Citation
Quiet eye duration as an index of cognitive processing

Material Information

Title:
Quiet eye duration as an index of cognitive processing the effect of task complexity and task duration on visual search patterns and performance in highly-skilled and lesser-skilled billiards players
Creator:
Frehlich, Shane G., 1968-
Publication Date:
Language:
English
Physical Description:
ix, 169 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Billiards ( jstor )
Cognitive psychology ( jstor )
Eye movements ( jstor )
Eyes ( jstor )
Mathematical dependent variables ( jstor )
Optical tracking ( jstor )
Performing artists ( jstor )
Search strategies ( jstor )
Task complexity ( jstor )
Visual fixation ( jstor )
Dissertations, Academic -- Health and Human Performance -- UF ( lcsh )
Health and Human Performance thesis, Ph.D ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1997.
Bibliography:
Includes bibliographical references (leaves 152-168).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Shane G. Frehlich.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
028627623 ( ALEPH )
38868661 ( OCLC )

Downloads

This item has the following downloads:


Full Text













QUIET EYE DURATION AS AN INDEX OF COGNITIVE PROCESSING: THE EFFECT OF TASK COMPLEXITY AND TASK DURATION ON VISUAL SEARCH PATTERNS AND PERFORMANCE IN HIGHLY-SKILLED
AND LESSER-SKILLED BILLIARDS PLAYERS








By

SHANE G. FREHLICH


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1997











ACKNOWLEDGMENTS


I would like to express my deepest of appreciations to the many individuals who

helped me in the development of this manuscript over the course of the past five years. To my committee chair, Dr. Robert Singer, I wish to heartily thank you for your mentoring efforts in helping me achieve my goals. From you, I have learned to appreciate and respect the values of commitment, desire, and dedication. The spirit and enthusiasm for life that you possess has provided a greater inspiration to me than you can ever know, and I hope to pass these values on to my own students someday. I would also like to extend my sincere gratitude to Dr. James Cauraugh for teaching me the joys of the scientific process, to Dr. Milledge Murphey for keeping my views grounded in the applied perspective, and to Dr. Ira Fischler for the many cognitive insights. You have all guided me towards the fulfillment of a significant goal in my life, and words cannot express how important each of you are to me. A special debt of gratitude is owed to Dr. Mark Williams, who graciously provided much of the equipment used in this study. I would also like to thank Doug and Laura Barba, Chris and Carol Janelle, Suzanne Broch, Lisa Pugliese, Paul Hubman, Sean Walsh, and Hjalmer Setzer for giving me all the love, help, and friendship anyone could ask for. Finally, my deepest and most heartfelt thanks go out to my family, Gary, Patricia, and Michelle Frehlich. Everything I have ever achieved in my life has been due to your unending love and support, and for that I will always be grateful.


ii
















TABLE OF CONTENTS




ACKNOW LEDGM ENTS ................................................................................... ii

LIST OF TABLES .............................................................................................. v

LIST OF FIGURES ............................................................................................. vii

ABSTRACT .................................................................................................... viii

CHAPTERS

1. INTRODUCTION ............................................................................... 1

Eye M ovement Research in Sport.............................................................. 5
Two Critical Questions Arise ................................................................... 7
Theoretical Approaches Emerge ............................................................. 10
Statement of the Problem ........................................................................ 13
Hypotheses .............................................................................................. 14
Assumptions ............................................................................................ 22
Limitations .............................................................................................. 23
Definition of Terms ................................................................................. 24
Significance of Study .............................................................................. 28

2. REVIEW OF LITERATURE ................................................................... 32

Expertise in Sport ................................................................................... 33
The Visual Search Paradigm in Sport
Research ............................................................................................ 38
Theories of Oculomotor Control in
Aiming Tasks ..................................................................................... 51
The Influence of Task Complexity ........................................................... 58
The Influence of Task Duration ............................................................... 60






iii











3. M ETHODS .............................................................................................. 64

Experiment 1 ......................................................................................... 64
Participants ............................................................................................ 64
Apparatus .............................................................................................. 66
Procedures .............................................................................................. 68
Dependent M easures ............................................................................... 76
Experiment 2 ......................................................................................... 78
Participants ............................................................................................ 79
Apparatus .............................................................................................. 79
Procedures .............................................................................................. 80
Dependent M easures ............................................................................... 82

4. RESULTS .............................................................................................. 83

Experiment 1 ......................................................................................... 83
Statistical Analyses ................................................................................. 83
Experiment 2 ......................................................................................... 98
Statistical Analyses ................................................................................. 98

5. DISCUSSION, SUMMARY, CONCLUSIONS, AND IMPLICATIONS
FOR FUTURE RESEARCH ............................................................... 113

Experiment 1 .......................................................................................... 114
Experiment 2 .......................................................................................... 132
Summ ary ................................................................................................ 142
Conclusions ............................................................................................ 145
Implications for Future Research ............................................................ 147

REFERENCES .................................................................................................. 152

BIOGRAPHICAL SKETCH .............................................................................. 169


iv











LIST OF TABLES


Table pUage

4-1 Performance means and standard deviations for highly-skilled and lesser-skilled participants for each level of complexity ....................... 84

4-2 Total duration of the billiards stroke for highly-skilled and lesser-skilled participants for each level of complexity ....................... 86

4-3 Total duration for each phase of the billiards stroke for highly-skilled and lesser-skilled participants collapsed across complexity level ........... 88

4-4 Mean number of fixations and average fixation durations for highly-skilled and lesser-skilled participants for the EC condition..... 90

4-5 Mean number of fixations and average fixation durations for highly-skilled and lesser-skilled participants for the IC condition........ 92

4-6 Mean number of fixations and average fixation durations for highly-skilled and lesser-skilled participants for the HC condition..... 95

4-7 Mean quiet eye durations for successful and unsuccessful shots for both highly-skilled and lesser-skilled participants for each level
of com plexity .................................................................................. 97

4-8 Performance means and standard deviations for highly-skilled and lesser-skilled participants for each duration level ............................... 99

4-9 Total duration of the billiards stroke for highly-skilled and lesser-skilled participants for each duration level ................................ 101

4-10 Mean number of fixations and average fixation durations for highly-skilled and lesser-skilled participants for the
25% constrained condition ................................................................. 106


v











4-11 Mean number of fixations and average fixation durations for
highly-skilled and lesser-skilled participants for the
50% constrained condition ................................................................. 109

4-12 Mean quiet eye durations for successful and unsuccessful shots for
both highly-skilled and lesser-skilled participants for each
duration level ..................................................................................... 111


vi











LIST OF FIGURES


Figure page

3-1 Initial perform ance test ........................................................................... 70

3-2 Required shot in the easy complexity condition ....................................... 73

3-3 Required shot in the intermediate complexity condition .......................... 74

3-4 Required shot in the hard complexity condition ....................................... 75

5-1 Quiet eye durations for each group and level of task complexity ............... 126

5-2 Quiet eye durations for each group and level of task duration ................... 139


vii









Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

QUIET EYE DURATION AS AN INDEX OF COGNITIVE PROCESSING: THE
EFFECT OF TASK COMPLEXITY AND TASK DURATION ON VISUAL
SEARCH PATTERNS AND PERFORMANCE IN HIGHLY-SKILLED
AND LESSER-SKILLED BILLIARDS PLAYERS By

Shane G. Frehlich

December 1997

Chairman: Robert N. Singer, Ph.D.
Major Department: Health and Human Performance


The role of visual control mechanisms in the production of manual aiming tasks has been examined recently in a variety of motor tasks. Using an expert-novice research paradigm, researchers have investigated the extent to which performers of different skill levels differed in the visual search strategy used to execute a response in a particular sport situation. The purpose of this study was to investigate potential expertise differences in visual search strategy profile between twelve highly-skilled (M = 9.1 years of playing experience) and twelve lesser-skilled (M = 2.67 years) billiards players while they performed a series of strokes in an ecologically valid environment. Specifically, the dependent measures in this study included performance outcome percentage, stroke and phase duration, number of fixations, average fixation duration, quiet eye duration, and number of eye blinks. The latter two variables were proposed to play key roles in the control of visual attention in Vickers' (1996) location-suppression hypothesis. A second


viii









purpose of this study was to determine whether quiet eye duration and number of eye blinks were related to performance between groups, and within participants of each skill level.

In Experiment 1, shots of three different levels of complexity (easy, intermediate, and hard difficulty levels) were presented to the performers to determine whether accompanying changes in visual search patterns would occur. In Experiment 2, the participants executed shots under three different time constrained conditions (unconstrained, 25% constrained, and 50% constrained). Results of both studies indicated that the highly-skilled players possessed visual search strategies that were more efficient, as they made fewer fixations of longer duration to their target while lesser-skilled players made significantly more fixations of shorter duration to their target.

Only partial support was provided for Vickers' (1996) location-suppression

hypothesis. The quiet eye variable was found to be significantly related to performance between and within the two groups, with the highly-skilled players demonstrating significantly greater durations than the lesser-skilled players throughout every condition of task complexity and task duration. For both groups, successful shots were characterized by significantly longer quiet eye durations than unsuccessful shots. The number of eye blinks did not differ between the groups. Thus, only the location aspect of the locationsuppression hypothesis was supported, and it was.argued that the quiet eye duration represented a critical period of cognitive programming in the aiming response.


ix















CHAPTER 1
INTRODUCTION


Due to its highly complex and often unpredictable nature, the sporting arena provides an excellent context in which to test the assumptions of the information processing model of human behavior. For example, in fast ball sports such as baseball and tennis, athletes often have as little as 200-300 ms in which to make decisions based on the intended direction and speed of the approaching ball (Hyllegard, 1991; Slater-Hammel & Stumpner, 1950, 1951), as well as to organize and execute the most appropriate movement to effectively coincide with the object in motion. In addition, such sports as billiards, darts, archery, and pistol shooting require the ability to program exquisitely precise aiming movements at the neural level (Vickers, 1996). In fact, errors of less than one millimeter can often be the difference between success and failure in these motor activities.

In each of these sports, the processing of visual information has been shown to be a key ingredient to the successful execution of appropriate motor responses. In fact, it has been argued that at least 80% of the information an athlete receives and attends to in the sport environment is visual in nature (Schmidt, 1988). Indeed, the importance of visual


1









2


information, and its role in the human information processing system, has been the focus of a large body of research in cognitive psychology.

In recent years, scholars have attempted to determine the extent to which eye

movement data are indicative of underlying cognitive processing (e.g., Just & Carpenter, 1976; Neisser, 1967; Posner, 1980; Viviani, 1990; Wright & Ward, 1994; Yarbus, 1967). Typically, the process of inductive reasoning has been used, such that stages of cognitive processing have been inferred to exist based on the observed pattern of eye movements while the individual engages in a task influenced by events or stimuli present in the visual environment.

For example, early investigations of reading tasks in which a moving windows paradigm was used demonstrated decrements in word recognition and comprehension performance when the window moved at faster speeds (e.g., Just & Carpenter, 1976). This seemed to suggest that time constrained covert stages of information processing, such as stimulus identification, could be marked or identified by the overt eye movements of the individual. Thus, early research appeared to confirm the notion that higher-order cognitive processes, in essence, controlled the location and duration of ocular fixations. To incorporate an analogy forwarded by Viviani (1990), the eyes were believed to be similar to television cameras in that they were "directed" to the important aspects of the visual environment by specific stages of cognitive processing. In fact, much of the research









3


in the study of eye movements has, either implicitly or explicitly, adhered to this belief in a strong "eye-mind" connection.

However, the extent to which eye movements are in fact related to more global cognitive processing (a strong eye-mind view) has been questioned by a number of theorists, with several lines of research indicating a significant degree of independence between eye movement data and cognitive activities (e.g., Fisher, Karsh, Breitenbach, & Barnette, 1983; Sigman & Coles, 1980; Viviani, 1990). For example, Viviani (1990) provided a cogent argument against what he termed the "central dogma" of eye movement research, which states that "eye movements can at the very least be considered as tags or experimentally accessible quantities that scientists can observe to understand underlying processes of cognition" (p. 354).

Coupled with the central dogma are two underlying assumptions. First, visual

scanning patterns are believed to be the overt manifestation of strictly serial or sequential cognitive processes, such that subsequent visual stimuli presented to the individual cannot be adequately attended to until previously presented stimuli have been processed. Second, it is assumed that the direction of an individual's line of sight coincides with the direction and allocation of attentional resources (Viviani, 1990).

In reviews of literature in which eye movement processes have been investigated, both Viviani (1990) and Theeuwes (1993) provide evidence suggesting that the role of









4

cognition may not be so clearly inferred from eye movement data. Numerous investigations have refuted the assumptions underlying the central dogma. For example, the area of the visual field where focal attention is concentrated may occasionally be dissociated from the foveal field (Klein, 1980, 1994). Under appropriate conditions of precuing, it appears as though visual attention can be directed almost anywhere in the visual field, irrespective of the actual direction of the line of sight (Posner, 1980; Posner & Cohen, 1980; Remington, 1980; Shaw, 1983). In addition, the serial nature of visual tracking has been challenged by more recent theories of cognition (such as Parallel Distributed Processing models), which claim that a number of concomitant visual processes occur in parallel, and are hierarchically coordinated (Marr, 1982; Rumelhardt & McClelland, 1986). Thus, the results of several eye movement studies may be more readily explained with a bottom-up or reflexive approach to behavior, rather than providing evidence for underlying cognitive structures (Theeuwes, 1993).

Although a number of studies have not provided explicit support for the central dogma, Viviani (1990) conceded that the observation of expert-novice differences in the scanning of a particular visual field does serve as evidence that some degree of cognitive processing is associated with specific patterns of eye movements. In a variety of cognitive and motor tasks, skilled experts have been observed to immediately concentrate fixations on pertinent details of a relevant image, whereas novices scan the image in a random, uniform manner. This result has been documented in skilled versus non-skilled chess









5


players (de Groot, 1978), automobile drivers (Mourant & Rockwell, 1972), and radiologists (Kundel & Nodine, 1983).

Taken together, the results of these studies tend to support the notion that eye

movement patterns may indeed reflect specific stages of cognitive processing that allocate attentional resources to particular visual stimuli. Given the dynamic and often timeconstrained nature of athletic competition in which attentional resources are frequently strained to their limits (Abernethy, 1993; Nideffer, 1993; Rotella & Lerner, 1993), it was only a matter of time before investigators began to conduct eye movement studies in sport.

Eye Movement Research in Sport

Implicitly adhering to the strong "eye-mind" connection, in which eye movements were interpreted to reflect higher order mental processing (e.g., Gardner, 1985; Just & Carpenter, 1976, 1980), researchers interested in the selective attention of athletes of differing skill levels began to focus on the visual search patterns of performers in sport task situations in the mid-1970's (e.g., Bard & Fleury, 1976, 1981; Bard, Fleury, & Carriere, 1975; Haase & Mayer, 1978). Specifically, these scholars have attempted to outline the gaze behaviors and visual tracking patterns that differentiate experts and novices in their ability to perceive and use vital information in the sport environment (e.g., Abernethy, 1990b, 1991; Goulet, Bard, & Fleury, 1989; Helsen & Pauwels, 1990; Ripoll,









6


1984; Shank & Haywood, 1987; Singer, Cauraugh, Chen, Steinberg, & Frehlich, 1996; Vickers, 1992; Williams, Davids, Burwitz, & Williams, 1994).

In the majority of these studies, experts as compared to novices made fewer

fixations to achieve successful response outcomes and exhibited lower search rates for sport-specific tasks. Also, experts made a greater number of fixations to the pertinent cues in the visual array than did novices, and had search rates that were typically more systematic and consistent (Abernethy, 1988). From these results, it would appear that successful performance in the highly-skilled was determined in part by where they looked in the sport environment. That is, experts seemed to possess a more efficient process of visual search, such that they attended to only the most important aspects of the sport situation while ignoring irrelevant stimuli.

Unfortunately, while a preponderance of the work described above supported the contention that visual search patterns adequately discriminate levels of expertise, more contemporary researchers have questioned the utility of this paradigm in sport research. The reason for this skepticism is quite straightforward; the expected pattern of expertnovice differences has not been documented in several recent studies, thus raising the question of whether visual behavior is indeed a critical factor in expert performance.

In other words, similar patterns of ocular fixations have been observed to be

directed to particular aspects of the visual display by both elite and non-elite athletes, such that the caliber of skill possessed by the performer could not be adequately discriminated









7


solely on the basis of where he or she looked in the sport environment (Abernethy, 1990a, 1991; Abernethy & Russell, 1987b). In addition, scholars have argued that statistically significant differences in fixation patterns observed between highly-skilled and lesserskilled athletes are likely of little practical value in explaining the differences in decisionmaking time between the two groups (Singer et al., 1996).

Two Critical Ouestions Arise

Given the equivocal findings that have been observed in comparing visual search patterns of expert and novice athletes over the past two decades, it appears as though research in which this paradigm has been used has come to a critical juncture in its development. In particular, two fundamental questions must be addressed for this paradigm to be of utility in describing and explaining athletic performance (Abernethy, 1991; Vickers, 1996). First, exactly why have some researchers reported expertise differences in visual search patterns, while others have not? And, perhaps more importantly, are visual search processes related in some manner to expert performance, or are they of no consequence at all?

The Need For Ecological Validity

In addressing the first question, one possible explanation for the lack of consensus lies in the nature of the task performed by the individual under investigation. Both Ripoln (1991) and Vickers (1996) have noted that those investigations reporting no expertise









8


differences in visual search rates typically require athletes to passively view photographic or videotaped displays (usually in a comfortable, seated position), and to make contrived motoric responses to these stimuli (e.g., pressing a button with the index finger; Abernethy, 1990b, 1991; Abernethy & Russell, 1987b; Singer et al., 1996).

In contrast, studies that require athletes to physically move around in their environment in a manner similar to that encountered in their sport while their gaze behaviors are recorded have demonstrated significant expert-novice differences in ocular fixation location and duration measures (Helsen & Pauwels, 1993; Ripoll, Bard, & Paillard, 1986; Ripoll, Papin, Guezennec, Verdy, & Philip, 1985; Vickers, 1992, 1996). Similarly, those experiments that have been conducted in real world settings, as opposed to using videotaped simulations, have also exhibited significant differences in search rates between highly-skilled and lesser-skilled performers (Ripoll et al., 1986; Vickers, 1992). In addition, intra-expert differences have also been demonstrated using this protocol (Vickers, 1996).

While the distinction between real-world and artificial experimental settings has been targeted as a possible explanation for equivocal results in visual search research, the question remains: Why does the setting make a difference to the outcome of the study? One answer to this question comes from an ecological approach to the study of human behavior, in which perception-action theorists argue that behavior can only be understood by the kinematic coupling between the human biomechanical system and information from









9


the environment (Bootsma, 1992; Davids, Handford, & Williams, 1994; Turvey, 1990). That is, these theorists argue that information from the environment is used "on line" by the performer to regulate motor activity, and that any investigation conducted in an artificial setting limits the ability of the individual to access this valuable information when engaging in decision-making processes (Davids et al., 1994; Turvey, 1992).

This stream of ecological psychology reacted strongly against the behaviorist and cognitive notion that the relationship between the person and the environment is contingent (Gibson, 1979), and contends that behavior is not solely due to processes under the control of the performer (van Wieringen, 1986, 1988). Rather, specific visual information available in the sport environment (visual flow fields) is equally vital to the production of optimal performance, because it is this information that allows for the possibility of different responses in the face of changing stimuli (the concept of affordances; Gibson, 1979; Williams, Davids, Burwitz, & Williams, 1992). Thus, the perceivable affordances from an object in the environment invite a specific action within a particular context from the performer, and this action is in turn dependent on the biomechanical characteristics of the individual (Davids et al., 1994).

These concepts of affordances and the symbiotic relationship between the person and the environment, then, may help explain why studies requiring movement of the participant in the actual sport environment document strong expert-novice differences,









10


while contrived studies using limited (e.g., two-dimensional) visual stimuli and simple motoric responses sometimes do not. In the present study, performers were tested under conditions highly similar to that encountered in the sport of billiards, and executed strokes identical to those required in real-world settings. The Need For Theory

A second explanation for the discrepant pattern of results observed in sportspecific research, and one that impacts on each of the critical questions discussed previously, is that these studies were descriptive in nature and were not conducted within a theoretical framework generating meaningful hypothesis testing. Further, investigations in which the expert-novice paradigm in sport has been used may be characterized as solely data-driven as opposed to theory-driven (Abernethy, Thomas, & Thomas, 1993; Vickers, 1992). Interestingly, this is a criticism that Viviani (1990) also levels at investigations of eye movements in more basic cognitive psychology tasks.

Only recently has a call has gone out for more theoretically based, deductive

approaches of hypothesis testing in the study of visual search (Ripoll et al., 1986; Vickers, 1996). Unfortunately, to date few theories have been developed in an attempt to explain oculomotor behavior from a motor control standpoint.

Theoretical Approaches Emerge

In reaction to these criticisms, Vickers (1996) proposed a location-suppression

hypothesis for aiming behaviors to far targets (beyond arm's length). Far aiming tasks, she









11


contends, are highly prevalent in such sport tasks such as baseball pitching, basketball free-throw shooting, archery, pistol shooting, bowling, and billiards.

In the location aspect of her hypothesis, Vickers (1996) proposes that fixations of relatively long durations must be made to specific target locations during the preparatory phase of the movement. The extended duration of fixation in this phase is essential for successful performance because it is believed that the programming of the parameters of the movement occur in this time period. That is, the parameters of the location and distance of the target are programmed, as are the forces, timing, and coordination of limbs necessary to produce the optimal movement. As the movement is first initiated (the impulse phase), relatively slow movements are required to maintain fixation on the target and complete the final structuring of the aiming commands.

During the execution phase of the movement, Vickers (1996) posits that a

suppression mechanism is used to reduce interfering visual information that may result as a consequence of the aiming movement (e.g., hands and ball appearing in front of the eye during a basketball free-throw, thus occluding the target). Expert performers, she contends, have developed the ability to divert their visual attention from the target during execution by blinking or orienting their gaze to other elements of the visual field. Since the parameters of the movement have already been planned in the preparation phase, and









12


modified in the impulse phase, visual attention is deemed unnecessary in the execution phase for successful completion of the aiming task.

Two critical components related to performance are outlined in Vickers' (1996)

location-suppression hypothesis. The first of these deals with the duration of fixation upon the target during the preparatory phase of the movement. Vickers (1996) termed this duration quiet eye, and defined it as that portion of the final fixation from onset to the first observable movement of the aiming limb. The longer the duration of quiet eye, it was hypothesized, the greater the performance in aiming to a far target. It is in this time period, Vickers (1996) contends, that the performer sets the final parameters of the movement to be executed. The key principle, she notes, is that quiet eye duration is directly associated with the amount of cognitive programming required for successful aiming to a target.

In numerous studies, it has been demonstrated that greater durations of reaction time and preparation time are required to adequately perform more complex motor responses (e.g., Henry, 1980; Henry & Rogers, 1960; Kerr, 1978; Klapp, 1977, 1980; Schmidt, 1988). Logically, it follows that if the quiet eye duration is directly related to a period of cognitive programming, more complex aiming behaviors should be characterized by longer quiet eye durations. Similarly, reductions in the time allotted for the participant to complete the task may reduce the time spent in the quiet eye phase, thus limiting the extent of cognitive programming and impairing performance. To date, however, no manipulations of task complexity or task duration have been conducted in studies









13


examining the relationship between quiet eye duration and performance in far aiming tasks.

In addition, Vickers (1996) argued that better performance (in this case, expert performance) would be characterized by the suppression of vision during the execution phase of the movement. Poorer performance, it was noted, was related to the maintenance of fixation on the target, while expert shooters tended to terminate their fixations on the target during the execution phase. This result is contrary to what many scholars would hypothesize, and contradicts virtually every instructional guideline for improving freethrow shooting performance (e.g., Schmidt, 1988; Wooden, 1988).

Statement of the Problem

In response to a series of critical fundamental questions raised by contemporary scholars, Vickers (1996) developed a location-suppression theory of oculomotor control proposed to account for expertise differences in far aiming tasks. Based upon specific hypotheses proposed in this theory, several aspects of the relationship between eye movements and performance in a billiards task were examined in this study.

First, the relationship between quiet eye duration and performance in highly-skilled and lesser-skilled billiards shooters was assessed, as was the question of whether the visual suppression aspect of the aiming movement was robust to billiards shooting. In addition, by investigating the effect on performance of manipulations in task complexity and









14


constraining the temporal aspects of the billiards shot, potential evidence was generated to determine (a) the importance of the location phase of Vickers' (1996) theory, and (b) whether there was support for a strong view of the eye-mind connection. More specifically, an attempt was made to resolve the issue of whether eye-movement processes are reflective of higher-order cognitive programming, and not merely the consequence of bottom-up processing.

A second purpose of this study was to assess potential differences between highlyskilled and lesser-skilled players in fixation location and fixation duration measures in an ecologically valid environment. It was expected that realistic data collection procedures would yield more accurate and valid measures of exactly where in the environment expert and novice players direct their visual attention (as revealed by their visual search patterns) than that generated by crude video simulation experiments performed in a contrived laboratory situation.

Hypotheses

Given the comprehensive scope of this study, a series of hypotheses were tested across two experiments. In the first experiment, the complexity of the billiards task to be performed was manipulated, such that participants were required to successfully execute shots that possessed easy, intermediate, and hard ratings of difficulty (Martin & Reeves, 1993). In Experiment 2, participants were asked to initiate their shot within particular time periods. Individual subjects experienced constraints of 25% and 50% on the time allotted









15


to execute shots of intermediate difficulty, based on their normal temporal stroke as recorded in Experiment 1.

Six hypotheses were expected to be consistent across both experiments, and they are presented next. Hypotheses specific to each experiment will follow.

1. It was hypothesized that the highly-skilled participants would possess a visual search strategy that was different than the lesser-skilled performers. This difference will be manifested in discrepancies in the dependent variables of fixation location and fixation durations to particular elements of the visual environment. These expected variations would be in accordance with previous research demonstrating expertise differences in visual search patterns when ecologically valid testing situations are used. For example, Vickers (1992) reported that highly-skilled golfers generated more fixations of longer duration to specific aspects of the putting situation, in particular to the target (hole), the golf ball, and the putting surface after the putter made contact with the ball. Less-skilled putters (as evidenced by their higher handicap) were observed to fixate more on the club head in the backswing phase and the ball in the flight phase.

It has been argued that expert performers are more likely to fixate only on the

pertinent aspects of the display, due to the valuable predictive information of these visual cues (e.g., Bard & Fleury, 1981; Singer et al., 1996). Indeed, this point has been supported in studies using fast-paced dynamic situations such as service reception in tennis









16

(Goulet, Bard, & Fleury, 1989; Singer et al., 1996), and offensive strategy in soccer (Helsen & Pauwels, 1993; Williams et al., 1994), and in such self-paced closed tasks as pistol shooting (Ripoll et al., 1985), basketball free-throw shooting (Ripoll et al., 1986; Vickers, 1996), and golf putting (Vickers, 1992). Thus, it was believed that differences in fixation locations and durations would be observed in the sport of billiards, such that the highly-skilled performers would fixate more on the cue ball, target ball, and target cushions than would lesser-skilled players.

2. It was hypothesized that highly-skilled shooters would demonstrate longer quiet eye durations, regardless of the level of task complexity or task duration in each experiment, than lesser-skilled players. Expert shooters have been shown to possess greater cognitive knowledge of the factors associated with successful stroke production (Abernethy, Neal, & Koning, 1994), and since quiet eye duration is believed to be reflective of a period of cognitive programming in which the parameters of the aiming movement are set, logically it follows that quiet eye durations would be greater in more advanced performers. This notion is supported by research investigating basketball freethrow shooting (Vickers, 1996), in which players with higher career shooting percentages possessed larger durations of quiet eye preparation than less-skilled shooters.

3. In a related vein, it was believed that successful shots for both groups would be characterized by longer quiet eye durations. As Vickers (1996) reported, successfully made shots by both highly-skilled and lesser-skilled free-throw shooters where









17


characterized by more time spent in quiet eye, as compared to unsuccessful shots. She hypothesized that longer durations of quiet eye reflected a more effective period of motor programming, such that each of the parameters of an ideal movement could be adequately addressed. When the participant spends too little time in quiet eye, she contends, the movement cannot be appropriately programmed, and poor performance results.

4. As a result of the visual suppression component of her theory, Vickers (1996) noted that expert players demonstrated a greater degree of visual suppression during the execution phase of the movement than the lesser-skilled players. This suppression was manifested in a greater number of eyeblinks and shifts of gaze to less pertinent (unidentified) areas of the visual environment during the foreswing and flight phases of the free throw by the better shooters.

However, this result contradicts virtually every other study investigating the role of oculomotor control in aiming tasks. In the majority of these studies, it was determined that the target was fixated on throughout the entire duration of the movement, and gaze was not directed to other areas of the display (Abrams et al., 1990; Guitton & Volle, 1987; Prablanc & Pellison, 1990). Thus, it was hypothesized that the suppression element of Vickers' (1996) theory would not be supported, as it was expected that both the highlyskilled and lesser-skilled billiards player would not divert fixation from the target in the foreswing and flight phases of the billiards stroke.









18


5. As an inherent aspect of their status, it was hypothesized that highly-skilled participants would demonstrate a greater percentage of successful shots than the lesserskilled players in every experimental condition. Indeed, this hypothesis is one that has been confirmed in virtually every study investigating expertise from a visual search standpoint, as experts have been shown to possess faster and more accurate decision-making capabilities than their less-skilled counterparts (e.g., Abernethy, 1988, 1991; Singer et al., 1994, 1996; Wright, Pleasants, & Gomez-Mesa, 1990).

6. In the billiards stroke, four distinct phases can be observed through kinematic

analysis. These phases include the preparation, backswing, foreswing, and flight aspects of the movement. Based on previous kinematic research on rapid aiming movements (e.g., Vickers, 1992, 1996), it was hypothesized that highly-skilled and lesser-skilled individuals would differ in the amount of time spent in the preparation phase, but not in the backswing, foreswing, or flight phases of the movement. Specifically, the more advanced participants were expected to demonstrate longer durations in the preparation phase, due to the increased amount of time devoted to the programming of the movement. This time difference spent in the preparation phase is also believed to exist as a result of the greater proportion of quiet eye time used by expert performers (Vickers, 1996). Experiment 1

In Experiment 1, the level of complexity of the task performed was manipulated, and two hypotheses were generated.









19


1. As the complexity of the task increases, it was hypothesized that successful shots would be characterized by concomitant increases in quiet eye duration. For both groups, it was believed that effective stroke outcomes would be produced in those instances when more time was spent in quiet eye in the preparation phase of the movement. Unsuccessful shots, therefore, would be marked by shorter quiet eye durations. Thus, quiet eye duration should be greatest in the hard condition, and shortest in the easy level of complexity.

Although the relationship between quiet eye and task complexity has never been examined, several lines of motor control research have indicated that reaction and movement times increase as a function of task complexity (Henry & Rogers, 1960; Sternberg, Monsell, Knoll, & Wright, 1978). Scholars have taken this pattern of results to indicate that manipulations in complexity influence the response programming stage of the human information processing system (Schmidt, 1988). In this stage, the parameters of the movement to be performed are delineated, such that the velocity, direction, force, and temporal sequencing of the motor response are set. More complex movements, it has been consistently demonstrated, require more time spent in the response programming stage (Henry & Rogers, 1960; Schmidt, 1988; Sternberg et al., 1978). As Vickers (1996) contends, quiet eye duration is a reflection of response programming time. Therefore, it






~.j.
~ ~


20

was expected that more complex shots would require longer quiet eye durations in order to successfully execute the task.

2. Decrements in performance were expected to occur as the complexity level of the required shot increased. Again, this finding is very robust in virtually every task reported in studies of response programming and oculomotor control (Henry & Rogers, 1960; Kerr, 1978; Klapp 1980; Schmidt, 1988). Experiment 2

The hypotheses generated in Experiment 1 were similar to those pertaining to Experiment 2, in which task duration was manipulated.

1. As the time apportioned for initiating the billiards stroke was reduced, it was believed that decrements in performance would be observed for both groups. Indeed, the percentage of successful shots was expected to diminish as the participant progresses from their normal temporal stroke to the 25% reduction, with the largest decrements in performance occurring in the 50% reduction condition. As Shapiro (1977, 1978) and Summers (1977) have demonstrated in their work on hand aiming movements, an increase in the percentage of errors occurred when participants were required to execute the movement under time constrained situations. These and other researchers (e.g., Vickers, 1996) have argued that reductions in time allotted to the task lead to direct constrictions in the response programming stages of the human information processing system. This, in turn, leads to poorer motor performance.









21


2. Although manipulations in task duration were assumed to lead to performance decrements, it was hypothesized that the relative ratio of durations spent in each of the four phases of the movement would remain constant. In essence, it has been argued that the same motor program is employed as time becomes constrained, while only the overall duration parameter is modified (Schmidt, 1988). Other researchers have supported this view, as studies by Shapiro (1977), Summers (1977), and Terzuolo and Viviani (1979) have demonstrated that the proportion of total movement time required to traverse each phase of a movement remained the same regardless of whether the movement was performed in normal or compressed conditions (see Gentner, 1987 for a comprehensive review).

It is important to note that decrements in performance were not expected to be due to changes in the relative proportion of durations between the phases of the billiards stroke (Gentner, 1987; Schmidt, 1988). If the proportions remain constant across conditions, then a stronger argument can be made for the invaluable role that the quiet eye variable plays in determining the quality of the aiming behavior. That is, impaired performance may likely be caused by insufficient time spent in quiet eye, thus leading to inadequate programming of the stroke. Thus, it was expected that successful shots will once again be characterized by longer quiet eye durations than unsuccessful attempts, even under externally imposed time constraints.









22


Assumptions

For the purposes of this study, the following assumptions were made:

1. The selection procedure of participants would be sufficient to adequately

discriminate highly-skilled billiards players from lesser-skilled players. To accomplish this goal, two methods were employed to ensure that playing ability and experience was different between groups. This distinction is important in determining whether the location-suppression hypothesis is robust to a billiards task, and whether it is associated with differences in shooting performance.

2. Vickers (1996) posits that quiet eye duration is a marker of response

programming time in motor activities requiring aiming to a far target. Based on results reported in other studies investigating the relationship between task complexity and response programming (e.g., Henry & Rogers, 1960), it was assumed that increases in the complexity of the billiards task would directly lead to concomitant changes in cognitive programming times. As such, it was also assumed that evidence for these changes would be documented by altered durations in the quiet eye variable of the location hypothesis.

3. Similarly, since scholars have demonstrated a direct relationship between task duration and response programming (Shapiro, 1977, 1978; Summers, 1977), it was assumed that constraints in the duration required to complete the billiards task would directly lead to reductions in the duration of the quiet eye variable and in each of the









23


phases of the movement. Performance was also expected to decrease as greater constraints were placed on the time allocated to the execution of the billiards stroke.

4. It was assumed that the testing situation possessed a high degree of ecological validity, since the participant was required to complete the billiards tasks using a normal biomechanical stroke with a regulation sized cue, on an actual pool table. In addition, the shots required in the billiards task, regardless of the level of complexity, were highly similar to those that could be encountered in the normal course of a game of nine-ball.

Limitations

Given the technological factors associated with this study, the following limitations may have existed and must be acknowledged when interpreting the results.

1. Due to the fact that the instrumentation used to measure participants' eye movements and visual tracking patterns relied on a corneal reflection technique, the apparatus was highly sensitive to lighting conditions. To generate a strong reflection, the testing environment needed to be more dimly lit than normally encountered in routine game situations. Thus, this potential reduction in the quality of the visual stimuli may lead to decrements in performance. As researchers have noted (Abrams, Meyer, & Kornblum, 1990; Elliot, Chua, & Pollack, 1994; Zelaznik, Hawkins, & Kisselburgh, 1983), increased visual occlusion of a target tends to lead to systematic deteriorations in performance, especially in targets that require more complex aiming behaviors. However, the targets









24


will be visible at all times, and since the impaired lighting conditions were experienced to an equal degree by each participant, this variable most likely did not contribute to any observed differences between the two groups.

2. The eye movement monitor collected data via a special piece of equipment

(weighing approximately 700 g) which was mounted on the participant's head. This added weight may have served as an inconvenience for the shooter, and may have detracted from the ecologically validity of the billiards situation. However, previous researchers using this equipment in other self-paced sport tasks have reported no significant reductions in performance on behalf of the participant, thus indicating that this measurement technique may not be inherently intrusive (Vickers, 1992; 1996).

3. The reductions in duration allotted for completion of each shot in Experiment 2 may have detracted from the ecological validity of the billiards task, since shooters do not normally face these externally imposed time constraints (although 30 sec shot clocks are used in most nine-ball tournaments). Again, however, these constraints were experienced by each participant, and therefore are not likely to produce any differential pattern of results between the groups.

Definition of Terms

To operationally define and standardize some of the terminology used in this study, each of the following terms are defined:









25


Carom Shot occurs when the cue ball glances off the object ball into a second ball (Martin & Reeves, 1993).

Cue Ball is the all-white, unnumbered ball which is the only legal ball struck with the cue stick (Martin & Reeves, 1993).

Cue Stick is the tapered wooden stick, usually 57 in (145 cm) in length and 17-21 oz (480-594 g) in weight, with which the player strikes the cue ball (Martin & Reeves, 1993).

Cushion refers to the cloth covered rubber buffers that line the inside rails of the billiards table (Martin & Reeves, 1993).

Cut Shot refers to the action of hitting the object ball with the cue ball at less than dead center, thus deflecting the object ball off at an angle (Martin & Reeves, 1993).

English is the spin of the cue ball either to the left or to the right, controlling the action of the cue ball before and after it strikes the object ball or cushion (Martin & Reeves, 1994).

Execution Phase is the final, error-correction phase of an aiming movement, in

which attempts are made to reduce apparent discrepancies between the current position of the limb and the movement goal as specified by the motor program (Abrams et al., 1990). This phase corresponded to the time period involving the foreswing and flight phases of the billiards stroke.









26


Eyeblinks, as was operationally recorded in this study, occurred when the

participant closed their eyes and occluded the optics of the system for 100 ms or more (3 or more video frames).

Eye-Mind Assumption refers to the belief that "eye movements can at the very least be considered as tags or experimentally accessible quantities that scientists can observe to understand underlying processes of cognition" (Viviani, 1990, p. 354). Stronger views of the eye-mind assumption believe that eye movements are reflective of higher-order cognitive processing (Just & Carpenter, 1976).

Fixations occur when the individual's eyes pause at a specific location such that information is stabilized on the high-acuity area of the retina, the fovea (Vickers, 1992). Consistent with previous research (e.g., Carl & Gellman, 1987; Optican, 1985; Vickers, 1996), fixations were recorded when the participant's gaze was fixed on a location for 100 ms or more (3 or more video frames).

Impulse Phase is described as a fairly rapid and continuous change in the position of the aiming limb as it traverses most of the distance between the starting position and the final location of the target (Abrams et al., 1990; Carlton, 198 1a, 198 1b). In this study, the impulse phase included the time period corresponding to the backswing of the cue stick until initiation of the foreswing.









27


Location-Suppression Hypothesis is a recent theory proposed to account for visual control mechanisms when an individual is engaged in aiming behaviors to a far target which is beyond arms reach (Vickers, 1996).

Nine-Ball is a billiards game played with a special diamond-shaped rack which

holds balls numbered 1 through 9. The object of the game is to sink the balls in numerical order (from 1 to 9), and the person legally sinking the 9 ball is declared the winner (Billiard Congress of America, 1996).

Object Ball refers to the ball the player wishes to hit with the cue ball (Martin & Reeves, 1993).

Preparation Phase refers to the time period in which the performer engages in behavior designed to familiarize themselves with the general direction of the target (Abrams et al., 1990), and to set the parameters required for successful execution of the aiming behavior (Vickers, 1996). In the present study, the preparation phase was operationally defined as the time recorded when the performer was positioned over the cue ball until the first observable movement towards the striking of the cue ball (i.e., backswing was initiated).

Quiet Eye Duration is defined as "that portion of the final fixation from onset to the first observable movement of the hands into the shooting action" (Vickers, 1996, p.348)









28


Visual Attention refers to the process by which individuals select information gathered by visual mechanisms to provide the basis for responding to particular stimuli (Theeuwes, 1993).

Visual Search may be defined as "the manner in which individuals move their eyes to take in available visual information while preparing and executing a movement" or engaging in a decision-making process (Vickers, 1996, p.342).

Visual Suppression refers to an aspect of the location-suppression hypothesis in which participants suppress their vision by using a blink behavior, thereby preserving the aiming commands and eliminating potential limb interference in the visual field (Vickers, 1996).

Significance of the Study

As is evidenced by the critical questions raised by contemporary scholars

concerning the continued use of the visual search paradigm in examinations of expertise in sport, it is clear that this paradigm is at a crossroads in terms of its development. Researchers could abandon this line of investigation, as some reviewers have suggested (e.g., Abernethy, 1993), or they could attempt to answer their critics with more theoretically based studies conducted in ecologically valid environments. It was the intent in the present study to follow the latter path, in the hope that greater insight could be gained into the relationship between eye movement behavior and optimal motor performance in far aiming tasks.









29


To this end, one of the main goals of this study was to further advance the

theoretical foundations of oculomotor control in aiming tasks, specifically by investigating aspects of the location-suppression hypothesis (Vickers, 1996). Although this theory has been supported in a study of basketball free-throw shooting (Vickers, 1996), no other line of research has attempted to critically evaluate it to determine if it is in fact robust to other far aiming tasks. In addition, if the location-suppression hypothesis is a critical factor in the organization and programming of neural substructures underlying the aiming response (Vickers, 1996), then manipulations in factors such as task complexity and task duration should result in alterations in performance.

In the present study, therefore, an attempt was made to determine whether the quiet eye and suppression aspects of Vickers' (1996) theory were present in the sport of billiards, and whether these factors were influenced by manipulations related to cognitive programming stages. The information gathered from this study may serve to either advance the location-suppression hypothesis, and in turn provide support for the use of visual search paradigms in expertise research, or modify specific aspects of the hypothesis that do not hold up to empirical testing. In addition, support for Vickers' (1996) theory would provide further evidence for a stronger view of the eye-mind assumption. If the manipulations were indeed effective in altering performance, then a more logical argument









30


could be made in favor of the notion that eye movements are reflective of cognitive processes.

In addition to the unique concept of theory testing that has been so limited in visual search research, the present study attempted to incorporate an ecologically valid environment in which the location-suppression hypothesis was tested. Again, very few studies have been conducted in situations that allow the performer to actively perceive information and respond with behaviors that would be typical of the true sporting environment. Thus, unlike virtually every other study examining eye movements in sport, this study investigated aiming behaviors in a real-world setting while the billiards players actively engaged in "on line" processing of information made available to them in their environment. This information included visual cues and knowledge of results pertaining to the outcomes of actions they initiated.

Finally, an attempt was made to delineate expertise differences in the sport of billiards from a visual search standpoint. Very little research of any kind has been conducted using a billiards task, and virtually nothing is known about the mechanisms highly-skilled players use to perform the precise movements required to achieve success in this sport. Thus, this study was significant in the fact that it represented an initial attempt to determine exactly what visual cues more advanced and less advanced players fixated on when actively engaged in the act of shooting. Potential expertise differences in cue usage









31


may then be used to initiate the development of cognitive training programs designed to improve billiards performance in players of all skill levels.















CHAPTER 2
REVIEW OF LITERATURE


Given the importance of the human information processing system in relation to sport performance (e.g., Magill, 1998; Schmidt, 1988), this literature review will focus on several aspects germane to the topic. First, factors associated with the development of expertise in sport, in particular hardware-software approaches, will be discussed. Next, a specific software approach in the study of expertise, the visual search paradigm, will be reviewed. This section will also include critiques of current methodology used to assess general expert-novice differences, with particular emphasis placed on research examining eye movement patterns.

Three specific theories of oculomotor control will then be discussed, as these

theories focus on aiming tasks which require a high degree of eye-hand coordination. The location-suppression hypothesis (Vickers, 1996) will be outlined in greater detail, as it is the most important theory under investigation in this study. Finally, because task complexity and task duration will be manipulated in the present study, an overview of the motor learning and control literature related to these topics will be addressed.


32









33


Expertise in Sport

The Hardware Approach

Early research in sport expertise focused on the physical aspects of the athlete, since it was believed that expert athletes differed from novices in that they possessed advanced psychophysical and mechanical properties of the central nervous system (Abernethy, 1991; Blundell, 1985; Burke, 1972; Nielsen & McGown, 1985). That is, proponents of this theory believe that experts have much faster overall reaction (RT) and movement (MT) times than do novices, and also possess greater optometric (static, dynamic, and mesopic acuity) and perimetric (horizontal and peripheral vertical range) attributes (Carlson, 1985; Revien, 1987). Not surprisingly, this perspective has been termed the "hardware" approach to expertise (Starkes, 1987; Starkes & Deakin, 1984).

However, research examining athletes' mechanical and optometric properties

(hardware) has yielded equivocal results. Blundell (1985) reported that static visual acuity (SVA) was only slightly related to expert performance, accounting for roughly 8-10% of the variance in a pool of athletes. Similarly, Starkes (1987) could find no significant relationship between SVA, dynamic visual acuity (DVA), or peripheral vision in her study of expert-novice field hockey players. This result was also obtained by Helsen (1994) in an examination of athletes from several different sports. Although Blundell (1985) and Sanderson and Whiting (1975) have reported that the role of DVA and peripheral vision









34


are important in interception tasks such as ball catching, support for their results has not been generated in other studies (e.g., Abernethy, 1986; Helsen, 1994).

Also, highly-skilled performers have been found to possess base rates of RT and MT that are equal to those of their lesser-skilled counterparts when assessed in non-sport related conditions (McLeod, 1987; Nielsen & McGown, 1985; Starkes, 1987), and as such their demonstrated performance superiority does not appear to be based on a faster ability to process and react to information. More evidence of this contention comes from a study conducted by Singer, Cauraugh, Chen, Steinberg, and Frehlich (1993), in which high level college and professional tennis players were compared to novice and intermediate (recreational) players on several variables related to RT, MT, and coincidence-anticipation capabilities. No significant differences were discovered between the groups on non-tennis related tests of foot and hand speed (RT and MT measures), and anticipation capability was similar for the expert and novice players on a Bassin timing task.

Other attempts to investigate differences in rates of visual information processing (VIP) have shown that elite performers may be faster processors of visual information in general, perhaps explaining in part how they achieved expert status (Adam & Wilberg, 1986). However, this result has not been replicated in more recent studies (e.g., Starkes, Allard, Lindley, & OReilly, 1994). Thus, it appears that advanced athletes can not be consistently discriminated from less-skilled participants solely on hardware aspects related to superior CNS and VIP capabilities.









35


The Software Approach

In contrast to the hardware theory of expertise, proponents of the "software" approach argue that experts have a much greater knowledge base of information pertaining to their particular sport, and that differences in expert performance as compared to novices is the result of a cognitive advantage, rather than a physical advantage (e.g., Singer et al., 1994; Starkes & Deakin, 1984). Elite athletes, it is believed, make faster and more appropriate decisions based on the cognitive processes of selective attention, anticipation, and pattern recognition (Abernethy, 1991). That is, highly-skilled performers appear to possess the ability to know which cues to focus their attention on in the sport environment, and understand the importance of these cues in predicting future actions (Abernethy & Russell, 1987a; Bard, Fleury, & Goulet, 1994; Singer et al., 1996).

Support for the software approach to explain expertise has been demonstrated in studies in which decision time and accuracy responses were assessed in sport specific situations (Bard & Fleury, 1976; Starkes, 1987), and for the recognition and recall of structured elements of game situations in sports such as baseball (Hyllegard, 1991), basketball (Allard & Burnett, 1985; Bard & Fleury, 1976), field hockey (Starkes, 1987), and volleyball (Borgeaud & Abernethy, 1987). In addition, experts have been found to possess more declarative (factual information) and procedural knowledge (how to do something) in their sport domain than do novices (Allard, Graham, & Paarsalu, 1980;









36


Allard & Starkes, 1980; Chase & Simon, 1973; Chi, Glaser, & Farr, 1988; Sandu, 1982; Starkes & Deakin, 1984). That is, elite performers were much faster and are more accurate in recognizing and recalling information related to structured game situations. This ability, however, appears to be task-specific, as highly-skilled and lesser-skilled athletes do not differ in recognition and recall performance for elements that are unrelated to their sport (Chi & Bjork, 1991; Chi, Glaser, & Farr, 1988). As Abernethy, Neal, and Koning, (1994) noted, "the more sport-specific the stimuli and response(s) used in the test task, and hence the more closely the processing demands of the test mimic those of the intact skill, the more probable it is that systematic expert-novice differences will be demonstrated" (p.186).

To determine the relative contributions of software and hardware variables on

expertise in the sport of billiards, Abernethy, Neal, and Koning (1994) compared 7 expert, 7 intermediate, and 15 novice Australian snooker players on general vision tests and sportspecific perceptual and cognitive tests. Results of their study indicated that no expertise differences were found on optometric tests of SVA, ocular muscle balance, color vision, and depth perception (hardware variables). However, expert and intermediate players differed from novices in their ability to recall and recognize rapidly presented slides of organized game situations, but these differences were not observed in randomly arranged situations. Finally, expert shooters were characterized by the ability to choose a greater number of shot options that were more appropriate, and planned more shots in advance









37


(M = 6.22) than did novices (M = 4.56). This finding is indicative of a greater depth of encoding for structured material, and supports the notion that expertise is developed not from general visual capabilities, but rather from acquired processing strategies.

Thus, a preponderance of the research supports the software view that expertise in sport is the result of a cognitive advantage, and not necessarily a physical advantage. But how is this cognitive advantage developed? That is, given the same game situation, what processes direct highly-skilled performers to select and execute the most appropriate responses while novices do not?

A likely source of the cognitive advantage comes from multiple experiences in the sport task. For example, Ericsson, Krampe, and Tesch-Romer (1993) and Ericsson and Chamess (1994) argue that the development of expertise is explicitly due to a dedicated regimen of deliberate practice, in which a set of highly structured activities are practiced. Here, the individual engages in effortful and attention-demanding endeavors that have the explicit goal of improving performance. In fact, Ericsson et al. (1993) argue for a monotonic relationship between practice and performance. That is, the more deliberate practice one undertakes, presumably the greater the performance and level of expertise. Finally, these researchers argue that true expertise cannot be developed until a minimum of 10 years of deliberate practice is undertaken.









38


However, the question remains: What exactly does the elite performer learn during this process of deliberate practice? Researchers supporting the software explanation of expertise have offered one possibility: Highly-skilled athletes have learned to effectively search their environment for pertinent cues that allow them to select and execute the most appropriate behavior in a particular situation (e.g., Bard & Fleury, 1981; Singer et al., 1996; Vickers, 1992). That is, processes such as the visual searching of one's environment were hypothesized to account for some of the differences observed to exist between elite and non-elite performers. The results of this line of research are discussed in more detail in the next section.

The Visual Search Paradigm in Sport Research

Recent advances in technology over the past two decades have allowed scholars to examine eye movements and tracking behaviors in simulated sport conditions which demand immediate attention (Bard & Fleury, 1976). The majority of research undertaken in this area has been concerned with the relationship between visual search and selective attention, and with the influence of these processes on decision making strategies in relation to filmed scenarios (Helsen & Pauwels, 1992, 1993; Singer et al., 1994). In each of these studies, it was assumed that the allocation of focal attention could be determined by examining the locations and durations of ocular fixation patterns. These patterns were assessed with eye-movement recording devices that measured ocular fixation via a cornea









39


reflection technique (e.g., Bard & Fleury, 1981) while athletes viewed slides or videotaped simulations of particular sport situations.

Reviews of the sport-specific literature on selective attention and visual search, even in simulated conditions, often demonstrates systematic differences in eye movement patterns between expert and novice performers (e.g., Abernethy, 1988, 1991). Sports of interest include baseball (Bahill & LaRitz, 1984; Shank & Haywood, 1987), fencing (Bard, Guezennec, & Papin, 1980; Haase & Mayer, 1978), golf (Vickers, 1992), gymnastics judging (Bard, Fleury, Carriere, & Halle, 1980; Vickers, 1988), ice hockey (Bard & Fleury, 1981), soccer (Tyldesley, Bootsma, & Bomhoff, 1982; Helsen, Pauwels, & D'Ydewalle, 1986; Williams et al., 1994), table tennis (Ripoll & Fleurance, 1985), tennis (Goulet, Bard, & Fleury, 1989; Ritzdorf, 1983; Singer et al., 1996), and volleyball (Neumaier, 1983; Ripoll, 1988).

Evidence in Support of Expert-Novice Differences

As mentioned earlier, most of these studies yielded three general results. Elite performers required fewer fixations on elements in the display to produce a successful response (e.g., Singer et al., 1996), allocated a greater number of fixations to pertinent aspects of a visual array (e.g., Ripoll, 1988), and were more systematic and consistent in their search patterns than novices (e.g., Ripoll et al., 1986; Vickers, 1992). For example, expert goaltenders fixated on the stick of ice hockey shooters more than novice









40


goaltenders, who tended to allocate more fixations to the puck (Bard & Fleury, 1981). In addition, experienced soccer players fixated more on pertinent aspects of the display than did less-skilled players, who had much more variable visual tracking patterns (Helsen & Pauwels, 1990; Williams et al., 1994).

When reviewing the particular sports that have been studied using the visual search paradigm, it becomes clear that much of the research has been directed towards dynamic, fast-paced tasks. This is understandable, since the ability to successfully fixate on advance cues that may be used to anticipate the intended actions of an opponent gives the participant a decided advantage in preparing an effective response under time constrained situations (Abernethy, 1991). For example, expertise differences were found in visual search patterns in an examination of expert and novice tennis players preparing to return a videotaped serve (Fleury, Goulet, and Bard, 1986; Goulet, Bard, & Fleury, 1989; Singer et al., 1996). Specifically, highly-skilled players tended to make more fixations on the shoulder and trunk areas of their opponent in the ritual phase of the serve, while lesserskilled participants fixated most often on the server's head. During the execution phase of the serve, experts demonstrated a greater number of fixations on the arm and racquet of the opponent, while novices tracked only the ball. No significant differences were found during the preparatory phase of the serve (Goulet, Bard, & Fleury, 1989). It was demonstrated that those cues fixated on by the experts were the most pertinent, because they provided important information about the direction and type of serve to be returned.









41


Similar results have also been obtained in the sport of baseball, where it was

demonstrated that successful hitters tended to focus more on the release point of the pitch, while less advanced batters tended to shift their focus from the pitcher's head, to the release point, and then back to the head (Shank & Haywood, 1987). It was determined that the experts' focus of attention (pitch release point) provided the most important cues for successful recognition of pitch velocity, type of pitch, and location of the ball before it crossed the plate (Shank & Haywood, 1987).

These studies, and many others, have documented the positive relationship

between expertise and visual search patterns in dynamic sport tasks. However, several investigations have also revealed expert-novice differences in search patterns for closed, self-paced tasks in which the performer has unlimited time to execute the movement under stable environmental conditions (Singer & Chen, 1994). Thus, it would appear that visual search processes play an important role in sports which require precise aiming behaviors.

For example, in a basketball free-throw shooting task, highly-skilled individuals oriented their gaze towards the basket sooner, and made a greater number of fixations of longer duration towards this target than lesser-skilled performers (Ripoll et al., 1986). In addition, more successful shots were characterized by longer durations of fixation towards the hoop during the preparation and flight phases of the shot than missed attempts (Vickers, 1996).









42


In pistol shooting, elite performers maintained their fixation on the target

throughout all phases of the aiming action, while less advanced shooters tracked the sight of the gun until they made a fixation on the target immediately prior to pulling the trigger (Ripoll et al., 1985).

Finally, in an analysis of the putting stroke used by expert and novice golfers, Vickers (1992) reported that low handicap golfers (experts) possessed an economy of gaze allocation when compared to high handicap golfers (novices). Specifically, experts made more express saccades, had quicker saccades between gaze locations, and demonstrated significantly more fixations of longer durations to the ball and target during execution of the putting stroke. Lesser-skilled putters, on the other hand, often tracked the golf club head immediately prior to contact with the ball, and allocated fewer fixations to the ball during the contact phase of the movement.

Taken together, much of the visual search and gaze control research indicates that expert athletes possess highly systematic and selective process of focusing their attention. Specifically, highly-skilled performers know exactly what to look at in the visual field to glean the most informative cues in dynamic as well as static sport situations, and make the most efficient and appropriate ocular movements to pick up this information. In comparison, novices are characterized by random and less uniform visual tracking patterns, and do not appear to understand the relative importance of the link between advance visual cues and effective response preparation (Abernethy, 1991). The robustness









43


of these views, however, have been questioned by scholars over the course of the past decade.

Evidence Against Expert-Novice Differences

Although the bulk of visual search research has demonstrated expertise differences, the expected pattern of results has not been observed in a few empirical studies. In a comparison of 20 expert (International level) and 35 novice (untrained undergraduate) badminton players, Abernethy and Russell (1987b) noted that fixation sequences, distributions, and durations were indistinguishable between the two groups. Both groups fixated more on the racquet and arm regions of the filmed opponent than the head, trunk, and leg regions. This search strategy, the investigators posited, represented a more proximal-to-distal transition in cue usage.

A similar arrangement of results have been observed in investigations of expertise involving squash players (Abernethy, 1990a, 1990b), in which skill differences were not paralleled by comparable contrasts in visual search patterns. Also, elite and non-elite tennis players were found to possess identical search rates when viewing the preparatory phase of an opponent's motion in a simulated tennis service reception task (Goulet et al., 1989; Singer et al., 1996). However, the expected differences did emerge during the execution phase of the serve, as experts centered their fixations on the racquet and the arm holding the racquet, and terminated their search on the opponent's racquet at the moment of ball









44


contact. Beginners, on the other hand, demonstrated search rates that were highly variable with fixations allocated to several regions of the server.

Thus, while the visual search paradigm has provided a fruitful means of assessing selective attention strategies used by athletes in sport situations, the employment of this paradigm is not without criticism. Before concluding this section on visual search, it is necessary to discuss some of the limitations and potential problems that currently exist in expert-novice eye movement research. These concerns, as outlined by Abernethy (1988, 1993) and Viviani (1990), are directed toward the assumptions of selective visual attention theory, current techniques of recording eye movements, and general methodological issues.

Criticisms of Expertise Research in Sport

In the study of expertise in sport, a number of criticisms have been aimed at a variety of aspects of the research program, including the selection and determination of participants, participant pool sizes, the nature of the task under study, the number of dependent measures used, the type of equipment used, the environment in which the study takes place, and the high rates of subject variability in performance. The typical approach to expert-novice research has been to select a number of athletes (usually at a college varsity level of expertise) and compare them to untrained individuals, usually undergraduate students (Abernethy, Thomas, & Thomas, 1993). These two groups are









45


then compared on a single dependent measure of interest, on tasks that are predominantly laboratory based (and are often artificial and contrived).

For example, in studies of expertise differences in pattern recognition, two

criticisms have been levied. Often only static displays are used (usually slides of game situations, Vickers, 1996; Williams et al., 1994), and a confound exists in that one cannot be sure that expertise differences are the result of actual expertise advantage, or merely an artifact of the number of years of experience in these situations. In knowledge-based studies, several problems have been exposed, including the use of self-report data (Nisbett & Wilson, 1979), which have often been shown to be unreliable and prone to errors in memory or influenced by events occurring after the one of interest. In addition, problems have been identified in terms of the ignoring of action in many tasks (participants only verbally report, or respond to pen and paper tests, but do not carry out actions).

In this vein, proponents of a more ecological approach to the study of motor behavior have soundly criticized many experiments as being too laboratory-based and reductionist in nature (Turvey & Carello, 1986; Williams et al., 1992). They argue that the need for heightened control serves to constrain the coupling between perception and action that is so vital for expert performance. Again, typical designs include the use of static visual displays (slides) or two dimensional videotaped images, and the performer is often required to make only verbal or simple finger responses to the incoming stimuli.









46


As Bootsma (1988) discovered, the act of having the individual perform only simple tasks yields vastly different results as compared to more real world tasks. In his study, three groups of participants were asked to hit a moving ball with a stick. In the first group, the individuals were required to press a button which then moved the stick into the path of the oncoming ball. The second group of participants were allowed to touch the stick with their index finger, but could only pull it back and release it to strike the ball. In the third group, performers were allowed to grasp the stick and swing it freely toward the ball. Significant differences in coincidence timing were observed, such that the most errors were observed in the first group, while the third group demonstrated the best overall performance. Bootsma (1988) argued that these results provided evidence for the notion that the experimental situation can influence the responses generated by the performer, thus highlighting the importance of testing individuals in a more ecologically valid environment.

In addition, Vickers (1996) and Ripoll (1991) both argue that removing the actor from the actual sport environment changes the results of the study dramatically, such that expert-novice differences may no longer appear in variables such as visual fixation locations and durations. Since ecological psychology theorists contend that attunements between the person and the environment help comprise expert performance, placing the athlete in contrived lab settings retards the level and effect of their expertise (Chi, Glaser, & Farr, 1988; Fowler & Turvey, 1978; Nougier, Stein, & Bonnel, 1991).









47


In addition, a question of motivation and arousal levels have been levied against the majority of expert-novice studies, since the participants are often placed in a seated position (which is drastically different from the ready position found in most sports), and are required to only passively react to presented stimuli (Singer et aL, 1994). Thus, it would appear that severe constraints impacting on arousal and motivational levels of athletes are removed in the typical lab study. Although little is known about the resulting impact of these constraints on performance, it has been argued that performance is altered quite extensively (Ripoll, 1991; Singer et al., 1994; Vickers, 1996).

Other problems can be found in expert-novice studies. For example, the vast

majority of researchers conduct only cross-sectional rather than longitudinal investigations that are characterized by small sample sizes (usually because few experts can be found or determined). To increase sample sizes, researchers often dilute the degree to which an expert is operationally defined, thus creating inconsistent results between studies since this independent variable is not held constant (Abernethy, 1989). Finally, novice subjects often may not be classified as true beginners, as they may vary extensively in their performance and experience in the task under study (Singer et al., 1994).

Each of these criticisms have, in some form or another, been directed toward the majority of studies examining expertise differences in visual search patterns of athletes. However, based on the theoretical and methodological approaches used in this area of









48


investigation, three particular critiques have been forwarded. Each of these will be discussed in more detail in the following section. Criticisms of the Visual Search Paradigm

The first major limitation of eye movement recordings lies in the assumption that visual search orientation is reflective of a person's actual allocation of attention (a strong eye-mind connection). Stated more succinctly, it is explicitly believed that visual fixation and attention are one in the same (where we look is where we attend). As has been discussed previously, this notion has been refuted in research by Posner (1980), Remington (1980), and Remington and Pierce (1984), where it was shown that attention could be allocated to areas other than the foveal fixation point. Indeed, attention can also be allocated to areas in peripheral vision, a mode that cannot be measured with current visual search equipment (Bard, Fleury, & Goulet, 1994; Buckholz, Martinelli, & Hewey, 1993; Davids, 1984, 1987, 1988).

This limitation of eye movement recording focuses on the issue of visual

orientation and information pick-up. As Abernethy (1993) noted, merely "looking" at visual information does not necessarily equate with "seeing" (or comprehending) this information. Thus, a performer may fixate upon pertinent cues in the visual array, but there is no guarantee that the individual is actually attending to or using these cues when preparing a response. This view may help to explain those studies in which no expertnovice differences were observed in visual search patterns (Abernethy, 1990a, 1990b;









49


Abernethy & Russell, 1987b). That is, even though elite and non-elite performers may in fact fixate on the same environmental stimuli, only the highly-skilled understand the importance of these cues in determining the resulting outcome of their opponent's actions (Abernethy, 1991, 1993). While novice performers may fixate on the most pertinent advance cues, they do not possess the cognitive link underlying the predictive nature of this information.

In response to these limitations, it should be noted that much of the criticism is aimed at studies in which dynamic, externally-paced sports such as baseball, squash, and tennis have been of interest. In these studies, it is highly likely that the ability to pick up advance cues from an opponent allows the expert performer to select and execute the most appropriate response in a much more rapid fashion than novices. Indeed, anecdotal evidence from highly-skilled athletes suggests a perception that they have "all the time in the world" to respond to an opponent's actions in such time constrained conditions, while less advanced performers report being rushed and struggling to generate effective reactions (Abernethy, 1991).

In sports such as archery and billiards, however, the aiming task to be performed is self-paced, and is not based on an a constantly changing environment. In these instances, it has been suggested that the ocular and motor systems are tightly coupled, such that the role of vision and gaze behaviors are highly correlated with the motor response to be









50


performed (Abrams, Meyer, & Kornblum, 1990; Guitton & Volle, 1987; Vickers, 1996; Zangemeister & Stark, 1982). Given this large degree of coupling between the two systems, researchers have speculated that foveal vision is directly associated with visual attention, and that this visual information is in fact used to program the aiming movement to be executed (e.g., Abrams et al., 1990). Thus, because the present study was designed, in part, to assess the degree to which eye movements were associated with the execution of aiming movements in the self-paced sport of billiards, Abernethy's (1991, 1993) criticisms of the issues of information pick-up and peripheral vision are less relevant.

The second limitation of the current visual search paradigm involves the high trialto-trial variability in search patterns that has been observed between participants of similar skill levels (e.g., Singer et al., 1996). These variable patterns make reliable conclusions about the relative importance of specific visual cues difficult, as it has been shown that even expert performers differ in the type of search sequence used. Related to this limitation is the fact that the majority of experiments examining expert-novice differences in search patterns have relatively low sample sizes (often n = 4 or 6), thus raising concerns pertaining to assumptions underlying the internal and external validity of the study.

To help reduce the impact of this limitation, the present study incorporated a sample size that was, in some instances, two to three times larger than that reported in previous investigations. Also, two methods of operationally defining highly-skilled and lesser-skilled billiards players were used, including a preliminary performance analysis









51


designed to test the shooting abilities of the players. A protocol of this nature has been included in only a handful of previous visual search studies (e.g., Vickers, 1992).

Finally, researchers investigating expertise differences in eye movement patterns have been critiqued for their blatant lack of regard for theory testing, with the majority of studies being rather descriptive in nature (Abernethy et al., 1993). In fact, this atheoretical approach has been recognized by more contemporary scholars (Abernethy, 1993; Ripoll, 1991; Vickers, 1996), who have suggested that the time for simple descriptive investigations is past. Instead, a call has gone out for initiating the processes of hypothesis testing, theory development, and deductive reasoning in this area of study. To this end, three innovative theories of oculomotor control have been developed, and each will be described in more detail below.

Theories of Oculomotor Control in Aiming Tasks The Eye-Head and Image-Retina Systems

In an extension of Whiting and his colleagues' seminal work on ball catching (e.g., Savelsbergh & Whiting, 1988, 1992; Savelsbergh, Whiting, Burden, & Bartlett, 1992; Sharp & Whiting, 1975; Whiting, 1972; Whiting, Savelsbergh, & Faber, 1988), Gregory (1990) developed two competing theories in an attempt to explain how individuals use visual information to intercept objects projected in space.









52


The first approach, termed the image-retina system, contends that the individual allocates fixations to the point of release of the object in space, and that successful interception of this object is dependent solely on the information detected via motion of the object on the retina. Thus, in order to intercept an object (e.g., a thrown ball), the image-retina system theory contends that all one needs to do is fixate on the release point of the ball. All other visual information is deemed superfluous for successful catching behaviors.

In contrast, the eye-head system (Gregory, 1990) proposes that information is picked up either by the motion of the projection of the object in relation to the background, or via proprioceptive information pertaining to the motion of the eyes and the head. In either case, successful execution of an interception response is dependent on the continuous visual tracking of the object with the eyes.

Although many theorists implicitly assumed that the eye-head system was a far

more cogent explanation of catching behaviors (e.g., Brancazio, 1985; McLeod & Dienes, 1993), Michaels and Oudejans (1992) and Oudejans, Michaels, Bakker, and Davids (in press) were the first researchers to empirically examine the eye movements of individuals as they attempted to catch balls projected to them in a manner similar to fielding fly balls in the sport of baseball. Indeed, the results of these studies provided support for the eyehead system, as catchers tracked the ball from release to the catching hand with smooth eye and head pursuit movements.









53


The Position, Movement, and Movement + Position Hypotheses

While the results reported by Michaels and Oudejans (1992) confirmed a

relationship between eye movements and interceptive behaviors that was intuitively assumed to exist, the importance of hypothesis testing and empirical confirmation cannot be discounted. However, the nature of the catching task under study is quite different than the self-paced tasks of many aiming movements required in a sport such as billiards. In an attempt to explain the relationship between visual feedback processing and the production of aimed limb movements, Abrams, Meyer, and Komblum (1990) forwarded three hypotheses proposed to account for the eye-hand coordination observed when individuals perform precise aiming behaviors to a target.

In the position-only hypothesis, Abrams et al. (1990) posited that visual

information on the position of the target is so essential for the accurate completion of the movement that the eyes fixate upon the target throughout the duration of the movement. This hypothesis has been supported by Hansen (1979) and Honda (1985), who observed accurate pointing responses when individuals used either saccadic or pursuit eye movements directed solely to the target (and not on the aiming limb).

The movement-only hypothesis contends that the eyes move in a coupled fashion with the limb, or limbs, performing the aiming action. Here, information generated from oculomotor commands and proprioceptive inflow from the eye muscles is required in









54


order for the movement to be performed successfully (Abrams et al., 1990). Thus, the actual movement of the eyes acts as the critical input for effective aiming movements. Support for this hypothesis has been generated by Festinger and Canon (1965) and Miller (1980), who documented differences in final gaze location and aiming movements following saccadic versus pursuit eye movements.

Finally, the movement-plus-position hypothesis posits that saccadic eye

movements are closely time locked with the initiation of the aiming movement, while fixations upon the target only occur during the later phases of the movement (i.e., during the error correction or homing-in phase). In a critical test of the three competing hypotheses, Abrams et al. (1990) determined that the data supported the movement-plusposition hypothesis. That is, when the participants were required to perform a wrist rotation to a target, the motor control system was posited to receive and use information about both the movement and the relative position of the eyes.

Although not a direct test of Abrams et al. (1990) hypotheses, Vickers (1992)

noted that highly-skilled and lesser-skilled golf putters differed in the extent to which they made saccadic and pursuit eye movements. Extrapolating from Vickers' (1992) data, it appeared that the more advanced putters exhibited a strategy of eye movements similar to the position-only explanation, while novice putters were characterized by a movementplus-position explanation. This result suggested a different strategy of optimal eye movement behavior than that offered by Abrams et al. (1990). Again, however, it must be









55


noted that the three hypotheses were not under critical evaluation in Vickers' (1992) study.

The Location-Suppression Hypothesis

In a more direct test of the Abrams et al. (1990) hypotheses, Vickers (1996)

examined the gaze behaviors of highly-skilled athletes performing a basketball free-throw shooting task. Sixteen elite players were classified as either expert or near-expert shooters based on their free-throw percentages accumulated over the course of a competitive season. Each shooter was required to perform consecutive free throws until 10 successful and 10 unsuccessful shots were obtained. Fixation and visual search patterns for the expert and near-expert players were compared, as were successful shots to unsuccessful shots, on such variables as number of fixations, fixation location, and fixation duration. Vickers (1996) expected to replicate the position-only and movement-plus-position hypothesis findings in order to interpret differences in expertise, and to shed light on processes that may have occurred in made versus missed shots. While the near-expert performers possessed a sequence of fixation patterns that were supportive of the movement-plusposition hypothesis, none of Abrams et al. (1990) hypotheses seemed to adequately explain the results she obtained for the expert shooters.

The expert shooters were characterized by an eye movement strategy that was

unlike any proposed by Abrams et aL (1990). Experts took significantly longer to prepare









56


the free throw (M = 1899 ms versus M = 1386 ms for near-experts), made fewer fixations than near-experts during the preparation and impulse phases of the shot, and generated a greater frequency of fixations during the execution phase. More importantly, the final fixation and quiet eye durations before the initiation of the movement was significantly longer in the expert shooters. In addition, experts suppressed their vision to a greater extent during the execution phase than did the near-experts. That is, they either blinked or diverted their gaze to areas other than the hoop, the ball, or their hands during the key propulsion element of the shooting action.

From this unexpected pattern of results, Vickers (1996) generated the locationsuppression hypothesis for aimed limb movements to distant targets. She operationalized aiming skills to far targets as those in which the individual maintained control of the object to be directed to the target only to the point of release. While much is known about the coordination of the hand and the eyes in near aiming tasks (e.g., Abrams, 1994; Carlton, 1981a, 1981b; Gauthier, Semmlow, Vercher, Pedrono, & Obrecht, 1991; Guitton & Volle, 1987; Prablanc & Pellison, 1990), very little research has been conducted with the express purpose of investigating eye-hand coordination in far aiming tasks.

As has already been discussed in the introduction of this paper, the location aspect of Vickers' (1996) hypothesis proposed that fixations of relatively long durations must be made to specific target locations during the preparatory phase of the movement. Here, the final fixation and quiet eye durations are invaluable for optimal aiming performance,









57


because it is believed that the programming of the parameters of the movement, such as the location and distance to the target, the optimal force and velocity used to initiate the movement, and the relative timing and coordination of the limbs occur in this time period. As evidenced by the expert shooters high free-throw shooting percentages, the longer the duration of the quiet eye variable, the better the performance.

An equally vital characteristic of expert free-throw shooting occurred during the execution phase of the movement. Vickers (1996) noticed that a suppression mechanism was used to possibly reduce interfering visual information that may have resulted as a consequence of the aiming movement. Expert performers, she claims, have developed the ability to divert their visual attention from the target during execution by blinking or orienting their gaze to other elements of the visual field. Since the parameters of the movement have already been planned in the preparation phase, and modified in the impulse phase, visual attention is deemed unnecessary in the execution phase for successful completion of the aiming task.

Thus, expert performance in far aiming tasks, such as those found in billiards,

darts, and archery is dependent on the important variables of quiet eye duration and visual suppression (Vickers, 1996). In fact, if quiet eye duration is critical to the organization and parameterization of underlying neural structures, as has been suggested, then experimental manipulations of this time should result in changes in performance.









58


Aside from direct occlusion techniques which restrict the performer's vision, but unfortunately undermines the ecological validity of the experimental situation, two methods are readily available which may influence quiet eye duration. A large body of evidence has demonstrated that manipulations in the variables of task complexity and task duration influence the information processing stages of response programming and response preparation (Schmidt, 1988), and therefore are likely to impact on quiet eye durations. The literature pertaining to each of these variables will be reviewed in the following section.

The Influence of Task Complexity

It has long been known that there is a direct relationship between human

performance capabilities and the informational load of a particular task or set of tasks (Fitts & Posner, 1967; Hick, 1952; Hyman, 1953). That is, as the level of response uncertainty (informational load) increases, so too does reaction time (RT). More importantly, laboratory research tends to indicate that RT to a single unanticipated visual stimulus is in the order of 180-220 ms, with this delay composed of latencies associated with stimulus detection, response preparation, and neural and muscular activity associated with a simple key press (e.g., Wood, 1977).

In numerous studies, greater durations of reaction time (RT) and preparation time are required to adequately perform more complex motor responses (e.g., Henry, 1980; Henry & Rogers, 1960; Kerr, 1978; Klapp, 1977, 1980; Schmidt, 1988). In Henry and









59


Rogers' (1960) seminal experiment, participants were required to perform a series of movements that varied in their level of complexity, from a simple finger lift to a single ball grasp (intermediate complexity) to a double ball strike (greatest complexity). Results indicated that RT and MT increased proportionally with increases in task complexity. Because the stimuli in each condition were held constant, Henry and Rogers (1960) argued that the more complex motor tasks required more time to program the movements in the response programming stage.

Concomitant increases in RT as a function of increasing task complexity have also been reported in studies in which key-pressing movements of different complexities have been used (Klapp, 1977; Klapp & Erwin, 1976; Klapp & Greim, 1979), in which subjects were required to produce a series of morse-code transmissions. More complex transmissions were characterized by longer RT's in the production of the movement than less complex tasks. Similar results have been observed by Sternberg, Monsell, Knoll, and Wright (1978) in their studies of both typed and spoken sequences of words. More complex sequences, it was hypothesized, required greater parameterization in the response programming stage, and therefore greater RT's were needed to complete these tasks.

In relation to manual aiming processes, both Falkenberg and Newell (1980) and Newell, Hoshizaki, Carlton, and Halbert (1979) used coincident timing tasks to assess the influence of increasing levels of complexity on RT. In particular, it was noted that as









60


movement velocities changed, so too did the participant's RT's. Thus, these researchers hypothesized that movement velocity was a key parameter in the programming of a motor response. Other parameters, such as direction and force of the movement, have been proposed to influence the programming time of a motor response to stimuli of varying complexity (Klapp & Greim, 1979; Spijkers, 1989; Spijkers & Steyvers, 1984; Temprado & Spijkers, 1996; Vickers, 1996).

In relation to the location aspect of Vickers (1996) hypothesis, it follows that if the quiet eye duration is directly related to a period of cognitive programming, more complex aiming behaviors should be characterized by longer quiet eye durations. To date, this relationship has never been empirically tested. Thus, one of the purposes in Experiment 1 of the present study was to investigate the influence of task complexity on quiet eye duration.

The Influence of Task Duration

Much in the same manner that changes in task complexity influenced RT's in the motor programming stage of information processing, researchers have demonstrated that changes in the duration in which the performer is required to execute a response impacts upon overall performance (Shapiro, 1977, 1978; Summers, 1977). Interestingly, although decrements in the precision of the response are observed, the relative temporal phasing and sequencing of the movement often remains constant (Gentner, 1987).









61


For example, Shapiro (1977) noted that the proportion of time spent in each phase of a learned sequence of wrist movements remained highly similar under time compressed trials and normal conditions. Carter and Shapiro (1984) have also documented this seemingly invariant temporal structure in electromyographic (EMG) measures of the phasing of relevant muscles when participants were required to speed up their gait in a running task.

This proportional timing phenomenon, in which movements are performed in

highly similar (although not exactly identical) relative phase durations, has been chronicled in such motor activities as typing (Terzuolo & Viviani, 1979), running (Shapiro, Zernicke, Gregor, & Diestel, 1981), jumping (Lee, Lishman, & Thomson, 1982), piano playing (Shaffer, 1980, 1984), and breathing (Clark & von Euler, 1972). Regardless of whether the required response was accelerated or extended in terms of overall movement time, phase ratios remained remarkably constant. Thus, it is expected that reductions in the time the performer is allowed to initiate a motor task, such as the billiards stroke, will not drastically alter the relative phases of the stroke. Indeed, the proportional duration model would predict that the preparation, backswing, foreswing, and contact phases of the stroke would not change in terms of their relative times to each other when constraints are placed on the time allotted to the execution of the shot.









62


However, in relation to the critical variable of quiet eye duration in Vickers (1996) theory, reductions in the time allotted for the participant to complete the task may reduce the time that he or she spends in the quiet eye phase, thus limiting the extent of cognitive programming and impairing performance. To date, however, no manipulations of task duration have been conducted in studies examining the relationship between quiet eye duration and performance in far aiming tasks. Therefore, this investigation was the focus of Experiment 2 in the present study.

In summary, examinations into the variables associated with expertise in motor activities has yielded mixed results. Typically, researchers have indicated that software components, such as visual tracking patterns, anticipatory capabilities, and declarative and procedural knowledge, are more predictive of athletic expertise than hardware components. However, much of this research is descriptive in nature, and does not adhere to the processes of stringent theory testing and development.

Vickers (1996) was one of the few researchers to develop a theory of visual

control when performers are required to aim at a far target. By examining the impact of variables (task complexity and task duration) known to influence the response programming stage of the human information processing system, the present study will assess whether Vickers' (1996) quiet eye variable is truly reflective of this underlying cognitive processing. In addition, an attempt was made to determine whether highly-









63


skilled billiards players differed from lesser-skilled players in their visual search processes as they executed their shots in a more ecologically valid testing environment.















CHAPTER 3
METHODS


Experiment 1

In this experiment, the complexity of the aiming task (shooting pool) was

manipulated to determine if accompanying changes in quiet eye duration would result. It has been posited that the quiet eye period reflects an index of cognitive activity, such that the parameters of the task to be performed are delineated and programmed (Vickers, 1996). Therefore, it was hypothesized that as the complexity of the task increases, so too would the concomitant period of cognitive programming (e.g., Henry & Rogers, 1960). Thus, quiet eye duration should be greater in more complex tasks versus less complex tasks. If this contention is supported, a stronger argument can be forwarded in favor of the eye-mind assumption of gaze behavior.

Participants

Twenty-four right-handed male participants were tested, and were categorized as either highly-skilled or lesser-skilled billiards players based on two criteria. These criteria included years of experience in the sport of billiards, as well as performance on an initial skills test. To be eligible for the highly-skilled category, players were required to have at least eight years of playing experience, including competition in at least one sanctioned billiards tournament. Members of this group were also required to complete the initial


64









65


skills test in 12 or fewer shots. Requirements for the lesser-skilled participants included less than three years of playing experience, no competitive tournament participation, and a score of 20 to 30 shots on the initial skills test. The initial skills test is described in more detail in the procedures section.

The highly-skilled group consisted of 12 players (M = 23.17 years of age, SD =

3.01 years) who had an average of 9.1 years of experience in the sport of billiards (SD =

2.64 years) and played 3.33 days per week on average (SD = 0.98 days). Members of this group also had competed in an average of 14.25 (SD = 14.97) sanctioned tournaments in their career, winning 2.25 of them (SD = 3.19). In addition to their experience, each of the 12 highly-skilled participants successfully completed the test in fewer than 12 shots (M = 10.67 shots, SD = 2.46 shots). The lesser-skilled group also consisted of 12 players (M = 21.83 years of age, SD = 1.85 years), averaging 2.67 years of playing experience (SD = 0.65 years). These participants had no competitive experience, played only 1.25 days per week (SD = 0.45 days) and scored an average of 26.67 shots (SD = 3.17 shots) to complete the initial skills test.

The number of participants used in this study was determined by Cohen's (1977) table for power and effect sizes. The values used in the tables were: a = .05 (level of significance), .1 = 18 [Skill Level (2-1) x Complexity Level (3-1) x Accuracy (2-1) x Trial (10-1)], f = .40 (effect size), and power = .80.









66


Apparatus

Eye Movement Recording

Eye movement data were collected with an Applied Sciences Laboratories (ASL; Waltham, MA) 4000SU mobile corneal reflection unit. This video-based monocular system measured the eye line of gaze with respect to a head-mounted eye camera by computing the relative positions of two features of the eye, the pupil and the corneal reflex (a reflection of a near-infrared light source from the surface of the cornea), in relation to the optics. Both the infrared beam and the image of the participant's eye were reflected from a visor mounted on the helmet, which was coated to be reflective in the near infrared region and transmissive to visible light. The line of gaze was computed by measuring the vertical and horizontal distances between the center of the pupil and the corneal reflection after correcting for second-order effects.

The resulting displacement data were recorded and processed by an external

Gateway 2000 486SX/166 microcomputer via a 30 m cable attached to the participant's waist. To record the field of view as observed by the participant, an Elmo MP481 color scene camera was positioned near the eye. This allowed for a view of the scene from the same position as would be observed by the participant, while avoiding the problems of parallax error and the mounting of stationary cameras in front of the subject. Since the participant was free to move about the environment, the use of the scene camera was









67


especially beneficial, as the scene constantly changed with each shift in the individual's position.

To assess the exact location of gaze, the 4000 SU processor superimposed a white cursor representing 1 deg of visual angle on the video image produced by the scene camera. These images were recorded via an Akai VS-X9EGN S-VHS video recorder, and used for data analysis. The 4000 SU possessed an accuracy of 1 deg in both the vertical and horizontal directions, and a precision of better than 0.5 deg. Finally, the system sampled at a rate of 60 Hz, and point of gaze was updated for each frame of video (every 33.3 ms).

Vision and Action Coupling

In an effort to more effectively couple vision with action, the participant's gaze

behaviors were recorded simultaneously with the action phases of the billiard stroke. This was done by interfacing the data from the 4000 SU system with an image recorded by an external Panasonic WV-PS03/B S-VHS video camera (positioned perpendicular to the direction of the arm motion of the player) using a Panasonic WJ-MXIO digital production mixer. The mixer created a split-screen effect, in which the lower left portion of the frame showed the participant performing the billiard stroke, while the right portion of the frame displayed his gaze position as recorded by the eye and scene cameras. This method of coupling vision and action has been successfully used by Vickers (1992, 1996) to









68


demonstrate the high degree of coordination between visual search patterns and motoric performance. Both the gaze locations and the temporal aspects of each phase of the stroke were then analyzed in a frame-by-frame fashion. Billiards Arrangements

The experimental conditions, including the billiards proficiency test of the highlyskilled and lesser-skilled participants, were performed on a Brunswick 4.5 ft x 9 ft (1.37 m x 2.74 m) billiards table, which had a playing surface area of 100 in x 50 in (254 cm x 127 cm). The corner pockets were 5 in (12.7 cm) in width, with the side pockets being slightly larger in width (5.5 in, 13.97 cm). Standard modern composition balls measuring 2.25 in (5.72 cm) in diameter and weighing 6 ounces (170 g) were used. To control for the possible confound of instrumentation (Campbell & Stanley, 1963), players were required to use a standard 57 in (144.78 cm) billiards cue, weighing 18 oz (510 g). The arrangement of balls for the initial skills test, as well as for each of the three tasks of varying complexity levels, was controlled by the experimenter by placing small colored markers on the felt surface of the billiards table.

Procedure

Initial Performance Test

Upon arrival at the Motor Behavior Laboratory, participants completed an

informed consent form and a short demographic questionnaire assessing variables such as age, years of billiards playing experience, and whether they had normal or corrected









69


vision. The players were then instructed as to what was expected of them throughout the course of the experiment. Once questions had been answered completely by the experimenter, each individual was given the billiards cue and led to the table, where an arrangement of balls could be seen similar to those that might occur in a typical game of nine-ball. This arrangement is illustrated in Figure 3-1. The participants were asked to pocket the balls (in order from one to nine) in as few shots as possible. The balls were arranged in such a fashion that high caliber players would be able to complete the initial test in less than 12 shots, while lesser-skilled participants would require more than 20 shots.

Performance Trials

Once the initial test was completed, the participants were fitted with the eye tracking system, and initial calibration procedures were performed. This calibration procedure consisted of having the person fixate on a sequence of nine equidistant points located on a target board placed on top of the billiards table. Upon completion of the calibration procedure, a series of shots of three levels of complexity (easy, intermediate, and hard) were performed. Each player was instructed on the correct type of speed and english to place on the cue ball, and three practice trials for each level of complexity were given. They then performed consecutive shots until they reach 10 successful and 10 unsuccessful outcomes per complexity condition, a research goal of which they were not












70













15 14 13 12 11 10







16 0 0 9



17 o 8




18 2*3 7







1 2 3 45


Figure 3-1. Initial performance test.









71


made aware. For each condition, performance was deemed successful only when the object ball was pocketed. Shots were considered unsuccessful if the object ball was not sunk or the cue ball was inadvertently pocketed (scratch shot).

After the participants had achieved 10 successful and 10 unsuccessful shots on a particular complexity level, a 3 min rest period was given, followed by a recalibration of the eye tracking system. The players then progressed to the next series of shots of a different complexity level. The sequence of task complexity was randomly assigned and counterbalanced between each participant (e.g., participant A progressed in a sequence of hard, easy, and intermediate shots, while participant B encountered a sequence of easy, intermediate, and hard shots). Total time to complete the task for each participant depended on the number of shots necessary to achieve 10 hits and misses. In general, each participant completed this experiment in 45 min or less. Shots of Varying Complexity

For the performance task, participants were required to pocket a series of balls

with shots of differing complexity. Specifically, three conditions of complexity were used. In the easy condition (EC), the object ball lay near the corner pocket, and could be made with a cut shot using bottom-left english (Martin & Reeves, 1993). This shot is illustrated in Figure 3-2. The intermediate condition (IC) featured a "cushion first" approach to sinking the object ball (Martin & Reeves, 1993), where the object ball lay near the cushion









72


while the cue ball sat on the opposite side of the table. The object ball was partially occluded by another ball, such that the player had no direct way of hitting the object ball directly. In order to sink the object ball the player needed to strike the cue ball with straight-follow english and hit the cushion prior to hitting the object ball. This shot is demonstrated in Figure 3-3.

Finally, the hard condition (HC) featured a "carom draw" shot, in which the cue ball needed to be struck sharply with straight-bottom english so that it caromed off the object ball into another ball located near the corner pocket (Martin & Reeves, 1993). The difficulty of the carom draw shot lies in the especially precise angle of the carom, and due to the fact that a number of other balls were located on the table limiting the options of the player. This shot is illustrated in Figure 3-4.

Given the fact that participants of differing skill level and experience were

performing these tasks, a number of alternative approaches to making the shots described above may be observed. In particular, it was expected that the highly-skilled players possessed the cognitive ability to select the most appropriate response for a given shot to a much greater extent than the lesser-skilled performers (Abernethy, Neal, & Koning, 1994). Therefore, instructions as to the most pertinent action that will lead to the correct execution of the shot were given prior to the start of the task.

























15


14


13


12 11 10


00 00 00


. S




.*_,


0 0)


,


K


0


1


0


2


3


6


5


0


9



8




7


6


Figure 3-2. Required shot in the easy complexity condition.


73


le 17




18


0


4


0


12


11


10










74













15 14 13 12 11 10





.6 9



18


18 .





2 3


Figure 3-3. Required shot in the intermediate complexity condition.











75








15 14 13 12 11 10










17 o I

e8 18 o 27

0

0 0000

1 2 3 46


Figure 3-4. Required shot in the hard complexity condition.









76


Dependent Measures

Performance

The performance measure consisted of the number of shots required to achieve the criterion level of 10 successful and 10 unsuccessful shots for each of the task complexity conditions. Specifically, the percentage of shots made to shots attempted was calculated, with higher values indicating greater performance. Visual Tracking Measures

For each participant, videotaped eye movement data was recorded for each shot. Although the participants were required to shoot until they made 10 successful and 10 unsuccessful shots of each complexity level, a review of the data indicated that some trials were unable to be analyzed due to calibration or video errors. Therefore, only five randomly selected successful and five unsuccessful shots for each level of complexity could be analyzed in a frame-by-frame fashion using an Akai VS-X9EGN S-VHS video recorder. Each frame constituted 33.3 ms of data, and a fixation was defined as three or more consecutive frames (99.9 ms or more) in which the cursor was located in the same space in the visual environment (e.g., Ripoll, 1991; Vickers, 1992, 1996; Williams et al., 1994). The variables of fixation location, average fixation duration, quiet eye duration, and the number of blinks present in each phase (the suppression aspect of the locationsuppression hypothesis) was then subjected to statistical analysis.









77


For fixation location, the number of fixations allocated to specific areas of the

visual display was recorded. In particular, fixations directed to four areas of interest were examined and subjected to statistical analysis: The cue ball, the object ball, the target cushion (or second object ball in the hard condition only), and the intended pocket of the object ball. Any remaining fixations (e.g., on the cue stick, felt surface, etc.) were defined in a category labeled "other" areas. For the dependent measure of average fixation duration, the time spent (in ms) allocating fixations to each of the predetermined locations of interest was calculated.

Of critical importance to this experiment, the quiet eye duration (in ms) in the

preparatory phase of the stroke was recorded and compared to differences in performance between the groups for each complexity condition. In addition, to test the Vickers (1996) visual suppression hypothesis, the number of eyeblinks as calculated by the absence of the cursor on the videotaped image, was assessed during the backswing, foreswing, and flight phases of the stroke.

Temporal Components of the Billiards Stroke

To assess the temporal components of the stroke, four distinct phases were

identified from the sagittal view of the performer as provided by the external camera: (1) the preparatory phase, defined as the time from the start of the task (when the performer positioned himself over the cue ball) until the first observable movement towards the









78


striking of the cue ball (e.g., the duration in which the billiards player leaned over the table to begin the shot until the moment he initiated the final backswing motion of the cue); (2) the backswing phase, defined as the duration from the final backswing movement until the cue began to move forward towards the cue ball; (3) the foreswing phase, defined as the time from foreswing initiation until the tip of the cue came into contact with the cue ball; and (4) flight phase, defined as the duration of the flight of the cue ball until it struck the object ball (e.g., includes cue ball contact with the object ball).

Total time (in ms) spent in each phase of the stroke between the highly-skilled and lesser-skilled performers was calculated for each of the varying complexity conditions. This was accomplished by analyzing the total number of video frames (each frame equated to 33.3 ms) corresponding to each phase of the stroke.

Experiment 2

In this experiment, the duration of the task was manipulated such that the participant performed a billiards shooting task under normal and time-constrained conditions. Again, the quiet eye duration was assessed, as were the temporal aspects of each phase of the aiming movement. It was hypothesized that quiet eye durations would be reduced because of the time constraints, thus leading to decrements in performance as compared to the normal condition. However, it was also hypothesized that the temporal ratios of each phase of the behavior would remain constant for each constrained condition as compared to the normal condition. Therefore, changes in performance may be explained









79


via changes in the time allotted for programming of the movement, rather than in the relative temporal changes in the movement itself. Again, it was expected that this result would provide support for the strong eye-mind assumption, indicating that constraints imposed on the cognitive processes underlying eye movement behaviors and the planning of the movement directly influence performance.

Participants

The individuals in this experiment were those who participated in Experiment 1. Again, the number of people to be tested in this study was determined by Cohen's (1977) table for power and effect sizes. The values used in the tables were: a = .05 (level of significance), u = 18 [Skill Level (2-1) x Duration (3-1) x Accuracy (2-1) x Trial (10-1)], f = .40 (effect size), and power = .80.

Apparatus

All apparati, including the eye tracking unit, video recording and mixing devices, and billiards table, cue, and balls were identical to those described in Experiment 1. However, only the arrangement of balls employed in the intermediate complexity condition

(IC) of the first experiment were included in Experiment 2.

In addition to these instruments, the countdown timing mechanism feature of the ReAction Coach (S.T.A.R.T. Technologies, NY) movement timing system was used to direct the participants as to the time they were allotted for the task. The ReAction Coach









80


is a portable electronic device (16 cm long and 12 cm wide) which uses auditory and visual cues to ostensibly promote quicker and more explosive motoric abilities (Singer, Cauraugh, Frehlich, Barba, Kelly, & Umans, 1993). Auditory cues can be made available to the participant in the form of a series of beeps, and these cues can be adjusted in volume and sensitivity by the experimenter. In particular, the countdown mode of the ReAction Coach was used in Experiment 2, in which three preliminary auditory beeps were presented (priming cues indicating that the trial is imminent), followed by a tone of longer duration (indicating the initiation of the trial). Once this tone sounded, the internal timer in the device was activated, and was terminated only when a loud sound was generated by the participant (i.e., the striking of the cue ball with the billiard cue). Durations (in ms) between onset of the final tone and termination of the timer were displayed on the liquid crystal display (LCD) function of the ReAction Coach, and were used to provide feedback to the participants.

Procedure

Upon completion of Experiment 1, participants were given instructions as to the goals and nature of the second experiment. Specifically, they were required to perform the IC complexity task in a randomly assigned and counterbalanced order of two different conditions. In the 25% constrained condition, the shooter was required to initiate their stroke within a duration of 75% of the average time spent in the initiation of the stroke as recorded in an unconstrained condition. The unconstrained condition times were assessed









81


during data analysis in Experiment 1, where the participant was allowed to take as much time as he wished to perform the IC task. For example, if a particular participant took an average of 4 sec to execute the shot in the unconstrained IC condition, he was required to perform the shot in 3 sec or less in the 25% constrained condition. The participant received three practice trials preceding the test trials.

In the 50% constrained condition, the player needed to execute the shot within a duration of 50% of the average time recorded to initiate the shot in the unconstrained condition. To use the same example, a player averaging 4 secs per shot in the initial testing session (IC condition in Experiment 1) had to initiate the shot within 2 sec or less in the 50% constrained condition. Again, three practice trials were given to the participant before test trials were recorded.

To aid the participants in determining when they would be allowed to initiate a

stroke, the countdown timing mechanism feature of the ReAction Coach movement timing system was used. Players stood in a ready position over the shot and closed their eyes, and were informed that an experimental trial was beginning upon hearing the initial warning tones produced by the ReAction Coach. After hearing the final auditory cue, the timer was activated and the participants were allowed to open their eyes and begin their preparatory motions used to execute the shot. Based on the times recorded by the ReAction Coach measuring device, the experimenter provided feedback to the participant after each trial as









82


to whether they were initiating the stroke within the allotted time frame. The players performed consecutive shots until they reach 10 successful and 10 unsuccessful outcomes per duration condition. Again, this was a research goal of which they were unaware. Trials initiated after the constrained period were considered as inaccurate, and were not included in the data analyses.

Dependent Measures

The dependent measures were recorded and analyzed in an identical fashion to

those described in Experiment 1. In particular, performance percentages, durations (in ms) spent in each phase of the billiard stroke, and the visual tracking measures of fixation location, average fixation duration, quiet eye duration, and the number of blinks present in each phase were assessed.















CHAPTER 4
RESULTS


Experimental Design

Statistical analyses were conducted with the SPSS for Windows 6.1.4 computer

applications package. For all statistical analyses in each of the two experiments, alpha level was set at p<.05. Simple main effects and Scheffd's post hoc tests were used to follow up significant analysis of variance (ANOVA) results when appropriate. In addition, for each repeated measures ANOVA design, violations of the assumptions of sphericity were assessed. In those instances when the assumptions were violated, Greenhouse-Geisser adjustments to the level of significance are reported.

Experiment 1

Statistical Analyses

Performance

The performance percentage was determined by dividing the number of shots made by the total number of shots taken to reach 10 hits and 10 misses for each group at each level of complexity. A 2 (Skill Level) x 3 (Complexity Level) ANOVA with repeated measures on the last factor was conducted to assess whether performance differed significantly between the highly-skilled and lesser-skilled participants, and indicated main


83









84


effects for Skill Level, F(1,22) = 75.99, p < .001, and Complexity Level, F(2,44) = 116.28, p <.001. Post hoc analyses revealed that highly-skilled players made significantly more shots (M = 67.53%, SD = 18.17) that did lesser-skilled players (M = 47.88%, SD = 21.24). In addition, there were significant reductions in performance for both groups as the level of complexity increased, from M = 74.13% (SD = 12.99) for the EC condition, to M = 65.01% (SD = 14.21) for the IC condition, and M = 47.88% (SD = 13.96) in the HC condition. It appears as though the highly-skilled players sank significantly more shots in all levels of complexity than did lesser-skilled players. Also, the manipulations in complexity affected members of both groups in a proportionate manner. None of the interactions were significant. Overall performance for each group is presented in Table 41.



Table 4-1. Performance means and standard deviations for highly-skilled and lesser-skilled participants for each level of complexity.

Complexity Highly-Skilled Highly-Skilled Lesser-Skilled Lesser-Skilled
M SD M SD
Easy 83.30% 6.74 64.97% 11.15
Intermediate 74.84% 6.17 55.19% 13.19
Hard 44.47% 7.52 23.47% 10.52









85


Temporal Components of the Billiards Stroke

To determine the effect of changes in complexity on the overall duration (in ms) of the billiards stroke, a 2 (Skill Level) x 3 (Complexity Level) x 2 (Accuracy) x 5 (Trials) ANOVA with repeated measures on the last three factors was conducted. Significant main effects were found for Accuracy, E(1,22) = 4.78, p< .05, and for Complexity Level, E(2,44) = 78.17, p < .001. Accurate shots were characterized by longer durations in both groups (M = 3093.06 ms, SD = 898.23) as compared to missed shots (M = 2885.69 ms, SI= 752.84). In addition, overall shot duration time increased significantly for both groups as complexity level increased (EC M = 2489.38 ms, SD = 615.33; IC M = 2656.67 ins, SD = 540.6; and HC M = 3822.08 ms, SQ = 600.44). No significant differences were found for Skill Level, as highly skilled mean duration was 2977.08 ms (SD = 870.7) and lesser-skilled mean duration was 3001.67 ms (SD = 798.1).

More importantly, a significant interaction between Complexity Level and Skill

Level was observed, F(2,44) = 5.33, p < .05. Within each skill level, total duration for the HC condition was significantly greater than either the EC or IC conditions. The overall stroke duration in the EC and IC conditions were not significantly different from each other. This interaction is displayed in Table 4-2.









86


Table 4-2. Total duration of the billiards stroke for highly-skilled and lesser-skilled participants for each level of complexity.

Complexity Highly-Skilled Highly-Skilled Lesser-Skilled Lesser-Skilled
_____DMM SD
Easy 2322.50 ms 435.6 2656.25 ms 795.1
Intermediate 2587.50 ms 511.3 2725.83 ms 569.9
Hard 4021.25 ms 568.2 3622.92 ms 632.7


From these results, it appears as though accuracy of the shot for both groups was related to the time taken to perform the shot. On average, successful shots were characterized by a 207.4 ms increase in total duration as compared to unsuccessful shots. In addition, for each level of complexity, time taken to perform the shot increased significantly for both groups in the most complex condition. Phase Durations

To determine whether manipulations in task complexity led to specific changes in duration for each of the four phases of the billiards stroke, a 2 (Skill Level) x 3 (Complexity Level) x 4 (Phase) x 2 (Accuracy) x 5 (Trials) ANOVA with repeated measures on the last four factors was conducted. A significant main effect was found for the variables of Phase, F(3,66) = 12.73, p<.01, and Complexity Level, F(2,44) = 5.65, p_< .05. Both groups spent significantly more time in the preparation phase of the stroke (M = 1731.46 ms, SD = 632.5) than in any of the other phases. Means for the other phases included: backswing phase (M = 491.84 ms, SD = 40.7), foreswing phase (M = 125.32









87


ms, SD = 9.27), and flight phase (M = 597.87 ms, SD = 65.9). In addition, the HC condition led to the greatest mean phase time (M = 955.52 ms) than did the EC (M = 603.85 ms) or IC (M = 650.50 ms). No other main effect for Accuracy, Skill Level, or Trial was significant.

More importantly, a significant Phase by Complexity Level interaction was observed, F(6,132) = 4.32, p < .03. Follow-up tests indicated that the backswing, foreswing, and flight phases for each level of complexity were similar in duration. However, durations of the preparation phase were significantly greater (M = 2532.16 ms) in the HC condition as compared to the EC or IC conditions (M = 1240.72 and M = 1421.52 ms, respectively). Thus, it appears as though both groups required significantly more time to prepare their shots in the most complex condition in Experiment 1 as compared to the other two levels of complexity.

Interestingly, a Skill Level by Phase interaction was not observed to be significant. That is, highly-skilled players spent a similar amount of time in each of the four phases of the stroke as compared to their lesser-skilled counterparts. Although not significant, the average phase durations collapsed across the three complexity levels for both skill levels are presented in Table 4-3.









88


Table 4-3. Total duration for each phase of the billiards stroke for highly-skilled and lesser-skilled participants collapsed across complexity level.

Phase Highly-Skilled Highly-Skilled Lesser-Skilled Lesser-Skilled
M SD M SD
Preparation 1763.93 ms 432.50 1698.99 ms 832.54
Backswing 490.02 ms 33.33 493.67 ms 48.05
Foreswing 120.00 ms 9.56 130.63 ms 10.73
Flight 566.68 ms 103.17 629.07 ms 245.27


It should be noted that a comparison of mean durations for each phase did not

differentiate successful from unsuccessful shots. None of the interactions except the Phase by Complexity Level interaction were significant. Overall, it appeared as though time spent in each phase of the billiards stroke did not differ between highly-skilled and lesserskilled players, and could not be used as a means of explaining shots that were successful as compared to shots that were unsuccessful. Although the more skilled players spent more time in the preparation phase for each level of complexity, these differences were not statistically significant.

Visual Tracking Measures

To assess whether changes in task complexity were associated with concomitant changes in visual tracking measures, the dependent variables of the number of fixations and average fixation duration (in ms) were analyzed via separate 2 (Skill Level) x 5









89


(Location) x 2 (Accuracy) x 5 (Trials) ANOVA's with repeated measures on the last three factors for each level of complexity.

Easy complexity level. For the EC condition, only three levels of the Location factor were activated in the study. Therefore, a 2 (Skill Level) x 3 (Location) x 2 (Accuracy) x 5 (Trials) ANOVA with repeated measures on the last three factors was conducted for this condition only. Significant main effects for Location, F(2,44) = 579.48, p < .001, and Skill Level, F(1,22) = 54.35, p <.01 were observed for the dependent variable of number of fixations. Specifically, more fixations were directed to the cue ball (M = 3.00) and the object ball (M = 2.83) than to the cushion (M = 0.00), the pocket (M = 0.00), or other areas of the display (M = 0.50). In addition, lesser-skilled players made more fixations (M = 1.48) on average to each location than did highly-skilled players (M =

1.06).

A significant interaction between Location and Skill Level was also observed for the number of fixations in the EC condition, F(2,44) = 20.44, p <.01. Lesser-skilled participants made significantly more fixations to the cue ball (M = 3.54) and object ball (M = 3.33) than did the highly-skilled participants (M = 2.46 and M = 2.33 fixations, respectively).

A similar pattern of results was observed for the dependent variable of average fixation duration in the EC condition. Significant main effects for Location, E(2,44) =









90


204.87, p < .001 indicated that greater average fixation durations were directed to the cue ball ( M = 396.02 ms) and the object ball (M = 418.26 ms) than to any of the other locations. The main effect for Skill Level was also significant, E(1,22) = 7.57, P < .05. On average, the highly-skilled performers made fixations of longer duration (M = 197.47 ms) than did less-skilled players (M = 161.24 ms).

More importantly, a significant interaction between Skill Level and Location was found, F(2,44) = 6.41, p < .05. Post hoc tests revealed that highly-skilled performers fixated on the target ball significantly longer (M = 498.96 ms) than did the lesser-skilled players (M = 337.57 ms). The overall pattern of results for both dependent variables fixation number and average fixation location for the EC condition is summarized in Table 4-4.



Table 4-4. Mean number of fixations and average fixation durations for highly-skilled and lesser-skilled participants for the EC condition. Location Highly-Skilled Highly-Skilled Lesser-Skilled Lesser-Skilled
Fix. Num. Avg. Fix. Dur. Fix. Num. Avg. Fix. Dur.
Cue Ball 2.46 418.82 ms 3.54 373.23 ms
Object Ball 2.33 498.96 ms 3.33 337.57 ms
Cushion 0.00 0.00 ms 0.00 0.00 ins
Pocket 0.00 0.00 ms 0.00 0.00 ms
Other 0.50 69.58 ms 0.50 95.42 ms









91


From these results, it can be inferred that highly-skilled players made fewer fixations of longer duration to specific areas of the display as compared to the lesserskilled participants. In particular, the more advanced group fixated longer on the object ball than did the less advanced group. However, these differences did not explain the discrepancy between successful versus unsuccessful shots, because neither a main effect nor any interaction was found for the independent variable of accuracy. Thus, for the EC condition, no differential pattern of visual tracking could be determined for successful and unsuccessful shots.

Intermediate complexity level. For the IC condition, significant main effects in the dependent variable of number of fixations were observed for Location, E(4,88) = 53.01, P <.001, and for Skill Level, E(1,22) = 22.56, p < .01. Follow-up tests revealed that lesserskilled players demonstrated a greater number of fixations (M= 1.49) on average than did their higher-skilled counterparts (M = 1.18), and for both groups more fixations were directed to the cue ball (M = 2.65), the cushion (M = 2.08), and the object ball (M = 1.33) than to the pocket (M = 0.25) or other areas of the display (M = 0.35).

In addition to these main effects, a significant Location by Skill Level interaction was found, E(4,88) = 5.18, p < .02. The lesser-skilled players directed significantly more fixations to the cue ball (M = 3.00) than did the highly-skilled players. Number of fixations




Full Text

PAGE 1

QUIET EYE DURATION AS AN INDEX OF COGNITIVE PROCESSING: THE EFFECT OF TASK COMPLEXITY AND TASK DURATION ON VISUAL SEARCH PATTERNS AND PERFORMANCE IN HIGHLY-SKILLED AND LESSER-SKILLED BILLIARDS PLAYERS By SHANE G. FREHLICH A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1997

PAGE 2

ACKNOWLEDGMENTS I would like to express my deepest of appreciations to the many individuals who helped me in the development of this manuscript over the course of the past five years. To my committee chair, Dr. Robert Singer, I wish to heartily thank you for your mentoring efforts in helping me achieve my goals. From you, I have learned to appreciate and respect the values of commitment, desire, and dedication. The spirit and enthusiasm for life that you possess has provided a greater inspiration to me than you can ever know, and I hope to pass these values on to my own students someday. I would also like to extend my sincere gratitude to Dr. James Cauraugh for teaching me the joys of the scientific process, to Dr. Milledge Murphey for keeping my views grounded in the applied perspective, and to Dr. Ira Fischler for the many cognitive insights. You have all guided me towards the fulfillment of a significant goal in my life, and words cannot express how important each of you are to me. A special debt of gratitude is owed to Dr. Mark Williams, who graciously provided much of the equipment used in this study. I would also like to thank Doug and Laura Barba, Chris and Carol Janelle, Suzanne Broch, Lisa Pugliese, Paul Hubman, Sean Walsh, and Hjalmer Setzer for giving me all the love, help, and friendship anyone could ask for. Finally, my deepest and most heartfelt thanks go out to my family, Gary, Patricia, and Michelle Frehlich. Everything I have ever achieved in my life has been due to your unending love and support, and for that I will always be grateful. u

PAGE 3

TABLE OF CONTENTS page ACKNOWLEDGMENTS ii LIST OF TABLES v LIST OF FIGURES vii ABSTRACT viii CHAPTERS 1. INTRODUCTION 1 Eye Movement Research in Sport 5 Two Critical Questions Arise 7 Theoretical Approaches Emerge 10 Statement of the Problem 13 Hypotheses 14 Assumptions 22 Limitations 23 Definition of Terms 24 Significance of Study 28 2. REVIEW OF LITERATURE 32 Expertise in Sport 33 The Visual Search Paradigm in Sport Research 38 Theories of Oculomotor Control in Aiming Tasks 51 The Influence of Task Complexity 58 The Influence of Task Duration 60 iii

PAGE 4

3. METHODS 64 Experiment 1 64 Participants 64 Apparatus 66 Procedures 68 Dependent Measures 76 Experiment 2 78 Participants 79 Apparatus 79 Procedures 80 Dependent Measures 82 4. RESULTS 83 Experiment 1 83 Statistical Analyses 83 Experiment 2 98 Statistical Analyses 98 5. DISCUSSION, SUMMARY, CONCLUSIONS, AND IMPLICATIONS FOR FUTURE RESEARCH 1 13 Experiment 1 114 Experiment 2 132 Summary 142 Conclusions 145 Implications for Future Research 147 REFERENCES 152 BIOGRAPHICAL SKETCH 169 IV

PAGE 5

LIST OF TABLES Table page 41 Performance means and standard deviations for highly-skilled and lesser-skilled participants for each level of complexity 84 4-2 Total duration of the billiards stroke for highly-skilled and lesser-skilled participants for each level of complexity 86 4-3 Total duration for each phase of the billiards stroke for highly-skilled and lesserskilled participants collapsed across complexity level 88 4-4 Mean number of fixations and average fixation durations for highly-skilled and lesser-skilled participants for the EC condition 90 4-5 Mean number of fixations and average fixation durations for highly-skilled and lesser-skilled participants for the IC condition 92 4-6 Mean number of fixations and average fixation durations for highly-skilled and lesser-skilled participants for the HC condition 95 4-7 Mean quiet eye durations for successful and unsuccessful shots for both highly-skilled and lesser-skilled participants for each level of complexity 97 4-8 Performance means and standard deviations for highly-skilled and lesser-skilled participants for each duration level 99 4-9 Total duration of the billiards stroke for highly-skilled and lesser-skilled participants for each duration level 101 4-10 Mean number of fixations and average fixation durations for highly-skilled and lesser-skilled participants for the 25% constrained condition 106 v

PAGE 6

4-11 Mean number of fixations and average fixation durations for highly-skilled and lesserskilled participants for the 50% constrained condition 109 4-12 Mean quiet eye durations for successful and unsuccessful shots for both highly-skilled and lesser-skilled participants for each duration level Ill vi

PAGE 7

LIST OF FIGURES Figure page 3-1 Initial performance test 70 3-2 Required shot in the easy complexity condition 73 3-3 Required shot in the intermediate complexity condition 74 3-4 Required shot in the hard complexity condition 75 5-1 Quiet eye durations for each group and level of task complexity 126 5-2 Quiet eye durations for each group and level of task duration 139 vn

PAGE 8

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy QUIET EYE DURATION AS AN INDEX OF COGNITIVE PROCESSING: THE EFFECT OF TASK COMPLEXITY AND TASK DURATION ON VISUAL SEARCH PATTERNS AND PERFORMANCE IN HIGHLYSKILLED AND LESSER-SKILLED BILLIARDS PLAYERS By Shane G. Frehlich December 1997 Chairman: Robert N. Singer, Ph.D. Major Department: Health and Human Performance The role of visual control mechanisms in the production of manual aiming tasks has been examined recently in a variety of motor tasks. Using an expert-novice research paradigm, researchers have investigated the extent to which performers of different skill levels differed in the visual search strategy used to execute a response in a particular sport situation. The purpose of this study was to investigate potential expertise differences in visual search strategy profile between twelve highly-skilled (M = 9.1 years of playing experience) and twelve lesser-skilled (M = 2.67 years) billiards players while they performed a series of strokes in an ecologically valid environment. Specifically, the dependent measures in this study included performance outcome percentage, stroke and phase duration, number of fixations, average fixation duration, quiet eye duration, and number of eye blinks. The latter two variables were proposed to play key roles in the control of visual attention in VickersÂ’ (1996) location-suppression hypothesis. A second vm

PAGE 9

purpose of this study was to determine whether quiet eye duration and number of eye blinks were related to performance between groups, and within participants of each skill level In Experiment 1, shots of three different levels of complexity (easy, intermediate, and hard difficulty levels) were presented to the performers to determine whether accompanying changes in visual search patterns would occur. In Experiment 2, the participants executed shots under three different time constrained conditions (unconstrained, 25% constrained, and 50% constrained). Results of both studies indicated that the highlyskilled players possessed visual search strategies that were more efficient, as they made fewer fixations of longer duration to their target while lesser-skilled players made significantly more fixations of shorter duration to their target. Only partial support was provided for VickersÂ’ (1996) location-suppression hypothesis. The quiet eye variable was found to be significantly related to performance between and within the two groups, with the highly-skilled players demonstrating significantly greater durations than the lesser-skilled players throughout every condition of task complexity and task duration. For both groups, successful shots were characterized by significantly longer quiet eye durations than unsuccessful shots. The number of eye blinks did not differ between the groups. Thus, only the location aspect of the locationsuppression hypothesis was supported, and it was argued that the quiet eye duration represented a critical period of cognitive programming in the aiming response. IX

PAGE 10

CHAPTER 1 INTRODUCTION Due to its highly complex and often unpredictable nature, the sporting arena provides an excellent context in which to test the assumptions of the information processing model of human behavior. For example, in fast ball sports such as baseball and tennis, athletes often have as little as 200-300 ms in which to make decisions based on the intended direction and speed of the approaching ball (Hyllegard, 1991; Slater-Hammel & Stumpner, 1950, 1951), as well as to organize and execute the most appropriate movement to effectively coincide with the object in motion. In addition, such sports as billiards, darts, archery, and pistol shooting require the ability to program exquisitely precise aiming movements at the neural level (Vickers, 1996). In fact, errors of less than one millimeter can often be the difference between success and failure in these motor activities. In each of these sports, the processing of visual information has been shown to be a key ingredient to the successful execution of appropriate motor responses. In fact, it has been argued that at least 80% of the information an athlete receives and attends to in the sport environment is visual in nature (Schmidt, 1988). Indeed, the importance of visual 1

PAGE 11

2 information, and its role in the human information processing system, has been the focus of a large body of research in cognitive psychology. In recent years, scholars have attempted to determine the extent to which eye movement data are indicative of underlying cognitive processing (e.g.. Just & Carpenter, 1976; Neisser, 1967; Posner, 1980; Viviani, 1990; Wright & Ward, 1994; Yarbus, 1967). Typically, the process of inductive reasoning has been used, such that stages of cognitive processing have been inferred to exist based on the observed pattern of eye movements while the individual engages in a task influenced by events or stimuli present in the visual environment. For example, early investigations of reading tasks in which a moving windows paradigm was used demonstrated decrements in word recognition and comprehension performance when the window moved at faster speeds (e.g.. Just & Carpenter, 1976). This seemed to suggest that time constrained covert stages of information processing, such as stimulus identification, could be marked or identified by the overt eye movements of the individual Thus, early research appeared to confirm the notion that higher-order cognitive processes, in essence, controlled the location and duration of ocular fixations. To incorporate an analogy forwarded by Viviani (1990), the eyes were believed to be similar to television cameras in that they were "directed" to the important aspects of the visual environment by specific stages of cognitive processing. In fact, much of the research

PAGE 12

3 in the study of eye movements has, either implicitly or explicitly, adhered to this belief in a strong "eye-mind" connection. However, the extent to which eye movements are in fact related to more global cognitive processing (a strong eye-mind view) has been questioned by a number of theorists, with several lines of research indicating a significant degree of independence between eye movement data and cognitive activities (e.g., Fisher, Karsh, Breitenbach, & Barnette, 1983; Sigman & Coles, 1980; Viviani, 1990). For example, Viviani (1990) provided a cogent argument against what he termed the "central dogma" of eye movement research, which states that "eye movements can at the very least be considered as tags or experimentally accessible quantities that scientists can observe to understand underlying processes of cognition" (p. 354). Coupled with the central dogma are two underlying assumptions. First, visual scanning patterns are believed to be the overt manifestation of strictly serial or sequential cognitive processes, such that subsequent visual stimuli presented to the individual cannot be adequately attended to until previously presented stimuli have been processed. Second, it is assumed that the direction of an individual's line of sight coincides with the direction and allocation of attentional resources (Viviani, 1990). In reviews of literature in which eye movement processes have been investigated, both Viviani (1990) and Theeuwes (1993) provide evidence suggesting that the role of

PAGE 13

4 cognition may not be so clearly inferred from eye movement data. Numerous investigations have refuted the assumptions underlying the central dogma. For example, the area of the visual field where focal attention is concentrated may occasionally be dissociated from the foveal field (Klein, 1980, 1994). Under appropriate conditions of precuing, it appears as though visual attention can be directed almost anywhere in the visual field, irrespective of the actual direction of the line of sight (Posner, 1980; Posner & Cohen, 1980; Remington, 1980; Shaw, 1983). In addition, the serial nature of visual tracking has been challenged by more recent theories of cognition (such as Parallel Distributed Processing models), which claim that a number of concomitant visual processes occur in parallel, and are hierarchically coordinated (Marr, 1982; Rumelhardt & McClelland, 1986). Thus, the results of several eye movement studies may be more readily explained with a bottom-up or reflexive approach to behavior, rather than providing evidence for underlying cognitive structures (Theeuwes, 1993). Although a number of studies have not provided explicit support for the central dogma, Viviani (1990) conceded that the observation of expert-novice differences in the scanning of a particular visual field does serve as evidence that some degree of cognitive processing is associated with specific patterns of eye movements. In a variety of cognitive and motor tasks, skilled experts have been observed to immediately concentrate fixations on pertinent details of a relevant image, whereas novices scan the image in a random, uniform manner. This result has been documented in skilled versus non-skilled chess

PAGE 14

players (de Groot, 1978), automobile drivers (Mourant & Rockwell, 1972), and radiologists (Kundel & Nodine, 1983). 5 Taken together, the results of these studies tend to support the notion that eye movement patterns may indeed reflect specific stages of cognitive processing that allocate attentional resources to particular visual stimuli. Given the dynamic and often timeconstrained nature of athletic competition in which attentional resources are frequently strained to their limits (Abemethy, 1993; Nideffer, 1993; Rotella & Lemer, 1993), it was only a matter of time before investigators began to conduct eye movement studies in sport. Eve Movement Research in Sport Implicitly adhering to the strong "eye-mind" connection, in which eye movements were interpreted to reflect higher order mental processing (e.g., Gardner, 1985; Just & Carpenter, 1976, 1980), researchers interested in the selective attention of athletes of differing skill levels began to focus on the visual search patterns of performers in sport task situations in the mid1970's (e.g., Bard & Fleury, 1976, 1981; Bard, Fleury, & Carriere, 1975; Haase & Mayer, 1978). Specifically, these scholars have attempted to outline the gaze behaviors and visual tracking patterns that differentiate experts and novices in their ability to perceive and use vital information in the sport environment (e.g., Abemethy, 1990b, 1991; Goulet, Bard, & Fleury, 1989; Helsen & Pauwels, 1990; Ripoll,

PAGE 15

6 1984; Shank & Haywood, 1987; Singer, Cauraugh, Chen, Steinberg, & Frehlich, 1996; Vickers, 1992; Williams, Davids, Burwitz, & Williams, 1994). In the majority of these studies, experts as compared to novices made fewer fixations to achieve successful response outcomes and exhibited lower search rates for sport-specific tasks. Also, experts made a greater number of fixations to the pertinent cues in the visual array than did novices, and had search rates that were typically more systematic and consistent (Abemethy, 1988). From these results, it would appear that successful performance in the highly-skilled was determined in part by where they looked in the sport environment. That is, experts seemed to possess a more efficient process of visual search, such that they attended to only the most important aspects of the sport situation while ignoring irrelevant stimuli. Unfortunately, while a preponderance of the work described above supported the contention that visual search patterns adequately discriminate levels of expertise, more contemporary researchers have questioned the utility of this paradigm in sport research. The reason for this skepticism is quite straightforward; the expected pattern of expertnovice differences has not been documented in several recent studies, thus raising the question of whether visual behavior is indeed a critical factor in expert performance. In other words, similar patterns of ocular fixations have been observed to be directed to particular aspects of the visual display by both elite and non-elite athletes, such that the caliber of skill possessed by the performer could not be adequately discriminated

PAGE 16

7 solely on the basis of where he or she looked in the sport environment (Abemethy, 1990a, 1991; Abemethy & Russell, 1987b). In addition, scholars have argued that statistically significant differences in fixation patterns observed between highly-skilled and lesserskilled athletes are likely of little practical value in explaining the differences in decisionmaking time between the two groups (Singer et al„ 1996). Two Critical Questions Arise Given the equivocal findings that have been observed in comparing visual search patterns of expert and novice athletes over the past two decades, it appears as though research in which this paradigm has been used has come to a critical juncture in its development. In particular, two fundamental questions must be addressed for this paradigm to be of utility in describing and explaining athletic performance (Abemethy, 1991; Vickers, 1996). First, exactly why have some researchers reported expertise differences in visual search patterns, while others have not? And, perhaps more importantly, are visual search processes related in some manner to expert performance, or are they of no consequence at all? The Need For Ecological Validity In addressing the first question, one possible explanation for the lack of consensus lies in the nature of the task performed by the individual under investigation. Both Ripoll (1991) and Vickers (1996) have noted that those investigations reporting no expertise

PAGE 17

8 differences in visual search rates typically require athletes to passively view photographic or videotaped displays (usually in a comfortable, seated position), and to make contrived motoric responses to these stimuli (e.g., pressing a button with the index finger; Abemethy, 1990b, 1991; Abemethy & Russell, 1987b; Singer et aL, 1996). In contrast, studies that require athletes to physically move around in their environment in a manner similar to that encountered in their sport while their gaze behaviors are recorded have demonstrated significant expert-novice differences in ocular fixation location and duration measures (Helsen & Pauwels, 1993; Ripoll, Bard, & Paillard, 1986; Ripoll, Papin, Guezennec, Verdy, & Philip, 1985; Vickers, 1992, 1996). Similarly, those experiments that have been conducted in real world settings, as opposed to using videotaped simulations, have also exhibited significant differences in search rates between highly-skilled and lesser-skilled performers (Ripoll et al., 1986; Vickers, 1992). In addition, intra-expert differences have also been demonstrated using this protocol (Vickers, 1996). While the distinction between real-world and artificial experimental settings has been targeted as a possible explanation for equivocal results in visual search research, the question remains: Why does the setting make a difference to the outcome of the study? One answer to this question comes from an ecological approach to the study of human behavior, in which perception-action theorists argue that behavior can only be understood by the kinematic coupling between the human biomechanical system and information from

PAGE 18

9 the environment (Bootsma, 1992; Davids, Handford, & Williams, 1994; Turvey, 1990). That is, these theorists argue that information from the environment is used "on line" by the performer to regulate motor activity, and that any investigation conducted in an artificial setting limits the ability of the individual to access this valuable information when engaging in decision-making processes (Davids et al., 1994; Turvey, 1992). This stream of ecological psychology reacted strongly against the behaviorist and cognitive notion that the relationship between the person and the environment is contingent (Gibson, 1979), and contends that behavior is not solely due to processes under the control of the performer (van Wieringen, 1986, 1988). Rather, specific visual information available in the sport environment (visual flow fields) is equally vital to the production of optimal performance, because it is this information that allows for the possibility of different responses in the face of changing stimuli (the concept of affordances; Gibson, 1979; Williams, Davids, Burwitz, & Williams, 1992). Thus, the perceivable affordances from an object in the environment invite a specific action within a particular context from the performer, and this action is in turn dependent on the biomechanical characteristics of the individual (Davids et al., 1994). These concepts of affordances and the symbiotic relationship between the person and the environment, then, may help explain why studies requiring movement of the participant in the actual sport environment document strong expert-novice differences.

PAGE 19

10 while contrived studies using limited (e.g., two-dimensional) visual stimuli and simple motoric responses sometimes do not. In the present study, performers were tested under conditions highly similar to that encountered in the sport of billiards, and executed strokes identical to those required in real-world settings. The Need For Theory A second explanation for the discrepant pattern of results observed in sportspecific research, and one that impacts on each of the critical questions discussed previously, is that these studies were descriptive in nature and were not conducted within a theoretical framework generating meaningful hypothesis testing. Further, investigations in which the expert-novice paradigm in sport has been used may be characterized as solely data-driven as opposed to theory-driven (Abemethy, Thomas, & Thomas, 1993; Vickers, 1992). Interestingly, this is a criticism that Viviani (1990) also levels at investigations of eye movements in more basic cognitive psychology tasks. Only recently has a call has gone out for more theoretically based, deductive approaches of hypothesis testing in the study of visual search (Ripoll et al., 1986; Vickers, 1996). Unfortunately, to date few theories have been developed in an attempt to explain oculomotor behavior from a motor control standpoint. Theoretical Approaches Emerge In reaction to these criticisms, Vickers (1996) proposed a location-suppression hypothesis for aiming behaviors to far targets (beyond arm's length). Far aiming tasks, she

PAGE 20

11 contends, are highly prevalent in such sport tasks such as baseball pitching, basketball free-throw shooting, archery, pistol shooting, bowling, and billiards. In the location aspect of her hypothesis, Vickers (1996) proposes that fixations of relatively long durations must be made to specific target locations during the preparatory phase of the movement. The extended duration of fixation in this phase is essential for successful performance because it is believed that the programming of the parameters of the movement occur in this time period. That is, the parameters of the location and distance of the target are programmed, as are the forces, timing, and coordination of limbs necessary to produce the optimal movement. As the movement is first initiated (the impulse phase), relatively slow movements are required to maintain fixation on the target and complete the final structuring of the aiming commands. During the execution phase of the movement, Vickers (1996) posits that a suppression mechanism is used to reduce interfering visual information that may result as a consequence of the aiming movement (e.g., hands and ball appearing in front of the eye during a basketball free-throw, thus occluding the target). Expert performers, she contends, have developed the ability to divert their visual attention from the target during execution by blinking or orienting their gaze to other elements of the visual field. Since the parameters of the movement have already been planned in the preparation phase, and

PAGE 21

12 modified in the impulse phase, visual attention is deemed unnecessary in the execution phase for successful completion of the aiming task. Two critical components related to performance are outlined in Vickers' (1996) location-suppression hypothesis. The first of these deals with the duration of fixation upon the target during the preparatory phase of the movement. Vickers (1996) termed this duration quiet eye, and defined it as that portion of the final fixation from onset to the first observable movement of the aiming limb. The longer the duration of quiet eye, it was hypothesized, the greater the performance in aiming to a far target. It is in this time period, Vickers (1996) contends, that the performer sets the final parameters of the movement to be executed. The key principle, she notes, is that quiet eye duration is directly associated with the amount of cognitive programming required for successful aiming to a target. In numerous studies, it has been demonstrated that greater durations of reaction time and preparation time are required to adequately perform more complex motor responses (e.g., Henry, 1980; Henry & Rogers, 1960; Kerr, 1978; Klapp, 1977, 1980; Schmidt, 1988). Logically, it follows that if the quiet eye duration is directly related to a period of cognitive programming, more complex aiming behaviors should be characterized by longer quiet eye durations. Similarly, reductions in the time allotted for the participant to complete the task may reduce the time spent in the quiet eye phase, thus limiting the extent of cognitive programming and impairing performance. To date, however, no manipulations of task complexity or task duration have been conducted in studies

PAGE 22

13 examining the relationship between quiet eye duration and performance in far aiming tasks. In addition, Vickers (1996) argued that better performance (in this case, expert performance) would be characterized by the suppression of vision during the execution phase of the movement. Poorer performance, it was noted, was related to the maintenance of fixation on the target, while expert shooters tended to terminate their fixations on the target during the execution phase. This result is contrary to what many scholars would hypothesize, and contradicts virtually every instructional guideline for improving freethrow shooting performance (e.g., Schmidt, 1988; Wooden, 1988). Statement of the Problem In response to a series of critical fundamental questions raised by contemporary scholars, Vickers (1996) developed a location-suppression theory of oculomotor control proposed to account for expertise differences in far aiming tasks. Based upon specific hypotheses proposed in this theory, several aspects of the relationship between eye movements and performance in a billiards task were examined in this study. First, the relationship between quiet eye duration and performance in highly-skilled and lesser-skilled billiards shooters was assessed, as was the question of whether the visual suppression aspect of the aiming movement was robust to billiards shooting. In addition, by investigating the effect on performance of manipulations in task complexity and

PAGE 23

14 constraining the temporal aspects of the billiards shot, potential evidence was generated to determine (a) the importance of the location phase of Vickers' (1996) theory, and (b) whether there was support for a strong view of the eye-mind connection. More specifically, an attempt was made to resolve the issue of whether eye-movement processes are reflective of higher-order cognitive programming, and not merely the consequence of bottom-up processing. A second purpose of this study was to assess potential differences between highlyskilled and lesser-skilled players in fixation location and fixation duration measures in an ecologically valid environment. It was expected that realistic data collection procedures would yield more accurate and valid measures of exactly where in the environment expert and novice players direct their visual attention (as revealed by their visual search patterns) than that generated by crude video simulation experiments performed in a contrived laboratory situation. Hypotheses Given the comprehensive scope of this study, a series of hypotheses were tested across two experiments. In the first experiment, the complexity of the billiards task to be performed was manipulated, such that participants were required to successfully execute shots that possessed easy, intermediate, and hard ratings of difficulty (Martin & Reeves, 1993). In Experiment 2, participants were asked to initiate their shot within particular time periods. Individual subjects experienced constraints of 25% and 50% on the time allotted

PAGE 24

15 to execute shots of intermediate difficulty, based on their normal temporal stroke as recorded in Experiment 1. Six hypotheses were expected to be consistent across both experiments, and they are presented next. Hypotheses specific to each experiment will follow. 1. It was hypothesized that the highlyskilled participants would possess a visual search strategy that was different than the lesser-skilled performers. This difference will be manifested in discrepancies in the dependent variables of fixation location and fixation durations to particular elements of the visual environment. These expected variations would be in accordance with previous research demonstrating expertise differences in visual search patterns when ecologically valid testing situations are used. For example, Vickers (1992) reported that highlyskilled golfers generated more fixations of longer duration to specific aspects of the putting situation, in particular to the target (hole), the golf ball, and the putting surface after the putter made contact with the ball. Less-skilled putters (as evidenced by their higher handicap) were observed to fixate more on the club head in the backswing phase and the ball in the flight phase. It has been argued that expert performers are more likely to fixate only on the pertinent aspects of the display, due to the valuable predictive information of these visual cues (e.g., Bard & Fleury, 1981; Singer et al., 1996). Indeed, this point has been supported in studies using fast-paced dynamic situations such as service reception in tennis

PAGE 25

16 (Goulet, Bard, & Fleury, 1989; Singer et al., 1996), and offensive strategy in soccer (Helsen & Pauwels, 1993; Williams et al., 1994), and in such self-paced closed tasks as pistol shooting (Ripoll et al., 1985), basketball free-throw shooting (Ripoll et al., 1986; Vickers, 1996), and golf putting (Vickers, 1992). Thus, it was believed that differences in fixation locations and durations would be observed in the sport of billiards, such that the highly-skilled performers would fixate more on the cue ball, target ball, and target cushions than would lesserskilled players. 2. It was hypothesized that highly-skilled shooters would demonstrate longer quiet eye durations, regardless of the level of task complexity or task duration in each experiment, than lesser-skilled players. Expert shooters have been shown to possess greater cognitive knowledge of the factors associated with successful stroke production (Abemethy, Neal, & Koning, 1994), and since quiet eye duration is believed to be reflective of a period of cognitive programming in which the parameters of the aiming movement are set, logically it follows that quiet eye durations would be greater in more advanced performers. This notion is supported by research investigating basketball freethrow shooting (Vickers, 1996), in which players with higher career shooting percentages possessed larger durations of quiet eye preparation than less-skilled shooters. 3. In a related vein, it was believed that successful shots for both groups would be characterized by longer quiet eye durations. As Vickers (1996) reported, successfully made shots by both highly-skilled and lesser-skilled free-throw shooters where

PAGE 26

17 characterized by more time spent in quiet eye, as compared to unsuccessful shots. She hypothesized that longer durations of quiet eye reflected a more effective period of motor programming, such that each of the parameters of an ideal movement could be adequately addressed. When the participant spends too little time in quiet eye, she contends, the movement cannot be appropriately programmed, and poor performance results. 4. As a result of the visual suppression component of her theory, Vickers (1996) noted that expert players demonstrated a greater degree of visual suppression during the execution phase of the movement than the lesser-skilled players. This suppression was manifested in a greater number of eyeblinks and shifts of gaze to less pertinent (unidentified) areas of the visual environment during the foreswing and flight phases of the free throw by the better shooters. However, this result contradicts virtually every other study investigating the role of oculomotor control in aiming tasks. In the majority of these studies, it was determined that the target was fixated on throughout the entire duration of the movement, and gaze was not directed to other areas of the display (Abrams et al., 1990; Guitton & Voile, 1987; Prablanc & Pellison, 1990). Thus, it was hypothesized that the suppression element of Vickers' (1996) theory would not be supported, as it was expected that both the highlyskilled and lesser-skilled billiards player would not divert fixation from the target in the foreswing and flight phases of the billiards stroke.

PAGE 27

18 5. As an inherent aspect of their status, it was hypothesized that highly-skilled participants would demonstrate a greater percentage of successful shots than the lesserskilled players in every experimental condition. Indeed, this hypothesis is one that has been confirmed in virtually every study investigating expertise from a visual search standpoint, as experts have been shown to possess faster and more accurate decision-making capabilities than their less-skilled counterparts (e.g., Abemethy, 1988, 1991; Singer et al., 1994, 1996; Wright, Pleasants, & Gomez-Mesa, 1990). 6. In the billiards stroke, four distinct phases can be observed through kinematic analysis. These phases include the preparation, backswing, foreswing, and flight aspects of the movement. Based on previous kinematic research on rapid aiming movements (e.g., Vickers, 1992, 1996), it was hypothesized that highly-skilled and lesser-skilled individuals would differ in the amount of time spent in the preparation phase, but not in the backswing, foreswing, or flight phases of the movement. Specifically, the more advanced participants were expected to demonstrate longer durations in the preparation phase, due to the increased amount of time devoted to the programming of the movement. This time difference spent in the preparation phase is also believed to exist as a result of the greater proportion of quiet eye time used by expert performers (Vickers, 1996). Experiment 1 In Experiment 1, the level of complexity of the task performed was manipulated. and two hypotheses were generated.

PAGE 28

19 1. As the complexity of the task increases, it was hypothesized that successful shots would be characterized by concomitant increases in quiet eye duration. For both groups, it was believed that effective stroke outcomes would be produced in those instances when more time was spent in quiet eye in the preparation phase of the movement. Unsuccessful shots, therefore, would be marked by shorter quiet eye durations. Thus, quiet eye duration should be greatest in the hard condition, and shortest in the easy level of complexity. Although the relationship between quiet eye and task complexity has never been examined, several lines of motor control research have indicated that reaction and movement times increase as a function of task complexity (Henry & Rogers, 1960; Sternberg, Monsell, Knoll, & Wright, 1978). Scholars have taken this pattern of results to indicate that manipulations in complexity influence the response programming stage of the human information processing system (Schmidt, 1988). In this stage, the parameters of the movement to be performed are delineated, such that the velocity, direction, force, and temporal sequencing of the motor response are set. More complex movements, it has been consistently demonstrated, require more time spent in the response programming stage (Henry & Rogers, 1960; Schmidt, 1988; Sternberg et al., 1978). As Vickers (1996) contends, quiet eye duration is a reflection of response programming time. Therefore, it

PAGE 29

20 was expected that more complex shots would require longer quiet eye durations in order to successfully execute the task. 2. Decrements in performance were expected to occur as the complexity level of the required shot increased. Again, this finding is very robust in virtually every task reported in studies of response programming and oculomotor control (Henry & Rogers, 1960; Kerr, 1978; Klapp 1980; Schmidt, 1988). Experiment 2 The hypotheses generated in Experiment 1 were similar to those pertaining to Experiment 2, in which task duration was manipulated. 1. As the time apportioned for initiating the billiards stroke was reduced, it was believed that decrements in performance would be observed for both groups. Indeed, the percentage of successful shots was expected to diminish as the participant progresses from their normal temporal stroke to the 25% reduction, with the largest decrements in performance occurring in the 50% reduction condition. As Shapiro (1977, 1978) and Summers (1977) have demonstrated in their work on hand aiming movements, an increase in the percentage of errors occurred when participants were required to execute the movement under time constrained situations. These and other researchers (e.g., Vickers, 1996) have argued that reductions in time allotted to the task lead to direct constrictions in the response programming stages of the human information processing system. This, in turn, leads to poorer motor performance.

PAGE 30

21 2. Although manipulations in task duration were assumed to lead to performance decrements, it was hypothesized that the relative ratio of durations spent in each of the four phases of the movement would remain constant. In essence, it has been argued that the same motor program is employed as time becomes constrained, while only the overall duration parameter is modified (Schmidt, 1988). Other researchers have supported this view, as studies by Shapiro (1977), Summers (1977), and Terzuolo and Viviani (1979) have demonstrated that the proportion of total movement time required to traverse each phase of a movement remained the same regardless of whether the movement was performed in normal or compressed conditions (see Gentner, 1987 for a comprehensive review). It is important to note that decrements in performance were not expected to be due to changes in the relative proportion of durations between the phases of the billiards stroke (Gentner, 1987; Schmidt, 1988). If the proportions remain constant across conditions, then a stronger argument can be made for the invaluable role that the quiet eye variable plays in determining the quality of the aiming behavior. That is, impaired performance may likely be caused by insufficient time spent in quiet eye, thus leading to inadequate programming of the stroke. Thus, it was expected that successful shots will once again be characterized by longer quiet eye durations than unsuccessful attempts, even under externally imposed time constraints.

PAGE 31

22 Assumptions For the purposes of this study, the following assumptions were made: 1. The selection procedure of participants would be sufficient to adequately discriminate highly-skilled billiards players from lesser-skilled players. To accomplish this goal, two methods were employed to ensure that playing ability and experience was different between groups. This distinction is important in determining whether the location-suppression hypothesis is robust to a billiards task, and whether it is associated with differences in shooting performance. 2. Vickers (1996) posits that quiet eye duration is a marker of response programming time in motor activities requiring aiming to a far target. Based on results reported in other studies investigating the relationship between task complexity and response programming (e.g., Henry & Rogers, 1960), it was assumed that increases in the complexity of the billiards task would directly lead to concomitant changes in cognitive programming times. As such, it was also assumed that evidence for these changes would be documented by altered durations in the quiet eye variable of the location hypothesis. 3. Similarly, since scholars have demonstrated a direct relationship between task duration and response programming (Shapiro, 1977, 1978; Summers, 1977), it was assumed that constraints in the duration required to complete the billiards task would directly lead to reductions in the duration of the quiet eye variable and in each of the

PAGE 32

23 phases of the movement. Performance was also expected to decrease as greater constraints were placed on the time allocated to the execution of the billiards stroke. 4. It was assumed that the testing situation possessed a high degree of ecological validity, since the participant was required to complete the billiards tasks using a normal biomechanical stroke with a regulation sized cue, on an actual pool table. In addition, the shots required in the billiards task, regardless of the level of complexity, were highly similar to those that could be encountered in the normal course of a game of nine-ball. Limitations Given the technological factors associated with this study, the following limitations may have existed and must be acknowledged when interpreting the results. 1. Due to the fact that the instrumentation used to measure participants' eye movements and visual tracking patterns relied on a corneal reflection technique, the apparatus was highly sensitive to lighting conditions. To generate a strong reflection, the testing environment needed to be more dimly lit than normally encountered in routine game situations. Thus, this potential reduction in the quality of the visual stimuli may lead to decrements in performance. As researchers have noted (Abrams, Meyer, & Komblum, 1990; Elliot, Chua, & Pollack, 1994; Zelaznik, Hawkins, & Kisselburgh, 1983), increased visual occlusion of a target tends to lead to systematic deteriorations in performance, especially in targets that require more complex aiming behaviors. However, the targets

PAGE 33

24 will be visible at all times, and since the impaired lighting conditions were experienced to an equal degree by each participant, this variable most likely did not contribute to any observed differences between the two groups. 2. The eye movement monitor collected data via a special piece of equipment (weighing approximately 700 g) which was mounted on the participant's head. This added weight may have served as an inconvenience for the shooter, and may have detracted from the ecologically validity of the billiards situation. However, previous researchers using this equipment in other self-paced sport tasks have reported no significant reductions in performance on behalf of the participant, thus indicating that this measurement technique may not be inherently intrusive (Vickers, 1992; 1996). 3. The reductions in duration allotted for completion of each shot in Experiment 2 may have detracted from the ecological validity of the billiards task, since shooters do not normally face these externally imposed time constraints (although 30 sec shot clocks are used in most nine-ball tournaments). Again, however, these constraints were experienced by each participant, and therefore are not likely to produce any differential pattern of results between the groups. Definition of Terms To operationally define and standardize some of the terminology used in this study, each of the following terms are defined:

PAGE 34

25 Carom Shot occurs when the cue ball glances off the object ball into a second ball (Martin & Reeves, 1993). Cue Ball is the all-white, unnumbered ball which is the only legal ball struck with the cue stick (Martin & Reeves, 1993). Cue Stick is the tapered wooden stick, usually 57 in (145 cm) in length and 17-21 oz (480-594 g) in weight, with which the player strikes the cue ball (Martin & Reeves, 1993). Cushion refers to the cloth covered rubber buffers that line the inside rails of the billiards table (Martin & Reeves, 1993). Cut Shot refers to the action of hitting the object ball with the cue ball at less than dead center, thus deflecting the object ball off at an angle (Martin & Reeves, 1993). English is the spin of the cue ball either to the left or to the right, controlling the action of the cue ball before and after it strikes the object ball or cushion (Martin & Reeves, 1994). Execution Phase is the final, error-correction phase of an aiming movement, in which attempts are made to reduce apparent discrepancies between the current position of the limb and the movement goal as specified by the motor program (Abrams et aL, 1990). This phase corresponded to the time period involving the foreswing and flight phases of the billiards stroke.

PAGE 35

26 Eyeblinks , as was operationally recorded in this study, occurred when the participant closed their eyes and occluded the optics of the system for 100 ms or more (3 or more video frames). Eve-Mind Assumption refers to the belief that "eye movements can at the very least be considered as tags or experimentally accessible quantities that scientists can observe to understand underlying processes of cognition" (Viviani, 1990, p. 354). Stronger views of the eye-mind assumption believe that eye movements are reflective of higher-order cognitive processing (Just & Carpenter, 1976). Fixations occur when the individual's eyes pause at a specific location such that information is stabilized on the high-acuity area of the retina, the fovea (Vickers, 1992). Consistent with previous research (e.g., Carl & Gellman, 1987; Optican, 1985; Vickers, 1996), fixations were recorded when the participant's gaze was fixed on a location for 100 ms or more (3 or more video frames). Impulse Phase is described as a fairly rapid and continuous change in the position of the aiming limb as it traverses most of the distance between the starting position and the final location of the target (Abrams et al, 1990; Carlton, 1981a, 1981b). In this study, the impulse phase included the time period corresponding to the backswing of the cue stick until initiation of the foreswing.

PAGE 36

27 Location-Suppression Hypothesis is a recent theory proposed to account for visual control mechanisms when an individual is engaged in aiming behaviors to a far target which is beyond arms reach (Vickers, 1996). Nine-Ball is a billiards game played with a special diamond-shaped rack which holds balls numbered 1 through 9. The object of the game is to sink the balls in numerical order (from 1 to 9), and the person legally sinking the 9 ball is declared the winner (Billiard Congress of America, 1996). Object Ball refers to the ball the player wishes to hit with the cue ball (Martin & Reeves, 1993). Preparation Phase refers to the time period in which the performer engages in behavior designed to familiarize themselves with the general direction of the target (Abrams et al., 1990), and to set the parameters required for successful execution of the aiming behavior (Vickers, 1996). In the present study, the preparation phase was operationally defined as the time recorded when the performer was positioned over the cue ball until the first observable movement towards the striking of the cue ball (i.e., backswing was initiated). Quiet Eve Duration is defined as "that portion of the final fixation from onset to the first observable movement of the hands into the shooting action" (Vickers, 1996, p.348)

PAGE 37

28 Visual Attention refers to the process by which individuals select information gathered by visual mechanisms to provide the basis for responding to particular stimuli (Theeuwes, 1993). Visual Search may be defined as "the manner in which individuals move their eyes to take in available visual information while preparing and executing a movement" or engaging in a decision-making process (Vickers, 1996, p.342). Visual Suppression refers to an aspect of the location-suppression hypothesis in which participants suppress their vision by using a blink behavior, thereby preserving the aiming commands and eliminating potential limb interference in the visual field (Vickers, 1996). Significance of the Study As is evidenced by the critical questions raised by contemporary scholars concerning the continued use of the visual search paradigm in examinations of expertise in sport, it is clear that this paradigm is at a crossroads in terms of its development. Researchers could abandon this line of investigation, as some reviewers have suggested (e.g., Abemethy, 1993), or they could attempt to answer their critics with more theoretically based studies conducted in ecologically valid environments. It was the intent in the present study to follow the latter path, in the hope that greater insight could be gained into the relationship between eye movement behavior and optimal motor performance in far aiming tasks.

PAGE 38

29 To this end, one of the main goals of this study was to further advance the theoretical foundations of oculomotor control in aiming tasks, specifically by investigating aspects of the location-suppression hypothesis (Vickers, 1996). Although this theory has been supported in a study of basketball free-throw shooting (Vickers, 1996), no other line of research has attempted to critically evaluate it to determine if it is in fact robust to other far aiming tasks. In addition, if the location-suppression hypothesis is a critical factor in the organization and programming of neural substructures underlying the aiming response (Vickers, 1996), then manipulations in factors such as task complexity and task duration should result in alterations in performance. In the present study, therefore, an attempt was made to determine whether the quiet eye and suppression aspects of Vickers' (1996) theory were present in the sport of billiards, and whether these factors were influenced by manipulations related to cognitive programming stages. The information gathered from this study may serve to either advance the location-suppression hypothesis, and in turn provide support for the use of visual search paradigms in expertise research, or modify specific aspects of the hypothesis that do not hold up to empirical testing. In addition, support for VickersÂ’ (1996) theory would provide further evidence for a stronger view of the eye-mind assumption. If the manipulations were indeed effective in altering performance, then a more logical argument

PAGE 39

30 could be made in favor of the notion that eye movements are reflective of cognitive processes. In addition to the unique concept of theory testing that has been so limited in visual search research, the present study attempted to incorporate an ecologically valid environment in which the location-suppression hypothesis was tested. Again, very few studies have been conducted in situations that allow the performer to actively perceive information and respond with behaviors that would be typical of the true sporting environment. Thus, unlike virtually every other study examining eye movements in sport, this study investigated aiming behaviors in a real-world setting while the billiards players actively engaged in "on line" processing of information made available to them in their environment. This information included visual cues and knowledge of results pertaining to the outcomes of actions they initiated. Finally, an attempt was made to delineate expertise differences in the sport of billiards from a visual search standpoint. Very little research of any kind has been conducted using a billiards task, and virtually nothing is known about the mechanisms highlyskilled players use to perform the precise movements required to achieve success in this sport. Thus, this study was significant in the fact that it represented an initial attempt to determine exactly what visual cues more advanced and less advanced players fixated on when actively engaged in the act of shooting. Potential expertise differences in cue usage

PAGE 40

31 may then be used to initiate the development of cognitive training programs designed to improve billiards performance in players of all skill levels.

PAGE 41

CHAPTER 2 REVIEW OF LITERATURE Given the importance of the human information processing system in relation to sport performance (e.g., Magill, 1998; Schmidt, 1988), this literature review will focus on several aspects germane to the topic. First, factors associated with the development of expertise in sport, in particular hardware-software approaches, will be discussed. Next, a specific software approach in the study of expertise, the visual search paradigm, will be reviewed. This section will also include critiques of current methodology used to assess general expert-novice differences, with particular emphasis placed on research examining eye movement patterns. Three specific theories of oculomotor control will then be discussed, as these theories focus on aiming tasks which require a high degree of eye-hand coordination. The locationsuppression hypothesis (Vickers, 1996) will be outlined in greater detail, as it is the most important theory under investigation in this study. Finally, because task complexity and task duration will be manipulated in the present study, an overview of the motor learning and control literature related to these topics will be addressed. 32

PAGE 42

33 Expertise in Sport The Hardware Approach Early research in sport expertise focused on the physical aspects of the athlete, since it was believed that expert athletes differed from novices in that they possessed advanced psychophysical and mechanical properties of the central nervous system (Abemethy, 1991; Blundell, 1985; Burke, 1972; Nielsen & McGown, 1985). That is, proponents of this theory believe that experts have much faster overall reaction (RT) and movement (MT) times than do novices, and also possess greater optometric (static, dynamic, and mesopic acuity) and perimetric (horizontal and peripheral vertical range) attributes (Carlson, 1985; Revien, 1987). Not surprisingly, this perspective has been termed the "hardware" approach to expertise (Starkes, 1987; Starkes & Deakin, 1984). However, research examining athletes' mechanical and optometric properties (hardware) has yielded equivocal results. Blundell (1985) reported that static visual acuity (SVA) was only slightly related to expert performance, accounting for roughly 8-10% of the variance in a pool of athletes. Similarly, Starkes (1987) could find no significant relationship between SVA, dynamic visual acuity (DVA), or peripheral vision in her study of expert-novice field hockey players. This result was also obtained by Helsen (1994) in an examination of athletes from several different sports. Although Blundell (1985) and Sanderson and Whiting (1975) have reported that the role of DVA and peripheral vision

PAGE 43

34 are important in interception tasks such as ball catching, support for their results has not been generated in other studies (e.g., Abemethy, 1986; Helsen, 1994). Also, highly-skilled performers have been found to possess base rates of RT and MT that are equal to those of their lesser-skilled counterparts when assessed in non-sport related conditions (McLeod, 1987; Nielsen & McGown, 1985; Starkes, 1987), and as such their demonstrated performance superiority does not appear to be based on a faster ability to process and react to information. More evidence of this contention comes from a study conducted by Singer, Cauraugh, Chen, Steinberg, and Frehlich (1993), in which high level college and professional tennis players were compared to novice and intermediate (recreational) players on several variables related to RT, MT, and coincidence-anticipation capabilities. No significant differences were discovered between the groups on non-tennis related tests of foot and hand speed (RT and MT measures), and anticipation capability was similar for the expert and novice players on a Bassin timing task. Other attempts to investigate differences in rates of visual information processing (VIP) have shown that elite performers may be faster processors of visual information in general, perhaps explaining in part how they achieved expert status (Adam & Wilberg, 1986). However, this result has not been replicated in more recent studies (e.g., Starkes, Allard, Lindley, & O'Reilly, 1994). Thus, it appears that advanced athletes can not be consistently discriminated from less-skilled participants solely on hardware aspects related to superior CNS and VIP capabilities.

PAGE 44

35 The Software Approach In contrast to the hardware theory of expertise, proponents of the "software" approach argue that experts have a much greater knowledge base of information pertaining to their particular sport, and that differences in expert performance as compared to novices is the result of a cognitive advantage, rather than a physical advantage (e.g.. Singer et al., 1994; Starkes & Deakin, 1984). Elite athletes, it is believed, make faster and more appropriate decisions based on the cognitive processes of selective attention, anticipation, and pattern recognition (Abemethy, 1991). That is, highly-skilled performers appear to possess the ability to know which cues to focus their attention on in the sport environment, and understand the importance of these cues in predicting future actions (Abemethy & Russell, 1987a; Bard, Fleury, & Goulet, 1994; Singer et al, 1996). Support for the software approach to explain expertise has been demonstrated in studies in which decision time and accuracy responses were assessed in sport specific situations (Bard & Fleury, 1976; Starkes, 1987), and for the recognition and recall of structured elements of game situations in sports such as baseball (Hyllegard, 1991), basketball (Allard & Burnett, 1985; Bard & Fleury, 1976), field hockey (Starkes, 1987), and volleyball (Borgeaud & Abemethy, 1987). In addition, experts have been found to possess more declarative (factual information) and procedural knowledge (how to do something) in their sport domain than do novices (Allard, Graham, & Paarsalu, 1980;

PAGE 45

36 Allard & Starkes, 1980; Chase & Simon, 1973; Chi, Glaser, & Farr, 1988; Sandu, 1982; Starkes & Deakin, 1984). That is, elite performers were much faster and are more accurate in recognizing and recalling information related to structured game situations. This ability, however, appears to be task-specific, as highly-skilled and lesser-skilled athletes do not differ in recognition and recall performance for elements that are unrelated to their sport (Chi & Bjork, 1991; Chi, Glaser, & Farr, 1988). As Abemethy, Neal, and Koning, (1994) noted, "the more sport-specific the stimuli and response(s) used in the test task, and hence the more closely the processing demands of the test mimic those of the intact skill, the more probable it is that systematic expert-novice differences will be demonstrated" (p.186). To determine the relative contributions of software and hardware variables on expertise in the sport of billiards, Abemethy, Neal, and Koning (1994) compared 7 expert, 7 intermediate, and 15 novice Australian snooker players on general vision tests and sportspecific perceptual and cognitive tests. Results of their study indicated that no expertise differences were found on optometric tests of SVA, ocular muscle balance, color vision, and depth perception (hardware variables). However, expert and intermediate players differed from novices in their ability to recall and recognize rapidly presented slides of organized game situations, but these differences were not observed in randomly arranged situations. Finally, expert shooters were characterized by the ability to choose a greater number of shot options that were more appropriate, and planned more shots in advance

PAGE 46

37 (M = 6.22) than did novices (M = 4.56). This finding is indicative of a greater depth of encoding for structured material, and supports the notion that expertise is developed not from general visual capabilities, but rather from acquired processing strategies. Thus, a preponderance of the research supports the software view that expertise in sport is the result of a cognitive advantage, and not necessarily a physical advantage. But how is this cognitive advantage developed? That is, given the same game situation, what processes direct highly-skilled performers to select and execute the most appropriate responses while novices do not? A likely source of the cognitive advantage comes from multiple experiences in the sport task. For example, Ericsson, Krampe, and Tesch-Romer (1993) and Ericsson and Chamess (1994) argue that the development of expertise is explicitly due to a dedicated regimen of deliberate practice, in which a set of highly structured activities are practiced. Here, the individual engages in effortful and attention-demanding endeavors that have the explicit goal of improving performance. In fact, Ericsson et al. (1993) argue for a monotonic relationship between practice and performance. That is, the more deliberate practice one undertakes, presumably the greater the performance and level of expertise. Finally, these researchers argue that true expertise cannot be developed until a minimum of 10 years of deliberate practice is undertaken.

PAGE 47

38 However, the question remains: What exactly does the elite performer learn during this process of deliberate practice? Researchers supporting the software explanation of expertise have offered one possibility: Highly-skilled athletes have learned to effectively search their environment for pertinent cues that allow them to select and execute the most appropriate behavior in a particular situation (e.g., Bard & Fleury, 1981; Singer et al., 1996; Vickers, 1992). That is, processes such as the visual searching of one's environment were hypothesized to account for some of the differences observed to exist between elite and non-elite performers. The results of this line of research are discussed in more detail in the next section. The Visual Search Paradigm in Sport Research Recent advances in technology over the past two decades have allowed scholars to examine eye movements and tracking behaviors in simulated sport conditions which demand immediate attention (Bard & Fleury, 1976). The majority of research undertaken in this area has been concerned with the relationship between visual search and selective attention, and with the influence of these processes on decision making strategies in relation to filmed scenarios (Helsen & Pauwels, 1992, 1993; Singer et aL, 1994). In each of these studies, it was assumed that the allocation of focal attention could be determined by examining the locations and durations of ocular fixation patterns. These patterns were assessed with eye-movement recording devices that measured ocular fixation via a cornea

PAGE 48

39 reflection technique (e.g., Bard & Fleury, 1981) while athletes viewed slides or videotaped simulations of particular sport situations. Reviews of the sport-specific literature on selective attention and visual search, even in simulated conditions, often demonstrates systematic differences in eye movement patterns between expert and novice performers (e.g., Abemethy, 1988, 1991). Sports of interest include baseball (Bahill & LaRitz, 1984; Shank & Haywood, 1987), fencing (Bard, Guezennec, & Papin, 1980; Haase & Mayer, 1978), golf (Vickers, 1992), gymnastics judging (Bard, Fleury, Carriere, & Halle, 1980; Vickers, 1988), ice hockey (Bard & Fleury, 1981), soccer (Tyldesley, Bootsma, & Bomhoff, 1982; Helsen, Pauwels, & D'Ydewalle, 1986; Williams et al„ 1994), table tennis (Ripoll & Fleurance, 1985), tennis (Goulet, Bard, & Fleury, 1989; Ritzdorf, 1983; Singer et al„ 1996), and volleyball (Neumaier, 1983; Ripoll, 1988). Evidence in Support of Expert-Novice Differences As mentioned earlier, most of these studies yielded three general results. Elite performers required fewer fixations on elements in the display to produce a successful response (e.g., Singer et al., 1996), allocated a greater number of fixations to pertinent aspects of a visual array (e.g., Ripoll, 1988), and were more systematic and consistent in their search patterns than novices (e.g., Ripoll et al., 1986; Vickers, 1992). For example, expert goaltenders fixated on the stick of ice hockey shooters more than novice

PAGE 49

40 goaltenders, who tended to allocate more fixations to the puck (Bard & Fleury, 1981). In addition, experienced soccer players fixated more on pertinent aspects of the display than did less-skilled players, who had much more variable visual tracking patterns (Helsen & Pauwels, 1990; Williams et al„ 1994). When reviewing the particular sports that have been studied using the visual search paradigm, it becomes clear that much of the research has been directed towards dynamic, fast-paced tasks. This is understandable, since the ability to successfully fixate on advance cues that may be used to anticipate the intended actions of an opponent gives the participant a decided advantage in preparing an effective response under time constrained situations (Abemethy, 1991). For example, expertise differences were found in visual search patterns in an examination of expert and novice tennis players preparing to return a videotaped serve (Fleury, Goulet, and Bard, 1986; Goulet, Bard, & Fleury, 1989; Singer et al., 1996). Specifically, highly-skilled players tended to make more fixations on the shoulder and trunk areas of their opponent in the ritual phase of the serve, while lesserskilled participants fixated most often on the server's head. During the execution phase of the serve, experts demonstrated a greater number of fixations on the arm and racquet of the opponent, while novices tracked only the ball. No significant differences were found during the preparatory phase of the serve (Goulet, Bard, & Fleury, 1989). It was demonstrated that those cues fixated on by the experts were the most pertinent, because they provided important information about the direction and type of serve to be returned.

PAGE 50

41 Similar results have also been obtained in the sport of baseball, where it was demonstrated that successful hitters tended to focus more on the release point of the pitch, while less advanced batters tended to shift their focus from the pitcher's head, to the release point, and then back to the head (Shank & Haywood, 1987). It was determined that the experts' focus of attention (pitch release point) provided the most important cues for successful recognition of pitch velocity, type of pitch, and location of the ball before it crossed the plate (Shank & Haywood, 1987). These studies, and many others, have documented the positive relationship between expertise and visual search patterns in dynamic sport tasks. However, several investigations have also revealed expert-novice differences in search patterns for closed, self-paced tasks in which the performer has unlimited time to execute the movement under stable environmental conditions (Singer & Chen, 1994). Thus, it would appear that visual search processes play an important role in sports which require precise aiming behaviors. For example, in a basketball free-throw shooting task, highly-skilled individuals oriented their gaze towards the basket sooner, and made a greater number of fixations of longer duration towards this target than lesser-skilled performers (Ripoll et al., 1986). In addition, more successful shots were characterized by longer durations of fixation towards the hoop during the preparation and flight phases of the shot than missed attempts (Vickers, 1996).

PAGE 51

42 In pistol shooting, elite performers maintained their fixation on the target throughout all phases of the aiming action, while less advanced shooters tracked the sight of the gun until they made a fixation on the target immediately prior to pulling the trigger (Ripoll et al„ 1985). Finally, in an analysis of the putting stroke used by expert and novice golfers, Vickers (1992) reported that low handicap golfers (experts) possessed an economy of gaze allocation when compared to high handicap golfers (novices). Specifically, experts made more express saccades, had quicker saccades between gaze locations, and demonstrated significantly more fixations of longer durations to the ball and target during execution of the putting stroke. Lesser-skilled putters, on the other hand, often tracked the golf club head immediately prior to contact with the ball, and allocated fewer fixations to the ball during the contact phase of the movement. Taken together, much of the visual search and gaze control research indicates that expert athletes possess highly systematic and selective process of focusing their attention. Specifically, highly-skilled performers know exactly what to look at in the visual field to glean the most informative cues in dynamic as well as static sport situations, and make the most efficient and appropriate ocular movements to pick up this information. In comparison, novices are characterized by random and less uniform visual tracking patterns, and do not appear to understand the relative importance of the link between advance visual cues and effective response preparation (Abemethy, 1991). The robustness

PAGE 52

43 of these views, however, have been questioned by scholars over the course of the past decade. Evidence Against Expert-Novice Differences Although the bulk of visual search research has demonstrated expertise differences, the expected pattern of results has not been observed in a few empirical studies. In a comparison of 20 expert (International level) and 35 novice (untrained undergraduate) badminton players, Abemethy and Russell (1987b) noted that fixation sequences, distributions, and durations were indistinguishable between the two groups. Both groups fixated more on the racquet and arm regions of the filmed opponent than the head, trunk, and leg regions. This search strategy, the investigators posited, represented a more proximal-to-distal transition in cue usage. A similar arrangement of results have been observed in investigations of expertise involving squash players (Abemethy, 1990a, 1990b), in which skill differences were not paralleled by comparable contrasts in visual search patterns. Also, elite and non-elite tennis players were found to possess identical search rates when viewing the preparatory phase of an opponent's motion in a simulated tennis service reception task (Goulet et al., 1989; Singer et al., 1996). However, the expected differences did emerge during the execution phase of the serve, as experts centered their fixations on the racquet and the arm holding the racquet, and terminated their search on the opponent's racquet at the moment of ball

PAGE 53

44 contact. Beginners, on the other hand, demonstrated search rates that were highly variable with fixations allocated to several regions of the server. Thus, while the visual search paradigm has provided a fruitful means of assessing selective attention strategies used by athletes in sport situations, the employment of this paradigm is not without criticism. Before concluding this section on visual search, it is necessary to discuss some of the limitations and potential problems that currently exist in expert-novice eye movement research. These concerns, as outlined by Abemethy (1988, 1993) and Viviani (1990), are directed toward the assumptions of selective visual attention theory, current techniques of recording eye movements, and general methodological issues. Criticisms of Expertise Research in Sport In the study of expertise in sport, a number of criticisms have been aimed at a variety of aspects of the research program, including the selection and determination of participants, participant pool sizes, the nature of the task under study, the number of dependent measures used, the type of equipment used, the environment in which the study takes place, and the high rates of subject variability in performance. The typical approach to expert-novice research has been to select a number of athletes (usually at a college varsity level of expertise) and compare them to untrained individuals, usually undergraduate students (Abemethy, Thomas, & Thomas, 1993). These two groups are

PAGE 54

45 then compared on a single dependent measure of interest, on tasks that are predominantly laboratory based (and are often artificial and contrived). For example, in studies of expertise differences in pattern recognition, two criticisms have been levied. Often only static displays are used (usually slides of game situations, Vickers, 1996; Williams et al., 1994), and a confound exists in that one cannot be sure that expertise differences are the result of actual expertise advantage, or merely an artifact of the number of years of experience in these situations. In knowledge-based studies, several problems have been exposed, including the use of self-report data (Nisbett & Wilson, 1979), which have often been shown to be unreliable and prone to errors in memory or influenced by events occurring after the one of interest. In addition, problems have been identified in terms of the ignoring of action in many tasks (participants only verbally report, or respond to pen and paper tests, but do not carry out actions). In this vein, proponents of a more ecological approach to the study of motor behavior have soundly criticized many experiments as being too laboratory-based and reductionist in nature (Turvey & Carello, 1986; Williams et al., 1992). They argue that the need for heightened control serves to constrain the coupling between perception and action that is so vital for expert performance. Again, typical designs include the use of static visual displays (slides) or two dimensional videotaped images, and the performer is often required to make only verbal or simple finger responses to the incoming stimuli.

PAGE 55

46 As Bootsma (1988) discovered, the act of having the individual perform only simple tasks yields vastly different results as compared to more real world tasks. In his study, three groups of participants were asked to hit a moving ball with a stick. In the first group, the individuals were required to press a button which then moved the stick into the path of the oncoming ball. The second group of participants were allowed to touch the stick with their index finger, but could only pull it back and release it to strike the ball. In the third group, performers were allowed to grasp the stick and swing it freely toward the ball. Significant differences in coincidence timing were observed, such that the most errors were observed in the first group, while the third group demonstrated the best overall performance. Bootsma (1988) argued that these results provided evidence for the notion that the experimental situation can influence the responses generated by the performer, thus highlighting the importance of testing individuals in a more ecologically valid environment. In addition, Vickers (1996) and Ripoll (1991) both argue that removing the actor from the actual sport environment changes the results of the study dramatically, such that expert-novice differences may no longer appear in variables such as visual fixation locations and durations. Since ecological psychology theorists contend that attunements between the person and the environment help comprise expert performance, placing the athlete in contrived lab settings retards the level and effect of their expertise (Chi, Glaser, & Farr, 1988; Fowler & Turvey, 1978; Nougier, Stein, & Bonnel, 1991).

PAGE 56

47 In addition, a question of motivation and arousal levels have been levied against the majority of expert-novice studies, since the participants are often placed in a seated position (which is drastically different from the ready position found in most sports), and are required to only passively react to presented stimuli (Singer et al„ 1994). Thus, it would appear that severe constraints impacting on arousal and motivational levels of athletes are removed in the typical lab study. Although little is known about the resulting impact of these constraints on performance, it has been argued that performance is altered quite extensively (Ripoll, 1991; Singer et al„ 1994; Vickers, 1996). Other problems can be found in expert-novice studies. For example, the vast majority of researchers conduct only cross-sectional rather than longitudinal investigations that are characterized by small sample sizes (usually because few experts can be found or determined). To increase sample sizes, researchers often dilute the degree to which an expert is operationally defined, thus creating inconsistent results between studies since this independent variable is not held constant (Abemethy, 1989). Finally, novice subjects often may not be classified as true beginners, as they may vary extensively in their performance and experience in the task under study (Singer et al., 1994). Each of these criticisms have, in some form or another, been directed toward the majority of studies examining expertise differences in visual search patterns of athletes. However, based on the theoretical and methodological approaches used in this area of

PAGE 57

48 investigation, three particular critiques have been forwarded. Each of these will be discussed in more detail in the following section. Criticisms of the Visual Search Paradigm The first major limitation of eye movement recordings lies in the assumption that visual search orientation is reflective of a person's actual allocation of attention (a strong eye-mind connection). Stated more succinctly, it is explicitly believed that visual fixation and attention are one in the same (where we look is where we attend). As has been discussed previously, this notion has been refuted in research by Posner (1980), Remington (1980), and Remington and Pierce (1984), where it was shown that attention could be allocated to areas other than the foveal fixation point. Indeed, attention can also be allocated to areas in peripheral vision, a mode that cannot be measured with current visual search equipment (Bard, Fleury, & Goulet, 1994; Buckholz, Martinelli, & Hewey, 1993; Davids, 1984, 1987, 1988). This limitation of eye movement recording focuses on the issue of visual orientation and information pick-up. As Abemethy (1993) noted, merely "looking" at visual information does not necessarily equate with "seeing" (or comprehending) this information. Thus, a performer may fixate upon pertinent cues in the visual array, but there is no guarantee that the individual is actually attending to or using these cues when preparing a response. This view may help to explain those studies in which no expertnovice differences were observed in visual search patterns (Abemethy, 1990a, 1990b;

PAGE 58

49 Abemethy & Russell, 1987b). That is, even though elite and non-elite performers may in fact fixate on the same environmental stimuli, only the highly-skilled understand the importance of these cues in determining the resulting outcome of their opponent's actions (Abemethy, 1991, 1993). While novice performers may fixate on the most pertinent advance cues, they do not possess the cognitive link underlying the predictive nature of this information. In response to these limitations, it should be noted that much of the criticism is aimed at studies in which dynamic, externally-paced sports such as baseball, squash, and tennis have been of interest. In these studies, it is highly likely that the ability to pick up advance cues from an opponent allows the expert performer to select and execute the most appropriate response in a much more rapid fashion than novices. Indeed, anecdotal evidence from highly-skilled athletes suggests a perception that they have "all the time in the world” to respond to an opponent's actions in such time constrained conditions, while less advanced performers report being rushed and struggling to generate effective reactions (Abemethy, 1991). In sports such as archery and billiards, however, the aiming task to be performed is self-paced, and is not based on an a constantly changing environment. In these instances, it has been suggested that the ocular and motor systems are tightly coupled, such that the role of vision and gaze behaviors are highly correlated with the motor response to be

PAGE 59

50 performed (Abrams, Meyer, & Komblum, 1990; Guitton & Voile, 1987; Vickers, 1996; Zangemeister & Stark, 1982). Given this large degree of coupling between the two systems, researchers have speculated that foveal vision is directly associated with visual attention, and that this visual information is in fact used to program the aiming movement to be executed (e.g., Abrams et al., 1990). Thus, because the present study was designed, in part, to assess the degree to which eye movements were associated with the execution of aiming movements in the self-paced sport of billiards, Abemethy's (1991, 1993) criticisms of the issues of information pick-up and peripheral vision are less relevant. The second limitation of the current visual search paradigm involves the high trialto-trial variability in search patterns that has been observed between participants of similar skill levels (e.g.. Singer et al., 1996). These variable patterns make reliable conclusions about the relative importance of specific visual cues difficult, as it has been shown that even expert performers differ in the type of search sequence used. Related to this limitation is the fact that the majority of experiments examining expert-novice differences in search patterns have relatively low sample sizes (often n = 4 or 6), thus raising concerns pertaining to assumptions underlying the internal and external validity of the study. To help reduce the impact of this limitation, the present study incorporated a sample size that was, in some instances, two to three times larger than that reported in previous investigations. Also, two methods of operationally defining highlyskilled and lesser-skilled billiards players were used, including a preliminary performance analysis

PAGE 60

51 designed to test the shooting abilities of the players. A protocol of this nature has been included in only a handful of previous visual search studies (e.g., Vickers, 1992). Finally, researchers investigating expertise differences in eye movement patterns have been critiqued for their blatant lack of regard for theory testing, with the majority of studies being rather descriptive in nature (Abemethy et al„ 1993). In fact, this atheoretical approach has been recognized by more contemporary scholars (Abemethy, 1993; Ripoll, 1991; Vickers, 1996), who have suggested that the time for simple descriptive investigations is past. Instead, a call has gone out for initiating the processes of hypothesis testing, theory development, and deductive reasoning in this area of study. To this end, three innovative theories of oculomotor control have been developed, and each will be described in more detail below. Theories of Oculomotor Control in Aiming Tasks The Eve-Head and Image-Retina Systems In an extension of Whiting and his colleagues' seminal work on ball catching (e.g., Savelsbergh & Whiting, 1988, 1992; Savelsbergh, Whiting, Burden, & Bartlett, 1992; Sharp & Whiting, 1975; Whiting, 1972; Whiting, Savelsbergh, & Faber, 1988), Gregory (1990) developed two competing theories in an attempt to explain how individuals use visual information to intercept objects projected in space.

PAGE 61

52 The first approach, termed the image-retina system, contends that the individual allocates fixations to the point of release of the object in space, and that successful interception of this object is dependent solely on the information detected via motion of the object on the retina. Thus, in order to intercept an object (e.g., a thrown ball), the image-retina system theory contends that all one needs to do is fixate on the release point of the ball. All other visual information is deemed superfluous for successful catching behaviors. In contrast, the eye-head system (Gregory, 1990) proposes that information is picked up either by the motion of the projection of the object in relation to the background, or via proprioceptive information pertaining to the motion of the eyes and the head. In either case, successful execution of an interception response is dependent on the continuous visual tracking of the object with the eyes. Although many theorists implicitly assumed that the eye-head system was a far more cogent explanation of catching behaviors (e.g., Brancazio, 1985; McLeod & Dienes, 1993), Michaels and Oudejans (1992) and Oudejans, Michaels, Bakker, and Davids (in press) were the first researchers to empirically examine the eye movements of individuals as they attempted to catch balls projected to them in a manner similar to fielding fly balls in the sport of baseball. Indeed, the results of these studies provided support for the eyehead system, as catchers tracked the ball from release to the catching hand with smooth eye and head pursuit movements.

PAGE 62

53 The Position, Movement, and Movement + Position Hypotheses While the results reported by Michaels and Oudejans (1992) confirmed a relationship between eye movements and interceptive behaviors that was intuitively assumed to exist, the importance of hypothesis testing and empirical confirmation cannot be discounted. However, the nature of the catching task under study is quite different than the self-paced tasks of many aiming movements required in a sport such as billiards. In an attempt to explain the relationship between visual feedback processing and the production of aimed limb movements, Abrams, Meyer, and Komblum (1990) forwarded three hypotheses proposed to account for the eye-hand coordination observed when individuals perform precise aiming behaviors to a target. In the position-only hypothesis, Abrams et al. (1990) posited that visual information on the position of the target is so essential for the accurate completion of the movement that the eyes fixate upon the target throughout the duration of the movement. This hypothesis has been supported by Hansen (1979) and Honda (1985), who observed accurate pointing responses when individuals used either saccadic or pursuit eye movements directed solely to the target (and not on the aiming limb). The movement-only hypothesis contends that the eyes move in a coupled fashion with the limb, or limbs, performing the aiming action. Here, information generated from oculomotor commands and proprioceptive inflow from the eye muscles is required in

PAGE 63

54 order for the movement to be performed successfully (Abrams et al., 1990). Thus, the actual movement of the eyes acts as the critical input for effective aiming movements. Support for this hypothesis has been generated by Festinger and Canon (1965) and Miller (1980), who documented differences in final gaze location and aiming movements following saccadic versus pursuit eye movements. Finally, the movement-plus-position hypothesis posits that saccadic eye movements are closely time locked with the initiation of the aiming movement, while fixations upon the target only occur during the later phases of the movement (i.e„ during the error correction or homing-in phase). In a critical test of the three competing hypotheses, Abrams et al. (1990) determined that the data supported the movement-plusposition hypothesis. That is, when the participants were required to perform a wrist rotation to a target, the motor control system was posited to receive and use information about both the movement and the relative position of the eyes. Although not a direct test of Abrams et al. (1990) hypotheses, Vickers (1992) noted that highly-skilled and lesser-skilled golf putters differed in the extent to which they made saccadic and pursuit eye movements. Extrapolating from Vickers' (1992) data, it appeared that the more advanced putters exhibited a strategy of eye movements similar to the position-only explanation, while novice putters were characterized by a movementplus-position explanation. This result suggested a different strategy of optimal eye movement behavior than that offered by Abrams et al. (1990). Again, however, it must be

PAGE 64

55 noted that the three hypotheses were not under critical evaluation in Vickers' (1992) study. The LocationSuppression Hypothesis In a more direct test of the Abrams et al. (1990) hypotheses, Vickers (1996) examined the gaze behaviors of highly-skilled athletes performing a basketball free-throw shooting task. Sixteen elite players were classified as either expert or near-expert shooters based on their free-throw percentages accumulated over the course of a competitive season. Each shooter was required to perform consecutive free throws until 10 successful and 10 unsuccessful shots were obtained. Fixation and visual search patterns for the expert and near-expert players were compared, as were successful shots to unsuccessful shots, on such variables as number of fixations, fixation location, and fixation duration. Vickers (1996) expected to replicate the position-only and movement-plus-position hypothesis findings in order to interpret differences in expertise, and to shed light on processes that may have occurred in made versus missed shots. While the near-expert performers possessed a sequence of fixation patterns that were supportive of the movement-plusposition hypothesis, none of Abrams et al. (1990) hypotheses seemed to adequately explain the results she obtained for the expert shooters. The expert shooters were characterized by an eye movement strategy that was unlike any proposed by Abrams et al. (1990). Experts took significantly longer to prepare

PAGE 65

56 the free throw (M = 1899 ms versus M = 1386 ms for near-experts), made fewer fixations than near-experts during the preparation and impulse phases of the shot, and generated a greater frequency of fixations during the execution phase. More importantly, the final fixation and quiet eye durations before the initiation of the movement was significantly longer in the expert shooters. In addition, experts suppressed their vision to a greater extent during the execution phase than did the near-experts. That is, they either blinked or diverted their gaze to areas other than the hoop, the ball, or their hands during the key propulsion element of the shooting action. From this unexpected pattern of results, Vickers (1996) generated the locationsuppression hypothesis for aimed limb movements to distant targets. She operationalized aiming skills to far targets as those in which the individual maintained control of the object to be directed to the target only to the point of release. While much is known about the coordination of the hand and the eyes in near aiming tasks (e.g., Abrams, 1994; Carlton, 1981a, 1981b; Gauthier, Semmlow, Vercher, Pedrono, & Obrecht, 1991; Guitton & Voile, 1987; Prablanc & Pellison, 1990), very little research has been conducted with the express purpose of investigating eye-hand coordination in far aiming tasks. As has already been discussed in the introduction of this paper, the location aspect of Vickers' (1996) hypothesis proposed that fixations of relatively long durations must be made to specific target locations during the preparatory phase of the movement. Here, the final fixation and quiet eye durations are invaluable for optimal ai min g performance,

PAGE 66

57 because it is believed that the programming of the parameters of the movement, such as the location and distance to the target, the optimal force and velocity used to initiate the movement, and the relative timing and coordination of the limbs occur in this time period. As evidenced by the expert shooters high free-throw shooting percentages, the longer the duration of the quiet eye variable, the better the performance. An equally vital characteristic of expert free-throw shooting occurred during the execution phase of the movement. Vickers (1996) noticed that a suppression mechanism was used to possibly reduce interfering visual information that may have resulted as a consequence of the aiming movement. Expert performers, she claims, have developed the ability to divert their visual attention from the target during execution by blinking or orienting their gaze to other elements of the visual field. Since the parameters of the movement have already been planned in the preparation phase, and modified in the impulse phase, visual attention is deemed unnecessary in the execution phase for successful completion of the aiming task. Thus, expert performance in far aiming tasks, such as those found in billiards, darts, and archery is dependent on the important variables of quiet eye duration and visual suppression (Vickers, 1996). In fact, if quiet eye duration is critical to the organization and parameterization of underlying neural structures, as has been suggested, then experimental manipulations of this time should result in changes in performance.

PAGE 67

58 Aside from direct occlusion techniques which restrict the performer's vision, but unfortunately undermines the ecological validity of the experimental situation, two methods are readily available which may influence quiet eye duration. A large body of evidence has demonstrated that manipulations in the variables of task complexity and task duration influence the information processing stages of response programming and response preparation (Schmidt, 1988), and therefore are likely to impact on quiet eye durations. The literature pertaining to each of these variables will be reviewed in the following section. The Influence of Task Complexity It has long been known that there is a direct relationship between human performance capabilities and the informational load of a particular task or set of tasks (Fitts & Posner, 1967; Hick, 1952; Hyman, 1953). That is, as the level of response uncertainty (informational load) increases, so too does reaction time (RT). More importantly, laboratory research tends to indicate that RT to a single unanticipated visual stimulus is in the order of 180-220 ms, with this delay composed of latencies associated with stimulus detection, response preparation, and neural and muscular activity associated with a simple key press (e.g.. Wood, 1977). In numerous studies, greater durations of reaction time (RT) and preparation time are required to adequately perform more complex motor responses (e.g., Henry, 1980; Henry & Rogers, 1960; Kerr, 1978; Klapp, 1977, 1980; Schmidt, 1988). In Henry and

PAGE 68

59 Rogers' (1960) seminal experiment, participants were required to perform a series of movements that varied in their level of complexity, from a simple finger lift to a single ball grasp (intermediate complexity) to a double ball strike (greatest complexity). Results indicated that RT and MT increased proportionally with increases in task complexity. Because the stimuli in each condition were held constant, Henry and Rogers (1960) argued that the more complex motor tasks required more time to program the movements in the response programming stage. Concomitant increases in RT as a function of increasing task complexity have also been reported in studies in which key-pressing movements of different complexities have been used (Klapp, 1977; Klapp & Erwin, 1976; Klapp & Greim, 1979), in which subjects were required to produce a series of morse-code transmissions. More complex transmissions were characterized by longer RT's in the production of the movement than less complex tasks. Similar results have been observed by Sternberg, Monsell, Knoll, and Wright (1978) in their studies of both typed and spoken sequences of words. More complex sequences, it was hypothesized, required greater parameterization in the response programming stage, and therefore greater RT's were needed to complete these tasks. In relation to manual aiming processes, both Falkenberg and Newell (1980) and Newell, Hoshizaki, Carlton, and Halbert (1979) used coincident timing tasks to assess the influence of increasing levels of complexity on RT. In particular, it was noted that as

PAGE 69

60 movement velocities changed, so too did the participant's RT's. Thus, these researchers hypothesized that movement velocity was a key parameter in the programming of a motor response. Other parameters, such as direction and force of the movement, have been proposed to influence the programming time of a motor response to stimuli of varying complexity (Klapp & Greim, 1979; Spijkers, 1989; Spijkers & Steyvers, 1984; Temprado & Spijkers, 1996; Vickers, 1996). In relation to the location aspect of Vickers (1996) hypothesis, it follows that if the quiet eye duration is directly related to a period of cognitive programming, more complex aiming behaviors should be characterized by longer quiet eye durations. To date, this relationship has never been empirically tested. Thus, one of the purposes in Experiment 1 of the present study was to investigate the influence of task complexity on quiet eye duration. The Influence of Task Duration Much in the same manner that changes in task complexity influenced RT's in the motor programming stage of information processing, researchers have demonstrated that changes in the duration in which the performer is required to execute a response impacts upon overall performance (Shapiro, 1977, 1978; Summers, 1977). Interestingly, although decrements in the precision of the response are observed, the relative temporal phasing and sequencing of the movement often remains constant (Gentner, 1987).

PAGE 70

61 For example, Shapiro (1977) noted that the proportion of time spent in each phase of a learned sequence of wrist movements remained highly similar under time compressed trials and normal conditions. Carter and Shapiro (1984) have also documented this seemingly invariant temporal structure in electromyographic (EMG) measures of the phasing of relevant muscles when participants were required to speed up their gait in a running task. This proportional timing phenomenon, in which movements are performed in highly similar (although not exactly identical) relative phase durations, has been chronicled in such motor activities as typing (Terzuolo & Viviani, 1979), running (Shapiro, Zemicke, Gregor, & Diestel, 1981), jumping (Lee, Lishman, & Thomson, 1982), piano playing (Shaffer, 1980, 1984), and breathing (Clark & von Euler, 1972). Regardless of whether the required response was accelerated or extended in terms of overall movement time, phase ratios remained remarkably constant. Thus, it is expected that reductions in the time the performer is allowed to initiate a motor task, such as the billiards stroke, will not drastically alter the relative phases of the stroke. Indeed, the proportional duration model would predict that the preparation, backswing, foreswing, and contact phases of the stroke would not change in terms of their relative times to each other when constraints are placed on the time allotted to the execution of the shot.

PAGE 71

62 However, in relation to the critical variable of quiet eye duration in Vickers (1996) theory, reductions in the time allotted for the participant to complete the task may reduce the time that he or she spends in the quiet eye phase, thus limiting the extent of cognitive programming and impairing performance. To date, however, no manipulations of task duration have been conducted in studies examining the relationship between quiet eye duration and performance in far aiming tasks. Therefore, this investigation was the focus of Experiment 2 in the present study. In summary, examinations into the variables associated with expertise in motor activities has yielded mixed results. Typically, researchers have indicated that software components, such as visual tracking patterns, anticipatory capabilities, and declarative and procedural knowledge, are more predictive of athletic expertise than hardware components. However, much of this research is descriptive in nature, and does not adhere to the processes of stringent theory testing and development. Vickers (1996) was one of the few researchers to develop a theory of visual control when performers are required to aim at a far target. By examining the impact of variables (task complexity and task duration) known to influence the response programming stage of the human information processing system, the present study will assess whether Vickers' (1996) quiet eye variable is truly reflective of this underlying cognitive processing. In addition, an attempt was made to determine whether highly-

PAGE 72

63 skilled billiards players differed from lesser-skilled players in their visual search processes as they executed their shots in a more ecologically valid testing environment.

PAGE 73

CHAPTER 3 METHODS Experiment 1 In this experiment, the complexity of the aiming task (shooting pool) was manipulated to determine if accompanying changes in quiet eye duration would result. It has been posited that the quiet eye period reflects an index of cognitive activity, such that the parameters of the task to be performed are delineated and programmed (Vickers, 1996). Therefore, it was hypothesized that as the complexity of the task increases, so too would the concomitant period of cognitive programming (e.g., Henry & Rogers, 1960). Thus, quiet eye duration should be greater in more complex tasks versus less complex tasks. If this contention is supported, a stronger argument can be forwarded in favor of the eye-mind assumption of gaze behavior. Participants Twenty-four right-handed male participants were tested, and were categorized as either highly-skilled or lesser-skilled billiards players based on two criteria. These criteria included years of experience in the sport of billiards, as well as performance on an ini tial skills test. To be eligible for the highly-skilled category, players were required to have at least eight years of playing experience, including competition in at least one sanctioned billiards tournament. Members of this group were also required to complete the initial 64

PAGE 74

65 skills test in 12 or fewer shots. Requirements for the lesser-skilled participants included less than three years of playing experience, no competitive tournament participation, and a score of 20 to 30 shots on the initial skills test. The initial skills test is described in more detail in the procedures section. The highly-skilled group consisted of 12 players (M = 23.17 years of age, SD = 3.01 years) who had an average of 9.1 years of experience in the sport of billiards (SD = 2.64 years) and played 3.33 days per week on average (SD = 0.98 days). Members of this group also had competed in an average of 14.25 (SD = 14.97) sanctioned tournaments in their career, winning 2.25 of them (SD = 3.19). In addition to their experience, each of the 12 highly-skilled participants successfully completed the test in fewer than 12 shots (M = 10.67 shots, SD = 2.46 shots). The lesser-skilled group also consisted of 12 players (M = 21.83 years of age, SD = 1.85 years), averaging 2.67 years of playing experience (SD = 0.65 years). These participants had no competitive experience, played only 1.25 days per week (SD = 0.45 days) and scored an average of 26.67 shots (SD = 3.17 shots) to complete the initial skills test. The number of participants used in this study was determined by CohenÂ’s (1977) table for power and effect sizes. The values used in the tables were: a = .05 (level of significance), u = 18 [Skill Level (2-1) x Complexity Level (3-1) x Accuracy (2-1) x Trial (10-1)], f = .40 (effect size), and power = .80.

PAGE 75

66 A pparatus Eve Movement Recording Eye movement data were collected with an Applied Sciences Laboratories (ASL; Waltham, MA) 4000SU mobile corneal reflection unit. This video-based monocular system measured the eye line of gaze with respect to a head-mounted eye camera by computing the relative positions of two features of the eye, the pupil and the corneal reflex (a reflection of a near-infrared light source from the surface of the cornea), in relation to the optics. Both the infrared beam and the image of the participant's eye were reflected from a visor mounted on the helmet, which was coated to be reflective in the near infrared region and transmissive to visible light. The line of gaze was computed by measuring the vertical and horizontal distances between the center of the pupil and the corneal reflection after correcting for second-order effects. The resulting displacement data were recorded and processed by an external Gateway 2000 486SX/166 microcomputer via a 30 m cable attached to the participant's waist. To record the field of view as observed by the participant, an Elmo MP481 color scene camera was positioned near the eye. This allowed for a view of the scene from the same position as would be observed by the participant, while avoiding the problems of parallax error and the mounting of stationary cameras in front of the subject. Since the participant was free to move about the environment, the use of the scene camera was

PAGE 76

especially beneficial, as the scene constantly changed with each shift in the individual's position. 67 To assess the exact location of gaze, the 4000 SU processor superimposed a white cursor representing 1 deg of visual angle on the video image produced by the scene camera. These images were recorded via an Akai VS-X9EGN S-VHS video recorder, and used for data analysis. The 4000 SU possessed an accuracy of + 1 deg in both the vertical and horizontal directions, and a precision of better than 0.5 deg. Finally, the system sampled at a rate of 60 Hz, and point of gaze was updated for each frame of video (every 33.3 ms). Vision and Action Coupling In an effort to more effectively couple vision with action, the participant's gaze behaviors were recorded simultaneously with the action phases of the b illiar d stroke. This was done by interfacing the data from the 4000 SU system with an image recorded by an external Panasonic WV-PS03/B S-VHS video camera (positioned perpendicular to the direction of the arm motion of the player) using a Panasonic WJ-MX10 digital production mixer. The mixer created a split-screen effect, in which the lower left portion of the frame showed the participant performing the billiard stroke, while the right portion of the frame displayed his gaze position as recorded by the eye and scene cameras. This method of coupling vision and action has been successfully used by Vickers (1992, 1996) to

PAGE 77

68 demonstrate the high degree of coordination between visual search patterns and motoric performance. Both the gaze locations and the temporal aspects of each phase of the stroke were then analyzed in a frame-by-frame fashion. Billiards Arrangements The experimental conditions, including the billiards proficiency test of the highlyskilled and lesser-skilled participants, were performed on a Brunswick 4.5 ft x 9 ft (1.37 m x 2.74 m) billiards table, which had a playing surface area of 100 in x 50 in (254 cm x 127 cm). The comer pockets were 5 in (12.7 cm) in width, with the side pockets being slightly larger in width (5.5 in, 13.97 cm). Standard modem composition balls measuring 2.25 in (5.72 cm) in diameter and weighing 6 ounces (170 g) were used. To control for the possible confound of instrumentation (Campbell & Stanley, 1963), players were required to use a standard 57 in (144.78 cm) billiards cue, weighing 18 oz (510 g). The arrangement of balls for the initial skills test, as well as for each of the three tasks of varying complexity levels, was controlled by the experimenter by placing small colored markers on the felt surface of the billiards table. Procedure Initial Performance Test Upon arrival at the Motor Behavior Laboratory, participants completed an informed consent form and a short demographic questionnaire assessing variables such as age, years of billiards playing experience, and whether they had normal or corrected

PAGE 78

69 vision. The players were then instructed as to what was expected of them throughout the course of the experiment. Once questions had been answered completely by the experimenter, each individual was given the billiards cue and led to the table, where an arrangement of balls could be seen similar to those that might occur in a typical game of nine-ball. This arrangement is illustrated in Figure 3-1. The participants were asked to pocket the balls (in order from one to nine) in as few shots as possible. The balls were arranged in such a fashion that high caliber players would be able to complete the initial test in less than 12 shots, while lesser-skilled participants would require more than 20 shots. Performance Trials Once the initial test was completed, the participants were fitted with the eye tracking system, and initial calibration procedures were performed. This calibration procedure consisted of having the person fixate on a sequence of nine equidistant points located on a target board placed on top of the billiards table. Upon completion of the calibration procedure, a series of shots of three levels of complexity (easy, intermediate, and hard) were performed. Each player was instructed on the correct type of speed and english to place on the cue ball, and three practice trials for each level of complexity were given. They then performed consecutive shots until they reach 10 successful and 10 unsuccessful outcomes per complexity condition, a research goal of which they were not

PAGE 79

70 Figure 3-1. Initial performance test.

PAGE 80

made aware. For each condition, performance was deemed successful only when the object ball was pocketed. Shots were considered unsuccessful if the object ball was not sunk or the cue ball was inadvertently pocketed (scratch shot). 71 After the participants had achieved 10 successful and 10 unsuccessful shots on a particular complexity level, a 3 min rest period was given, followed by a recalibration of the eye tracking system. The players then progressed to the next series of shots of a different complexity level. The sequence of task complexity was randomly assigned and counterbalanced between each participant (e.g., participant A progressed in a sequence of hard, easy, and intermediate shots, while participant B encountered a sequence of easy, intermediate, and hard shots). Total time to complete the task for each participant depended on the number of shots necessary to achieve 10 hits and misses. In general, each participant completed this experiment in 45 min or less. Shots of Varying Complexity For the performance task, participants were required to pocket a series of balls with shots of differing complexity. Specifically, three conditions of complexity were used. In the easy condition (EC), the object ball lay near the comer pocket, and could be made with a cut shot using bottom-left english (Martin & Reeves, 1993). This shot is illustrated in Figure 3-2. The intermediate condition (IC) featured a "cushion first" approach to sinking the object ball (Martin & Reeves, 1993), where the object ball lay near the cushion

PAGE 81

72 while the cue ball sat on the opposite side of the table. The object ball was partially occluded by another ball, such that the player had no direct way of hitting the object ball directly. In order to sink the object ball the player needed to strike the cue ball with straight-follow english and hit the cushion prior to hitting the object ball. This shot is demonstrated in Figure 3-3. Finally, the hard condition (HC) featured a "carom draw" shot, in which the cue ball needed to be struck sharply with straight-bottom english so that it caromed off the object ball into another ball located near the comer pocket (Martin & Reeves, 1993). The difficulty of the carom draw shot lies in the especially precise angle of the carom, and due to the fact that a number of other balls were located on the table limiting the options of the player. This shot is illustrated in Figure 3-4. Given the fact that participants of differing skill level and experience were performing these tasks, a number of alternative approaches to making the shots described above may be observed. In particular, it was expected that the highly-skilled players possessed the cognitive ability to select the most appropriate response for a given shot to a much greater extent than the lesser-skilled performers (Abemethy, Neal, & Koning, 1994). Therefore, instructions as to the most pertinent action that will lead to the correct execution of the shot were given prior to the start of the task.

PAGE 82

73 Figure 3-2. Required shot in the easy complexity condition.

PAGE 83

74 Figure 3-3. Required shot in the intermediate complexity condition.

PAGE 84

75 Figure 3-4. Required shot in the hard complexity condition.

PAGE 85

76 Dependent Measures Performance The performance measure consisted of the number of shots required to achieve the criterion level of 10 successful and 10 unsuccessful shots for each of the task complexity conditions. Specifically, the percentage of shots made to shots attempted was calculated, with higher values indicating greater performance. Visual Tracking Measures For each participant, videotaped eye movement data was recorded for each shot. Although the participants were required to shoot until they made 10 successful and 10 unsuccessful shots of each complexity level, a review of the data indicated that some trials were unable to be analyzed due to calibration or video errors. Therefore, only five randomly selected successful and five unsuccessful shots for each level of complexity could be analyzed in a frame-by-frame fashion using an Akai VS-X9EGN S-VHS video recorder. Each frame constituted 33.3 ms of data, and a fixation was defined as three or more consecutive frames (99.9 ms or more) in which the cursor was located in the same space in the visual environment (e.g., Ripoll, 1991; Vickers, 1992, 1996; Williams et al., 1994). The variables of fixation location, average fixation duration, quiet eye duration, and the number of blinks present in each phase (the suppression aspect of the locationsuppression hypothesis) was then subjected to statistical analysis.

PAGE 86

77 For fixation location, the number of fixations allocated to specific areas of the visual display was recorded. In particular, fixations directed to four areas of interest were examined and subjected to statistical analysis: The cue ball, the object ball, the target cushion (or second object ball in the hard condition only), and the intended pocket of the object ball. Any remaining fixations (e.g., on the cue stick, felt surface, etc.) were defined in a category labeled “other” areas. For the dependent measure of average fixation duration, the time spent (in ms) allocating fixations to each of the predetermined locations of interest was calculated. Of critical importance to this experiment, the quiet eye duration (in ms) in the preparatory phase of the stroke was recorded and compared to differences in performance between the groups for each complexity condition. In addition, to test the Vickers (1996) visual suppression hypothesis, the number of eyeblinks as calculated by the absence of the cursor on the videotaped image, was assessed during the backswing, foreswing, and flight phases of the stroke. Temporal Components of the Billiards Stroke To assess the temporal components of the stroke, four distinct phases were identified from the sagittal view of the performer as provided by the external camera: (1) the preparatory phase, defined as the time from the start of the task (when the performer positioned himself over the cue ball) until the first observable movement towards the

PAGE 87

78 striking of the cue ball (e.g., the duration in which the billiards player leaned over the table to begin the shot until the moment he initiated the final backswing motion of the cue); (2) the backswing phase, defined as the duration from the final backswing movement until the cue began to move forward towards the cue ball; (3) the foreswing phase, defined as the time from foreswing initiation until the tip of the cue came into contact with the cue ball; and (4) flight phase, defined as the duration of the flight of the cue ball until it struck the object ball (e.g., includes cue ball contact with the object ball). Total time (in ms) spent in each phase of the stroke between the highly-skilled and lesser-skilled performers was calculated for each of the varying complexity conditions. This was accomplished by analyzing the total number of video frames (each frame equated to 33.3 ms) corresponding to each phase of the stroke. Experiment 2 In this experiment, the duration of the task was manipulated such that the participant performed a billiards shooting task under normal and time-constrained conditions. Again, the quiet eye duration was assessed, as were the temporal aspects of each phase of the aiming movement. It was hypothesized that quiet eye durations would be reduced because of the time constraints, thus leading to decrements in performance as compared to the normal condition. However, it was also hypothesized that the temporal ratios of each phase of the behavior would remain constant for each constrained condition as compared to the normal condition. Therefore, changes in performance may be explained

PAGE 88

79 via changes in the time allotted for programming of the movement, rather than in the relative temporal changes in the movement itself. Again, it was expected that this result would provide support for the strong eye-mind assumption, indicating that constraints imposed on the cognitive processes underlying eye movement behaviors and the planning of the movement directly influence performance. Participants The individuals in this experiment were those who participated in Experiment 1. Again, the number of people to be tested in this study was determined by Cohen's (1977) table for power and effect sizes. The values used in the tables were: a = .05 (level of significance), u = 18 [Skill Level (2-1) x Duration (3-1) x Accuracy (2-1) x Trial (10-1)], f = .40 (effect size), and power = .80. A pparatus All apparati, including the eye tracking unit, video recording and mixing devices, and billiards table, cue, and balls were identical to those described in Experiment 1. However, only the arrangement of balls employed in the intermediate complexity condition (IC) of the first experiment were included in Experiment 2. In addition to these instruments, the countdown timing mechanism feature of the ReAction Coach (S.T.A.R.T. Technologies, NY) movement timing system was used to direct the participants as to the time they were allotted for the task. The ReAction Coach

PAGE 89

80 is a portable electronic device (16 cm long and 12 cm wide) which uses auditory and visual cues to ostensibly promote quicker and more explosive motoric abilities (Singer, Cauraugh, Frehlich, Barba, Kelly, & Umans, 1993). Auditory cues can be made available to the participant in the form of a series of beeps, and these cues can be adjusted in volume and sensitivity by the experimenter. In particular, the countdown mode of the ReAction Coach was used in Experiment 2, in which three preliminary auditory beeps were presented (priming cues indicating that the trial is imminent), followed by a tone of longer duration (indicating the initiation of the trial). Once this tone sounded, the internal timer in the device was activated, and was terminated only when a loud sound was generated by the participant (i.e„ the striking of the cue ball with the billiard cue). Durations (in ms) between onset of the final tone and termination of the timer were displayed on the liquid crystal display (LCD) function of the ReAction Coach, and were used to provide feedback to the participants. Procedure Upon completion of Experiment 1, participants were given instructions as to the goals and nature of the second experiment. Specifically, they were required to perform the IC complexity task in a randomly assigned and counterbalanced order of two different conditions. In the 25% constrained condition, the shooter was required to initiate their stroke within a duration of 75% of the average time spent in the initiation of the stroke as recorded in an unconstrained condition. The unconstrained condition times were assessed

PAGE 90

81 during data analysis in Experiment 1, where the participant was allowed to take as much time as he wished to perform the IC task. For example, if a particular participant took an average of 4 sec to execute the shot in the unconstrained IC condition, he was required to perform the shot in 3 sec or less in the 25% constrained condition. The participant received three practice trials preceding the test trials. In the 50% constrained condition, the player needed to execute the shot within a duration of 50% of the average time recorded to initiate the shot in the unconstrained condition. To use the same example, a player averaging 4 secs per shot in the initial testing session (IC condition in Experiment 1) had to initiate the shot within 2 sec or less in the 50% constrained condition. Again, three practice trials were given to the participant before test trials were recorded. To aid the participants in determining when they would be allowed to initiate a stroke, the countdown timing mechanism feature of the ReAction Coach movement timing system was used. Players stood in a ready position over the shot and closed their eyes, and were informed that an experimental trial was beginning upon hearing the initial warning tones produced by the ReAction Coach. After hearing the final auditory cue, the timer was activated and the participants were allowed to open their eyes and begin their preparatory motions used to execute the shot. Based on the times recorded by the ReAction Coach measuring device, the experimenter provided feedback to the participant after each trial as

PAGE 91

82 to whether they were initiating the stroke within the allotted time frame. The players performed consecutive shots until they reach 10 successful and 10 unsuccessful outcomes per duration condition. Again, this was a research goal of which they were unaware. Trials initiated after the constrained period were considered as inaccurate, and were not included in the data analyses. Dependent Measures The dependent measures were recorded and analyzed in an identical fashion to those described in Experiment 1 . In particular, performance percentages, durations (in ms) spent in each phase of the billiard stroke, and the visual tracking measures of fixation location, average fixation duration, quiet eye duration, and the number of blinks present in each phase were assessed.

PAGE 92

CHAPTER 4 RESULTS Experimental Design Statistical analyses were conducted with the SPSS for Windows 6.1.4 computer applications package. For all statistical analyses in each of the two experiments, alpha level was set at p<.05. Simple main effects and ScheffdÂ’s post hoc tests were used to follow up significant analysis of variance (ANOVA) results when appropriate. In addition, for each repeated measures ANOVA design, violations of the assumptions of sphericity were assessed. In those instances when the assumptions were violated, Greenhouse-Geisser adjustments to the level of significance are reported. Experiment 1 Statistical Analyses Performance The performance percentage was determined by dividing the number of shots made by the total number of shots taken to reach 10 hits and 10 misses for each group at each level of complexity. A 2 (Skill Level) x 3 (Complexity Level) ANOVA with repeated measures on the last factor was conducted to assess whether performance differed significantly between the highly-skilled and lesser-skilled participants, and indicated main 83

PAGE 93

84 effects for Skill Level, F(l,22) = 75.99, p < .001, and Complexity Level, F(2,44) = 116.28, £ <.001. Post hoc analyses revealed that highly-skilled players made significantly more shots (M = 67.53%, SD = 18.17) that did lesser-skilled players (M = 47.88%, SD = 21.24). In addition, there were significant reductions in performance for both groups as the level of complexity increased, from M = 74.13% (SD = 12.99) for the EC condition, to M = 65.01% (SD = 14.21) for the IC condition, and M = 47.88% (SD = 13.96) in the HC condition. It appears as though the highly-skilled players sank significantly more shots in all levels of complexity than did lesser-skilled players. Also, the manipulations in complexity affected members of both groups in a proportionate manner. None of the interactions were significant. Overall performance for each group is presented in Table 41 . Table 4-1. Performance means and standard deviations for highly-skilled and lesser-skilled participants for each level of complexity. Comnlexitv HiehlvSkilled M HiehlvSkilled SD LesserSkilled M Lesser-Skilled SD Easy 83.30% 6.74 64.97% 11.15 Intermediate 74.84% 6.17 55.19% 13.19 Hard 44.47% 7.52 23.47% 10.52

PAGE 94

85 Temporal Components of the Billiards Stroke To determine the effect of changes in complexity on the overall duration (in ms) of the billiards stroke, a 2 (Skill Level) x 3 (Complexity Level) x 2 (Accuracy) x 5 (Trials) ANOVA with repeated measures on the last three factors was conducted. Significant main effects were found for Accuracy, F(l,22) = 4.78, p< .05, and for Complexity Level, F(2,44) = 78.17, p < .001. Accurate shots were characterized by longer durations in both groups (M = 3093.06 ms, SD = 898.23) as compared to missed shots (M = 2885.69 ms, SD = 752.84). In addition, overall shot duration time increased significantly for both groups as complexity level increased (EC M = 2489.38 ms, SD = 615.33; IC M = 2656.67 ms, SD = 540.6; and HC M = 3822.08 ms, SD = 600.44). No significant differences were found for Skill Level, as highly skilled mean duration was 2977.08 ms (SD = 870.7) and lesser-skilled mean duration was 3001.67 ms (SD = 798.1). More importantly, a significant interaction between Complexity Level and S kill Level was observed, F(2,44) = 5.33, p < .05. Within each skill level, total duration for the HC condition was significantly greater than either the EC or IC conditions. The overall stroke duration in the EC and IC conditions were not significantly different from each other. This interaction is displayed in Table 4-2.

PAGE 95

Table 4-2. Total duration of the billiards stroke for highlyskilled and lesser-skilled participants for each level of complexity. 86 Comnlexitv HiahlvSkilled M HighlvSkilled SD Lesser-Skilled M Lesser-Skilled SD Easy 2322.50 ms 435.6 2656.25 ms 795.1 Intermediate 2587.50 ms 511.3 2725.83 ms 569.9 Hard 4021.25 ms 568.2 3622.92 ms 632.7 From these results, it appears as though accuracy of the shot for both groups was related to the time taken to perform the shot. On average, successful shots were characterized by a 207 .4 ms increase in total duration as compared to unsuccessful shots. In addition, for each level of complexity, time taken to perform the shot increased significantly for both groups in the most complex condition. Phase Durations To determine whether manipulations in task complexity led to specific changes in duration for each of the four phases of the billiards stroke, a 2 (Skill Level) x 3 (Complexity Level) x 4 (Phase) x 2 (Accuracy) x 5 (Trials) ANOVA with repeated measures on the last four factors was conducted. A significant main effect was found for the variables of Phase, F(3,66) = 12.73, p<.01, and Complexity Level, F(2,44) = 5.65, p_< .05. Both groups spent significantly more time in the preparation phase of the stroke (M = 1731.46 ms, SD = 632.5) than in any of the other phases. Means for the other phases included: backswing phase (M = 491.84 ms, SD = 40.7), foreswing phase (M = 125.32

PAGE 96

87 ms, SD = 9.27), and flight phase (M = 597.87 ms, SD = 65.9). In addition, the HC condition led to the greatest mean phase time (M = 955.52 ms) than did the EC (M = 603.85 ms) or IC (M = 650.50 ms). No other main effect for Accuracy, Skill Level, or Trial was significant. More importantly, a significant Phase by Complexity Level interaction was observed, F(6,132) = 4.32, p < .03. Follow-up tests indicated that the backswing, foreswing, and flight phases for each level of complexity were similar in duration. However, durations of the preparation phase were significantly greater (M = 2532.16 ms) in the HC condition as compared to the EC or IC conditions (M = 1240.72 and M = 1421.52 ms, respectively). Thus, it appears as though both groups required significantly more time to prepare their shots in the most complex condition in Experiment 1 as compared to the other two levels of complexity. Interestingly, a Skill Level by Phase interaction was not observed to be significant. That is, highly-skilled players spent a similar amount of time in each of the four phases of the stroke as compared to their lesser-skilled counterparts. Although not significant, the average phase durations collapsed across the three complexity levels for both skill levels are presented in Table 4-3.

PAGE 97

Table 4-3. Total duration for each phase of the billiards stroke for highly-skilled and lesser-skilled participants collapsed across complexity level. 88 Phase HiehlvSkilled M HiehlvSkilled SD LesserSkilled M LesserSkilled SD Preparation 1763.93 ms 432.50 1698.99 ms 832.54 Backswing 490.02 ms 33.33 493.67 ms 48.05 Foreswing 120.00 ms 9.56 130.63 ms 10.73 Flight 566.68 ms 103.17 629.07 ms 245.27 It should be noted that a comparison of mean durations for each phase did not differentiate successful from unsuccessful shots. None of the interactions except the Phase by Complexity Level interaction were significant. Overall, it appeared as though time spent in each phase of the billiards stroke did not differ between highly-skilled and lesserskilled players, and could not be used as a means of explaining shots that were successful as compared to shots that were unsuccessful Although the more skilled players spent more time in the preparation phase for each level of complexity, these differences were not statistically significant. Visual Tracking Measures To assess whether changes in task complexity were associated with concomitant changes in visual tracking measures, the dependent variables of the number of fixations and average fixation duration (in ms) were analyzed via separate 2 (Skill Level) x 5

PAGE 98

89 (Location) x 2 (Accuracy) x 5 (Trials) ANOVA’s with repeated measures on the last three factors for each level of complexity. Easy complexity level. For the EC condition, only three levels of the Location factor were activated in the study. Therefore, a 2 (Skill Level) x 3 (Location) x 2 (Accuracy) x 5 (Trials) ANOVA with repeated measures on the last three factors was conducted for this condition only. Significant main effects for Location, F(2,44) = 579.48, p < .001, and Skill Level, F(L22) = 54 . 35 , £ < .01 were observed for the dependent variable of number of fixations. Specifically, more fixations were directed to the cue ball (M = 3.00) and the object ball (M = 2.83) than to the cushion (M = 0 . 00 ), the pocket (M = 0 . 00 ), or other areas of the display (M = 0.50). In addition, lesser-skilled players made more fixations (M = 1.48) on average to each location than did highlyskilled players (M = 1.06). A significant interaction between Location and Skill Level was also observed for the number of fixations in the EC condition, F(2,44) = 20.44, p < .01. Lesserskilled participants made significantly more fixations to the cue ball (M = 3.54) and object ball (M = 3.33) than did the highly-skilled participants (M = 2.46 and M = 2.33 fixations, respectively). A similar pattern of results was observed for the dependent variable of average fixation duration in the EC condition. Significant main effects for Location, F(2,44) =

PAGE 99

90 204.87, £ < .001 indicated that greater average fixation durations were directed to the cue ball ( M = 396.02 ms) and the object ball (M = 418.26 ms) than to any of the other locations. The main effect for Skill Level was also significant, F(l,22) 7 . 57 , ^ < .05. On average, the highly-skilled performers made fixations of longer duration (M = 197.47 ms) than did lessskilled players (M = 161.24 ms). More importantly, a significant interaction between Skill Level and Location was found, F(2,44) = 6.41, p < .05. Post hoc tests revealed that highly-skilled performers fixated on the target ball significantly longer (M = 498.96 ms) than did the lesser-skilled players (M = 337.57 ms). The overall pattern of results for both dependent variables fixation number and average fixation location for the EC condition is summarized in Table 4-4. Table 4-4. Mean number of fixations and average fixation durations for highly-skilled and lesser-skilled participants for the EC condition. Location HiehlvSkilled Fix. Num. HiehlvSkilled Avg. Fix. Dur. Lesser-Skilled Fix. Num. Lesser-Skilled Avg. Fix. Dur. Cue Ball 2.46 418.82 ms 3.54 373.23 ms Object Ball 2.33 498.96 ms 3.33 337.57 ms Cushion 0.00 0.00 ms 0.00 0.00 ms Pocket 0.00 0.00 ms 0.00 0.00 ms Other 0.50 69.58 ms 0.50 95.42 ms

PAGE 100

91 From these results, it can be inferred that highly-skilled players made fewer fixations of longer duration to specific areas of the display as compared to the lesserskilled participants. In particular, the more advanced group fixated longer on the object ball than did the less advanced group. However, these differences did not explain the discrepancy between successful versus unsuccessful shots, because neither a main effect nor any interaction was found for the independent variable of accuracy. Thus, for the EC condition, no differential pattern of visual tracking could be determined for successful and unsuccessful shots. Intermediate complexity level. For the IC condition, significant main effects in the dependent variable of number of fixations were observed for Location, F(4,88) = 53.01, p < .001, and for Skill Level, F(l,22) = 22.56, p < .01. Follow-up tests revealed that lesserskilled players demonstrated a greater number of fixations (M= 1.49) on average than did their higherskilled counterparts (M = LI 8), and for both groups more fixations were directed to the cue ball (M = 2.65), the cushion (M = 2.08), and the object ball (M = 1-33) than to the pocket (M = 0.25) or other areas of the display (M = 0.35). In addition to these main effects, a significant Location by Skill Level interaction was found, F(4,88) = 5.18, p < .02. The lesser-skilled players directed significantly more fixations to the cue ball (M = 3.00) than did the highly-skilled players. Number of fixations

PAGE 101

92 to the other four locations was similar for both groups. This pattern of results is illustrated in Table 4-5. For the dependent variable of average fixation duration, the only significant finding for the IC condition was a main effect for Location, F(4,88) = 156.77, p < .001. Again, both groups fixated longer on the cushion (M = 445.66 ms), the cue ball (M = 406.32 ms), and the object ball (M = 364.06 ms) than they did on the pocket (M = 51.56 ms) or other areas (M = 53.54 ms). No other main effect for Accuracy, Skill Level, or Trials was significant, nor were any of the interaction terms. Mean average durations for each location for both skill levels is displayed in Table 4-5. Table 4-5. Mean number of fixations and average fixation durations for highly-skilled and lesser-skilled participants for the IC condition. Location Hiehlv-Skilled Fix. Num. Hiehlv-Skilled Avg. Fix. Dur. Lesser-Skilled Fix. Num. LesserSkilled Avg. Fix. Dur. Cue Ball 2.29 428.47 ms 3.00 384.17 ms Object Ball 1.01 397.08 ms 1.67 331.04 ms Cushion 2.04 470.14 ms 2.13 421.18 ms Pocket 0.21 38.33 ms 0.29 64.58 ms Other 0.33 46.67 ms 0.38 60.42 ms Closer examination of the IC visual tracking dependent measures revealed that lesser-skilled participants made more fixations to locations in the display (specifically the

PAGE 102

93 cue ball) than the more skilled players. However, the groups did not differ in the average amount of time that they spent fixating on each location. Similar to the EC condition, the visual tracking measures were not useful indicators of performance, as no significant main effects or interactions were found for the independent variable of accuracy. That is, both successful and unsuccessful shots were characterized by similar visual tracking patterns. Hard complexity level. An analysis of the dependent measures of number of fixations and average fixation duration for the HC condition yielded si milar results to the other complexity conditions. The threeand four-way interactions were not significant for either dependent variable. For number of fixations, significant main effects were demonstrated for Location, F(4,88) = 377.41, p < .001, and for Skill Level, F(l,22) = 17.09, p < .01. Again, lesser-skilled performers made a greater number of fixations (M = 1.78) than did more highly-skilled billiards players (M = 1-50). In addition, the greatest number of fixations was directed the cue ball (M = 3.60), then to the object ball (M = 2.71), then to the second target ball (M = 1.04). Only a minimal number of fixations were allocated to the pocket (M = 0.44) or to other areas (M = 0.40). ScheffdÂ’s post hoc analysis indicated that each of these means were significantly different from each other, except for the comparison between the pocket and other areas. A significant Location by Skill Level interaction, F(4,88) = 8.26, p < .03, revealed that the lesser-skilled participants generated more fixations to the cue ball (M = 4.04) and

PAGE 103

94 to the object ball (M = 3.00) than did the more skilled players (M = 3.17 and M = 2.42 fixations, respectively). These results are displayed in Table 4-6. For the average fixation duration dependent variable, significant main effects were also found for Location, F(4,88) = 113.6, £ < .001, and for S kill Level, F(l,22) = 14.87, p < .01. The more advanced players possessed longer fixation durations on average (M = 347.56 ms) than did the lesser-skilled (M = 274.99 ms), with the longest fixation for both groups directed at the cue ball (M = 582.21 ms), the object ball (M = 447.38 ms), and the second target ball (M = 365.10 ms). Only short fixation durations were found for the pocket (M = 89. 17 ms) and other areas (M = 72.50 ms). Post hoc analyses revealed that each of the means were significantly different from each other, save for the pocket and other areas comparison. Finally, the Location by Skill Level interaction was again observed to be significant, F(4,88) = 3.57, p < .05. The highlyskilled players had longer mean fixations to the cue ball (M = 678.47 ms) and the object ball (M = 504.93 ms) than did the novice players (M = 485.95 ms and M = 389.83 ms). The means for each location are presented in more detail in Table 4-6.

PAGE 104

95 Table 4-6. Mean number of fixations and average fixation durations for highly-skilled and lesser-skilled participants for the HC condition. Location Hiehlv-Skilled Fix. Num. Hiehlv-Skilled Avg. Fix. Dur. Lesser-Skilled Fix. Num. Lesser-Skilled Avg. Fix. Dur. Cue Ball 3.17 678.47 ms 4.04 485.95 ms Object Ball 2.42 504.93 ms 3.00 389.83 ms Second Ball 1.04 376.88 ms 1.04 353.33 ms Pocket 0.46 102.92 ms 0.42 75.42 ms Other 0.42 74.58 ms 0.38 70.42 ms A review of the visual tracking measures in the HC condition revealed that the highly-skilled billiards players made fewer fixations of longer duration to the cue ball and object ball locations than less-skilled players. However, as in the EC and IC conditions, successful and unsuccessful shots were not discriminated by the variables of fixation number and average duration. Again, no significant main effect or interactions were observed for the accuracy independent variable. Thus, while significant differences existed in several visual tracking measures, they did not appear to be related to the changes in performance that were observed within each group. Quiet Eye Duration and Number of Eve Blinks Vickers (1996) noted that the two critical components of the location-suppression hypothesis were quiet eye duration and number of blin ks performed by the performer. To test this hypothesis, both quiet eye duration (in ms) and suppression of gaze (as defined as

PAGE 105

96 the number of blinks) were assessed using a 2 (Skill Level) x 3 (Complexity Level) x 2 (Accuracy) x 5 (Trials) ANOVA with repeated measures on the last three factors. It was expected that main effects would be observed for Skill Level and Complexity Level. More importantly, it was hypothesized that the interaction between these factors would be significantly different, such that quiet eye durations will be greater for the higher-skilled players at each level of task complexity. For the quiet eye dependent variable, significant main effects were found for Skill Level, F(l> 22 ) = 68.41, p < .001, Complexity Level, F(2,44) = 76.25, p < .001, and Accuracy F(l>22) = 165.19, p < .001. Successful shots were characterized by longer quiet eye durations (M = 561.94 ms) than unsuccessful shots (M = 213.61 ms), highly-skilled players had significantly longer quiet eye measures (M = 499.86 ms) than lesser-skilled players (M = 275.69 ms), and quiet eye duration increased linearly with increases in task complexity (M = 229.79 ms for EC, M = 314.17 ms for IC, and M = 619.38 ms for HC conditions). In addition, a significant three-way interaction of Accuracy by Complexity Level by Skill Level was observed, F(2,44) = 5.05, p < .05. The mean values of this three-way interaction are displayed in Table 4-7.

PAGE 106

97 Table 4-7. Mean quiet eye durations for successful and unsuccessful shots for both highlyskilled and lesser-skilled participants for each level of complexity. Complexitv Hiehlv-Skilled Quiet Eye Hits Hiehlv-Skilled Quiet Eye Misses Lesser-Skilled Quiet Eye Hits Lesser-Skilled Quiet Eye Misses Easy 379.17 ms 170.13 ms 271.67 ms 98.33 ms Intermediate 585.50 ms 236.67 ms 286.67 ms 148.33 ms Hard 1233.33 ms 395.50 ms 615.83 ms 233.17 ms For the both the highly-skilled and lesser-skilled groups, quiet eye durations were significantly longer for accurate shots when compared to inaccurate shots for each level of complexity. The largest quiet eye durations were observed for successful shots in the HC condition for both groups. Also, for accurate shots in the IC and HC conditions, highlyskilled players possessed quiet eye durations that were nearly double those of their lesserskilled counterparts. Therefore, it was inferred that complexity level had a direct influence on the amount of time spent in quiet eye for both skill levels. In addition, quiet eye durations were observed to be greater for the highly-skilled players over each level of complexity, particularly in the successful shots. Finally, and perhaps most important, quiet eye duration (unlike the other dependent variables) was significantly related to shot performance. Best performances for both groups were characterized by longer quiet eye durations; in some instances quiet eye for successful shots was three times longer than

PAGE 107

98 unsuccessful shots. These results were consistent with the pattern found by Vickers (1996) in her study of basketball ffee-throw shooters. In appears, therefore, that the location aspect of her theory is robust to the far aiming task of b illiar ds. To determine whether the suppression phase was also robust to the sport of billiards, the number of eye blinks during each shot was measured and statistically analyzed. However, there were no significant main effects or interactions for the eye blink variable. In fact, very few eye blinks were observed throughout the experiment, with the highly-skilled group averaging 0.17 blinks per shot and the lesser-skilled group averaging 0.24 blinks per shot. It appears as though blink rate was not related to expertise level, performance accuracy, or level of complexity. These results were not consistent with Vickers (1996) suppression aspect of the location-suppression hypothesis, but rather tended to support many other theories of visual control (e.g., Abrams, Meyer, & Komblum, 1990). Experiment 2 Statistical Analyses Performance As in Experiment 1, the performance percentage was determined by dividing the number of shots made by the total number of shots taken to reach 10 hits and 10 misses for each group at each duration level. A 2 (Skill Level) x 3 (Duration Level) ANOVA with repeated measures on the last factor indicated main effects for Skill Level, F(l,22) =

PAGE 108

99 32.32, g < .001, and Duration Level, F(2,44) = 6.83, £ <.003. Post hoc analyses revealed that highly-skilled players made significantly more shots (M = 67.39%, SD = 10.31) that did lesserskilled players (M = 52.07%, SD = 14.01). In addition, there were significant reductions in performance for both groups as the duration level decreased, from M = 65.01% (SD = 14.21) for the 100% (unconstrained) condition, to M = 61.11% (SD = 13.40) for the 25% constrained condition, and M = 53.06% (SD = 13.57) in the 50% constrained condition. The interaction was not significant. It appears as though the highlyskilled players sank significantly more shots in all duration levels than did lesser-skilled players. Also, the manipulations in duration affected members of both groups in a proportionate manner. Overall performance for each group is presented in Table 4-8. Table 4-8. Performance means and standard deviations for highly-skilled and lesserskilled participants for each duration level. Duration HiehlvSkilled M Hiehlv-Skilled SD Lesser-Skilled M Lesser-Skilled SD Unconstrained 74.84% 6.17 55.19% 13.18 25% Const. 68.00% 7.74 54.23% 14.58 50% Const. 59.33% 10.42 46.78% 13.83

PAGE 109

100 Temporal Components of the Billiards Stroke To determine the effect of manipulations of duration level on the overall time spent (in ms) to execute the billiards stroke, a 2 (Skill Level) x 3 (Duration Level) x 2 (Accuracy) x 5 (Trials) ANOVA with repeated measures on the last three factors was conducted. Significant main effects were found for Accuracy, F(l,22) = 4.48, p< .04, and for Duration Level, F(2,44) = 207.13, p < .001. Longer durations led to more accurate shots in both groups (M = 1854.31 ms, SD = 770.39) as compared to mis sed shots (M = 1720.14 ms, SD = 728.27). In addition, overall shot duration time decreased significantly for both groups as more constraints were placed upon the participants (100% unconstrained M = 2656.67 ms, SD = 540.6; 25% constrained M = 1592.08 ms, SD = 327.27; and 50% constrained M = 1 1 12.92 ms, SD = 192.22). No significant differences were found for Skill Level, as highly skilled mean duration was 1794.44 ms (SD = 719.31) and lesser-skilled mean duration was 1780 ms (SD = 784.54). However, none of the interaction terms were found to be significant. For descriptive purposes, the total time of the billiards stroke for each s kill level and duration condition is presented in Table 4.9.

PAGE 110

101 Table 4-9. Total duration of the billiards stroke for highly-skilled and lesser-skilled participants for each duration level. ComDlexitv HiehlvSkilled M HiehlvSkilled SD LesserSkilled M Lesser-Skilled SD Unconstrained 2587.50 ms 511.3 2725.83 ms 569.9 25% Const. 1673.75 ms 301.2 1510.42 ms 354.2 50% Const. 1122.08 ms 126.7 1 103.75 ms 257.7 Similar to the pattern of results found in Experiment 1, it appeared as though accuracy of the shot for both groups was related to the time taken to perform the shot. Successful shots were characterized by a 134.2 ms increase in total duration of the billiards stroke as compared to unsuccessful shots. In addition, total time spent in the execution of the stroke was significantly decreased in both groups as the participants moved from the unconstrained to the 25% and 50% constrained conditions. This result was to be expected, since the goal of the experiment was to determine the effect of time constraints on shooting performance. However, the lack of a significant accuracy by duration level interaction indicated that successful shots required longer total stroke duration than did less successful shots, regardless of the time constrained condition. Phase Durations To determine whether constraining the time allowed to complete the billiard stroke led to specific changes in duration for each of the four phases of the stroke, a 2 (Skill

PAGE 111

102 Level) x 3 (Duration Level) x 4 (Phase) x 2 (Accuracy) x 5 (Trials) ANOVA with repeated measures on the last four factors was conducted. Significant main effects were found for the variables of Phase, F(3,66) = 9.90, pc.Ol, and Duration Level, F(2,44) = 7.29, p_< .02. Both groups spent significantly more time in the preparation phase of the stroke (M = 742.51 ms) than in any of the other phases. Means for the other phases included: backswing phase (M = 397.48 ms), foreswing phase (M = 1 18.53 ms), and flight phase (M = 537.66 ms). In addition, the 100% unconstrained condition led to the greatest mean phase time (M = 664.17 ms) than did the 25% constrained (M = 398.02 ms) or 50% constrained conditions (M = 284.95 ms). No other main effects for Accuracy, Skill Level, or Trials were significant. More importantly, a significant Phase by Duration Level interaction was observed, F(6,132) = 4.68, p < .05. Follow-up tests indicated that the backswing, foreswing, and flight phases for each level of complexity were similar in duration. However, durations of the preparation phase were significantly greater (M = 1476.19 ms) in the 100% unconstrained condition as compared to the 25% constrained or 50% constrained conditions (M = 548.00 ms and M = 203.33 ms, respectively). No other interactions were significant. From these results, with increasing time constraints it appeared as though reductions in the overall duration of the billiards stroke came at the expense of the preparation phase only. The backswing, foreswing, and flight phases all remained

PAGE 112

103 relatively similar for each duration level condition. In addition, the decrements in preparation phase time were observed in both the highly-skilled and lesser-skilled groups. As in Experiment 1, a Skill Level by Phase interaction was not observed to be significant. That is, highly-skilled players spent a similar amount of time in each of the four phases of the stroke when compared to the lesser-skilled participants. Similar to the results found for the complexity manipulations in Experiment 1, successful shots could not be differentiated from unsuccessful shots when examining the specific phases of the billiards stroke. None of the interaction terms except the Phase by Duration Level interaction were significant. Overall, it appeared as though time spent in each phase of the billiards stroke did not differ between highly-skilled and lesser-skilled players, and could not be used as a means of explaining shots that were successful as compared to shots that were unsuccessful. Visual Tracking Measures To assess whether manipulations in task duration were associated with concomitant changes in visual tracking measures, the dependent variables of the number of fixations and average fixation duration were analyzed via separate 2 (Skill Level) x 5 (Location) x 2 (Accuracy) x 5 (Trials) ANOVAÂ’s with repeated measures on the last three factors for each of the three duration levels (100% unconstrained, 25% constrained, and 50% constrained).

PAGE 113

104 Unconstrained condition. Analyses and results for the 100% unconstrained condition were identical to those conducted on the IC condition in Experiment 1. Significant main effects in the dependent variable of number of fixations were observed for Location, F(4,88) = 53.01, p < .001, and for Skill Level, F(l,22) = 22.56, p < .01. In addition to these main effects, a significant Location by Skill Level interaction was found, F(4,88) = 5.18, p < .04. For the dependent variable of average fixation duration, the only significant finding for the 100% unconstrained condition was a main effect for Location, F(4,88) = 156.77, p < .001. Mean number of fixations and average durations for each location for both skill levels are displayed in Table 4-5 of Experiment 1. Closer examination of the visual tracking dependent measures in the 100% unconstrained condition revealed that lesser-skilled participants made significantly more fixations to the cue ball than the more skilled players. However, the groups did not differ in the average amount of time that they spent fixating on each location. Again, the visual tracking measures observed in this duration level were not useful indicators of performance, as no significant main effects or interactions were found for the independent variable of accuracy. For both groups, successful and unsuccessful shots were characterized by similar visual tracking patterns. Constrained condition (25%). For the 25% constrained condition, a significant main effect for Location was found, F(4,88) = 182.01, p < .001, for the dependent variable of number of fixations. Specifically, more fixations were directed to the cue ball

PAGE 114

105 (M = 2.23) than to any other area. In addition, more fixations were made to the cushion (M = 1-50) and the object ball (M = 1-27) than to the pocket (M = 0.17), or other areas of the display (M = 0.25). In contrast to the 100% unconstrained condition, lesser-skilled players did not make more fixations (M = 1.1 1) on average to each location than did highly-skilled players (M = 1.09). A significant interaction between Location and Skill Level was observed for the number of fixations in the 25% constrained condition, F(4,88) = 15 . 17 , g < . 01 . Lesserskilled participants made significantly more fixations to the cue ball (M = 2.42) and the object ball (M = 1-54) than did the highly-skilled participants (M = 2.04 and M = 1-00 fixations, respectively). The highly-skilled players allocated more fixations to the cushion (M = 1 88 ) than did the lesser-skilled performers (M = 1.13). A similar pattern of results was observed for the dependent variable of average fixation duration in the 25% constrained condition. Significant main effects for Location, F(4,88) = 130.58, p < .001, indicated that greater average fixation durations were directed to the cue ball ( M = 292.36 ms), the object ball (M = 307.08 ms), and the cushion (M = 287.19) than to the pocket (M = 28.75 ms) or other areas (M = 38.75 ms). The main effect for Skill Level was significant, F(l,22) = 8.24, g < .01. On average, the highlyskilled performers made fixations of longer duration (M = 206.96 ms) than did less-skilled players (M = 174.69 ms).

PAGE 115

106 More importantly, a significant interaction between Skill Level and Location was found, F(4,88) = 9.58, p < .01. Post hoc tests revealed that highly-skilled performers fixated on the cushion significantly longer (M = 366.67 ms) than did the lesser-skilled players (M = 207.71 ms). The overall pattern of results for both dependent variables fixation number and average fixation location for the 25% constrained condition is summarized in Table 4-10. Table 4-10. Mean number of fixations and average fixation durations for highly-skilled and lesser-skilled participants for the 25% constrained condition. Location Hiehlv-Skilled Fix. Num. Hiehlv-Skilled Avg. Fix. Dur. Lesser-Skilled Fix. Num. Lesser-Skilled Avg. Fix. Dur. Cue Ball 2.04 315.63 ms 2.42 269.10 ms Object Ball LOO 290.83 ms 1.54 323.33 ms Cushion 1.88 366.67 ms 1.13 207.71 ms Pocket 0.21 36.25 ms 0.13 21.25 ms Other 0.17 25.42 ms 0.33 52.08 ms From these results, it can be inferred that highly-skilled players made fewer fixations of longer duration to specific areas of the display as compared to the lesserskilled participants. In particular, the more advanced group fixated longer on the cushion than did the less advanced group, in that it fixated more often on the cue ball and object ball. However, these differences did not explain the discrepancy between successful versus

PAGE 116

107 unsuccessful shots within each group, because neither a main effect nor any interaction was found for the independent variable of accuracy. So once again, no differential pattern of visual tracking was determined for successful and unsuccessful shots in this level of task duration within each group. Constrained condition (50%). An analysis of the dependent measures of number of fixations and average fixation duration for the 50% constrained condition yielded similar results to the other time constrained conditions. The threeand four-way interactions were not significant for either dependent variable. But, for number of fixations, significant main effects were demonstrated for Location, F(4,88) = 134 . 42 , p < .001, and for Skill Level, F(l,22) = 12.82, p < .01. Lesserskilled performers had a greater mean number of fixations (M = 1.12) than did more highly-skilled billiards players (M = 0.87). In addition, the greatest number of fixations were directed the cue ball (M = 2.14), then to the object ball (M = 1-33) and the cushion (M = 1.13). Very few fixations were allocated to the pocket (M = 0.15) or to other areas (M = 0.17). SchefKÂ’s post hoc analysis indicated that the number of fixations to the cue ball was significantly different from all other locations. The number of fixations to the object ball and the cushion was not different from each other, but they were significantly greater than the number of fixations to the pocket and other areas. The final two locations were not significantly different from each other in terms of the number of fixations allocated to them.

PAGE 117

108 A significant Location by Skill Level interaction, F(4,88) = 1 1.86, £ < .01, revealed that the lesser-skilled participants generated more fixations to the cue ball (M = 2.56) and to the object ball (M = 1.63) than did the more skilled players (M = 1.71 and M = 1.04 fixations, respectively). The highly-skilled players, on the other hand, made more fixations to the cushion (M = 1.33) than did the less-skilled performers (M = 0.92). These results are displayed in Table 4-11. For the average fixation duration dependent variable, significant main effects were also found for Location, F(4,88) = 1 12.48, £ < .001, and for Skill Level, F(l,22) = 31.66, P < .01. The more advanced players possessed longer fixation durations on average (M = 170.13) than did the lesser-skilled (M = 1 18.65), with the longest fixation for both groups directed at the cushion (M = 239.69 ms), the cue ball (M = 222.53 ms), and the object ball (M = 212.78 ms). Only short fixation durations were made towards the pocket (M = 18.72 ms) and other areas (M = 23.96 ms). Post hoc analyses revealed that the cushion, cue ball, and object ball means were significantly different from the pocket and other areas locations. Finally, the Location by Skill Level interaction was again observed to be significant, F(4,88) = 16.49, £ < .001. The highly-skilled players had longer mean fixations to the cue ball (M = 260.00 ms) and the cushion (M = 336.46 ms) than did the novice players (M = 186.57 ms and M = 142.92 ms). The means for each location are presented in more detail in Table 4-11.

PAGE 118

109 Table 4-11. Mean number of fixations and average fixation durations for highlyskilled and lesser-skilled participants for the 50% constrained condition. Location Hiehlv-Skilled Fix. Num. Hiehlv-Skilled Avg. Fix. Dur. LesserSkilled Fix. Num. Lesser-Skilled Avg. Fix. Dur. Cue Ball 1.71 260.00 ms 2.56 186.57 ms Object Ball 1.04 222.08 ms 1.63 203.47 ms Cushion 1.33 336.46 ms 0.92 142.92 ms Pocket 0.08 9.17 ms 0.22 28.70 ms Other 0.17 22.92 ms 0.17 25.00 ms A review of the visual tracking measures in the 50% constrained condition revealed that the highly-skilled billiards players made fewer fixations of longer duration to the cushion and the cue ball locations than less-skilled players. In addition, total fixation duration was much greater for the expert players, as the novices tended to be more variable in their scan paths. Experts made few fixations during the limited amount of time they had to execute the stroke in this condition, and they held these fixations on the particular locations for longer periods of time. Novice players, on the other hand, possessed a greater number of fixations, but they did not hold them on any particular location for periods of time equal to that of the more advanced players. In addition, as was noted in the other two conditions tested in this experiment, successful and unsuccessful shots were not discriminated by the variables of fixation number and average fixation duration. Again, no significant main effects or interactions

PAGE 119

110 were observed for the accuracy independent variable. Thus, while significant differences existed in several visual tracking measures, they did not appear to be related to the changes in performance that were observed within each group. Quiet Eve Duration and Number of Eve Blinks The dependent variables of quiet eye duration (in ms) and suppression of gaze (as defined as the number of blinks) were assessed using a 2 (Skill Level) x 3 (Duration Level) x 2 (Accuracy) x 5 (Trials) ANOVA with repeated measures on the last three factors. It was expected that main effects would be observed for Skill Level and Duration Level. More importantly, it was hypothesized that the interaction between these factors would be significantly different, such that quiet eye durations will be longer for the higher-skilled players at each level of task duration. For the quiet eye dependent variable, significant main effects were found for Skill Level, F(l,22) = 72.21, p < .001, Duration Level, F(2,44) = 61.50, p < .001, and Accuracy V{\,22) = 208.55, p < .001. Successful shots were characterized by longer quiet eye durations (M = 310.42 ms) than unsuccessful shots (M = 118.06 ms), highly-skilled players had significantly longer quiet eye measures (M = 270.83 ms) than lesser-skilled players (M = 157.64 ms), and quiet eye duration decreased linearly with increases in time constraints (M = 314.17 ms for 100% unconstrained, M = 190.63 ms for 25% constrained, and M = 137.92 ms for 50% constrained conditions).

PAGE 120

Ill In addition, a significant three-way interaction of Accuracy by Duration Level by Skill Level was observed, F(2,44) = 9.91, p < .01. The mean values of this three-way interaction are displayed in Table 4-12. Table 4-12. Mean quiet eye durations for successful and unsuccessful shots for both highly-skilled and lesser-skilled participants for each duration level. Duration Hiehlv-Skilled Quiet Eye Hits Hiehlv-Skilled Quiet Eye Misses Lesser-Skilled Quiet Eye Hits LesserSkilled Quiet Eye Misses Unconstrained 585.00 ms 236.67 ms 286.67 ms 148.33 ms 25% Const. 373.33 ms 102.50 ms 214.17 ms 72.50 ms 50% Const. 249.17 ms 78.33 ms 154.17 ms 72.00 ms For the both the highlyskilled and lesser-skilled groups, quiet eye durations were significantly longer for accurate shots when compared to inaccurate shots for each level of complexity. The largest quiet eye durations were observed for successful shots in the 100% unconstrained condition for both groups. Also, for accurate shots in the 25% and 50% constrained conditions, highly-skilled players possessed significantly greater quiet eye durations that were nearly double those of their lesser-skilled counterparts. For both skill levels, missed shots in the 25% and 50% constrained conditions were characterized by similarly low quiet eye durations. Finally, in the highly-skilled group, missed shots had

PAGE 121

112 mean quiet eye durations that were less than half the quiet eye duration associated with successful shots. From these results, it was clear that time constraints on the execution of the billiards stroke had a direct influence on the amount of time spent in quiet eye for both skill levels. In addition, quiet eye durations were observed to be greater for the highlyskilled players over each duration level, particularly in the successful shots. Finally, and perhaps most important, quiet eye duration (unlike the other visual tracking dependent variables) was significantly related to shot performance. Best performances for both groups were characterized by longer quiet eye durations; in some instances quiet eye for successful shots was three times longer than unsuccessful shots. These results are consistent to those found in Vickers (1996) work, and also in Experiment 1 of the present study. To determine whether the suppression phase was present during the execution of the stroke in Experiment 2, the number of eye blinks during each shot was measured and statistically analyzed. As in Experiment 1, there were no significant main effects or interactions for the eye blink variable. Clearly, the suppression aspect of Vickers (1996) theory was not shown to be robust to the sport of billiards. The highlyskilled group averaged only 0.18 blinks per shot, while the lesserskilled group averaged 0.15 blinks per shot. Again, blink rate was not related to expertise level, performance accuracy, or level of complexity.

PAGE 122

CHAPTER 5 DISCUSSION, SUMMARY, CONCLUSIONS, AND SUGGESTIONS FOR FUTURE RESEARCH The two experiments conducted in the present study represented an attempt to further advance an understanding of the role of visual attention in the far aiming task of billiards shooting. Several questions related to mechanisms of visual motor control were investigated, each stemming from a series of theoretically driven hypotheses. Specifically, the first purpose of this study was to determine whether highly-skilled and lesser-skilled billiards players possessed different visual search strategies while they performed a series of shots in an ecologically valid environment. Second, an attempt was made to determine whether the two critical elements of Vickers (1996) location-suppression hypothesis, namely quiet eye duration and visual suppression, were robust to the manual aiming task found in the sport of billiards. Third, manipulations in the complexity and the duration of the task to be performed were imposed on highly-skilled and lesser-skilled players to determine whether these manipulations affected performance, visual search strategy, quiet eye duration, and movement phase duration. The final purpose of this study was to determine whether quiet eye duration was in fact an index of cognitive programming, as suggested by Vickers (1996). 113

PAGE 123

114 Several hypotheses were generated to test the purposes of this study. Each of the hypotheses pertaining to the two experiments are discussed in more detail in the following sections. Experiment 1 The purpose of Experiment 1 was to investigate the effect of manipulations in task complexity on the dependent variables of performance, temporal components of the billiards stroke, the visual search patterns of the performer, and on quiet eye duration and blink rate. A comprehensive review of the data generated in Experiment 1 indicated that several of the expected hypotheses were in fact supported. However, evidence for some of the hypotheses specifically related to measures of visual tracking was not obtained. Each of the six general hypotheses presented in the introduction are discussed in more detail in the subsequent sections, as are the two hypotheses specifically related to Experiment 1. Hypotheses Related to Performance Two specific hypotheses were generated relative to performance outcomes for both skill level groups. First, it was expected that the highly-skilled performers would demonstrate a greater percentage of successful shots than the lesser-skilled shooters, regardless of the level of task complexity. This hypothesis was clearly supported by the data, as the more advanced players sank an average of 67.53% of their shots, while the less advanced players only sank 47.88% of their attempts. Given the inherent nature of their status as highly-skilled players, this result was not surprising. Indeed, virtually every

PAGE 124

115 researcher investigating the relationship between athletic expertise and performance outcome has demonstrated the superiority of experts in measures such as reaction time (Singer et aL, 1996), response time (Abemethy, 1991), response accuracy (Shank & Haywood; Vickers, 1996), and pattern recognition and recall (Abemethy, Neal, & Koning, 1994). The second hypothesis stated that greater decrements in performance would occur as the complexity level of the shot to be performed increased. For both skill level groups, the best performances were expected in the EC condition, while the lowest percentage of successful shots was expected in the HC condition. This performance hypothesis was supported by the data collected in Experiment 1. The success rate was highest in the EC condition, where both groups collectively sank an average of 74.13% of their shot attempts. The percentage of successful shots significantly dropped to 65% in the IC condition, and to 47.88% in the HC condition. The data gathered in the present study replicated results found in the majority of experiments investigating the relationship between task complexity and motor performance. Since Henry and Rogers (1960) seminal work on the proposed “memory drum” theory of motor control, many researchers have made the claim that task / complexity plays an influential role in the response programming stage of the human information-processing system (Kerr, 1978; Klapp, 1980; Schmidt, 1988). Indeed, it has

PAGE 125

116 been repeatedly demonstrated that performance decrements in measures such as reaction time, movement time, and accuracy accompany increases in the complexity of the motor task to be performed (Anson, 1982; Christina, Fischman, Lambert, & Moore, 1985; Fischman, 1984; Magill, 1998). The present study, therefore, supported the robustness of this relationship in a billiards shooting task. While the performance outcomes observed in this study corresponded quite strongly to those observed in previous research efforts, the main impetus of Experiment 1 was on the examination of the visual tracking strategies employed by highly-skilled and lesser-skilled players. As will be discussed in the next section, the relationship between expertise, task complexity, and visual tracking measures was not as explicit as originally believed. Hypotheses Related to Visual Tracking Measures Fixation location and duration . The first hypothesis was concerned with determining whether the visual tracking patterns of the performers were associated with their level of expertise. Previous research involving more ecologically valid testing conditions has demonstrated consistent expert-novice differences in the number of fixations and duration of fixations directed towards particular elements in the sport environment (e.g., Ripoll et al., 1986; Vickers, 1992). The same pattern was observed to some extent in the present study. For each level of complexity, a Skill by Location interaction was found for the dependent variables of fixation location and average fixation

PAGE 126

117 duration, indicating that the highly-skilled and lesser-skilled performers differed in how they visually searched the pattern of balls placed before them on the billiards table. Throughout each complexity level, the more advanced players generated fewer fixations of longer average duration than did their less skilled counterparts. This pattern has been documented quite frequently in the visual search sport literature (Abemethy, 1988, 1993; Vickers, 1992), and appears to indicate that the expert performer is characterized by a more “efficient” search strategy (Vickers, 1992). In the present study, both the highlyskilled and lesser-skilled participants seemed to fixate on only a few important locations on the billiards table. Typically, both groups followed the same pattern: Fixation was initiated on the cue ball, was then directed to the object ball (or target cushion in the IC condition), then shifted back to the cue ball This sequence of shifts in gaze was repeated more often for the novice group, which made more fixations of shorter duration to these areas. Often times the fixations in the lesser-skilled group were rapidly alternated between the cue and object balls. The more advanced players, on the other hand, did not repeat this sequence as often, and held their gaze on particular locations for a significantly longer period of time. Thus, although the highly-skilled and lesser-skilled players fixated on the same cues in the environment, the more experienced group was characterized by less rapid and more efficient scanpaths. The economy of gaze behavior in the highly-skilled players will be discussed in more detail later in the paper.

PAGE 127

118 The results also indicated that relatively few fixations were made to areas other than the cue ball, the object ball, and the target cushion. That is, both groups focused on the most important elements in the arrangement of balls on the table, and virtually ignored the other aspects of the situation (Le„ pocket, cue tip, other balls on the table). This strategy of visually searching the environment for only the most pertinent or predictive cues has commonly been associated only with expert performers (e.g., Helsen & Pauwels, 1990; Shank & Haywood, 1987; Vickers, 1992; Williams et al., 1994). Previous studies supported the notion that search strategies of novice performers were random in nature, and visual attention was often allocated to areas that did not provide meaningful information related to the outcome of the performance (Abemethy, 1988, 1993). However, this was not the case in the present study, as both groups allocated their visual attention to only a few specific elements of the display. Therefore, expertise differences were not manifested in where the performer focused his gaze, but rather in the number and duration of fixations directed to these locations. In the EC condition, the highly-skilled performers made fewer fixations on average than did the novice players, and maintained fixations on the object ball for a significant longer period of time (over 150 ms longer on average). There were no significant differences in fixation duration to any of the other fixation locations. Given that the more advanced group sank 18% more shots than did the lesserskilled players, the case can be made that perhaps longer fixations oriented to the object ball led, in part, to the

PAGE 128

119 superiority in performance. A similar pattern was observed in the HC condition, where the higher-skilled players held their fixations on the object and cue balls for significantly longer periods of time. The lesser-skilled players, on the other hand, seemed to allocate their visual attention to the cue ball, the object ball, and the second target ball in an equal manner. Again, the more advanced group made a significantly greater number of shots in this condition, successfully pocketing 20% more of their shot attempts. It may be argued that an economical and efficient search strategy, coupled with longer fixation durations to only two key areas of the display, may have led to increased performance. However, this pattern of results was not observed in the IC condition. Although the highly-skilled players had a success rate of 74.84%, as compared to the lesser-skilled rate of 55.19%, the two groups did not differ in the average duration of fixations allocated to the five location areas. The less skilled individuals did make significantly more fixations to the cue ball than did the more advanced players, but total time on this location was tempered by the highly-skilled participants having an average fixation duration that was 44 ms longer on average to the cue ball. Thus, it did not appear that differences in search patterns were responsible for the expertise differences observed in performance. Finally, the visual tracking dependent measures of number of fixations and average fixation duration were not related to changes in performance within each group. Across all three complexity conditions, the visual search patterns of highly-skilled and lesserskilled

PAGE 129

120 individuals were similar in shots that were made and shots that were missed. There were no significant main effects or interactions related to the accuracy variable. Thus, it may be concluded that alterations in performance within each group was dependent on factors other than the manner in which they visually scanned the arrangement of balls on the billiards table. Since changes in performance were not associated with differential visual search patterns within each group, the question was raised as to what other variables may have been influential in discriminating performance between and within the two skill level groups. According to VickersÂ’ (1996) location-suppression hypothesis, differences in performance may have been due to the period of time each group spent in their final fixation during the preparation phase of the movement (quiet eye duration). The importance of this dependent variable is discussed in the next section. Quiet eye duration. The second general hypothesis in this study was concerned with the duration of quiet eye that was recorded for both skill level groups. Vickers and her colleagues (Vickers, 1996, 1997; Vickers & Adolphe, 1997) have reported that experts, performing tasks such as basketball ffee-throw shooting and volleyball service reception, had distinctly longer periods of quiet eye than their less skilled counterparts. For basketball shooting, elite players had quiet eye durations of almost 900 ms, while lessskilled shooters demonstrated quiet eye durations of less than 400 ms. Similarly, expert passers recorded a quiet eye of 472 ms when receiving a volleyball serve, while lesser-

PAGE 130

121 skilled passers had quiet eye durations near zero. In both instances, elite players held their final fixation on the target for roughly 500 ms longer than the less advanced players. In the present study, it was hypothesized that a similar pattern of quiet eye durations between the highlyskilled and lesser-skilled participants would be observed, regardless of the level of task complexity. When the data were collapsed across task complexity, this hypothesis was supported, as the highly-skilled players did indeed demonstrate significantly longer durations of quiet eye. That is, they held their final fixation on the target for an average of roughly 500 ms, while the lesserskilled players had quiet eye durations in the order of 275 ms. It should be noted that the actual final fixation location did not differ between the two groups. For the EC condition, five members of the highly-skilled group fixated on the cue ball immediately prior to initiation of the backswing, while seven members of this group fixated on the object ball. For the lesser-skilled group, six players focused on the cue ball, and sue fixated on the object ball. For the IC condition, seven highly-skilled players had final fixations on the cue ball, and five had final fixations on the target cushion. The novice players reported similar numbers, with seven players fixating on the cue ball, four on the target cushion, and one on the object ball. In the HC condition, nine members of the highly-skilled group and eight members of the lesser-skilled group fixated the cue ball immediately prior to backswing initiation, while the rest directed their gaze to the

PAGE 131

122 target ball Therefore, the actual location of the performerÂ’s final fixation was not different between the groups. What was different, however, was the duration of this final fixation (quiet eye duration) before backswing initiation. Vickers (1996) contended that the quiet eye period is a reflection of a cognitive programming period, in which the parameters of the movement to be executed are finetuned and set. Importantly, Abemethy et al. (1994) noted that expert snooker players possessed greater cognitive knowledge of the mechanical and geometrical factors associated with successful production of the billiards stroke. When one takes these findings into account, the longer quiet eye durations in the highly-skilled group observed in the present study can be more readily explained. Given the greater cognitive knowledge and experience that the highlyskilled players were likely to possess in regards to how to most successfully execute the stroke, such as the proper cue grip and swing, and the specific type of speed, english, and force applied to the cue ball, longer periods of quiet eye may have been necessary to parameterize and set each of these factors. Perhaps the lesser-skilled players did not demonstrate quiet eye durations as lengthy as their more advanced counterparts because, in part, they did not possess the cognitive knowledge related to the importance of factor parameterization to achieve successful stroke outcomes. Another hypothesis related to the quiet eye dependent variable focused on the relationship between quiet eye duration and shot accuracy. Based on results obtained in

PAGE 132

123 the initial work of Vickers (1996), it was expected that, for both skill levels, more successful shots would be characterized by longer quiet eye durations than unsuccessful shots. The rationale for this hypothesis stemmed from VickersÂ’ (1996) supposition that quiet eye duration reflected a critical period of parameterization in manual far aiming tasks (Vickers, 1996, 1997). Thus, regardless of skill level or experience with the task, successful aiming movements required at least some minimum amount of time spent in quiet eye on behalf of the performer. This hypothesis was supported by the data obtained in Experiment 1. For both groups, quiet eye durations averaged 561.94 ms for successful shots and only 213.61 ms for unsuccessful shots. In other words, quiet eye duration was more than doubled for shots that were pocketed as opposed to shots that were missed. Perhaps the key point to be taken from this result is the notion that quiet eye duration played an important role in the accuracy of performance for both lesser-skilled and highly-skilled individuals. This result must be acknowledged, and indicates that the duration of quiet eye may have important implications for the quality of a response in the production of a far aiming movement. Moreover, these implications may be directed not only to expert performers, but to novice performers as well. One final hypothesis for the quiet eye variable examined the relationship between task complexity and quiet eye duration. Although the relationship between these two variables has never been tested before in the motor control literature, it was hypothesized

PAGE 133

124 that increases in task complexity would lead to concomitant increases in quiet eye duration. The rationale from this assumption came from several lines of motor control research which demonstrated increases in reaction time and movement time for more complex tasks (e.g., Klapp, 1980; Sternberg et al, 1978). Since quiet eye duration was presumed to reflect a period of cognitive response programming (Vickers, 1996), it follows that more complex tasks requiring more preparation time should also yield longer quiet eye times. This point was strongly supported in the present study. For both groups, time spent in quiet eye increased linearly with task complexity. Quiet eye duration in the EC condition averaged 230 ms, increased to 314 ms in the IC condition, and peaked at an average of almost 620 ms in the HC condition. In addition, the highly-skilled players possessed quiet eye durations that were significantly longer as compared to the less skilled performers at each level of task complexity. Thus, quiet eye duration appeared to play a significant role in the expertise differences that were evident in performance. Perhaps the most important results in Experiment 1 were that quiet eye duration was significantly related to successful performance in two key ways. First, the highlyskilled players possessed greater quiet eye durations for all task complexity levels. These differences may help to explain discrepancies in performance observed between the two skill levels groups. Second, performance within each group was significantly impacted by quiet eye duration. Within each group, successful shots were characterized by longer quiet

PAGE 134

125 eye durations, while unsuccessful shots were associated with significantly shorter quiet eye durations. This result was observed within each group across each level of task complexity. This interaction is illustrated in Figure 5-1. Therefore, unlike the other visual tracking dependent measures of number of fixations or average fixation duration, quiet eye duration appeared to be a key factor in explaining differences in successful performances between and within each of the skill level groups. For both the highlyskilled and lesser-skilled players, successful and unsuccessful shots did not seem to be due to differences in the visual search strategy employed by the performer, but rather in the amount of time spent in quiet eye in the preparation phase of the movement. This pattern of results replicated Vickers (1996) observations of expert and near-expert basketball free-throw shooters. Search patterns and fixation locations did not greatly differ between the groups in her study, but rather the amount of time each group spent in quiet eye appeared to better explain differences in performance between and within the two groups of shooters. In addition, as the difficulty of the shots increased in this study, so too did the quiet eye duration associated with successful shot outcomes. The most distinct example of this relationship was observed for the highly-skilled participants in the HC condition. Successful shots were characterized by quiet eye durations of 1233 ms, while misses were associated with quiet eye durations of only 395 ms. Simply stated, when they successfully

PAGE 135

126 1400 1200 -1000 -e o 'S3 U a Q V W 800 -200 -El Highly-Skilled Hits Highly-Skilled Misses Lesser-Skilled Hits @ Lesser-Skilled Misses Easy Intermediate Complexity Level Hard Figure 5-1. Quiet eye durations for each group and level of task complexity.

PAGE 136

127 pocketed the target ball in the HC condition, highly-skilled performers maintained their final fixation on a particular location for a time period that was more than three times greater than when they missed the shot. A similar pattern of results, only on a smaller scale, was also demonstrated by the lesser-skilled players, who had quiet eye durations of 616 ms for successful shots in the HC condition as opposed to quiet eye durations of 233 ms for missed shots. Thus, it appeared as though quiet eye duration played a key role in performance throughout every aspect of Experiment 1. Number of eve blinks. According to Vickers (1996), the second critical component to successful performance in far aiming tasks is the suppression of vision during the execution phase of the movement. She noted that expert free-throw shooters suppressed their gaze immediately following the initiation of the movement towards the target, either by blinking or shifting their visual attention to areas away from the target. Less skilled shooters, she observed, maintained continuous fixation on the target until the movement was completed (through the flight phase). These results were in stark contrast to the preponderance of evidence from several studies suggesting that optimal performance in manual aiming tasks was contingent on the participant fixating on the target throughout all phases of the aiming movement (Abrams et al„ 1990; Guitton & Voile, 1987; Roy & Elliott, 1986). However, some degree of support for Vickers (1996) suppression hypothesis can be inferred from experiments demonstrating that intermittent occlusion of

PAGE 137

128 vision did not disrupt performance to a large extent in walking (Assaiante, Marchand, & Amblard, 1989) and ball catching tasks (Elliott, Zuberec, & Milgram, 1994). In Experiment 1, it was hypothesized that the suppression aspect of Vickers (1996) location-suppression hypothesis would not be robust to the sport of billiards. This hypothesis was supported by the data presented in the Results section. The number of eyeblinks generated after the initiation of the stroke was minimal for both groups, across all three levels of task complexity. In fact, only very few eye blinks or shifts of gaze to locations other than the cue ball, object ball, or target cushion were observed in this study. Regardless of skill level, participants appeared to fixate only on the pertinent aspects of the display, and did not make any eye blinks throughout each phase of the billiards stroke. Once the backswing phase was initiated, both groups appeared to focus on either the cue or object ball, and then tracked the cue ball in the flight phase until it came in contact with the object ball. Rather than suppressing gaze to these pertinent areas, both groups followed the flight of the cue ball and object ball until the object ball went into the pocket (or missed the pocket). The results in Experiment 1 supported the conclusion that vision was not suppressed, but rather was maintained on the target locations throughout each phase of the billiards stroke. This view is consistent with the majority of results observed in the literature examining visual attention in aiming tasks (e.g., Abrams et al., 1990). The question arises, however, as to why Vickers (1996) detected a suppression of gaze in her

PAGE 138

129 particular sample of expert basketball free-throw shooters. Rather than being a robust phenomenon, it is likely that two possible factors may have contributed to her observations. First, Vickers (1996) reported that several of the participants used a jump shot technique in order to propel the basketball 15 ft to the target. Given the sensitivity of the ASL 4000SU eye registration system, sudden head movements or shifts in the position of the reflective visor may have led to brief interruptions in the corneal reflection signal, thus giving the impression that the performer was blinking when in fact they were not. Thus, the suppression response reported by Vickers (1996) may have been due merely to an equipment bias. The second, and more likely, explanation for the suppression of gaze in Vickers (1996) sample lies in the nature of the free-throw shooting task. For the vast majority of players performing this task, the hands, ball, and arms all come into view of the eyes and occlude the target (the hoop) during the execution phase of the shot. Perhaps these features created potential noise in the neural system which would divert valuable attentional resources away from the target. In this case, the expert players may have blinked or shifted their gaze from the target in order to avoid this noise, thus allowing their effectors to perform the action without the distracting visual afferent information. If this explanation is indeed plausible, one would expect that visual suppression would be robust to other sports in which a body part or piece of equipment would interfere with the

PAGE 139

130 participantÂ’s view of the target. However, the hands, arms, and cue stick do not occlude the line of view to the target in the sport of billiards. As was determined in the present study, VickersÂ’ (1996) suppression aspect of the locationsuppression hypothesis did not pertain to the sport of billiards. Hypotheses Related to the Temporal Components of the Billiards Stroke In her study of basketball free-throw shooters, Vickers (1996) noted that expert players did not differ in the overall amount of time they took to complete the shot when compared to near-expert players. However, the experts spent significantly more time in the preparation phase of the movement than did the near-experts. Vickers (1996) contended that this extra time spent in the preparation phase was related to successful shooting performance, presumably because this was the phase in which longer quiet eye duration was recorded for the expert participants. The present experiment investigated whether highly-skilled billiards players took more time to execute their shots as compared to the less skilled players (overall duration), and also examined the relative durations for each of the four phases of the billiards stroke. For both groups, more accurate shots were characterized by longer overall movement durations, with successful shots taking approximately 210 ms longer to complete. In addition, the level of complexity influenced the overall duration of the task, with shots in the HC condition possessing the highest overall duration. The pattern of time spent in each complexity level differed by skill level however, as less skilled players took more time

PAGE 140

131 to execute the shot in the EC condition, while the highly-skilled players took 500 ms more to complete shots in the HC condition. Given the increased time taken to execute strokes of greater complexity, an analysis of each phase of the movement was conducted to determine whether these increases were uniform for each phase. For both groups, the durations of the backswing, foreswing, and flight phases of the stroke were highly similar across all three levels of complexity. The increases in the total duration of the movement, then, occurred solely in the preparation phase. This result supports the literature relating reaction time measures to task complexity (e.g., Klapp, 1980), and suggests that the changes in task complexity presented in this study influenced the response programming stage of the human information-processing system. Thus, it was clear that the increases in total duration of the shot observed in conditions of higher task complexity were due solely to the performer spending more time preparing the movement to be performed. Interestingly, although not statistically significant, the highly-skilled players spent more time in the preparation phase of the stroke for each level of complexity than did the lesser-skilled players. This may indicate that they understood the increased complexity of the situation (particularly in the HC condition), and took more time to prepare an appropriate response. In many respects, the phase duration dependent variable did not reveal significant differences between the highly-skilled and lesser-skilled performers for each level of

PAGE 141

132 complexity. In addition, differences in performance within each group did not appear to be related to differential amounts of time spent in each of the phases. Thus, this dependent variable was not as highly related to performance as was initially expected. Several of the hypotheses generated and tested in Experiment 1 were included in the second experiment conducted in the present study. In addition, two hypotheses were specific to the manipulations presented in Experiment 2. Each of the hypotheses will be discussed in more detail in the next section. Experiment 2 The purposes of Experiment 2 were similar in principle to those of the first experiment. While Experiment 1 was designed to investigate the effects of manipulations of task complexity on a variety of dependent measures related to visual attentional control in billiards performance, the purpose of Experiment 2 was to examine the effect of manipulations in time constraints imposed on the performer on the same dependent measures. Specifically, both skill level groups were faced with conditions of increased time constraints, in which they were limited in the amount of time they were allowed to execute the billiards stroke. Again, hypotheses specific to each dependent variable assessed in this study are reviewed in more detail in the following section. Hypotheses Related to Performance As in the first experiment, two specific hypotheses were generated relative to performance outcomes for both skill level groups. Regardless of task duration level, it was

PAGE 142

133 hypothesized that the highly-skilled performers would demonstrate a greater percentage of successful shots than lesserskilled performers. This contention was supported, as the more experienced shooters sank an average of 67.4% of their shots, while the less experienced players were successful on 52.1% of their attempts. It is interesting to note that the expert players pocketed two-thirds of their shots, on average, in both Experiment 1 and Experiment 2. The novices in this experiment were slightly more successful in the overall percentage of shots made as compared to Experiment 1. The second hypothesis related to performance was also supported in the second experiment. Specifically, it was expected that decrements in performance would be observed with reductions in the time apportioned to execute the billiards stroke. For both groups, this was indeed the case, as the percentage of successful shots declined from 65% in the unconstrained condition to 53% in the 50% time constrained condition. Although the highly-skilled group achieved a significantly higher percentage of shots made in all duration conditions, the performance of both groups was clearly influenced by the amount of time they were allotted to execute the shot. This pattern of results indicated that the experimental manipulation was effective, and led to decrements in performance accuracy that has been demonstrated in other similar studies (e.g., Shapiro, 1977, 1978; Summers, 1977).

PAGE 143

134 Hypotheses Related to Visual Tracking Measures The main premise of Experiment 2 was to investigate the influence of task duration on the visual search strategy of highly-skilled and lesser-skilled players. Specifically, of interest was whether the ocular variables of number of fixations, average fixation duration, quiet eye duration, and eye blink rate were dependent on time constraints imposed on the performer as opposed to an unconstrained situation. Fixation location and duration. A similar pattern of results to those found in the first experiment were repeated in Experiment 2. Specifically, throughout each of the three duration conditions, the lesser-skilled players made significantly more fixations to the identified locations than did the highly-skilled players. In addition, the average duration of these fixations was significantly less than those observed in the more advanced group. Even under conditions of increasing time constraints, the highly-skilled participants possessed a visual search strategy that was more economical and efficient than the less advanced group. Again, the pattern of fixations was similar between the two groups, but the expert players did not shift their gaze between the cue ball, object ball, and target cushion nearly as often. Instead, they made fewer fixations, and held them on their targets for longer periods of time. This efficient search strategy appeared to be si milar to that observed in experts engaged in such motor tasks as golf putting (Vickers, 1992), baseball batting (Bahill & LaRitz, 1984), and volleyball service reception (Vickers & Adolphe, 1997).

PAGE 144

135 In addition to the more efficient manner in which the highly-skilled players scanned the billiards table, they also seemed to fixate longer on a different cue location than did the lesser-skilled players. In the unconstrained condition, the lesser-skilled players oriented their gaze to the cue ball significantly more often than did the better players, who fixated equally on the cue ball and the target cushion. In the 25% constrained condition, the highly-skilled directed significantly more fixations to the target cushion, while the novice group fixated the cue ball and object ball to a much greater extent. This pattern was repeated for both groups in the 50% condition. Thus, it appears as though differences in performance between the groups may have been explained by the differences in locations upon which the expert and novice players focused their attention. In this experiment, only the IC task was used for each of the three duration levels. The object ball was occluded by other balls on the table, and therefore could not be struck directly. The only way to successfully pocket the object ball was to first bank the cue ball off of the cushion at a precise angle. The cushion position, therefore, provided the most valuable information as to the appropriate target line to the object ball. It appeared as though the highly-skilled group possessed this cognitive knowledge, because the majority of their fixations were directed to this spot on the cushion. The lesser-skilled players, on the other hand, maintained more fixations on the cue and object balls, even though they

PAGE 145

136 were instructed during practice trials as to the most appropriate approach to successfully pocket the object ball. Therefore, the differential search strategy employed by the highly-skilled group may have explained their superiority in performance over the novice group. However, the dependent variables of the number of fixations and average fixation durations to particular cue locations were not significantly related to differences in performance within each group. In other words, the visual search strategies for successful shots were identical to those for unsuccessful shots within each skill level. Obviously, something other than these variables was related to within group differences. Quiet eye duration. In the first experiment, the dependent variable of quiet eye duration was significantly different between skill level groups, and also appeared to play a critical role in performance differences observed within each group. All of the hypotheses specific to Experiment 2 were related to those outlined in the first experiment. Specifically, it was hypothesized that quiet eye duration would be greater in the highlyskilled participants as compared to the lesser-skilled players, regardless of task duration level Also, across both groups, it was expected that successful shots would have longer mean quiet eye durations than unsuccessful shots. Finally, since constraints in task duration have been shown to influence the response programming stage of the human information-processing model (e.g., Schmidt, 1988; Terzuolo & Viviani, 1979), it was

PAGE 146

137 hypothesized that quiet eye durations would decrease in both groups as greater time constraints were placed on the performers. As in the first experiment, all three hypotheses were supported. The expert billiards shooters possessed significantly greater quiet eye durations in each level of task duration. This difference was manifested in an almost 200 ms increase in time spent in quiet eye for the highly-skilled group in the unconstrained condition. Even under the most severe time constraints (50% constrained condition), the more advanced group possessed mean quiet eye durations that were over 50 ms greater on average as compared to the lesserskilled participants. In addition, task duration had a significant effect on the time spent in quiet eye for both groups, as linear decreases were observed with increasing time constraints. Mean quiet eye values for both groups dropped from 314 ms in the unconstrained condition to 138 ms in the 50% constrained condition. Again, for reasons similar to those outlined in the first experiment, quiet eye appeared to be related to an index of cognitive programming. That is, when manipulations were made in an independent variable known to influence the response programming stage (task duration), concomitant changes in the duration of quiet eye were observed. At the very least, this seems to indicate that Vickers (1996) was correct when she stated that quiet eye duration was reflective of some form of cognitive parameterization of a manual aiming movement.

PAGE 147

138 Finally, the importance of quiet eye duration on performance was delineated in the examination of differences in performance within each group. For both the highlyskilled and lesser-skilled performers, successful performance was marked by significantly longer quiet eye durations than unsuccessful performance. This pattern occurred across all three duration conditions, and in some instances quiet eye duration was two to three times longer in successful shots as compared to shots that were unsuccessful. From these results, and those obtained in Experiment 1, it is quite obvious that quiet eye duration was significantly related to performance not only between groups of different skill levels, but also within each group as well. The relationship between quiet eye duration, task duration condition, and performance is illustrated most clearly in Figure 5-2. Number of eveblinks. Again, one of the primary hypotheses was to determine whether VickersÂ’ (1996) visual suppression hypothesis was robust to the manual aiming task of billiards shooting. Although she noted that expert basketball players suppressed their vision upon initiation of the shooting movement, this particular search strategy was not supported in Experiment 2. Throughout each level of task duration, very few eye blinks were recorded in both the expert and novice billiards players. This was not surprising, given that greater time constraints often led to shots that were performed in under one second. For shots of this small duration, it was unlikely that any of the

PAGE 148

139 600 500 1 400 § 300 fi | 200 & 100 0 Unconstrained 25% 50% Constrained Constrained Duration Level Figure 5-2. Quiet eye durations for each group and level of task duration.

PAGE 149

140 participants felt a need to blink or divert their gaze away from the billiards table. Thus, while the location aspect of Vickers (1996) location-suppression hypothesis was strongly supported in this study, the suppression phase was not. Hypotheses Related to the Temporal Components of the Billiards Stroke. In the second experiment, it was hypothesized that constraints in task duration would lead to concomitant reductions in the overall duration of the billiards stroke. This hypothesis was supported, as mean durations for both groups declined significantly and linearly with increasing time constraints. Highest overall stroke durations were observed in the unconstrained condition, while the lowest average stoke duration was documented in the 50% constrained condition. Given the nature and the purpose of Experiment 2, these results indicated that the experimental manipulation did indeed have an effect of overall stroke duration. More importantly, based on previous literature (e.g., Gentner, 1987; Shapiro, 1977), it was hypothesized that the relative ratio of the time durations spent in each of the four phases of the movement would remain constant for each of the three levels of task duration. In other words, even though the performers were spending less overall time in executing the stroke in the more constrained situations, the ratios of the phases would remain the same. This hypothesis was not supported by the data observed in Experiment 2. Throughout each of the three task duration conditions, the average durations for the backswing, foreswing, and flight phases remained the same. Conversely, linear reductions

PAGE 150

141 in the amount of time spent in the preparation phase of the movement were observed with increased time constraints. Thus, it appeared as though the decline in overall stroke duration was due solely to reductions in the preparation phase, and not in any of the other phases. This result conflicts with many previous investigations examining the effects on performance of modifications in the overall duration parameter of a motor task (e.g., Schmidt, 1988; Summers, 1977; Terzuolo & Viviani, 1979). These researchers argued that changes in the overall duration parameter led to proportional changes in each phase of the movement, such that the relative ratio between these phases remained constant. This was clearly not the case in the present study. For both skill level groups, time decrements were observed only in the preparation phase of the billiards stroke. Thus it appeared as though reductions in performance over the three task duration conditions could not be explained by reductions in the biomechanical phases of the stroke (e.g., backswing and foreswing phases stayed the same). Rather, performance decrements were likely the result of the performer not having enough time to adequately prepare the most efficient stroke in the preparation phase. This observation lends itself rather strongly to the notion that quiet eye duration is critical to the performance outcome of a manual aiming task. Because the performers spent less time in the preparation phase at each level of increased time constraints, they

PAGE 151

142 therefore were limited in the amount of time that they could spend in quiet eye. Again, this time period is believed to be responsible for the programming and fine tuning of each of the parameters critical to performance in the aiming task (Vickers, 1996). It follows, then, that the reductions in performance obtained in each of the increased time constrained conditions was likely due to errors in the programming of the billiards stroke (e.g., the preparation phase), rather than in the actual movement production of the stroke (e.g., the backswing and foreswing phases of the movement). Summary In the present study, two main purposes related to visual attentional control in the far aiming task of billiards were investigated. First, an assessment was made of whether potential differences in visual search strategies between twelve highly-skilled and twelve lesser-skilled billiards players could be detected while they executed shots in an ecologically valid environment. The highly-skilled players were of similar age to the lesserskilled players, but possessed much more competitive and recreational playing experience, were significantly better on an initial skills test, and played more times per week on average. Within this first purpose, VickersÂ’ (1996) location-suppression hypothesis was tested to determine whether the critical variables of quiet eye duration and visual suppression were robust to the sport of billiards. The second purpose of the study was to determine the importance of the location aspect of VickersÂ’ (1996) theory by manipulating the complexity and duration of the task to be performed.

PAGE 152

143 In Experiment 1, the participants were required to execute shots of three different complexity levels (easy, intermediate, and hard difficulty levels) until they successfully pocketed 10 attempts and missed 10 attempts. In the second experiment, time constraints of 25% and 50% were imposed on the players while they performed shots of intermediate complexity. Results from these externally-imposed constraints were compared to data recorded in the unconstrained intermediate complexity condition experienced by the participant in Experiment 1. Dependent measures in both experiments included performance outcome percentage, stroke and phase duration, number of fixations, average fixation duration, quiet eye duration, and number of eye blinks. From these results, evidence supporting or rejecting the strong eye-mind assumption that eye movement processes are reflective of higher-order cognitive programming could be gathered. For both experiments conducted in the present study, it appeared as though the highly-skilled participants had visual search strategies that were more efficient that the lesser-skilled players. Expert players made fewer fixations of longer average duration to specific elements in the billiards environment, while the novice players engaged in a tracking strategy that included a greater number of fixations of shorter duration that were directed mostly towards the cue ball and the object ball The advanced players fixated on the target line more often, and for longer periods of time, than did the less advanced players in the IC condition. In all other conditions the two skill level groups appeared to

PAGE 153

144 focus their attention on the same cues. Although the two groups did not differ greatly in what they looked at on the billiards table, they were consistently different in the efficiency and economy of their search patterns. The location aspect of VickersÂ’ (1996) hypothesis was supported in both experiments, as expertise differences were observed for the independent variable of quiet eye duration. Specifically, it appeared as though the important parameters of the billiards stroke such as force, direction, and velocity may have been delineated and fine tuned during this quiet eye period. Evidence for this notion comes from the observation that manipulations in task duration and task complexity led to changes in quiet eye duration, and that reduced quiet eye durations in both groups were associated with significant decrements in performance. Thus it appears as though quiet eye duration may reflect the amount of time spent in the response programming stage of the information-processing model, and may serve as evidence that higher-order cognitive processes control gaze behavior. Distinct expertise differences for overall stroke duration and time spent in each phase of the billiards stroke were not discovered. The lesser-skilled players took more time to execute the shot in the EC condition, while the highly-skilled performers possessed significantly greater overall stroke durations in the HC condition. In all other conditions, the time spent in each phase of the movement was similar for members of both groups. Therefore, the temporal elements of the billiards stroke did not appear to be significantly

PAGE 154

145 associated to differences in expertise. In the second experiment, increased time constraints led to reductions only in the preparation phase of the movement for both skill level groups. This result appears to add further support for the importance of quiet eye in successful performance outcomes, as the duration of this variable is directly influenced by the length of the preparation phase. No significant time decrements were observed in the backswing, foreswing, or flight phases of the stroke, thus suggesting that biomechanical disruptions did not cause the observed performance changes across both skill level groups. Finally, VickersÂ’ (1996) visual suppression theory was not supported in the present study. Rather than suppress or shift fixations away from the important elements of the display during the execution of the billiards stroke, both skill level groups made very few eye blinks and tracked the cue ball, object ball, or target cushion during the backswing, foreswing, and flight phases of the stroke. This aspect of VickersÂ’ (1996) theory was clearly not robust to the sport of billiards, and may not be appropriate for far aiming tasks in which visual occlusion of the target does not occur. Conclusions From the results of the present study, the following conclusions appear to be warranted: 1. Expertise differences in visual search strategy are evident in the sport of billiards. Highly-skilled performers possess more economical and efficient visual search

PAGE 155

146 strategies, making fewer fixations of significantly longer duration to particular locations in the environment as compared to their novice counterparts. 2. The location aspect of VickersÂ’ (1996) theory of visual control is robust to the sport of billiards. Indeed, quiet eye duration is a key variable in determining differences in performance between and within each of the two skill level groups. Highly-skilled players exhibit longer quiet eye durations than lesser-skilled players. In addition, successful shots within both skill level groups are characterized by significantly longer quiet eye durations than unsuccessful shot attempts. 3. Increases in task complexity lead to concomitant increases in quiet eye duration for both skill levels, although increases in quiet eye are significantly greater for the highlyskilled players. In addition, increased constraints on the amount of time allotted to the performer leads to decrements in quiet eye duration for both groups of participants. Quiet eye duration, therefore, appears to be a reflection of some underlying cognitive process that is highly influential in the preparation of effective billiard stroke responses. 4. Quiet eye duration may be used as an index of cognitive response processing time. That is, quiet eye duration appears to provide evidence for the strong eye-mind assumption that eye movement patterns are reflective of underlying cognitive processes. 5. The suppression aspect of VickersÂ’ (1996) theory is not present in the sport of billiards. Relatively few blinks or shifts of gaze away from the billiards table are made

PAGE 156

147 when players of both skill levels execute their shots. Therefore, visual suppression is not a viable or robust aspect of successful billiards performance. 6. Decrements in performance for both groups occur when time constraints are imposed. This is due primarily to insufficient time spent in the preparation phase of the movement, thus reducing the potential for adequate quiet eye durations in the performer. No differences are observed in the backswing, foreswing, or flight phases of the movement when increased time constraints are imposed upon the performer. Implications For Future Research The present study represented an attempt to further understand the role of visual attentional control mechanisms in an ecologically valid sport environment using a theoretically-driven method of hypothesis testing. Based on the results obtained, several implications for future research are offered. 1. Quiet eye duration was found to be an important variable in terms of performance outcome for both highly-skilled and lesserskilled billiards players. In addition, it appears that quiet eye may be used as an index of response programming, since manipulations in task complexity and task duration led to concomitant changes in quiet eye duration. However, more investigation is needed to determine the exact nature of this quiet eye duration. Vickers (1996) contends that the parameterization of factors associated with successful far aiming movements are delineated during the quiet eye period. These factors

PAGE 157

148 may include temporal, distance, force, and velocity measures. It is still unclear, though, as to whether the quiet eye duration is indeed reflective of underlying cognitive stages of information processing. Although the present study suggests that quiet eye duration is influenced by factors known to impact upon the programming of a motor response, further research investigating manipulations in each parameter of the aiming movement will provide more detailed information as to the exact nature of the quiet eye phenomenon. 2. In addition to manipulations in those aiming movement parameters thought to impact on the response programming stage (and therefore on quiet eye duration), researchers may wish to investigate the effect of direct manipulations of quiet eye on performance. Advances in technology, including the ASL 4000SU mobile eye tracking system and the development of specially designed glasses that produce varying degrees of visual occlusion in less than 1 ms (e.g., Elliott et al„ 1994), allow for the possibility that quiet eye can be directly influenced and selectively controlled by the experimenter. Unlike the indirect manipulations of quiet eye performed in the present study, direct manipulation and the heightened degree of experimental control may provide more concrete evidence in support of the significance of quiet eye duration to manual aiming performance. 3. The visual suppression aspect of Vickers’ (1996) theory was not supported in this study. Although this phenomenon was observed in a basketball free-throw shooting task (Vickers, 1996), it was not discovered in a billiards shooting task. Future research, therefore, must be conducted to discriminate specific far aiming tasks in which suppression

PAGE 158

149 is significantly related to performance. It has been suggested that suppression may be associated only with tasks that produce potentially distracting visual afferent information during the execution of the movement. This is a logical supposition, and the performer may blink or shift gaze away from the distracting information in order to maintain the integrity of aiming commands delineated in the preparation phase of the movement (Vickers, 1996, 1997). However, when the actual production of the movement does not lead to visual distraction or occlusion of the target, valuable information concerning the relationship between the target and the aiming limb can be gathered from maintaining gaze on the target throughout the entire duration of the aiming movement (Abrams, 1990; Abrams et aL, 1990). 4. Given the importance of visual search strategy and quiet eye duration on performance outcome in the present study, future investigators may wish to develop and implement programs designed to train and enhance visual attention control. Given that expert-novice differences have been found primarily for the software aspects of performance, and not the hardware aspects (Abemethy, 1987; Singer et al„ 1994; Starkes & Deakin, 1987), researchers may attempt to determine whether these software properties can be developed and enhanced in novice, intermediate, and elite athletes. From visual search and cue occlusion research, it has been found that experts seemingly know where to look in their environment for advanced cues as to the actions of an opponent, and they

PAGE 159

150 also understand the link between these cues and the eventual outcome of the opponents action (Abemethy, 1990, 1993; Singer et al., 1996). In addition, highly-skilled players use mechanisms of visual control that are more efficient (Bahill & LaRitz, 1984; Vickers, 1992), and seem to prepare responses in the motor programming stage more effectively through the use of longer quiet eye durations (Vickers, 1996, 1997; Vickers & Adolphe, 1997). To this end, the training of selective attention and anticipation s kills would seem to be vital in improving the performance of novice and intermediate sport performers (Burroughs, 1984; Christina, Barresi, & Shaffher, 1991; Day, 1980; Haskins, 1965; Singer et al., 1994). Preliminary work in this area has already begun, as Adolphe, Vickers, and Laplante (1997) developed a six week visual attention training program designed to improve ball tracking and forearm passing skills in highly-skilled volleyball players. Using information gathered from a previous study of elite volleyball players (Vickers & Adolphe, 1997), a training method of fixating only on the most pertinent cues in the volleyball environment was given to another sample of high caliber players. Results indicated that the six week program led to improvements in faster tracking onset times, longer ball tracking durations, and longer quiet eye durations. In addition, these improvements were still observed after a threeyear follow-up period. Thus, it appears that training methods can be developed to improve the visual attentional control mechanisms of performers of all skill levels.

PAGE 160

151 Given the results of these studies, there is strong evidence that improvements can be made in the visual control and anticipatory skills of novice and more advanced athletes, such that performance may be enhanced in a quicker, more complete fashion than regular practice schedules would dictate. However, more research needs to be conducted in this area, and the results need to be demonstrated not merely at the laboratory level, but also in the actual environments in which the athlete performs. Thus, like the present study, more ecologically valid environments need to be incorporated into studies attempting to train mechanisms of anticipation and visual controL Given the important role that vision appeared to play in performance in this particular study, this avenue of exploration is critical in furthering our understanding of the mechanisms related to expertise differences in the visual attention control of far aiming movements.

PAGE 161

152 REFERENCES Abemethy, B. (1986). Enhancing sports performance through clinical and experimental optometry. Clinical and Experimental Optometry. 69 . 189-196. Abemethy, B. (1988). Visual search in sport and ergonomics: Its relationship to selective attention and performer expertise. Human Performance. 1 . 205-235. Abemethy, B. (1989). Expert-novice differences in perception: How expert does the expert have to be? Canadian Journal of Sport Sciences. 14 . 27-30. Abemethy, B. (1990a). Anticipation in squash: Differences in advance cue utilization between expert and novice players. Journal of Sport Sciences. 8 . 17-34. Abemethy, B. (1990b). Expertise, visual search, and information pick-up in squash. Perception. 19 . 63-77. Abemethy, B. (1991). Visual search strategies and decision-making in sport. International Journal of Sport Psychology. 22 . 189-210. Abemethy, B. (1993). Attention. In R.N. Singer, M. Murphey, & L.K. Tennant (Eds.), Handbook of research in snort psychology (pp. 127-169). New York: Macmillan. Abemethy, B., Neal, R.J., & Koning, P. (1994). Visual-perceptual and cognitive differences between expert, intermediate, and novice snooker players. Applied Cognitive Psychology. 8 . 185-212. Abemethy, B„ & Russell, D.G. (1984). Advance cue utilization by skilled cricket batsmen. Australian Journal of Science and Medicine in Sport. 16 . 2-10. Abemethy, B„ & Russell, D.G. (1987a). Expert-novice differences in an applied selective attention task. Journal of Sport Psychology. 9 . 326-345. Abemethy, B., & Russell, D.G. (1987b). The relationship between expertise and visual search strategy in a racquet sport. Human Movement Science. 6 . 283-319. Abemethy, B., Thomas, K.T., & Thomas, J.T. (1993). Strategies for improving understanding of motor expertise. In J.L. Starkes & F. Allard (Eds.), Cognitive issues in motor expertise (pp. 317-356). Amsterdam: Elsevier Science.

PAGE 162

153 Abrams, R.A. (1994). Eye-hand coordination: Spatial localization after saccadic and pursuit eye movements. Journal of Motor Behavior. 26 . 215-224. Abrams, R.A., Meyer, D„ & Komblum, S. (1990). Eye-hand coordination: Oculomotor control in rapid aimed limb movements. Journal of Experimental Psychology: Human Perception and Performance. 16 . 248-267. Adolphe, R.M., Vickers, J.N., & Laplante, G. (1997). The effects of training visual attention on gaze behavior and accuracy: A pilot study. International Journal of Sports Vision. 4 . 28-33. Adam, J.J. & Wilberg, R.B. (1986). Rate of visual information processing and the prediction of performance excellence in professional ice-hockey and varsity downhill skiing. Toronto, ON: Proceedings of the Canadian Society for Psycho mo tor Learning and Sport Psychology. Allard, F., & Burnett, N. (1985). Skill in sport. Canadian Journal of Psychology. 32, 294-312. Allard, F., Graham, S., & Paarsalu, M.E. (1980). Perception in sport: Basketball. Journal of Sport Psychology. 2 . 14-21. Allard, F., & Starkes, J.L. (1980). Perception in sport: Volleyball. Journal of Sport Psychology, 2 . 22-33. Anson, J.G. (1982). Memory drum theory: Alternative tests and explanations for the complexity effects on simple reaction time. Journal of Motor Behavior. 14 . 228-246. Assaiante, C., Marchand, A.R., & Amblard, B. (1989). Discrete visual samples may control locomotor equilibrium and foot positioning in man. Journal of Motor Behavior. 21 . 72-91. Bahill, L.T., & LaRitz, T. (1984). Why can't batters keep their eyes on the ball? American Scientist. 72 . 249-253. Bard, C., & Fleury, M. (1976). Analysis of visual search activity in sport problem situations. Journal of Human Movement Studies. 3 . 214-222.

PAGE 163

154 Bard, C., & Fleury, M. (1981). Considering eye movement as a predictor of attainment. In I.M. Cockerill & W.W. MacGillivary (Eds.), Vision and sport (pp. 28-41). Cheltenham: Stanley Thornes. Bard, C., Fleury, C, & Carriere, L. (1975). Perceptual strategy and motor performance. Mouvement. 10 . 163-183. Bard, C., Fleury, M., Carriere, L., & Halle, M. (1980). Analysis of gymnastics judges' visual search. Research Quarterly for Exercise and Sport. 51 . 267-273. Bard, C., Fleury, M., & Goulet, C. (1994). Relationship between perceptual strategies and response adequacy in sport situations. International Journal of Sport Psychology. 25 . 266-281. Bard, C., Guezennec, Y., & Papin, J.P. (1980). Fencing: Analysis of visual exploration. Medecine du Sport. 55 . 22-28. Billiard Congress of America (1996). The official rules and record book for all pocket and carom billiard games . Chicago: Billiard Congress. Blundell, N.L. (1985). The contribution of vision to the learning and performance of sports skills: Part 1: The role of selected visual parameters. Australian Journal of Science and Medicine in Sport. 17 . 3-11. Bootsma, R.J. (1988). The timing of rapid interceptive actions . Amsterdam: Free University Press. Bootsma, RJ. (1992). Predictive information and the control of action: What you see is what you get. International Journal of Sport Psychology. 22 . 271-278. Bourgeaud, P., & Abemethy, B. (1987). Skilled perception in volleyball defence. Journal of Sport Psychology. 9 . 400-406. Brancazio, P.J. (1985). Looking into Chapman's homer: The physics of judging a fly ball. American Journal of Physics. 53 . 849-855. Buckholz, E., Martinelli, B., & Hewey, D. (1993). The influence of practice on the reaction time of ice hockey players to peripheral vision stimuli: Stationary inputs. International Journal of Sports Vision. 1 . 15-24.

PAGE 164

155 Burke, T.R. (1972). Athletes, athletic performance, and conditions in the environment. Quest. 17 . 56-60. Burroughs, W. (1984). Visual simulation training of baseball batters. International Journal of Sport Psychology, 15 . 117-126. Carl, J., & Gellman, R. (1987). Human smooth pursuit: Stimulus-dependent responses. Journal of Neurophysiology. 57 . 1446-1463. Carlson, N.J. (1985). Sports vision guidebook, vol. II . St. Louis, MO: American Optometric Association. Carlton, L.G. (1981a). Processing visual feedback information for movement control Journal of Experimental Psychology: Human Perception and Performance. 7 . 1019-1030. Carlton, L.G. (1981b). Visual information: The control of aiming movements. Quarterly Journal of Experimental Psychology: Human Perception and Performance. 33A . 87-93. Carter, M.C., & Shapiro, D.C. (1984). Control of sequential movements: Evidence for generalized motor programs. Journal of Neurophysiology. 52 . 787-796. Chase, W.G., & Simon, H.A. (1973). The mind’s eye in chess. In W.G. Chase (Ed.), Visual information processing , (pp. 215-281). New York: Academic Press. Chi, M.T.H., & Bjork, R.A. (1991). Modeling expertise. In D. Druckman (Ed.), In the mind's eve: Enhancing human performance (pp.57-79). Hillsdale, NJ: Lawrence Erlbaum. Chi, M.T.H., Glaser, R„ & Farr, M.J. (1988). The nature of expertise . Hillsdale, NJ: Lawrence Erlbaum. Christina, R.W., Barresi, J.V., & Shaffner, P. (1990). The development of response selection accuracy in a football linebacker using video training. The Sport Psychologist. 4 . 11-17.

PAGE 165

156 Christina, R.W., Fischman, M.G., Lambert, A.L., & Moore, J.F. (1985). Simple reaction time as a function of response complexity: Christina et al. (1982) revisited. Research Quarterly for Exercise and Sport. 56 . 316-322. Clark, F.J., & von Euler, C. (1972). On the regulation of depth and rate of breathing. Journal of Physiology. 222 . 267-295. Cohen, J. (1977). Statistical power analysis for the behavioral sciences . New York: Academic Press. Davids, K. (1984). The role of peripheral vision in ball games: Some theoretical and practical notions. Physical Education Review. 7 . 26-40. Davids, K. (1987). The development of peripheral vision in ballgames: An analysis of singleand dual-task paradigms. Journal of Human Movement Studies. 13 . 275-284. Davids, K. (1988). Developmental differences in the use of peripheral vision during catching performance. Journal of Motor Behavior. 20 . 39-51. Davids, K., Handford, C., & Williams, M. (1994). The natural physical alternative to cognitive theories of motor behavior: An invitation for interdisciplinary research in sports science? Journal of Sports Sciences. 12 . 495-528. Day, L. (1980). Anticipation in junior tennis players. In J.L. Groppel & R. Sears (Eds.), Proceedings of the international symposium on the effective teachings of racquet sports (pp. 107116). Champaign, IL: University of Illinois. de Groot, A.D. (1978). Thought and choice in chess . The Hague: Mouton. Elliott, D., Chua, R., & Pollock, B.J. (1994). The influence of intermittent vision on manual aiming. Acta Psvchologica. 85 . 1-13. Elliott, D., Zuberec, S., & Milgram, P. (1994). The effects of periodic visual occlusion on ball catching. Journal of Motor Behavior. 26 . 113-122. Ericsson, K.A., & Chamess, N. (1994). Expert performance: Its structure and acquisition. American Psychologist. 49 . 725-747. Ericsson, K.A., Krampe, R.T., & Tesch-Romer, C. (1993). The role of deliberate practice in the acquisition of expert performance. Psychological Review. 100 . 363-406.

PAGE 166

157 Falkenberg, L.E., & Newell, K.M. (1980). Relative contribution of movement time, amplitude, and velocity to response initiation. Journal of Experimental Psychology: Human Perception and Performance. 6 . 760-768. Festinger, L., & Canon, L.K. (1965). Information about spatial location based on knowledge about efference. Psychological Review. 72 . 373-384. Fischman, M.G. (1984). Programming time as a function of number of movement parts and changes in movement direction. Journal of Motor Behavior. 16 . 405-423. Fisher, D.F., Karsh, R., Breitenbach, F., & Barnette, B.D. (1983). Eye movements and picture recognition: Contribution or embellishment. In R. Groner, C. Menz, D.F. Fisher & R.A. Monty (Eds.), Eve movements and psychological functions: International views (pp. 193-210). Hillsdale, NJ: Lawrence Erlbaum Associates. Fitts, P.M., & Posner, M.I. (1967). Human performance . Belmont, CA: Brooks/Cole. Fleury, M., Goulet, C., & Bard, C. (1986). Eye fixations as visual indices of programming of service return in tennis. Psychology of motor behavior and sport abstracts (p. 17). Scottsdale, AZ: NASPSPA Fowler, C.A., & Turvey, M.T. (1978). Skill acquisition: An event approach with special reference to searching for the optimum of a function of several variables. In G.E. Stelmach (Ed.), Information processing in motor control and learning (pp. 1-40). New York: Academic Press. Gardner, H. (1985). The mind's new science: A history of the cognitive revolution . New York: Basic Books. Gauthier, G.M., Semmlow, J.L., Vercher, C„ Pedrono, C., & Obrecht, G. (1991). Adaptation of eye and head movements to reduced peripheral vision. In R. Schmid and D. Zambarbieri (Eds.), Oculomotor control and cognitive processes (pp. 179-196). New York: Elsevier Science. Gentner, D.R. (1987). Timing of skilled motor performance: Tests of the proportional duration model. Psychological Review. 94 . 255-276.

PAGE 167

158 Gibson, JJ. (1979). The ecological approach to visual perception . Boston, MA: Houghton Mifflin. Goulet, C., Bard, M., & Fleury, C. (1989). Expertise differences in preparing to return a tennis serve: A visual search information processing approach. Journal of Sport and Exercise Psychology. 11 . 382-398. Gregory, R.L. (1990). Eve and brain: The psychology of seeing (4th ed.). Princeton, NJ: Princeton University Press. Guitton, D., & Voile, M. (1987). Gaze control in humans: Eye-head coordination during orienting movements to targets within and beyond the oculomotor range. Journal of Neurophysiology. 58 . 427-459. Haase, H., & Mayer, H. (1978). Strategies of visual orientation of fencers. Leistungssport. 8 . 191-200. Hansen, R.M. (1979). Spatial localization during pursuit eye movements. Vision Research. 19 . 1213-1221. Haskins, M. (1965). Development of response recognition training films in tennis. Perceptual and Motor Skills. 21 . 202-211. Helsen, W. (1994). Perception in sport: An integrated approach to s kille d performance. Journal of Sport and Exercise Psychology. 16 . 21-22. Helsen, W., & Pauwels J.M. (1990). Analysis of visual search activity in solving tactical game problems. In D. Brogan (Ed.), Visual search (pp. 177-184). London: Taylor and Francis. Helsen, W„ & Pauwels J.M. (1992). A cognitive approach to visual search in sport. In D. Brogan & K. Carr (Eds.), Visual search II (pp. 173-185). London: Taylor and Francis. Helsen, W., & Pauwels J.M. (1993). The relationship between expertise and visual information processing in sport. In J.L. Starkes & F. Allard (Eds.), Cognitive issues in motor expertise (pp. 109-134). New York: North-Holland.

PAGE 168

159 Helsen, W„ Pauwels, J.M., & d'Ydewalle, G. (1986). Analysis of visual search patterns during sportgames. In R. Andersen (Ed.), Beitrage zur sportspielforschung (pp. 69-81). Hamburg: Czwalina. Henry, F.M. (1980). Use of simple reaction time in motor programming studies: A reply to Klapp, Wyatt, and Lingo. Journal of Motor Behavior. 12 . 163-168. Henry, F.M., & Rogers, D.E. (1960). Increased response latency for complicated movements and a "memory drum" theory of neuromotor reaction. Research Quarterly. 31 . 448-458. Hick, W.E. (1952). On the rate of gain of information. Quarterly Journal of Experimental Psychology. 4 . 11-26. Honda, H. (1985). Spatial localization in saccade and pursuit-eye-movement conditions: A comparison of perceptual and motor measures. Perception & Psvchophvsics. 28, 41-46. Hyllegard, R. (1991). The role of the baseball seam pattern in pitch recognition. Journal of Sport and Exercise Psychology. 13 . 80-84. Hyman, R. (1953). Stimulus information as a determinant of reaction time. Journal of Experimental Psychology. 45 . 188-196. Issacs, L.D., & Finch, A.E. (1983). Anticipatory timing of beginning and intermediate tennis players. Perceptual and Motor Skills. 57 . 451-454. Jones, C.M., & Miles, T.R. (1978). Use of advance cues in predicting the flight of a lawn tennis balL Journal of Human Movement Studies. 4 . 231-235. Jonides, J. (1981). Voluntary versus automatic control over the mind's eyes movement. In J. Long & A. Baddeley (Eds.), Attention and performance. IX (pp. 187203). Hillsdale, NJ: Lawrence Erlbaum. Just, M.A., & Carpenter, P. (1976). Eye fixations and cognitive processes. Cognitive Psychology. 8 . 441-480.

PAGE 169

160 Just, M.A., & Carpenter, P. (1980). A theory of reading: From eye fixation to comprehension. Psychological Review. 87 . 441-480. Kerr, B. (1978). Task factors that influence selection and preparation for voluntary movements. In G.E. Stelmach (Ed.), Information processing in motor control and learning (pp. 55-69). New York: Academic Press. Klapp, S.T. (1977). Reaction time analysis of programmed control. Exercise and Sport Sciences Reviews. 5 . 231-253. Klapp, S.T. (1980). The memory drum theory after twenty years: Comments on Henry's note. Journal of Motor Behavior. 12 . 169-171. Klapp, S.T., & Erwin, C.I. (1976). Relation between programming time and duration of the response being programmed. Journal of Experimental Psychology: Human Perception and Performance. 2 . 591-598. Klapp, S.T., & Greim, D.M. (1979). Programmed control of aimed movement revisited: The role of target visibility and symmetry. Journal of Experimental Psychology: Human Perception and Performance. 5 . 509-521. Klein, R.M. (1980). Does ocular motor readiness mediate cognitive control of visual attention? In R.S. Nickerson (Ed.), Attention & performance. VIII (pp. 127-145). Hillsdale, NJ: Lawrence Erlbaum. Klein, R.M. (1994). Perceptual-motor expectancies interact with covert visual orienting under conditions of endogenous but not exogenous control. Canadian Journal of Experimental Psychology, 48 . 167-181. Kundel, H.L., & Nodine, C.F. (1983). A visual concept shapes image perception. Radiology. 146 . 363-368. Lee, D.N., Lishman, J.R., & Thomson, J.A. (1982). Regulation of gait in long jumping. Journal of Experimental Psychology: Human Perception and Performance. 8 . 448-459. Magill, R.A. (1998). Motor learning: Concepts and applications (5th ed.). Dubuque, IO: William C. Brown.

PAGE 170

161 Marr, D. (1982). Vision: A computational investigation into the human representation and processing of visual information . San Francisco, CA: Freeman. Martin, R„ & Reeves, R. (1993). The 99 critical shots in pool . New York: Times Books. McLeod, P. (1987). Visual reaction times and high speed ballgames. Perception. 16, 49-59. McLeod, P., & Dienes, Z. (1993). Running to catch the ball. Nature. 362 . 23. Michaels, C.F., & Oudejans, R.R.D. (1992). The optics and actions of catching fly balls: Zeroing out optical acceleration. Ecological Psychology. 4 . 199-222. Miller, J.M. (1980). Information used by the perceptual and oculomotor systems regarding the amplitude of saccadic and pursuit eye movements. Vision Research. 20 . 5968 . Mourant, R.R., & Rockwell, T.H. (1972). Strategies of visual search by novice and experienced drivers. Human Factors. 14 . 325-335. Neisser, U. (1967). Cognitive psychology . New York: Appleton-Century-Crofts. Neumaier, A. (1983). Observation strategies and anticipation of defensive actions in volleyball. Leistungssport. 13 . 5-10. Newell, K.M., Hoshizaki, L.E.F., Carlton, M.J., & Halbert, J.A. (1979). Movement time and velocity as determinants of movement timing accuracy. Journal of Motor Behavior. 11 . 49-58. Nideffer, R.N. (1993). Attention control training. In R.N. Singer, M. Murphey, & L.K. Tennant (Eds.), Handbook of research in sport psychology (pp. 542-556). New York: Macmillan. Nielsen, D„ & McGown, C. (1985). Information processing as a predictor of offensive ability in baseball. Perceptual and Motor Skills. 60 . 775-781.

PAGE 171

162 Nisbett, R.E., & Wilson, T.D. (1977). Telling more than we can know: Verbal reports on mental processes. Psychological Review. 84 . 231-254. Nougier, V., Stein, J.F., & Bonnel, A.M. (1991). Information processing in sport and orienting of attention. International Journal of Sport Psychology, 22 . 307-327. Optican, L.M. (1985). Adaptive properties of the saccadic system. In A. Berthoz & M. Melville-Jones (Eds.), Adaptive mechanisms in gaze control: Facts and theories (pp. 71-79). New York: Elsevier Science. Oudejans, R.R.D., Michaels, C.F., Bakker, F.C., & Davids, K. (in press). Running to catch in the dark: Perceptual mechanisms for catching fly balls. Journal of Human Movement Studies . Posner, M.I. (1980). Orienting of attention. Quarterly Journal of Experimental Psychology, 32 . 3-27. Posner, M.I., & Cohen, Y. (1980). Attention and the control of movement. In G.E. Stelmach and J. Requin (Eds.), Tutorials in motor behavior (pp. 243-258). Amsterdam: North-Holland. Prablanc, C., & Pellison, D. (1990). Gaze saccade orienting and hand pointing are locked to their goal by quick internal loops. In M. Jeannerod (Ed.), Attention and performance. XIII . (pp. 653-676). Hillsdale, NJ: Lawrence Erlbaum. Remington, R.W. (1980). Attention and saccadic eye movements. Journal of Experimental Psychology: Human Perception and Performance. 6 . 726-744. Remington, R. & Pierce, L. (1984). Moving attention: Evidence for time-invariant shifts of visual selective attention. Perception and Psychophysics. 35 . 393-399. Revien, L. (1987). Visual skill training for sports. Scholastic Coach. 57 . 74-76. Ripoll, H. (1984). The study of visuo-motor coordination in rapid fire pistol In D. Landers (Ed.), Sport and elite performers (pp. 153-162). Champaign, IL: Human Kinetics. Ripoll, H. (1988). Analysis of visual scanning patterns of volleyball players in a problem solving task. International Journal of Sport Psychology. 19 . 9-25.

PAGE 172

163 Ripoll, H. (1991). The understanding-acting process in sport: The relationship between the semantic and the sensorimotor visual fimction. International Journal of Sport Psychology. 22 . 221-243. Ripoll, H., Bard, C., & Paillard, J. (1986). Stabilization of the head and eyes on the target as a factor in successful basketball shooting. Human Movement Science. 5 . 47-58. f Ripoll, H., & Fleurance, P. (1988). What does keeping one's eye on the ball mean? Ergonomics. 31 . 1647-1654. Ripoll, H„ Papin, J., Guezennec, J„ Verdy, J., & Philip, M. (1985). Analysis of visual scanning patterns of pistol performers. Journal of Sport Sciences. 3 . 93-101. Ritzdorf, V. (1983). Anticipation in sport: Investigation of the tennis ground stroke. Leistungssport. 13 . 5-9. Rotella, R.J., & Lemer, J.D. (1993). Responding to competitive pressure. In R.N. Singer, M. Murphey, & L.K. Tennant (Eds.), Handbook of research in sport psychology (pp. 528-541). New York: Macmillan. Roy, E.A., & Elliott, D. (1986). Manual asymmetries in visually directed aiming. Canadian Journal of Psychology, 40 . 109-121. Rumelhart, D.E., & McClelland, J.L. (1986). Parallel distributed processing: Explorations in the microstructure of cognition . Cambridge, MA: MIT Press. Sanderson, F.G., & Whiting, H.T.A. (1975). Dynamic visual acuity and performance in a catching task. Journal of Motor Behavior. 6 . 87-94. Sandu, G.S. (1982). Selective attention in competitive volleyball. In J.H. Salmela & T. Orlick (Eds.), New paths of sport learning and excellence (pp. 30-32). Ottawa, ON: Sport in Perspective, Inc. Savelsbergh, G.J.P., & Whiting, H.T.A. (1988). The effect of skill level, external frame of reference and environmental changes on one handed catching. Ergonomics. 31 . 1655-1663.

PAGE 173

164 Savelsbergh, G.J.P., & Whiting, H.T.A. (1992). The acquisition of catching under monocular and binocular conditions. Journal of Motor Behavior. 24 . 320-328. Savelsbergh, G.J.P., Whiting, H.T.A., Burden, A.M., & Bartlett, R.M. (1992). The role of predictive visual temporal information in the coordination of muscle activity in catching. Experimental Brain Research. 89 . 223-228. Schmidt, R.A. (1988). Motor control and learning (2nd ed.). Champaign, IL: Human Kinetics. Shaffer, L.H. (1980). Analyzing piano performance: A study of concert pianists. In G.E. Stelmach & J. Requin (Eds.), Tutorials in motor behavior (pp. 443-455). Amsterdam: North-Holland. Shaffer, L.H. (1984). Timing in solo and duet piano performances. Quarterly Journal of Experimental Psychology. 16A . 577-595. Shank, M.D., & Haywood, K.M. (1987). Eye movements while viewing a baseball pitch. Perceptual and Motor Skills. 64 . 1191-1197. Shapiro, D.C. (1977). A preliminary attempt to determine the duration of a motor program. In D.M. Landers & R.W. Christina (Eds.), Psychology of motor behavior and sport (pp. 17-24). Champaign, IL: Human Kinetics. Shapiro, D.C. (1978). The learning of generalized motor programs . Unpublished doctoral dissertation. University of Southern California, Los Angeles. Shapiro, D.C., Zemicke, R.F., Gregor, R.J., & Diestel, J.D. (1981). Evidence for generalized motor programs using gait pattern analysis. Journal of Motor Behavior. 13 . 33-47. Sharp, R.H., & Whiting, H.T.A. (1975). Information-processing and eye movement behavior in a ball-catching skill. Journal of Human Movement Studies. 1 . 124131. Shaw, M.L. (1983). Division of attention among spatial locations: A fundamental difference between detection of letters and detection of luminance increments. In H. Bouma and D.G. Bouwhuis (Eds.), Attention & performance. X (pp. 109-121). Hillsdale, NJ: Lawrence Erlbaum.

PAGE 174

165 Sigman, M., & Coles, P. (1980). Visual scanning during pattern recognition in children and adults. Journal of Experimental Child Psychology. 30 . 267-276. Singer, R.N., Cauraugh, J.H., Chen, D., Steinberg, G.M., Frehlich, S.G. (1993). Mental quickness in tennis: Project I: Comparisons of advanced and beginning players. Project II: A quickness training program . Technical report submitted to the United States Tennis Association (contract #37025). Singer, R.N., Cauraugh, J.H., Chen, D„ Steinberg, G.M., Frehlich, S.G. (1996). Visual search, anticipation, and reactive comparisons between highlyskilled and beginning tennis players. Journal of Applied Sport Psychology. 8 . 9-26. Singer, R.N., Cauraugh, J.H., Chen, D., Steinberg, G.M., Frehlich, S.G., & Wang, L. (1994). Training mental quickness in beginning/intermediate tennis players. The Sport Psychologist. 8 . 305-318. Singer, R.N., Cauraugh, J.H., Frehlich, S.G., Barba, D.A., Kelly, G.K, & Umans, M. J. (1993). Training quickness and explosiveness: The ReAction Coach . Technical report submitted to START Technologies, New York. Singer, R.N., & Chen, D. (1994). A classification scheme for cognitive strategies: Implications for learning and teaching psychomotor skills. Research Quarterly for Exercise and Sport. 65 . 143-151. Slater-Hammel, A.T., & Stumpner, R.L. (1950). Batting reaction time. Research Quarterly, 21 . 353-356. Slater-Hammel, A.T., & Stumpner, R.L. (1951). Choice batting reaction time. Research Quarterly. 22 . 377-380. Spijkers, W.A.C. (1989). Effect of average movement velocity on reaction time and spatiotemporal accuracy in single aiming and rapid-timing movement tasks. