables is reported as a standardized effect (the difference in means divided by the pooled standard deviations), and the relation between impacts as a standardized coefficient (b). The outcome of the analyses is shown in Tables 3–5. Tables 3 and 4 show that perceived mood accounted for a significant portion of variance in judgments about warmth; the impact on perceived warmth was not significant any longer after the addition of the impact on mood, whereas the effects relating to mood remained significant. The bias-corrected and accelerated bootstrapped 95% confidence interval [48] for the indirect effect via perceived mood is 0.54 to 0.05, providing support for the hypothesis that perceived mood contributes to the impact of pain behaviors on evaluations of warmth.
Ò
C.E. Ashton-James et al. / PAIN 155 (2014) 2656–2661 Table 3 Contribution of perceived mood and physical fitness to evaluations of warmth and competence (pain vs no-pain video): impact of pain behavior on mediators. Variable
Effect
a
t
0.46a 0.89a
Mood Physical fitness
2.49 4.91
df
P (1-sided)
55 55
.008
PAIN 155 (2014) 2656–2661
www.elsevier.com/locate/pain
Impact of pain behaviors on evaluations of warmth and competence Claire E. Ashton-James a,b,⇑, Daniel C. Richardson c, Amanda C. de C. Williams d, Nadia Bianchi-Berthouze e, Peter H. Dekker a a
Department of Social and Organizational Psychology, VU University Amsterdam, The Netherlands Department of Oral and Maxillofacial Surgery/Oral Pathology, VU University Medical Center/Academic Centre for Dentistry Amsterdam (ACTA), Amsterdam, The Netherlands c Department of Experimental Psychology, University College London, London, UK d Research Department of Clinical, Educational & Health Psychology, University College London, London, UK e UCL Interaction Center, University College London, London, UK b
Sponsorships or competing interests that may be relevant to content are disclosed at the end of this article.
a r t i c l e
i n f o
Article history: Received 7 July 2014 Received in revised form 21 September 2014 Accepted 23 September 2014
Keywords: Observers’ social judgments Pain estimates Perceived mood Body movement
a b s t r a c t This study investigated the social judgments that are made about people who appear to be in pain. Fiftysix participants viewed 2 video clips of human figures exercising. The videos were created by a motion tracking system, and showed dots that had been placed at various points on the body, so that body motion was the only visible cue. One of the figures displayed pain behaviors (eg, rubbing, holding, hesitating), while the other did not. Without any other information about the person in each video, participants evaluated each person on a variety of attributes associated with interpersonal warmth, competence, mood, and physical fitness. As well as judging them to be in more pain, participants evaluated the person who displayed pain behavior as less warm and less competent than the person who did not display pain behavior. In addition, the person who displayed pain behavior was perceived to be in a more negative mood and to have poorer physical fitness than the person who did not, and these perceptions contributed to the impact of pain behaviors on evaluations of warmth and competence, respectively. The implications of these negative social evaluations for social relationships, well-being, and pain assessment in persons in chronic pain are discussed. Ó 2014 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved.
1. Introduction Difficulties with social relationships are a major problem in chronic pain [21,30,55]. Children, adolescents, and adults with chronic pain report feeling misunderstood, stigmatized, and excluded by others because of their pain [29,44,52]. Scientific explanations of this experience include social withdrawal due to embarrassment about pain [55,56], reduced ability to participate in work, school, and social activities [4,37], and reduced sociability because of pain-induced negative mood [16,25,36,43]. These explanations suggest that the source of relationship dysfunction is the person with pain. Empirically, the role of the interaction partner has received less attention. People with pain consistently report that others evaluate them negatively [39,40,45], and some studies of doctor–patient and family interactions support that view [11,13,36]. However, because
chronic pain is also associated with a negative social information processing bias [10,22], the extent of actual stigmatization remains unclear. We aimed to clarify the impact of pain behavior1 on interpersonal evaluations that are critical to social relationships—namely, evaluations of warmth (trustworthiness, friendliness, or sincerity) and competence (intellect, skill, or capability) [20]. These core interpersonal judgments are made spontaneously on the basis of very little information [61], and predict social motivations to approach or to avoid others [1,19,20,32,38,49,62]. Judgments of high warmth and high competence elicit uniformly positive, affiliative behaviors (eg, admiration, respect, helping, cooperation), whereas judgments of low warmth and low competence are associated with uniformly negative, disaffiliating social responses (eg, contempt, disgust, harm, neglect) [8,9]. Understanding the impact of pain behaviors on warmth and competence evaluations may
⇑ Corresponding author at: Department of Social and Organizational Psychology, VU University Amsterdam, van der Boechorststraat 1, 1081BT Amsterdam, The Netherlands. Tel.: +31 639781224. E-mail address: [email protected] (C.E. Ashton-James).
