Journal of Nonverbal Behavior

, Volume 34, Issue 1, pp 1–13

Tackling Diving: The Perception of Deceptive Intentions in Association Football (Soccer)

Authors

    • Department of PsychologyUniversity of Portsmouth
  • David Lewis
    • Department of PsychologyUniversity of Portsmouth
Original Paper

DOI: 10.1007/s10919-009-0075-0

Cite this article as:
Morris, P.H. & Lewis, D. J Nonverbal Behav (2010) 34: 1. doi:10.1007/s10919-009-0075-0

Abstract

The three studies reported examine judgment about the attempts of footballers (soccer players) to deceptively exaggerate the effect of a tackle. Study one reveals that non-professional participants agree about which players were attempting deception and those that were not; there was also agreement about the tackles in which the intentions were ambiguous. Study two demonstrates that the intentions of tackled players match the judgment of their intentions by observers. Study three provides a taxonomy of behaviors that are associated with deceptive and non deceptive intentions. We conclude that deceptive intentions in this context are to a degree manifest in behavior and are observable.

Keywords

Deceptive intentionsPerception intentionsDeceptionSport perception

Introduction

For players, referees, pundits and football (soccer) fans there is often tremendous controversy about whether a player has attempted to exaggerate the effect of a tackle (i.e., ‘taking a dive’, or ‘simulation’ as the Fédération Internationale de Football Association [FIFA] name it) in order to deceive the referee into awarding them a free kick or penalty. The diving player attempts to conceal their deceptive intention from the referee. Judgments about the intention of the tackled player can have a decisive effect on a game due to a decision of the referee, the responses of other players, or even the responses of the crowd. The purpose of our studies is to explore the reliability and validity of judgments of deception by non-experts in this context. We will also describe the behaviors associated with deception and non deception.

A number of existing studies are relevant to the question of what factors influence judgments about whether a player is exaggerating the effect of a tackle or not. One set of studies has as its background the social cognition literature focusing on person perception (e.g., Frank and Gilovich 1988). The second set of studies is more technical in nature and focus mainly on the anticipation of the force and direction of the shot by an opponent during games such as badminton (Abernethy and Zawi 2007).

There is a substantial literature on the factors that influence perception of people and social events. Many of such studies have identified biases that can lead to inaccurate perception (e.g., the fundamental attribution error; gender, race and class stereotyping; implicit personality theories; target based and category based expectation effects; self serving attributions; defensive attributions; resistance to change; attributional style; false consensus bias; false uniqueness bias; just world bias; halo effect (Fiske and Taylor 1991; Nisbett and Ross 1980; Schneider 2005). However, other studies have revealed that in many contexts people are in fact good at making judgments about other people with reasonable accuracy. Vrij (2008) lists a wide range of psychological attributes (e.g., status, dominance, sexual orientation) that people judge with reasonable accuracy.

The studies examining social perception within the sporting context have in general focused on error and bias rather than accuracy. Plessner and Haar (2006) state that, “Ample anecdotal and empirical evidence indicates that sports performance judgments are at least as prone to systematic errors (biases) as other social judgments” (p. 572). For example, there have been a number of studies examining the influence of color of uniforms on judgments of aggression in sport (Frank and Gilovich 1988; Greenlees et al. 2005; Hagemann et al. 2008; Matsumoto et al. 2007); other studies have examined the effect of prior reputation (Jones et al. 2002) and the home crowd (Dawson et al. 2007) on the perception of aggression. Clothing has also been found to affect expectations of competence (Greenlees et al. 2008). There is also evidence that referees favor the home team in a number of sports (Balmer et al. 2005). Dohmen (2008), based on analysis of refereeing decisions in the Bundesliga (the highest German football league), stated that referees were more likely to award penalties and free kicks to the home side. The clear inference to be taken from the evidence reviewed, is that there may be a number of biases that could influence judgment of deceptive intentions. For example, referees could be more likely to judge home players as less deceptive than players of the away team.

There have been a number of studies that have examined the effect of different behaviors on judgment rather than the effect of incidental variables on judgment. However, such studies have tended to focus on basic aspects of perception such as judgment of weight and direction of shot rather than higher order psychological variables such as deception. Such perceptual studies focus on the ability of players to judge and sometimes anticipate the actions and intentions of another player (such as direction and weight of shot by a squash player or type of delivery by a cricket bowler) based on a fine grained analysis of the relevant behavior (e.g., Abernethy 1990; Abernethy et al. 2001; Abernethy and Zawi 2007; Farrow and Abernethy 2003; Műller et al. 2006). The results of the studies indicate that the behavior of an opponent can be used to anticipate many aspects of a future shot and that the anticipation of genuine experts (e.g., world class competitors) is superior to that of amateurs. One existing study has focused on deceptive intent. Jackson et al. (2006) found that expert rugby players were able to predict direction of change of an opponent on a video, even when the opponent tried to deceive as to the intended direction of movement. The deception involved in this example is relatively straightforward; however, the footballer exaggerating the effect of a tackle is engaged in a more complex deception.

The footballer deceptively exaggerating the effect of a tackle is attempting to deceive the viewer about the effect of the tackle; the effect of the behavior of the tackler; and to some degree the intent of the tackler. In Dennet’s (1989) terms, the rugby study is dealing with second order intentionality, where the runner is attempting to manipulate the tackler’s perception of the runner’s intentions. In the case of the tackled footballer, there is an attempt to manipulate an observer’s judgment about the physical effect of a tackle, the intentions of the tackled player and to some degree the intention of the tackler. Therefore in our study there are potentially three levels of intentionality involving the tackled player, the tackler, and the judge.

In summary, existing studies of higher order psychological variables such as judgment of aggression in sport are subject to a number of errors and biases. Studies of more basic aspects of perception in sport have demonstrated high levels of accuracy at least in experts. The studies we report combine aspects of the two literatures discussed. We examine a higher order psychological variable (i.e., the attempt of the tackled player to deceive observers) but in paradigms that are designed to explore basic issues of reliability and validity of judgments of behavior. Furthermore, we describe the behaviors associated with deceptive and non-deceptive tackles.

In the first study our central question was: do judges agree about which tackled players are ‘taking a dive’ (i.e., trying to deceive the viewer about the nature and effect of the contact with the tackling player) and which players are not ‘taking a dive’? This is an important starting point as we know nothing about how different people would judge intentions in a variety of tackles. It is important to stress in this study we are looking at between-subject consistency of judgments; we make no claims that a high level of consistency is in itself evidence of the validity of such judgments. We merely observe that validity in the absence of reliability may be less likely. In the second study, footballers were instructed to deceptively exaggerate the effect of a tackle to varying degrees; participants were then asked to judge the degree of exaggeration. We could thus relate the actual intention of the tackled player to the perception of that event. In the final study we establish a taxonomy of behaviors associated with deception.

Study 1

Method

Participants

An opportunity sample of 315 university students was employed. Some took part as partial fulfilment of a course requirement, others were paid or volunteered.

Design

Groups of 15 participants were randomly assigned to judge a single video clip of a tackle out of 21 such clips. Participants selected one of three classifications for the behavior of the tackled player (1) the player attempted to deceive by exaggerating the effect of the tackle (2) not sure of the intentions of the tackled player (3) the player did not exaggerate the effect of the tackle and did not attempt deception. All clips were presented in black and white rather than their original color with no sound. Monochrome presentation was used to eliminate color of shirts as a source of variation.

Selection of Stimuli

Twenty-one football clips were selected from recordings of televised live games (as opposed to edited highlights) from English professional football from a variety of sources (e.g., Sky, ITV, BBC).1 The experimenters selected a number of foul tackles (7 out of the 21 presented) on the basis that the experimenters judged the behavior of the tackled player as intentionally deceptive. The remainder of the fouls were selected unsystematically.

In all tackles the referee awarded the free kick to the tackled players’ team. The selection criteria included that the referee should have given a free kick for foul play; based on the video evidence actual contact appeared to have been made between the tackling and tackled player; furthermore there was a clear view of the tackle. The mean duration of all clips was 3.94 second (SD = .54).

Apparatus

All editing used the Adobe Premier Pro system digital video editing system. Presentation of stimuli was on a standard computer monitor.

Procedure

Participants were asked to view a single video clip of a tackle and classify the behavior of the player into one of three predefined categories. Each clip was presented twice with a pause of 2-s between each presentation.

Results

Reliability was assessed using Fleiss’ kappa (Fleiss 1971). Participants demonstrated a statistically significant level of agreement with a moderate to large effect size, kappa = .56,p < .01. ‘Not sure ‘was the most frequently used category (nine out of the 21 clips), ‘deception’ was the next most frequent category (eight clips) and ‘no deception’ was the least frequently used category (four clips) (see Table 1 for data).
Table 1

Categorization of intentions for each tackle

Tackle number

Category of tackle

Deception

Not sure

No deception

1

1

13

1

2

1

3

11

3

1

2

12

4

2

12

1

5

14

1

0

6

0

3

12

7

4

8

3

8

3

11

1

9

14

1

0

10

8

7

0

11

1

11

3

12

12

3

0

13

12

3

0

14

0

10

5

15

0

8

7

16

12

2

1

17

0

13

2

18

0

12

3

19

13

2

0

20

0

2

13

21

13

2

0

Note: Tackles in bold judged by experimenters as potentially deceptive tackles

Discussion

There was widespread agreement by participants in their classification of the tackles. It is noteworthy that for a number of clips ‘not sure’ was the most frequently used category indicating that participants were making a distinction between clips where deceptive intentions (or not) were clear and clips where intentions were not clear. The frequent use of the ‘not sure’ category probably accounts for the controversy surrounding whether a player has attempted deception or not. We suspect that it is in this category of tackles that other factors, such as allegiance to a particular team, or the effect of home advantage may have the most significant role in determining judgment about the event.

Reliability does not of course imply that the classification is necessarily accurate; we can only conclude that participants have a similar perception of the action. However, a reasonable level of reliability is a prerequisite of validity.

In the second study we investigated the issue of validity. Our participants may have been entirely mistaken about the actual intentions of the tackled players in the video clips. We asked players to deceptively exaggerate (i.e., as to deceive a referee about the effect of the tackle) the effects of a tackle to varying degrees. Participants then viewed videos of the tackles and were asked to judge the level of exaggeration (if any) of the tackled player. Thus, we could establish whether the judgment of the participants was related to the intent of the tackled player. Furthermore, we examined the effect of knowledge of football on the consistency of the classification of deceptive intent.

Study 2

Method

Participants

Forty-five undergraduates took part in the study in partial fulfillment of a course requirement. Participants self rated their level of football knowledge. Three groups were produced on the basis of this self rating, an extensive knowledge group, a moderate knowledge group and a little knowledge group. However, it should be noted that these groupings are relative to the knowledge of the entire group of participants, and not one of the participants could be considered expert.

Creation of the Scenarios Used as Stimuli

The scenario employed to create the stimuli was taken from a Football Association coaching manual (Cook 2004) in which an attacker practices dribbling past an approaching defender. One player dribbles with the ball and attempts to get past a defender in a channel 10.97 m long by 16.46 m wide. Players wore a standard kit of white football shirt, black shorts, black socks, shin pads, and boots. The defender was instructed to tackle the attacker using a standard block tackle. The attacker was instructed to dribble the ball around the defender without being tackled by the defender. However, if the defender did manage to tackle the attacker, he was instructed to exaggerate the effect of the tackle in a convincing manner. We used five levels of exaggeration 0%, 25%, 50%, 75%, 100%, where 0% was no exaggeration, and 100% was taken to mean the maximum possible exaggeration that could reasonably be expected in a real match. These are of course not absolute quantitative measures but we found it the easiest way to communicate to our players what we wanted. The players used were experienced amateur players. In each sequence the same attacking player and the same defending player were used. The five tackles were filmed in the same location on the same day to produce the five video clips (Mean length = 4.61 seconds, SD = 0.48).

Design

A 5 × 3 mixed factorial design was used. The repeated measures factor was the degree of exaggeration of the tackle (5 levels, 0%, 25%, 50%, 75%, 100%); the independent groups factor was level of football knowledge (3 levels, none at all, some knowledge, extensive knowledge). Participants viewed the five clips in randomized orders. The dependent variable was a score on a Likert scale (1 no exaggeration––5 extreme exaggeration).

Procedure

Participants viewed each clip twice before indicating the perceived level of exaggeration on the Likert scale. Participants were tested individually.

Results

An alpha level of .05 was employed in all statistical tests. A 5 (level of exaggeration) × 3 (level of football knowledge) mixed univariate ANOVA was conducted on the data. There was a clear increase in the perceived level of exaggeration as a function of the intended exaggeration in the clips (see Table 2), F(3.01, 126.53) = 18.49, p < .001, η2 = .31 (degrees of freedom adjusted for degree of violation of sphericity using Greenhouse Geisser correction (Greenhouse and Geisser 1959). The increase in perceived exaggeration was also a significant linear trend, F(1, 42) = 95.86, p < 001, η2 = .69.
Table 2

Ratings of judged deceptive exaggeration as function of tackle

 

M

SD

Clip One

2.24a

0.96

Clip Two

2.49ab

1.10

Clip Three

2.80b

1.04

Clip Four

3.16b

1.02

Clip Five

3.73c

0.89

Note: Only mean scores with a different subscript differ significantly from each other (p < .05, Bonferroni adjusted)

There was no effect of level of football knowledge, F(2, 42) = 0.73, ns, and no interaction between level of football knowledge and degree of exaggeration, F(6.03, 126.65) = 0.95, ns.

Discussion

There was a very powerful and consistent relationship between the intentions of the tackled player and the evaluations of those intentions by the participants. The findings do lend support to the view that the judgment of deceptive intent in this study are in fact valid. Furthermore, this is evidence for the argument that the judgments in the previous studies may also be valid. This study provides evidence of ‘ground truth’ comparable to that found in studies of lie detection in forensic psychology (Porter and Yuille 1996; Tye et al. 1999). We know prior to the study about what is going on in each video clip.

We did not observe an effect of knowledge in the results. This contradicts a well established finding in the literature that expertise has a powerful effect on judgment of intention in sport (Abernethy and Zawi 2007). However, such studies have often compared amateur players with world class players. We suggest that the discrepancy between our findings and previous findings is that our groups of participants did not include a group of experts comparable to those found in other studies. Furthermore, we classified participants according to their self reported knowledge of football which may be a poor indicator of perceptual-motor proficiency.

In the two previous studies evidence has been gathered that suggests a degree of reliability and validity in the judgment of deceptive intentions. However, there has been no exploration of the behaviors associated with deceptive and non-deceptive intentions. The previous studies provide only a limited sample of tackles that involve deceptive and non-deceptive intent. The purpose of the final study is to develop a taxonomy of deceptive and non-deceptive behavior. The format of the study involved behavioral coding of a large number of tackles (forty) in which there was a clear deceptive intent and tackles where there was clearly no intention to deceive.

Study 3

Method

Selection of Stimuli

Twenty tackles were selected from the same sources used in study one (tackles used in study one were not reused in the new study), however, the criterion employed was that the tackles were clear examples of attempted deception; a further twenty tackles were selected using the criterion that they were clear examples of no attempt at deception. There was perfect agreement between two raters on the classification of the tackles into deceptive and non-deceptive categories (the perfect level of agreement is not surprising given the selection criteria).

Coding of Behaviors

A coding scheme was developed that involved a number of classes of behavior. One set of categories referred to behaviors or acts that were present or absent (temporal contiguity; ballistic continuity; contact consistency) but were not related to a particular or limited set of limb, torso, and head movements. The remaining code related to a specific combination of limb, torso, and head movements associated with deception (this combination of limb movements we call the ‘archer’s bow’).

Category One: Lack of Temporal Contiguity (Or Not)

This category is defined by a time lag between the contact of the players in the tackle and the effect on the tackled player. Example: A player was tackled but takes another two strides where he is fully in control before falling over.

Category Two: Lack of Ballistic Continuity (Or Not)

This category is defined by whether the effect of the tackle only involves the momentum of the player or is the movement supplemented by the tackled player. Example: the players was tackled and his momentum caused him to roll as he hits the ground, however, the player engaged in additional rolls that were not part of the momentum of the fall.

Category Three: Lack of Contact Consistency (Or Not)

This category is defined by the absence of a relationship between the point of impact between the players and the point of injury indicated by the tackled player.

Example: the player was hit by an elbow in the chest but falls to the ground clutching his face.

Category Four: The ‘Archers Bow’, Limb, Torso and Head Movements Associated with Deception

The three categories reported previous to the ‘archer’s bow’ are similar in that the behaviors (in terms of limb, torso and head movements) observed were not in themselves indicative of deception; deception was manifest in how the behaviors integrate into the interaction between the tackled player and the tackler. In contrast, in the new category the behaviors observed were unique to deception. We name the unique behavior the ‘archer’s bow’ because in its most complete form the tackled player resembles a drawn bow: the chest is thrust out; the head is back; the arms are fully raised and pointing upwards and back; the legs are raised off the ground and bent at the knee (see Fig. 1 for an actual example). However, less complete forms of this behavior were also observed in which the tackled player used a subset of the head, torso, arm and leg movements.
https://static-content.springer.com/image/art%3A10.1007%2Fs10919-009-0075-0/MediaObjects/10919_2009_75_Fig1_HTML.jpg
Fig. 1

An example of the full ‘archer’s bow’

Reliability

As the coding scheme did not produce mutually exclusive categories we conducted four separate reliability analyses. Each tackle was parsed for lack of temporal contiguity or not, lack of ballistic continuity or not, lack of contact consistency or not and whether there were any elements of the ‘archer’s bow’. Cohen’s kappa values of above .8 were observed for the reliability of the coding of each behavioral category.

Results

The frequency of occurrence of temporal contiguity, ballistic continuity and contact consistency across deceptive and non-deceptive tackles are presented in Table 3. It is clear that deceptive behaviors occur far more frequently in the deceptive tackles than non-deceptive tackles. We could not conduct inferential tests (due to the lack of variance) comparing the frequency of occurrence of the classes of behavior between deceptive and non-deceptive tackles as the discrimination was almost perfect.
Table 3

Frequency of occurrence of classes of deceptive behavior in the twenty deceptive and twenty non-deceptive tackles

 

Deceptive*

Non-deceptive

Temporal contiguity

14

0

Ballistic continuity

15

0

Contact consistency

8

0

Note: The column totals sum to more than 20 because more than one class of deception can occur in a single tackle

With reference to the archer’s bow motion, in the nondeceptive tackles, only one scene contained even a single element of the motion; however, among the 20 deceptive tackles, 5 scenes contained two elements; 1 scene contained 3 elements; and 5 scenes contained all four elements of the archer’s bow.

We emphasize that the four behaviors are not mutually exclusive and more than one type of deceptive behavior can be observed in a single tackle (in two tackles 1 class of behavior was observed, in eleven tackles 2 classes of behavior were observed, in four tackles 3 classes of behavior were observed and in three tackles all 4 classes of behavior were observed).

Discussion

A number of categories of behavior associated with deception were observed. Three of the four categories (temporal contiguity, ballistic continuity, and contact consistency) involved behaviors that in themselves are not necessarily associated with deception. The behaviors are categorized as deceptive because they do not fit the behavioral logic of the interaction between the tackler and the tackled player. If the pattern of the interaction were different, then the behaviors in terms of the head, torso, arm and leg movements could be associated with non-deception. Whereas, a set of movements we term the ‘archer’s bow’ were associated with deception and were not observed in non-deception (except for one element [arms raised] in one of the non-deceptive tackles). On the basis of the results we conclude that there are classes of behavior that are regularly associated with deception in this context.

We make no claim that the frequency of occurrence of behaviors we report in the deceptive tackles represent the frequencies that occur in all deceptive tackles; our sampling procedures do not warrant such a claim. We would claim merely that the behaviors we describe are likely to be present in a proportion of deceptive tackles.

Many of the deceptive behaviors we coded were easy to observe. Furthermore, the coding scheme we use does not give a picture of the quality of some of the deceptive behaviors involved. Our impression is that some of the behaviors are performed in an almost theatrical manner. It is legitimate to ask why players seem such poor deceivers? We explain poor deception as a result of the need of the tackled player to ensure that an observer can clearly see that a tackle has had a substantial effect. The tackled player has a difficult challenge. If the tackled player is to win a free kick he needs to demonstrate that the tackle has had an effect, but he also needs to conceal the fact that the effect is of his own manufacture. Some of the players used all four types of categories of deceptive behavior in a single tackle. Such extreme attempts at deception may be rewarded on occasion because the referee may be some distance away, the tackle is happening at high speed and the referee does not have the benefit of viewing the challenge again. Nevertheless, there is a balance to be struck by the player as under FIFA rules simulation is an offense.

General Discussion

In the first study participants were consistent in their judgments about the deceptive intentions (or not) of tackled players. Furthermore, there was agreement about tackles in which it was difficult to make a clear judgment about the intent of the tackled player. The second study provided evidence of the validity of such judgments as there was a very powerful and consistent relationship between the intentions of a tackled player and judgment of those intentions by viewers. The third study revealed categories of behavior that were consistently associated with deception.

How do we reconcile our findings that there is a high degree of reliability and even validity, with the results of previous studies that have found significant effects of a variety of biases (Plessner and Haar 2006) on judgment? We do not doubt that biases, such as the color of shirts or reputation of a team or particular player, could have an influence on the judgment of intention. However, we suggest that some core information about intentions is specified in behavior and is observable. We further suggest that bias may play an enhanced role in those tackles which unbiased observers agree cannot be easily classified as deceptive or non-deceptive. The results of the first study reveal that there may be a substantial number of such tackles. The frequency of such difficult to classify tackles may explain why there is on occasion such argument about whether a tackle contains deception or not. In summary, we think that our results are compatible with the previous literature discussed. We suspect that a number of sources of bias could be found to account for a measurable amount of variance in judgment in this context (particularly in tackles where there may be ambiguity), however, our results indicate that behavior itself may also be an important source of variance. Future studies could of course address the issue of the relative contribution of behavior and incidental variables.

We did not find expertise effects in study two which does not accord with the findings of previous research. We explain our failure to find expertise effects as the result of our failure to use genuine experts. Some previous studies have used experts who were of world class in their sport (Abernethy and Zawi 2007). All of our participants would have been considered non-experts in the relevant previous studies. Therefore we should emphasise that our study revealed only that amateurs with different levels of expertise did not differ in performance. Genuine expertise would of course be an interesting factor to explore in future research. Experts may for example be able to increase the number of tackles that could be classified as either definite deception or definite non deception.

In terms of how we explain the perceptual availability of deceptive behavior we suggest that the work of Michotte (1946/1963) provides a coherent theoretical account. Michotte claimed that there is high sensitivity to violations of the laws of causation in the material world. Many subsequent studies have demonstrated such sensitivity in adult humans (Scholl and Tremoulet 2000) and infant humans (Leslie and Keeble 1987; Schlottman and Surian 1999). It has even been demonstrated that pigeons can successfully learn a discrimination task involving causal and non-causal collisions (Young et al. 2006). We suggest that for a number of the categories of behavior (e.g. lack of temporal contiguity and lack of ballistic continuity) participants are detecting the fact that the tackled player is imposing behavior on top of the effect of a simple collision. The intentions are manifest in the violations of the laws of cause and effect. The detection of deception in this context may be a relatively low level perceptual process. Our results of course do not discriminate between such an explanation and more conventional accounts of the judgment of intentions (Dassser et al. 1989). We do suggest that the former explanation is probably the more parsimonious.

The origin of the set of behaviors we name the ‘archer’s bow’ is to a degree puzzling. The other behaviors we observed are variations on what is observed in non-deceptive tackles. However, the ‘archer’s bow’ is unique to deception. We suggest that the ‘archer’s bow’ does serve one important function for the tackled player in that the behavior is clearly noticeable. But it is difficult to conceive of a collision in any context where the resultant impact on the body would be the ‘archer’s bow’. Most football tackles involve primary contact below the waist so conservation of momentum would involve the body going forward over the legs, the head does not go back but in fact stays forward in order to spot the landing, and arms go and out and or down to steady the body and/or prepare for impact. The image that the ‘archer’s bow’ most strongly resembles is an exaggerated form of Frank Capa’s famous photograph of the death of a loyalist soldier, although it surprising that the behavior depicted in this picture is less extreme than many of the behaviors observed in tackled players. We speculate that the ‘archer’s bow’ is used by the player to convey the extreme nature of the collision; the collision is so extreme that all the normal self-protection mechanisms involved with preparing for the fall cannot be utilized.

The findings could have an application in professional football. Players can appeal against disciplinary sanctions applied by the referee during a match (i.e., a booking or a sending off) on the grounds that the referee made a mistake. A tackler who had been disciplined for a serious foul challenge could appeal because he could demonstrate that the tackled player had exaggerated the effect of a tackle, and a tackled player who had been disciplined for simulation could demonstrate that he had not been attempting to use deception. Such appeals would carry more weight if an objective set of criteria for deception could be developed. At a theoretical level we suggest that the findings are of interest because they reveal indications of a lawful relationship between behavior and intentions. Wilkerson (1999) argues that psychologists have focused on intentions as internal mental events and have neglected how they are manifest in behavior. We suggest that future studies address the problems of establishing a precise linkage between judgment of deception and the critical frames in video clips, and how incidental factors previously shown to be important, such as color of clothing, interact with behavior to affect judgment.

Footnotes
1

We selected games from the lower leagues and used old recordings to avoid the participants knowing the particular footballers involved. We did not allow participants to watch clips of their ‘own’ team and we did not use important league games or cup semi-final or finals. No participants reported a particular interest in any of the games or players viewed.

 

Copyright information

© Springer Science+Business Media, LLC 2009