the influence of self-generaed emotions on pshisical performance

7/30/2019 The Influence of Self-generaed Emotions on Pshisical Performance

1/15

197

Journal of Sport & Exercise Psychology, 2013, 35, 197-210

2013 Human Kinetics, Inc.

Marco Rathschlag and Daniel Memmert are with the Institute

o Cognitive and Team/Racket Sport Research, German Sport

University, Cologne, Germany.

The Influence of Self-Generated Emotionson Physical Performance: An Investigation ofHappiness, Anger, Anxiety, and Sadness

Marco Rathschlag and Daniel Memmert

German Sport University

The present study examined the relationship between sel-generated emotions and physical perormance.All participants took part in ve emotion induction conditions (happiness, anger, anxiety, sadness, and anemotion-neutral state) and we investigated their infuence on the orce o the nger musculature (Experiment1), the jump height o a counter-movement jump (Experiment 2), and the velocity o a thrown ball (Experi-ment 3). All experiments showed that participants could produce signicantly better physical perormances

when recalling anger or happiness emotions in contrast to the emotion-neutral state. Experiments 1 and 2also revealed that physical perormance in the anger and the happiness conditions was signicantly enhancedcompared with the anxiety and the sadness conditions. Results are discussed in relation to the Lazarus (1991a,2000a) cognitive-motivational-relational (CMR) theory ramework.

Keywords: emotions, emotion-perormance relationship, physical perormance, happiness, anger, anxiety,sadness

Athletes are always searching or possibilities toimprove their physical perormance in training andespecially in competition situations. Emotions and theirregulation are a undamental part o perormance and theocus o considerable research in sports (Hanin, 2000).For example, in a weightliting competition, where only a

ew kilograms decides on whether an athlete gets a medal,athletes need strategies to regulate their emotional levelin an intended direction directly beore liting the weightsto maximize their probability o success.

A primary unction o emotion is the preparation oraction, and recent evidence suggests that manipulatingemotional states preceding or during movement leadsto consistent and repeatable alterations in overt motorbehavior (Coombes, Gamble, Cauraugh, & Janelle,2008). A wide range o emotions has been investigated insports to illuminate the emotion-perormance relationship(Coombes et al., 2008; Jones & Uphill, 2011; Robazza& Bortoli, 2007; Sve, Ria, Poizat, Saury, & Durand,2007; Woodman, Davis, Hardy, Callow, Glasscock, &

Yuill-Proctor, 2009), and numerous methods have beenused in the past to induce specic emotions, such as lms(Rottenberg, Ray, & Gross, 2007), sounds (Bradley, &Lang, 1999), imagery scripts (Woodman et al., 2009),and pictures with emotional content (Lang, Bradley, &Cuthbert, 2005).

However, the weightliting competition poses oneproblem: to induce a desirable emotional state directlybeore liting the weights, it is very dicult to useexternal stimuli, such as lms or pictures, which havebeen the method o choice or inducing emotions inexperiments. Thus, it remains unclear whether athletes

can induce a desired emotion on their own to improveperormance in a competition situation. Thereore, weused sel-generated emotions as the emotion inductionmethod in the present research. In studies using sel-generated emotions, participants typically are asked torecall and reexperience personal emotional episodes(e.g., Damasio et al., 2000). To our knowledge, therewere no studies in the literature that have investigatedthe emotion-perormance relationship by using sel-generated emotions as an emotion induction method. Weinvestigated sel-generated emotions and their infuenceon physical perormance by drawing on the rameworko Lazaruss (1991a, 2000a) cognitive-motivational-relational (CMR) theory.

Lazaruss (1991a, 2000a) CMR theory suggests thatthe specic emotions o an athlete are each guided by acore relational theme. For example, the core relationaltheme o anger is a demeaning oence against me andmine (Lazarus, 2000a, p. 242). The core relationaltheme is a description o the interaction between theathlete and the environment and is a summary o di-erent appraisal judgments which are brought together.Each core relational theme o an emotion has a biologi-cally derived action tendency or impulse that is dicult

JOURNAL OF

SPORT EXERCISE

PSYCHOLOGY

Official Journal of NASPSPAwww.JSEP-Journal.com

ORIGINAL RESEARCH
http://www.naspspa.org/http://www.jsep-journal.com/http://www.jsep-journal.com/http://www.naspspa.org/


2/15

198 Rathschlag and Memmert

to inhibit. The action tendency or anger is a powerulimpulse to counterattack (Lazarus, 2000a, p. 243).