1 We make no distinction here between protective (such as guarding) or communicative (such as facial expression) pain behavior [57], since protective pain behavior is also communicative [7] and, from an evolutionary perspective, communicative pain behavior is also protective by eliciting help from others [59].
http://dx.doi.org/10.1016/j.pain.2014.09.031 0304-3959/Ó 2014 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved.
Ò
C.E. Ashton-James et al. / PAIN 155 (2014) 2656–2661
provide insight into the social relationship difficulties experienced by people with pain. A person’s warmth is often inferred from that individual’s perceived mood. Positive mood is used as a cue for warmth, prompting a desire for social connection, whereas the perception or anticipation of another’s negative mood motivates avoidance of social contact [8]. Pain is associated with the experience and expression of negative mood [3,16,23,24,42,53]. To the extent that people perceive pain behavior to be indicative of negative mood, therefore, we predict that pain behavior may cue judgments of lower warmth. Competence is inferred from evolutionarily profitable traits such as cleverness, physical strength, and social status, whereas low competence is inferred from signs of physical, mental, social, or psychological vulnerability [9,46,47]. Because pain behaviors can signal physical vulnerability [41], we predict that pain behaviors will elicit less favorable judgments of competence than pain-free behavior, and that this relationship will be mediated by perceptions of physical fitness. We tested these predictions in a within-subjects experiment in which participants evaluated a person who displayed pain behavior and another who displayed none. Participants’ evaluations were based on body motion alone, avoiding the potential confounds inherent in social interaction and controlling for perceptual cues such as facial expressions, gender, and appearance.
2657
a forward bend (right and left) side bends, a back bend, and sitting. Reflective markers were placed on the joints and limbs of the actor and they were filmed by 6 high-speed infrared cameras. The locations of the markers in 3D space were reconstructed by the Vicon Nexus motion tracking system (see Appendix for video stills). The marker positions were then displayed as dots onscreen, and the recordings were edited to produce short video clips of the actor. Hence, the figures that participants viewed were composed only of point-light displays [34], with no facial features or expression visible, nor other individuating features that may influence social evaluations such as skin tone or clothing. The pain and no-pain videos were designed to be as similar as possible. The same actor performed the same actions in the same order, with indistinguishable levels of expressiveness or animation. Hence, the only observable difference between the figures in each video was the presence or absence of pain behaviors. Specifically, the pain behaviors were as follows: hesitation on all movements, and consequently the figure in the pain video took longer to perform the movements, but the difference was small; the actor performed a variety of guarding actions, for instance, when moving to sit, stand, and bend; using support when sitting from standing, and bracing of the back when standing from sitting, and in general, movements were performed more stiffly. The actor, who had experienced chronic pain herself in the past, was trained in pain behavior by a physiotherapist experienced in treating pain for maximum realism in both conditions.2
2. Methods 2.1. Participants and design A total of 56 university students (44 female) participated in a within-subjects experiment for university course credit or payment. Participants watched 2 video clips of an unidentifiable actor performing a series of exercises. In 1 of the 2 video clips, the person showed pain behaviors (eg, rubbing, guarding, hesitating) while completing the exercises, and in the other the exercises were performed normally. The order of presentation was counterbalanced. 2.2. Procedure Participants were recruited on a university campus by flyers advertising the opportunity to participate in research. After signing a consent form to participate in a study of ‘‘person judgments,’’ they were led to private cubicles to complete a computer-based study. Their first task was to watch two 2-minute videos of human figures performing a series of exercises (eg, side stretch, leaning forward, sitting down, standing up). Immediately after watching each video, participants were asked to evaluate the person in the video on a variety of different attributes related to warmth, competence, mood, and physical condition. The task instructions explained to participants, ‘‘We are interested in the impressions that people form of others based only on their bodily movements. For example, when you meet someone for the very first time, you often make a quick evaluation of their personality as you see them walking towards you.’’ After watching each video and completing each set of evaluative judgments, participants reported their perception of the level of pain experienced by the person in each video and reported their age and gender. Before participants were compensated and debriefed, they completed a variety of additional surveys as part of an ongoing research.
2.3.2. Evaluative judgments After watching each video, participants were given the following instructions: ‘‘Please give your impressions of the person that you saw in the video using the scales provided. We understand that you have very little information about this person, but we are interested in how much you can pick up based only on the information you received.’’ Participants then used a 7-point scale (1 = strongly disagree, 7 = strongly agree), to indicate their impression of the person in the video on a variety of attributes related to warmth (warm, friendly, cooperative, good natured, trustworthy, sincere, tolerant, honest, easy going, carefree; reverse-score items: mean, irritable, impatient), competence (confident, skillful, intelligent, competent, capable, prestigious, ambitious, lazy [reversed], economically successful, well educated, hardworking, persistent), mood (happy, depressed [reversed], relaxed, anxious [reversed]), and physical fitness (healthy, unfit [reversed], energetic, athletic). The order in which these attributes were presented was randomized to control for potential question order effects. The traits that together captured impressions of warmth and competence were selected based on their demonstrated validity and reliability in previous research (reviewed by Cuddy et al. [9]). Finally, participants reported how much pain they thought that the person in each video was experiencing (pain intensity) on an 11-point scale (0 = no pain at all, 10 = worst pain imaginable). 2.4. Data reduction and analytic approach
2.3. Materials and measures
Mean evaluations of warmth, competence, mood, and physical fitness were computed for evaluations of the person displaying pain behavior and the person showing no pain behavior. Reliabilities were high with 1 exception: warmth: pain behavior a = 0.82, no pain behavior a = 0.80; competence: pain behavior a = 0.83, no pain behavior a = 0.88; mood: pain behavior a = 0.80, no pain behavior a = 0.64; physical fitness: pain behavior a = 0.86, no pain behavior a = 0.89. We expect that there is greater heterogeneity in
2.3.1. Video stimuli Each video portrayed a human figure executing 5 simple exercises (3 times each) at a slow, steady pace: (right and left) leg lifts,
2 The video materials developed for this study are available upon request from the first author.
2658
Ò
C.E. Ashton-James et al. / PAIN 155 (2014) 2656–2661
negative affect ratings pertaining to the figure who showed no pain behavior because of the absence of affective (ie, pain) cues in this video. After checking for normal distributions, a multivariate repeated-measures analysis of variance (ANOVA) was conducted to examine whether evaluations of warmth, competence, mood, and physical fitness varied as a function of the presence or absence of pain behavior. Effect sizes were calculated using partial etasquared (g2p). Initial analyses included the order of evaluation (whether participants evaluated the person who displayed pain behavior first or second) as a between-subjects factor in the model. This analysis revealed no significant main or interaction effects involving order of evaluation or participant gender, so these factors are omitted from the model tested and presented in Section 3. Difference scores in participants’ evaluations of the warmth, competence, mood, and physical fitness of the person who showed pain versus no pain were calculated. These difference scores were used to investigate the correlations between the impact of pain behavior on evaluations of warmth, competence, mood, and physical fitness. The extent to which the impact of the pain versus no-pain video on perceived mood and physical fitness mediated the impact on perceptions of warmth and competence, respectively, was evaluated using multilevel modeling. Thereafter, the indirect effects were tested using the bootstrapping technique [54].
Table 1 Mean evaluations of warmth, competence, pain intensity, mood, and physical fitness as a function of pain behaviors.
Pain intensitya Warmth Competence Mood Physical fitness
All participants completed the study. No participants were excluded from analyses. One participant did not complete the perceived pain intensity rating scale. Because of a programming error, participants were unable to report their age, and hence this information is not available. 3.1. Impact of pain behavior on perceived pain intensity A repeated-measures analysis of variance (ANOVA) revealed that participants estimated the intensity of pain experienced by the person with pain behaviors as significantly greater (mean = 5.84, standard deviation [SD] = 2.53, range = 1–9) than the pain experienced by the person with no pain behavior (mean = 2.55, SD = 1.86, range = 1–10), F1,54 = 59.20, P < .001, g2p = 0.523. 3.2. Impact of pain behavior on evaluative judgments The omnibus multivariate repeated measures test revealed a significant main effect of the presence versus absence of pain behavior on evaluative judgments (F4,52 = 7.94, P < .001, g2p = 0.38). As presented in Table 1, compared to the person who exhibited no pain, the person who exhibited pain behaviors was evaluated as being significantly less warm (F1,55 = 6.96, P = .01, g2 = 0.11), less competent (F1,55 = 13.48, P = .001, g2 = 0.20), in a less positive mood (F1,55, = 6.19, P = .02, g2 = 0.10), and in worse physical condition (F1,55 = 24.08, P < .001, g2 = 0.31). 3.3. Interrelationships between the effects of pain behaviors on evaluative judgments As shown in Table 2, participants who estimated larger differences between the pain of the 2 video figures tended to perceive smaller differences between figures’ competence, mood, and physical fitness. This direction of difference was also apparent for perceptions of warmth but was not statistically significant. This means that, to the extent that participants interpreted pain behavior as a sign of pain intensity, they did not evaluate the behavior as a sign of competence, mood, or physical fitness.
No pain behavior
t(55)
5.84 4.41 4.18 4.51 3.65
2.55 4.74 4.69 4.96 4.91
7.69*** 2.64* 3.67** 2.49* 4.91***
(2.54) (.71) (.80) (1.11) (1.49)
(1.86) (.61) (.78) (.85) (1.32)
Numbers in parentheses are standard deviations. Higher scores reflect higher pain ratings, more perceived warmth and competence, better (more positive) mood, and more physical fitness. a df = 54. * P < .05. ** P < .01. *** P < .001.
Table 2 Correlations between difference scores reflecting relationships between the effects of pain behavior on evaluative judgments.
Pain intensity Warmth Competence Mood *
3. Results
Pain behavior
**
Warmth
Competence
0.18
*
–
0.29 0.19 –
Mood
Physical fitness **
0.45 0.55** 0.50**
0.50** 0.19 0.65** 0.66**
P < .05 (2-tailed). P < .01 (2-tailed).
The impact of pain behaviors on perceptions of warmth was correlated with the impact of pain behaviors on perceptions of mood but not physical fitness. This means that, to the extent that pain behaviors were used as a cue for evaluating warmth, they were also used to evaluate mood. However, the impact of pain behavior on evaluations of warmth was independent of the impact of pain behavior on evaluations of physical fitness and competence. The impact of pain behavior on perceptions of competence was associated with the impact of pain behavior on both physical fitness and mood. In other words, to the extent that participants interpreted pain behavior as a sign of (poor) physical fitness and (less positive) mood, they also used pain behavior as a cue for evaluating competence. 3.4. Exploring the impact of pain behaviors on judgments of warmth and competence It was hypothesized that the impact of observed pain behaviors on perceptions of a person’s mood and physical fitness would contribute to judgments regarding a person’s warmth and competence. This hypothesis was investigated using multilevel modeling. The repeated measurement character of the design is accounted for by introducing an unstructured (co-)variance matrix of the 2 sets of measurements (pain vs no-pain video). The impact of the pain versus no-pain video on variables is reported as a standardized effect (the difference in means divided by the pooled standard deviations), and the relation between impacts as a standardized coefficient (b). The outcome of the analyses is shown in Tables 3–5. Tables 3 and 4 show that perceived mood accounted for a significant portion of variance in judgments about warmth; the impact on perceived warmth was not significant any longer after the addition of the impact on mood, whereas the effects relating to mood remained significant. The bias-corrected and accelerated bootstrapped 95% confidence interval [48] for the indirect effect via perceived mood is 0.54 to 0.05, providing support for the hypothesis that perceived mood contributes to the impact of pain behaviors on evaluations of warmth.
Ò
C.E. Ashton-James et al. / PAIN 155 (2014) 2656–2661 Table 3 Contribution of perceived mood and physical fitness to evaluations of warmth and competence (pain vs no-pain video): impact of pain behavior on mediators. Variable
Effect
a
t
0.46a 0.89a
Mood Physical fitness
2.49 4.91
df
P (1-sided)
55 55
.008
