Does Motor or Visual Experience Enhance the Detection of Deceptive Movements in Football?

International Journal of Sports Science & Coaching Volume 7 · Number 2 · 2012 269

Does Motor or Visual Experience Enhance the Detection of

Deceptive Movements in Football?Alexandra Pizzera and Markus Raab

German Sport University Cologne, Institute of Psychology, German Sport University Cologne, Am Sportpark Müngersdorf 6,

50933 Cologne, GermanyE-mail: [email protected]

ABSTRACT

This study addressed the question of whether motor and visual

experiences enhance the detection of deceptive and non-deceptive

actions in football. Using a pre–post–retention test design, we conducted

an intervention study with 40 football players, manipulating specific motor

and visual experiences through a training intervention, with the motor

group learning to fake fouls and the visual group watching the training. Prior

general motor and visual experiences were also assessed via a

questionnaire. Referee-like decision-making performance, involving the

detection of deceptive actions, was measured using a video test. Separate

4 × 3 (Group × Video Test) repeated-measures analyses of variance were

calculated, with decision accuracy and decision time as the dependent

variables, representing decision-making performance. The results revealed

only slight effects of the intervention, with the motor group reporting long-

term effects, as reflected by the greater performance enhancement from

post- to retention test. Prior general visual experience in football was

shown to enhance deception detection for all three groups. Sport

associations could use these results in their referee education programs by

including prior general experiences as well as motor and visual training in

deceptive and non-deceptive actions to enhance accurate judgments of

these situations on the field.

Key words: Association Football, Deceptive Movement, Decision

Making, Perceptual Judgement, Refereeing, Soccer

INTRODUCTIONSport officials face tough demands during competitions, such as having to make the correctdecision in a time-constrained environment. Especially in ball games such as football,referees can be influenced by crowd noise [1], the referee assistant’s position [2], or thecalibration effect [3], which can bias their judgments. In addition, referees can be influenced

Reviewers: Damian Farrow (Victoria University, Australia)Steve Cobley (University of Sydney, Australia)

by players, who may engage in deceptive movements. Nonetheless, referees must stillcorrectly call fouls and give sanctions accordingly.

Deceptive movements are defined as movements that purposely mislead the observer oropponent as to the attacker’s true intention or direction of movement [4, 5]. As seen recentlyin the 2010 World Cup football tournament in South Africa, referees were occasionallyconfused about the actual intention of the players’ behavior. In some situations, theycorrectly—as inferred from review of the slow-motion replay after the foul—gave sanctionsto football players simulating having been fouled, but in other situations they gave sanctionsfor simulations although the player really had been fouled, or they called a foul although theplayer had simulated.1 Other studies found that match referees’ decisions did not tally withdecisions of expert panels [6, 7]. Such problems with varying refereeing decisions, togetherwith the above-mentioned influences on referees’ decisions, indicate the need to providereferees with training tools that will enhance their decisions. This can support referees inenforcing the rules and regulations of the Fédération Internationale de Football Association(FIFA) when dealing with the behavior of players in modern football.

Coaching referees in an attempt to educate them during their development and enhanceand optimize their working decisions is a highly challenging mission that has beenpreviously discussed [8, 9]. Several attempts have been made to develop training toolsfocusing on different aspects of a referee’s task. For instance, video-training methodscombined with appropriate feedback have been shown to enhance offside decisions [10, 11]and decision making in potential foul situations in football [12]. In another study [13],basketball referees were instructed to focus on defensive fouls. Although the referees did notshow significant improvement in the video-based infraction detection task, the study showedthat focusing in more detail on different abilities, such as the detection of fouls and violationscould enhance the decision making of sports officials. What makes the detection of fouls andviolations so difficult is the fact that players often set out to intentionally deceive theobserver [5]. Therefore, one aspect of coaching referees in decision making is to guide theirattention to the deceptive behavior of athletes.

It has been shown that the detection of deceptive movements is influenced by differentfactors, especially motor and perceptual expertise. In a study on rugby players [4], expertplayers were better than novices in detecting the direction of movement of rugby players whofaked movements in one direction while moving in the other. Similar results were found intwo studies on handball players. Participants were asked to watch temporally occluded videoclips of 7-m penalty situations in handball and indicate whether the penalty taker shot orfaked a shot at the goal. Skilled handball players and goalkeepers outperformed novices indiscriminating actual shots from faked shots [14]. The individual contributions of motor andvisual experiences to the recognition of deceptive intentions were examined by manipulatingthe viewing perspective [15]. With the same task of distinguishing actual shots from fakeshots, participants watched video clips of 7-m penalty situations in handball from a frontview and a side view. As the front view shows the penalty taker from a goalkeeper’scustomary viewing perspective, skilled goalkeepers are visually more familiar with thisperspective than are other players. The side view is seen as a more neutral perspective, sincethere is no task-specific experience with this perspective in either group. An influence ofmotor experience would predict that the perception of deceptive intentions is not influencedby the viewing perspective; an influence of perceptual experience would predict an

270 Deceptive Actions in Football

1Simulation is defined as “attempts to deceive the referee by feigning injury or pretending to have been fouled”(Law 12, Laws of the Game of the Fédération Internationale de Football Association).

advantage for goalkeepers in the frontal view perspective. Results indicate that neither motornor perceptual experience alone can explain successful recognition of deceptive intentions.However, as both handball players and goalkeepers outperformed novices in discriminatingactual shots from faked shots, a combination of motor and perceptual experience is thoughtto greatly contribute to the recognition of deceptive actions.

Different explanations have been proposed for the superiority of experts in detectingdeceptive movements [16]. Accumulated perceptual (especially visual) experience is thoughtto explain this superiority, as athletes have accumulated extensive visual experience over theyears by watching hundreds of situations involving deceptive and non-deceptive actions.Alternatively, or perhaps in addition to visual expertise, motor expertise has been suggestedto positively influence the detection of deception. Motor experts can draw upon their motorrepertoire that has been developed through substantial practice in the respective sport. It hasbeen argued that motor experts are more accurate and faster in predicting movements byreading body kinematics [17]. The ability to read body kinematics therefore helps peoplepredict movements as well as detect deception; this was shown in a study with basketballplayers [16]. In this experiment, point-light videos2 were used to investigate if motor expertswere able to detect another’s intention from kinematic information alone, without focusingon postural cues (foot and hand position) or facial expression. The results revealed that motorexperts distinguished deceptive intentions (fake passes) from veridical intentions (truepasses) better than novices did. Overall, evidence suggests that both motor and visualexpertise contribute to the visual analysis of other people’s actions, although with a slightlygreater influence of motor expertise [18]. Participants were better at identifying anddiscriminating point-light displays of themselves and their friends (second best) than ofstrangers performing various actions. The factor that contributes to this finding indistinguishing own movements is considered to be motor experience, as people usually donot see themselves. Yet watching friends perform various actions produces an enhancedvisual experience of these actions. Therefore, the results of the study indicate that actionperception is positively influenced by motor and visual experiences.

Although a body of evidence shows that expert athletes outperform novices indiscriminating deceptive from non-deceptive movements, so far this idea has not beenapplied to a sport referee’s task. This seems rather surprising, since one of the main tasks ofa referee is to judge the intentions of athletes correctly, that is, to know whether they are“simulating.” In this context, the question of how referees cope with the deceptive behaviorof football players arises. Given the existing research and the fact that most referees haveplayed and watched football in their past, do they use their motor and visual experiences tosupport the detection of the actual intentions of players and give sanctions accordingly? It isassumed that a referee’s ability to judge a certain action, with no specific required reactionto the movement (other than deciding), is influenced by the same processes while performingand perceiving this action.

Our main goal in the present study was to examine whether football experts use their ownsituation-specific and prior general motor and visual experiences to judge deceptive actionsin one-on-one situations correctly. Note that football experts were used instead of footballreferees to find out whether motor and visual experiences on their own, as opposed toadditional referee experience, contribute to better decision-making performance in areferee’s task. Further, we wanted to examine the efficiency of a training program for theacquisition and improvement of judging in referee-specific tasks, starting with a relatively


2Point-light displays represent a moving body using dots that are attached to key joints of the body.

low judging level. Based on the results in the field of motor and visual expertise [17] anddeceptive movements [14], we hypothesized that personal bodily information serves as asource to enhance decision making in perceptual tasks involving deceptive movements. Forthe purpose of our study, we define motor and visual experiences as the practical experiencesas an athlete or spectator. We manipulated specific motor and visual experiences through anintervention in which some football players learned to fake fouls (motor experts) and otherswatched this training (visual experts). Respectively, we examined their decision-makingperformance in judging one-on-one situations in football with a video test. We tested thehypotheses that playing experts are more accurate and faster in detecting deceptivemovements in this test than visual experts and that visual experts are more accurate and fasterthan novices. Furthermore, we hypothesized that prior general motor and visual experiencesin football (experience as an athlete or spectator accumulated over years) lead to moreaccurate and faster decisions. To test our hypotheses, we manipulated specific motor andvisual experiences through an intervention, captured prior general motor and visualexperiences via a questionnaire, and tested decision-making performance in a football-specific video test in a pre–post–retention test design.

METHODPARTICIPANTSParticipants were 40 football players aged between 20 and 32 years (M = 24.3, SD = 2.15) whowere randomly assigned to three groups. Because all the participants were active football playersin different clubs during the experiment, some suffered injuries and were not able to completethe study. This led to an uneven distribution of participants in the different groups. The motorgroup (n = 11, all male) participated in an intervention directed at training action simulations infootball. The visual group (n = 12, 9 males, 3 females of equal football experience in terms ofyears) participated in an intervention directed at observing action simulations. The control group(n = 17, all male) did not participate in the intervention. Groups were matched in accumulatedyears of football playing experience, with the motor group averaging 18.36 years (SD = 2.50),


Table 1. Means and Standard Deviations of Prior General Motor and VisualExperiences of the Participants and Their Accuracy and Decision Times inthe Pretest

Motor group Visual group Control groupM SD M SD M SD

Accumulated motor 18.36 2.50 15.83 6.25 17.75 6.64experience in yearsMaximum level of motor 4.50 1.58 5.10 1.37 5.82 0.75experiencea

Visual experience frequency 3.36 1.29 3.83 1.90 5.41 3.02Accumulated visual 16.82 1.83 16.83 2.89 16.94 3.58experience in yearsAccuracy in the pretest 26.73 3.10 26.17 4.09 27.65 3.06(max. 41 points)Decision time (in ms) 751.14 288.38 876.73 462.36 715.33 241.97in the pretest

aMotor experience level: 0 = hobby to 7 = provincial competition.

the visual group averaging 15.83 years (SD = 6.25), and the control group averaging 17.75 years(SD = 6.64). For an overview of the participants’ experiences, see Table 1. The participantsvoluntarily took part in this study and provided informed consent prior to the study. The studywas approved by the University’s Ethics Committee.

MATERIALSQuestionnaireTo capture prior general motor and visual experiences, we used a shortened and modifiedversion of a questionnaire developed and piloted by Pizzera and Raab [19]. Concerning themotor experience scale, participants were asked about their experience as athletes in thesport, such as the number of years spent playing (accumulated motor experience in years)and the highest level played in the sport (maximum level of motor experience). The variablemaximum level of motor experience will be referred to as motor experience level throughoutthe remaining text. The visual experience scale assessed their experience as spectators, suchas the number of times per week/month/year they watched football on TV or in a stadium(visual experience frequency) and the total number of years they had watched football(accumulated visual experience in years). However, as watching football in a stadium wasfound to be very infrequent, the data were not considered for further analyses, which is whyvisual experience frequency and accumulated years mostly refers to visual experience on TV.

Video Test We used a decision-making video test that consisted of 41 video clips and four additionallead-in clips (one for each category) designed to familiarize the participants with the task.The participants were asked to categorize the video clips into (1) no foul, (2) foul, (3) fouland yellow card, and (4) foul and red card by pressing the appropriate key on a keyboard.Four different colored keys on the keyboard indicated the four response categories. The clipswere displayed on an 18-inch computer monitor (1,280 × 1,024 pixels). Inquisit 3(Millisecond Software, Seattle, Wash.) was used to control video clip presentation. Theduration of each video clip was approximately 8.56 s (range: 7–12 s). The number of videoclips per category was approximated to their mean appearance in the countries’ first footballleague as indicated by the national statistics data base (n = 19, n = 11, n = 7, and n = 4 forthe categories no foul, foul, foul and yellow card, and foul and red card, respectively). Dataused were the number of red cards, yellow cards, fouls, and tackles in the 2008/2009 seasondivided by the number of games in that season. The test–retest reliability of the video testrevealed significant correlations of accuracy (r = .75, p = .001, n = 17) and decision time (r= .50, p = .041, n = 17) for the control group. These reliability measures are consideredacceptable with respect to the participants being novices in a refereeing-specific task.

PROCEDUREInterventionWe manipulated specific motor and visual experiences through an intervention consisting ofeight training sessions. Over four weeks, with two training sessions per week, the motorgroup learned to fake fouls in different categories. In each training session, each of whichlasted about 30 min, participants performed twelve trials, six as an attacker and six as adefender. In 1-on-1 situations, the attacker in possession of the ball had to try to get past thedefender, who in turn was asked to try to get the ball. As soon as the two players made bodycontact, the offensive player was asked to fake (simulate) having been fouled. They weregiven instructions for three different categories of faked fouls in three different situations that


represent typical scenarios during football games. The foul categories matched thoserepresented in the video clips, with participant goals being to

1. Use a slide tackle of the defensive player to pretend that you have been kicked ortripped by the defender’s feet.

2. Pretend you have been tripped up by tripping over your own feet.3. Use the contact of the opponent’s upper body to simulate a body check.

The participants were asked to learn all three categories of foul in three different situations,always with the offensive player being in possession of the ball (see Figure 1):

a. Start in a V-like position and meet in the middle of the field.b. Start side by side and simulate a sprint challenge for the ball.c. Start on opposite sides of the field, running toward each other.

After the trials, the participants received feedback from the experimenters as to whetherthe faked fouls were good (hard to differentiate from a real foul) or bad (easily recognizedas a faked foul) and instructions on how to enhance the deceptiveness of the action. Ataxonomy of behaviors associated with deceptive and non-deceptive intentions [5] was usedto provide appropriate feedback. Among the most frequent categories used to determinewhether an action is deceptive or not are temporal contiguity (time lag between tacklingcontact and the effect on the tackled player), ballistic contiguity (whether body rolls werepart of the momentum of the fall or additional), and contact consistency (whether the pointof impact between the two players and the point of injury match).

The visual group attended the training sessions, watched and rated the motor groupconcerning the quality of their faked fouls, and heard the feedback given to the motor group.They were placed on the side of the field so that they saw the actual “fake” from slightlybehind the players. This viewing position was chosen after receiving feedback from the


Figure 1. Overview of the Three Situations Trained in the Intervention OP = offensive player, DP = defensive player, xo = position of the player inpossession of the ball, x = position of the player without possession of theball, v = visual group participants. Players’ routes indicated by dashed linesand arrows.

expert referees in the pilot study, who stated that this is the referees’ typical viewingperspective on the field. The visual group watched each participant of the motor group onceper training session in a specific designated presentation field, to ensure equal amount ofwatching/performing the incidents. If the participants of the visual group missed a trainingsession, they caught up by coming into the lab and watching the video-recorded session. Thethird group served as a control group and only attended the video tests.

Video TestBefore the first video test, all participants filled out the questionnaire on their motor andvisual experiences in football. Subsequently, they were individually tested. First, they wereinstructed to read a short version of Law 12 (fouls and misconduct) of the official FIFA“Laws of the Game.” Then, the participants were presented with the instructions on thescreen. They were seated approximately 55 cm from the computer monitor. In each videotest, the first four clips were treated as lead-in items to familiarize participants with the task.Following these four video clips, the 41 video clips of the test were randomly presented. Weused a pre–post–retention test design, with an additional test halfway through theintervention phase (hereafter referred to as the midtest) and a retention test 4 weeks after theposttest, in order to measure learning (see Figure 2). The same 41 video clips were used forall four video tests to ensure equivalence in test difficulty. The participants of the motor andvisual groups attended the video tests and the intervention training sessions; the controlgroup only attended the video tests.

PILOT STUDYIn a pilot study with 26 participants, we tested the reliability of the treatment (training andwatching the training of faking fouls in football) and the video test. Participants wererandomly assigned to the two treatment groups (the motor or visual group) or the controlgroup. The motor training to fake fouls revealed that faking fouls can be taught in differentcategories by providing accurate and detailed descriptions accompanied by feedback.Independent expert referees judged the faked fouls to be representative of non-foul situationsor simulations according to the laws of the game in the FIFA regulations [20]. Furthermore,participants in the treatment groups showed slightly enhanced decision-making performancein the video test, compared to the control group. Training sessions were videotaped toprovide participants of the visual group the opportunity to catch up if they missed trainingsessions. The participants then watched the missed training session on a video in the lab,instead of outside on the field. As the faked fouls were rated similarly by the participants ofthe visual group on the field and the participants of the visual group watching the videotaped


Figure 2. Summary of the Experimental Design and ProcedureTS = training session.

clips in the lab, this was considered a valid procedure. For the video test itself, the video clipswere taken from a previously developed “referee decision training method” [21]. Theyconsisted of short scenes of national and international matches. All video clips were judgedfor the correct responses by the National Football Association’s referee board. For the finalvideo test, 41 video clips were chosen from a larger sample of 120 clips, after being pretestedby using difficulty and discrimination indexes. The reliability of the dependent variablesaccuracy and decision time was then considered to be acceptable. The test–retest reliabilityof the video test in the pilot study revealed significant correlations for accuracy (r = .79, p =.004, n = 11) and decision time (r = .88, p < .001, n = 11) for the control group.

DATA ANALYSESAccuracy and decision time served as the measures of decision-making performance in thevideo test. The scoring of one video test resulted in a maximum of 41 points. A response wasconsidered correct when the video clip was categorized appropriately. Decision time wasmeasured in milliseconds. To test our hypotheses, we used analyses of variance (ANOVAs)with group (motor, visual, and control) as the between-subjects factor and video test (pre,mid, post and retention) as the within-subject factor. For detailed analyses we conductedseparate ANOVAs for accuracy and decision time. When the sphericity assumption wasviolated, the Greenhouse–Geisser correction was used. Effect sizes were calculated as partialeta-squared values (hp

2) and are reported only for F > 1. In a first step, the results of the pretest were analyzed to control for baseline effects. The

results show that the groups differed in their accuracy as well as in their decision time.Indeed, a separate 3 (motor, visual, control group) × 3 (mid, post, retention test) analysis ofcovariance (ANCOVA) with pretest as a covariate showed that the results for accuracy anddecision time were additionally moderated by the pretest, F(1, 36) = 52.53, p < .001, hp

2 =.59 and F(1, 36) = 16.65, p <.001, hp

2 = .32, respectively. Therefore, in further analyses, thedata were adjusted according to the pretest scores. The pretest scores were set at 100% withall other video test scores representing the performance differences (in %) in relation to thepretest scores.

To test influences of the prior general experiences on the video test, we first calculatedseparate ANOVAs for each prior general experience variable. We then dichotomized theentire sample into a low and a high group according to prior experience via the median splittechnique. Using independent-samples t-tests with pretest as the dependent variable, wecalculated differences between the two groups. A significance criterion of p = .05 wasestablished for all results reported.

RESULTSACCURACYWe first hypothesized that participants in the motor and visual groups would showsignificantly enhanced decision accuracy from pre- to retention test, compared to the controlgroup. A 3 (Group) × 4 (Video Test) ANOVA with repeated measures on the first factor andaccuracy as the dependent variable showed that the three groups did not significantly differin their accuracy, F(3, 111) = 1.24, p = .299, hp

2 = .03, achieved power = .99 (see Figure 3).Further, there was no group effect, F(2, 37) = 0.37, p = .696, achieved power = .81, and nointeraction effect of Group × Video Test, F(6, 111) = 1.59, p = .157, hp

2 = .08, achievedpower > .99. Although not significant, the motor group showed the strongestposttest–retention test improvement compared to the other two groups.


Calculating separate ANCOVAs for each factor of the participants’ prior general motorand visual experiences revealed no main effects for the covariate motor experience level,F(1, 27) = 1.49, p = .233, hp

2 = .05, achieved power = .61, and no main effects for thecovariate accumulated visual experience in years, F(1, 36) = 1.63, p = .210, hp

2 = .04,achieved power = 0.51. To test for effects of prior general motor experience on decision-making accuracy in the video test, the entire sample was dichotomized into a low and a highgroup according to number of years of experience via the median split technique. Althoughthe group with the greater accumulated motor experience in years was more accurate (meandifference 1.85), independent-samples t-tests with pretest as the dependent variable revealedno significant differences between the two groups, t(29.05) = 1.72, p = .096, d = .57,achieved power = .68. Similar results were found for the motor experience level, where thehigher level group was shown to be better (mean difference 1.84), but the difference did notreach significance, t(29) = 1.47, p = .153, d = .53, achieved power = .67. Concerning thevisual experience frequency, the group with a greater average reported frequency showedbetter decision-making accuracy (mean difference 1.60), but again the group difference wasnot significant, t(30.94) = 1.52, p = .138, d = .48, achieved power = .66. However, theamount of accumulated visual experience in years revealed significant differences.Participants with a higher amount of accumulated visual experience in years weresignificantly better in the pretest than those with a lower amount (mean difference 3.33),t(38) = 3.40, p = .002, d = 1.07, achieved power = .59.

DECISION TIMEOur second hypothesis was that the motor and visual groups would significantly improvetheir decision time. The 3 (Group) × 4 (Video Test) ANOVA with repeated measures on thefirst factor and decision time as the dependent variable revealed that all groups became


Figure 3. Differences in Accuracy (as Percentage of Correct Responses)for Each Group and Each Video TestThe pretest scores were set at 100% with all other video test scoresrepresenting the performance differences in relation to the pretest scores.Error bars represent standard errors.

significantly faster, F(1.76, 64.96) = 12.48, p < .001, hp2 = .25, achieved power > .99, but

again, there was no group effect, F(2, 37) = .36, p = .701, achieved power = .77, and nointeraction effect, F(3.51, 64.96) = .34, p = .824, hp

2 = .02, achieved power = .96 (see Figure4). However, calculating separate paired-samples t-tests for each group from pretest toretention test revealed that the overall significant decision time improvements were mainlydue to the visual group, t(11) = 3.84, p = .003, d = 1.11, achieved power = .67. The resultsfor the motor and control groups were t(10) = 1.83, p = .098, d = .55, achieved power = .67,and t(16) = 1.75, p = .099, d = .42, achieved power = .66, respectively.

Separate ANCOVAs, with the prior general motor and visual experiences of theparticipants as covariates, revealed a significant main effect for accumulated motorexperience in years, F(1, 35) = 8.03, p = .008, hp

2 = .19, achieved power = .60, but no effectfor visual experience frequency, F(1, 36) = 2.08, p = .158, hp

2 = .06, achieved power = 0.57.To test for effects of prior general motor experience on decision time in the video test, the

entire sample was dichotomized into a low and a high group based on experience via themedian split technique. Although the group with a greater amount of accumulated motorexperience in years made faster decisions (mean difference 148.03 ms), independent-samplest-tests with pretest as the dependent variable revealed no significant differences between thetwo groups, t(37) = 1.44, p = .158, d = .47, achieved power = .68. Similar results were foundfor the motor experience level, where the higher level group was shown to be faster (meandifference 181.56 ms), but the difference did not reach significance, t(29) = 1.47, p = .151, d= .53, achieved power = .66. Concerning the visual experience frequency, the group with ahigher frequency made significantly faster decisions (mean difference 260.76 ms), t(27.60)= 2.67, p = .013, d = .84, achieved power = .63. The amount of accumulated visual


Figure 4. Improvement in Decision Time (as percentage) for Each Groupand Each Video TestThe pretest scores were set at 100% with all other video test scoresrepresenting the performance differences in relation to the pretest scores.Points are offset horizontally so that the error bars representing standarderrors are visible.

experience in years revealed near significant differences. Participants with a greater amountof accumulated visual experience in years made faster decisions (mean difference 210.90ms), t(38) = 2.02, p = .051, d = .64, achieved power = .61.

DISCUSSIONOur main goal in the present study was to examine whether football experts use their ownsituation-specific and prior general motor and visual experiences to judge deceptive and non-deceptive actions in one-on-one situations correctly. We predicted specific and prior generalmotor and visual experiences would enhance decision making in these perceptual tasks. Totest our hypotheses, we manipulated specific motor and visual experiences through anintervention and captured prior general motor and visual experiences via a questionnaire.

ACCURACYThe results show that providing motor and visual experience did not significantly enhanceaccuracy in the video test. The groups showed fluctuations in performance, in particular,declines in retention for the visual group. This finding is inconsistent with previous researchon the effects of visual experience on perceptual judgments. Yet, previous research focusedon prior general visual experience, while the current study examined very specific visualexperience, which does not seem to show long-term effects for accuracy. This finding needsfurther consideration in future research to understand the underlying mechanisms. However,the motor group improved their performance from pretest to retention test by 2.88%, with arelative improvement in relation to the control group of about 2.30%. Given that footballreferees make about 137 observable decisions in one football game [22], 2.88% can mean anenhancement of about four correct decisions, with as little as eight training sessions. Asrecently seen in the 2010 FIFA World Cup in South Africa, even one decision can influencethe outcome of a game. One should be careful with interpretations of these results, however,as a sample of expert referees and a bigger sample size might provide significant differences.Nevertheless, the effect sizes revealed small to medium effects, indicating some practicalrelevance with good reliability according to the power analysis. Therefore, the current studyprovides a basis from which further ideas and studies incorporating the link between actionand cognition in football referees can be developed. Interestingly, the motor group showedthe strongest retention test improvement in accuracy of the three groups. We suggest thathaving motor experience (i.e., bodily experiencing events usually experienced by those theyjudge) as opposed to visually perceiving the events or not training at all, may have led to thegroup being more accurate, in the long run, in the detection of deceptive and non-deceptiveactions. Given that there is great interest in coaching football referees to improve theirdecision-making accuracy in the long term, this result seems quite important. However, theeffects were only found for the retention test, with even a decline in performance in theposttest. The current experimental data does not allow further explanation of this finding,which is why this needs careful consideration for future studies.

Further, analyses revealed that the participants’ prior general visual experiences may haveled to enhanced decision making. Participants with a greater amount of accumulated visualexperience in years showed higher accuracy scores in the video test. Considering thatspecific visual experience even showed a decline in performance, it seems that accumulatedvisual experience is more important for decision accuracy. This extends findings in a studyon expert football referees, where the visual experience of referees was also shown topositively correlate with their decision-making performance [19].


DECISION TIMEAs football referees not only have to judge correctly but must do so in a time-constrainedenvironment, decision time was also measured. All participants made increasingly fasterdecisions from pretest to retention test, but our hypothesis that the motor and visual groupswould show superiority over the control group was not confirmed. Nevertheless, analysesrevealed that accumulated motor experience in years influences the decision-timeimprovement from pre-test to retention test.

Taking into account the learning process, the main enhancement in decision time wasshown to be between midtest and post test. Further studies are needed at this point to explainthe underlying mechanisms for this finding. However, paired-samples t-tests revealed thatonly the visual group showed significant decision-time improvements. This was furtherunderlined by analyses of prior general visual experience, where participants with a highervisual experience frequency reported faster decisions. Therefore, it was shown that specificand prior general visual experiences seem to be associated with faster decisions. In thisarticle, the results from studies of motor and visual expertise [16, 17] and deceptivemovements [14] conducted in sports such as rugby, basketball, and handball were extendedto the sport of football. Sensorimotor experience gained as an athlete or spectator in footballseems to be an important factor for perceptual tasks involving deceptive and non-deceptiveactions in football. Bodily information—both the amount and the frequency—provides abasis for judging actions. Considering our current results, we follow the opinion that domain-specific experience is important for the development of the knowledge necessary to makecognitive judgments [23]. Previous studies have concluded that enhanced declarativeknowledge is a characteristic of skill (motor experience) rather than a by-product of (forinstance) visual experience alone [24]. However, the question arises as to how specific thisexperience needs to be. The current study shows that prior general motor and visualexperiences provide a good basis from which further experience for perceptual judgmentscan be trained. In terms of accuracy, a higher amount of prior general visual experience inthe sport goes along with more accurate decisions. Furthermore, football players seemed toillustrate small retained benefits from additional specific motor training (training to fakefouls), in that training tended to enhance decision accuracy in a cognitive task involvingdeceptive movements in football.

In terms of decision time, a greater amount of prior general motor and visual experiencewas associated with faster decisions. Additionally, specific visual training (watching thetraining of faking fouls in football) tended to result in faster decisions. We therefore contendthat in addition to increasing their awareness and sensitivity to sources of bias [25], aspiringfootball referees, many of whom are in a role of transitioning from player to official, shouldalso learn to recognize deceptive behavior of players. In a special issue of Psychology ofSport and Exercise [26], different approaches from social and general psychology as well aseconomics were introduced to address the topic of judgment and decision making in sports.We propose extending these to include a more body-centric approach, where motor andvisual experiences are taken into account. To the best of our knowledge, this is one of thefirst studies training referee-specific situations on the field to enhance decision performance.

METHODOLOGICAL LIMITATIONSWe are aware of some methodological limitations in the current study. First, we used arelatively small sample, with an unequal number of participants per groups. As pointed out,due to the current activities of the participants in their football clubs, some suffered injuriesand could not finish the study. In addition, familiarity effects due to exposure to the same


video clips over the four video tests cannot be fully ruled out. In our study we wanted toensure that the video clips were equally difficult. Future studies should address this issue byusing different video clips in order to check for transferability to other contexts andsituations.

In this study, one also has to take into account that motor and visual experiences could beslightly confounded. Although we designed the experiment to separate the two types ofexperience, we cannot fully rule out that participants of the motor group also saw otherparticipants perform the simulations. Future research could try to overcome this problem bychoosing a movement that could also be learned while blindfolded, as demonstratedelsewhere [31]. In addition, it would be of interest to conduct a similar study with referees,to test the transferability to actual sport officials. A similar design to the one used in the studyby MacMahon et al. [30], in which the intervention was tested on three groups (referees,players, and spectators) could be used. Finally, the next step should be to examine whetherperformance in the video test actually does transfer to performance on the field.

PRACTICAL IMPLICATIONSIf future studies reliably replicated our preliminary study by carefully considering themethodological limitations as stated above, some practical implications could be derived.First, the associations governing the education of referees could use these results to enhancethe process of selecting aspiring referees by taking their prior general motor and visualexperiences into account. As already proposed by the Professional Footballers Association[27, 28], former players should be encouraged to become referees. To facilitate this process,the Association has planned a “fast-track system” to halve the period of time it currentlytakes for an official to reach the top league, which implicitly acknowledges the contributionthat experience makes to performance in a new role. Linked to this idea, a suggestion wasput forward to employ ex-players to review “bad tackles” in order to retrospectively penalizeplayers [29]. Currently, such reviewing systems are performed by a panel of former expertreferees; however, former expert players are not involved. Until now, there has been a lackof scientific evidence to support such ideas. One study confirmed this idea of role transitions(ex-players becoming referees) by revealing that past experience as a player can providelasting influences on current performance as a referee [30]. The current study also attemptedto shed light on this area of former football players becoming referees. The accumulatednumber of years and the level at which the participants had played football, which we usedas indicators of prior motor experience, did not seem to influence performance in the videotest. However, in combination with the intervention, the participants with a higher amount ofaccumulated motor experience in years showed better decision time improvements after theintervention. Linking these results to the fast-track system, we would suggest that ex-playerswith a high amount of motor experience in years should be selected, and they should takepart in an intervention, before becoming referees. However, the results should first bereplicated with a bigger sample and with referees as participants.

Second, situation-specific training on the field could be provided to enhance decisionmaking in ambiguous situations, such as judging deceptive and non-deceptive actions ofplayers correctly. Therefore, in addition to the video-based training tools for referees [10–13]already proposed, on-field training linked with prior bodily experiences could be developedto train situation-specific decisions involving bodily information.

Certainly, other approaches exist to address the problem of enhancing decision accuracyof referees. Some approaches focus on the performance of referees; others refer to changesin the environment, such as support through technical devices or changes in rules and


regulations, in order to change player behavior. However, with regard to the current status ofrules and contexts, our study proposes to additionally take into account the bodily andpractical experiences of referees. This could help in the detection of deceptive actions, andadequate sanctions could be instituted to prevent this form of cheating in football in thefuture.

CONCLUSIONThis study extends previous research by using a motor learning paradigm to examine directeffects of bodily experiences on perceptual judgments. We conclude that domain-specificexperience, both motor and visual and both prior general and specific, could help refereesunderstand the true intentions of football players and call fouls or simulations and givesanctions accordingly, to penalize deceptive actions in football. Taking this into accountwhen developing referee coaching concepts could be a first step toward improving thestandards of refereeing.

ACKNOWLEDGMENTSWe thank Michael Denkewitz for his help in the recruitment of participants and datacollection as well as the football players who participated in this study.

REFERENCES1. Nevill, A.M., Balmer, N.J. and Williams, A.M., The Influence of Crowd Noise and Experience upon

Refereeing Decisions in Football, Psychology of Sport and Exercise, 2002, 3, 261-272.

2. Oudejans, R.R.D., Bakker, F.C., Verheijen, R., Gerrits, J.C., Steinbrückner, M. and Beek, P.J., How Positionand Motion of Expert Assistant Referees in Soccer Relate to the Quality of their Offside Judgments duringActual Match Play, International Journal of Sport Psychology, 2005, 36, 3-21.

3. Unkelbach, C. and Memmert, D., Game Management, Context Effects, and Calibration: The Case of YellowCards in Soccer, Journal of Sport and Exercise Psychology, 2008, 30(1), 95-109.

4. Jackson, R.C., Warren, S. and Abernethy, B., Anticipation Skill and Susceptibility to Deceptive Movement,Acta Psychologica, 2006, 123, 355-371.

5. Morris, P.H. and Lewis, D., Tackling Diving: The perception of Deceptive Intentions in Association Football(Soccer), Journal of Nonverbal Behavior, 2010, 34, 1-13.

6. Gilis, B., Weston, M., Helsen, W.F., Astrid, J. and Dvorak, J., Interpretation and Application of the Laws ofthe Game in Football Incidents Leading to Player Injuries, International Journal of Sport Psychology, 2006,37, 121-138.

7. Mascarenhas, D., Button, C., O’Hare, D., and Dicks, M., Physical Performance and Decision Making inAssociation Football Referees: A Naturalistic Study, The Open Sports Sciences Journal, 2009, 2, 1-9.

8. MacMahon, C., Helsen, W.F., Starkes, J.L. and Weston, M., Decision-Making Skills and Deliberate Practicein Elite Association Football Referees, Journal of Sports Sciences, 2007, 25(1), 65-78.

9. MacMahon, C. and Plessner, H., The Sport Official in Research and Practice, in: Farrow, D., Baker, J. andMacMahon, C., eds., Developing Sport Expertise: Researchers and Coaches put Theory into Practice,Routledge, London, 2008, 172-188.

10. Catteeuw, P., Gilis, B., Wagemans, J. and Helsen, W.F., Perceptual-Cognitive Skills in Offside DecisionMaking: Expertise and Training Effects, Journal of Sport and Exercise Psychology, 2010, 32, 828-844.

11. Catteeuw, P., Gilis, B., Jaspers, A., Wagemans, J. and Helsen, W.F., Training of Perceptual-Cognitive Skillsin Offside Decision Making, Journal of Sport and Exercise Psychology, 2010, 32, 845-861.

12. Schweizer, G., Plessner, H., Kahlert, D. and Brand, R., A Video-Based Training Method for ImprovingSoccer Referees’ Intuitive Decision-Making Skills, Journal of Applied Sport Psychology, 2011, 23(4), 429-442.


13. MacMahon, C., Starkes, J.L. and Deakin, J., Referee Decision Making in a Video-Based Infraction DetectionTask: Application and Training Considerations, International Journal of Sports Science and Coaching, 2007,2(3), 257-265.

14. Cañal-Bruland, R. and Schmidt, M., Response Bias in Judging Deceptive Movements, Acta Psychologica,2009, 130, 235-240.

15. Cañal-Bruland, R., van der Kamp, J. and van Kesteren, J., An Examination of Motor and PerceptualContributions to the Recognition of Deception from Others’ Actions, Human Movement Science, 2010, 29,94-102.

16. Sebanz, N. and Shiffrar, M., Detecting Deception in a Bluffing Body: The Role of Expertise, PsychonomicBulletin & Review, 2009, 16(1), 170-175.

17. Aglioti, S.M., Cesari, P., Romani, M. and Urgesi, C., Action Anticipation and Motor Resonance in EliteBasketball Players, Nature Neuroscience, 2008, 11, 1109-1116.

18. Loula, F., Prasad, S., Harper, K. and Shiffrar, M., Recognizing People From Their Movement, Journal ofExperimental Psychology: Human Perception and Performance, 2005, 31(1), 210-220.

19. Pizzera, A. and Raab, M., Perceptual Judgments of Sports Officials are Influenced by their Motor and VisualExperience, Journal of Applied Sport Psychology, 2012, 24(1), 59-72.

20. Fédération Internationale de Football Association (FIFA), The Laws of the Game, 2008, FIFA, Zürich,Switzerland.

21. Brand, R., Plessner, H. and Schweizer, G., Conceptual Considerations about the Development of a Decision-Making Training Method for Expert Soccer Referees, in: Araújo, D., Ripoll, H. and Raab, M., eds.,Perspectives on Cognition and Action in Sport, Nova Science Publishers, New York, 2009, 181-190.

22. Helsen, W.F. and Bultynck, J.-B., Physical and Perceptual-Cognitive Demands of Top-Class Refereeing inAssociation Football, Journal of Sports Sciences, 2004, 22, 179-189.

23. Starkes, J.L. and Allard, F., eds., Cognitive Issues in Motor Expertise, Elsevier Science, Amsterdam, 1993.

24. Williams, M. and Davids, K., Declarative Knowledge in Sport: A By-Product of Experience or aCharacteristic of Expertise?, Journal of Sport & Exercise Psychology, 1995, 17, 259-275.

25. Plessner, H. and Haar, T., Sports Performance Judgments from a Social Cognitive Perspective, Psychologyof Sport and Exercise, 2006, 7, 555-575.

26. Bar-Eli, M. and Raab, M., Judgment and Decision Making in Sport and Exercise, Psychology of Sport andExercise, 2006, 7(6).

27. James, S., Wanted: Players to Referee, in The Guardian, Guardian News and Media Limited, 2006, RetrievedMarch 4, 2011, from http://www.guardian.co.uk/football/2006/ feb/08/newsstory.sport7.

28. Biggs, A., Hackett Plans to Fast-track Players into Referees, in The Guardian, Guardian News and MediaLimited, 2006, Retrieved March 4, 2011, from http://www.guardian.co.uk/football/2006/jan/11/newsstory.sport4.

29. Dixon, L., Lee Dixon: How about a Weekly Panel of Ex-players, Refs and Managers to Review ‘BadTackles’?, in The Independent, Independent News and Media, 2010, Retrieved March 4, 2011, fromhttp://www.independent.co.uk/sport/football/news-and-comment/lee-dixon-how-about-a-weekly-panel-of-explayers-refs-and-managers-to-review-bad-tackles-2132721.html.

30. MacMahon, C., Starkes, J.L., and Deakin, J., Differences in Processing of Game Information in BasketballPlayers, Coaches and Referees, International Journal of Sport Psychology, 2009, 40, 403-423.

31. Casile, A. and Giese, M.A., Nonvisual Motor Training Influences Biological Motion Perception, CurrentBiology, 2006, 16, 69-74.


Does Motor or Visual Experience Enhance the Detection of Deceptive Movements in Football?

Documents