One idea o Lazaruss (1991a, 2000a) CMR theoryis that the core relational theme o the respective emo-tion and its associated action tendency will infuencethe perormance o an athlete depending on the complexrelationship between the athlete and the situation. Forexample, in a tennis match, anger may negatively impactperormance i players are angry with themselves or aseries o missed points in that their anger draws resourcesaway rom the primary task. However, i the physicalskill requires a lashing out motion toward an aggressoror opponent, perormance may be enhanced due to itsclose association with angers action tendency (Lazarus,2000b). For example, the direct extension o the armsin weightliting could be interpreted as a lashing outmotion, and thus the induction o the emotion o angermay increase the perormance o the weightliter. Assuch, we believed that the complex emotion-perormance-relationship can be explained within the ramework oLazaruss CMR theory.

Although Lazaruss CMR theory is well respected,there is some evidence that emotions may not neces-sarily be associated with speciic action tendencies(Fredrickson, 2001), particularly in more ecologicallyvalid situations that may constrain or inluence thedisplay o emotion, particularly or positive emotions(Fredrickson, 2001). Erez and Isen (2002) ound thathappiness was positively related to eective problemsolving but that happiness was motivational only whenthe task had reached a certain degree o diculty. Thus,positive emotions may provide both sucient resourcesand sucient motivation to pursue a demanding task(Fredrickson, 2001).

To the best o our knowledge, Woodman and col-

leagues (2009) were the rst to investigate the emotion-perormance-relationship within this ramework andound some support or CMR theory. We ound theargumentation o Woodman and colleagues (2009) orexplaining their results within this ramework veryconvincing, and thus we intended to continue their workabout the emotion-perormance-relationship with the aido CMR theory. Instead o the sel-generated emotions oour study, Woodman et al. (2009) used imagery scriptsor the purpose o inducing happiness, anger, and anemotionally neutral aect. They ound that participantsperormance on a dynamometer was signicantly greaterin the anger condition than in the happiness and the emo-tionally neutral conditions. The authors interpreted theseresults in line with Lazaruss (2000b) suggestion thatanger may acilitate physical perormance i the requiredskill is similar to the action tendency in anger (i.e., tolash out). One limitation o the Woodman et al. (2009)study was that they only investigated the infuence oemotions on one physical skill (dynamometer task), andthus it is not clear whether the eect can be generalizedto other physical skills. Totterdell (2000) ound supportor the idea that happiness can also enhance physicalperormance. In his study, players rom two proessional

cricket teams were asked to give ratings o their moodsand perormances three times a day or our days duringa competitive match between the teams. Pooled time-series analysis showed signicant associations betweenthe players moods and subjective perormances, and theassociations were independent o hassles and avorablestanding in the match, and happiness was positivelyrelated to cricket batting average. Coombes et al. (2008)investigated the extent to which pleasant and unpleasantemotional states, induced via pictures rom the Interna-tional Aective Picture System (IAPS; Lang et al., 2005),infuenced the perormance o participants in a pinchgrip task. The pictures rom the IPAS were rated on thedimensions o arousal (ranging rom calm to excited)and valence (ranging rom pleasant to unpleasant) andwere not categorized in dierent emotions. The employedpictures in the study o Coombes et al. (2008) representedthree categories: erotic couples (excited and pleasant),mutilation (excited and unpleasant), and pictures with aneutral content. Researchers ound that viewing pictureso both erotic couples and pictures o mutilation led to

greater pinch grip orce production in contrast to neutralpictures. Taken together, it seems that not only positiveemotions like happiness (Totterdell, 2000) but also nega-tive emotional states like anger (Woodman et al., 2009)or seeing mutilation pictures (Coombes et al., 2008),which might be associated with the emotion o disgust,can increase perormance depending on the requiredphysical skill.

Instead o dierent emotions and their valence,physical perormance can also be infuenced by emo-tional arousal. For example, physiological arousal waspositively associated with perormance on aerobic tasks(Partt, Hardy, & Pates, 1995), and strength tasks (Per-kins, Wilson, & Kerr, 2001), but negatively associated in

tasks that require ne motor control (Noteboom, Fleshner,& Enoka, 2001). The majority o research examiningemotional arousal has ocused on anxiety, and it remainsunclear how other negative emotions (e.g., anger, sadness)or positive emotions (e.g., happiness) that display similaror dierent physiological activation (Lazarus, 2000b)aect perormance. For example, the arousal o the emo-tions anger and anxiety is oten very similar (e.g., Russell& Mehrabian, 1974) and thus, it is not clear whether theseemotions might aect perormance dierently (Robazza& Bortoli, 2007).

The Present Research

The purpose o this study was to contribute to researchon the emotion-perormance relationship by exploringthe links between dierent sel-generated emotions andtheir eect on perormance in dierent physical skills. Toour knowledge, there are no studies that have investigatedthe infuence o sel-generated emotions on perormancewithin the ramework o Lazaruss CMR theory. In thisarticle, we will present the results o three experimentsreerring to the emotion-perormance relationship. InExperiment 1, we investigated this relationship in a


3/15

Self-Generated Emotions and Physical Performance 199

non-sport-specic physical task o nger strength. Inthis experiment, participants were asked to hold twongers together as well as possible against a mechanicalorce ater inducing dierent sel-generated emotions.There are well-established anatomical connectionsbetween the motor cortex and the limbic system (e.g.,Mogenson, Jones, & Yim, 1980; Groenewegen, 2007),which is involved in many o our emotions. O urtherrelevance, the model o the motor homunculus (Peneld& Rasmussen, 1950), which is a somatotopic representa-tion o dierent body parts in the primary motor cortex,has demonstrated that the ngers represent one o thelargest areas in the primary motor cortex. Thus, webelieved that i dierent emotions infuence participantsorce, this eect might be most pronounced in the ngermusculature.

To urther examine whether dierent emotionsinfuence participants orce not only in a non-sport-specic physical task or the ngers, we subsequentlyinvestigated the emotion-perormance relationship ina sport-specic lower body movement in Experiment 2

and in a sport-specic upper body movement in Experi-ment 3. Participants were asked to jump as high as theycan (Experiment 2) and to throw a ball with maximumvelocity toward a goal (Experiment 3) ater inducingsel-generated emotions.

Focusing now on Experiment 1, we explored theeect o dierent kinds o emotions on the orce o thenger musculature, by examining the ollowing veemotion induction conditions: happiness, anger, neutral,anxiety, and sadness. First o all, we simply wanted toveriy that the emotion manipulations had been successulin inducing the appropriate emotions. We hypothesizedthat eelings o anger (happiness, anxiety, sadness) willbe signicantly enhanced in the anger (happiness, anxiety,

sadness) condition compared with all other conditions.Furthermore, we expected that the arousal will not diersignicantly between emotions o anger, happiness,anxiety, and sadness but to the emotion-neutral condi-tion. For the hedonic tone we expected that the valuesor pleasantness will be greater in happiness, comparedwith the emotion-neutral condition and the anger, anxiety,and sadness conditions. For physical perormance wehypothesized, in accordance with the results o Wood-man et al. (2009), that participants who sel-generate theemotion anger will show the highest perormance on apurely physical task. Specically, because the requiredskill is similar to the action tendency in anger (i.e., toclench ones st), we hypothesized that anger will benetperormance on such a task. The core relational themeor happiness is making reasonable progress toward therealization o a goal (Lazarus, 2000a, p. 234), and thusapproach is an action tendency in happiness (Lazarus,1991b). Specically, because the required skill (holdingtwo ngers together) could be interpreted as an approachmovement, and given the research o Totterdell (2000),who ound that the subjective perormance o cricketplayer is enhanced when players are happier, we believedthat happiness will also acilitate perormance in our task.

The action tendency or anxiety is avoidance or escape(Lazarus, 2000b) and the uncertain threat makes us eelmore or less powerless (Lazarus, 2000b, p. 57). Forsadness, the action tendency is inaction and withdrawal(Lazarus, 2000b, p. 57). Consequently, we hypothesizedthat the emotions anxiety and sadness would decreaseperormance in our physical task in Experiment 1, inwhich participants were asked to hold two ngers togetheras well as possible against a mechanical orce.

Experiment 1

Method

Participants

Thirteen male and 12 emale athletes with an average ageoM = 24.60 years (SD = 2.29), and ages ranging rom22 to 29 voluntarily took part in this study. Participantswere recruited rom a large university in Germany, andall o them were sport students at an amateur to semi-

proessional level in their respective sports and novicesin the required physical task. Participants received nocompensation or participation, and inormed consent wasobtained beore commencing the study. The study wascarried out in accordance with the Helsinki Declarationo 1975. None o the participants knew about the purposeo the study, the hypotheses, or CMR theory ramework.

Emotion Induction

We used a method in which ve emotion conditions (hap-piness, anger, neutral, anxiety, sadness) were induced byrecalling personal emotional episodes. In the happinesscondition, or example, participants had to imagine a veryhappy moment in their lie. For the emotion-neutral state,we suggested participants to imagine themselves brushingtheir teeth. This image was similar to the induction o thisstate via an emotion-neutral script outlining the process obrushing ones teeth (see Kavanagh & Hauseld, 1986).All participants participated in the ve emotion condi-tions, which were counterbalanced across all participants.

Manipulation Check

We used Likert scales (LS) to assess the degree to whichparticipants experienced the dierent emotions. Thiswas done to veriy that the emotion manipulations weresuccessul. For each emotion condition, participants retro-spectively rated the induction o the respective emotions,

valence, and arousal using a 9-point Likert scale (emotioninduction: 1 = no happiness / anger / anxiety / sadness to9 = most happiness / anger / anxiety / sadness; valence:1 = most pleasantto 9 = most unpleasant; arousal: 1 =not arousing to 9 = most arousing).

Physical Task

Strength o the nger musculature was measured viaa machine (see Figure 1) that represents an objective


4/15


measurement o the Bi-Digital O-Ring-Test (BDORT)developed by Omura (1985). The BDORT was originallydeveloped as a noninvasive diagnostic procedure ormedical problems in which a patient has to orm a ringwith the thumb and the index nger (Omura, 1985). Thediagnostician subjectively evaluates the patients healthaccording to their nger strength as the diagnosticiantries to pry them apart. Our machine allows us to objec-tiy the pulling orce in the BDORT. We used this kindo measurement instead o a pinch grip task (Coombeset al., 2008) or a dynamometer (Woodman et al., 2009)

because it was a part o a larger project. In this project wewant to investigate a new treatment specically designedor anxiety disorders but also or improving perormancein sport-speciic contexts, the so-called Wingwavemethod developed by Besser-Siegmund and Siegmundin 2001. This treatment combines elements o eye move-ment desensitization and reprocessing (EMDR), neuro-linguistic programming (NLP), and the BDORT. Oneidea o the developer is that the patients orce o ngermusculature in the BDORT is dierent depending onwhich kind o emotion they sel-generate and how well

Figure 1 Experimental setup. Top: posture o arm, orearm, and especially o index nger and thumb during the task. Bottomlet: posture o index nger and thumb rated as closed ring coded with 2. Bottom right: posture o index and thumb rated asunclosed ring coded with a 1.


5/15

Self-Generated Emotions and Physical Performance 201

patients can deal with this emotion. The aim o thisarticle is not to explain the Wingwave method in moredetail (or detailed explanations, see Besser-Siegmund,& Siegmund, 2010); we merely wanted to make clear thatwe chose this kind o measurement as a rst step in theevaluation process o the Wingwave method developedby Besser-Siegmund and Siegmund in 2001.

Our machine or the objective measurement othe BDORT generated a pulling orce that separatesthe index nger and the thumb when they touch eachother to orm a ring. A regulator controlled the strengtho the pulling orce. At rst, the maximal strength othe participants was measured using the one repetitionmaximum. The one repetition maximum was denedas the highest pulling orce at which participants canstill hold the ring o index nger and thumb together.Ater amiliarization with the equipment, the strengtho the pulling orce was added in small increments(0.51.0 bar), resting 30 s between measurements, untilthe subject could no longer hold the ring o index ngerand thumb together. The position o the ngers was

standardized or all participants (also see Figure 1). Allo the measurements in the dierent emotion condi-tions in the study were tested at 90% o participantsindividual maximum voluntary contraction (MVC).The measurements were lmed by a digital camera, andthe lmed material was observed by three raters whohad to decide independently whether the ring o indexnger and thumb was open or closed. We used a blinddesign: the rater was not inormed about the respectiveemotion condition that participants had induced. Thecoding system was the ollowing: 1.0 = unclosed ring,1.3 = approximately unclosed ring, 1.7 = approximatelyclosed ring, 2.0 = closed ring. Ater we assessed inter-rater reliability o the three dierent subjective orce

ratings, the mean o the three rater judgments or eachemotion condition (mean o the six trials per condition)was used or analysis.

Procedure

We inormed par ticipants, who attended the testing ses-sions individually, that the experiment was an examina-tion o physical perormance under dierent emotionconditions, and we provided them with instructionson how to complete the physical task. Ater providingdemographic inormation and written consent, par-ticipants were amiliarized with the machine or theobjective measurement o the BDORT and we tested theindividual MVC o the participants. Then participantswere seated at a desk, and the experimenter outlinedthe emotion that was to be induced during the rstemotion condition and asked participants to think o asituation in which they had experienced this emotion(e.g., happiness). When participants conrmed that theyhad a situation in mind, they had 1 min to sel-generatethe corresponding emotion. Immediately aterward,participants put their thumb and their index ngerin the machine or the objective measurement o the

BDORT and perormed six measurements o the orceo the nger musculature (90% MVC) under the sameemotion condition with breaks o 30 s in between eacho the six trials. The moment at which the machinebegan generating the pulling orce was announced byan acoustic signal 3 s in advance. From that momenton, participants were asked to hold the ring o indexnger and thumb together with their maximum orceand go on with sel-generating the emotion. Ater onetrial, participants were asked to relax their ngers inthe machine until the next acoustic signal but go onwith sel-generating the respective emotion in the restintervals between the trials. Par ticipants completed sixtrials under one emotion condition. Immediately aterone emotion condition with six trials, par ticipants wereasked to indicate how they were eeling retrospectivelyon the LS. Ater a rest o 3 min, we again measuredindividuals MVC in our task to control or atigueailures and, i necessary, the ollowing measures othe orce o the nger musculature within the next emo-tion condition were made at a new value o 90% o the

MVC. Then participants were asked to think o a situ-ation in which they had experienced the next emotion(e.g., sadness), and the same procedure started againor the new emotion condition. The procedure was thesame or each condition (i.e., happiness, anger, neutral,anxiety, and sadness), and we chose a resting time o3 min between each o the conditions to minimize anycarryover eect rom one emotion condition to the next.The order o presentation o the emotion conditionswas balanced and randomized across all participants.Since our objective measurement o the BDORT is anovel physical task, we repeated all measurements 2weeks later, counterbalancing the experimental condi-tions between the two times o measurements (T1 and

T2) to be sure that the results in our task do not dierbetween two times o measurements.

Data Analysis

First o all, we assessed the sel-generated emotionsecacy with respect to an induction o the respectiveemotions (happiness, anger, neutral, anxiety, sadness)during the physical task. Thereore, data were analyzedusing 5 (emotion condition: happiness vs. anger vs. neu-tral vs. anxiety vs. sadness) 2 (time o measurement:T1 vs. T2) ANOVAs with repeated measures on bothactors to examine the dierent LSs.

We then assessed the interrater reliability o thethree dierent strength ratings by calculating intraclasscorrelation coecients (ICC; Shrout & Fleiss, 1979)separately or each emotion condition (happiness, anger,neutral, anxiety, sadness), and or each time o measure-ment (rst measurement and second measurement).Subsequent to this analysis, we compared the ve emo-tion conditions. Thereore, data were analyzed using a5 (emotion condition: happiness vs. anger vs. neutral vs.anxiety vs. sadness) 2 (time o measurement: T1 vs. T2)ANOVA with repeated measures on the second actor.


6/15


Results

Manipulation Check

Happiness, Anger, Anxiety, and Sadness. There wasa signicant happiness dierence across emotion condi-tions, F(4, 96) = 198.53, p < .001, 2 = .89. Follow-up

Bonerroni-corrected pairwise comparisons revealedthat participants experienced signicantly more happi-ness in the happiness condition compared with all otherconditions (p < .001; see Table 1). Furthermore, therewas an anger dierence across emotion conditions, F(4,96) = 187.13,p < .001, 2 = .86. Participants experiencedsignicantly more anger in the anger condition comparedwith all other conditions (p < .001; see Table 1). We alsoound a signicant anxiety dierence across emotionconditions, F(4, 96) = 174.76, p < .001, 2 = .81. Par-ticipants experienced signicantly more anxiety in theanxiety condition compared with all other conditions (p< .001; see Table 1). There was also a signicant sadnessdierence across emotion conditions, F(4, 96) = 206.17,

p < .001, 2

= .91. Participants experienced signicantlymore sadness in the sadness condition compared with allother conditions (p < .001; see Table 1).

Arousal and Pleasantness (Hedonic Tone). Analysesrevealed a signicant arousal dierence across emotionconditions, F(4, 96) = 17.58,p < .001, 2 = .61. Follow-upBonerroni-corrected pairwise comparisons showed thatparticipants experienced signicantly less arousal in theneutral condition in relation to all other conditions (p

the influence of self-generaed emotions on pshisical performance

Documents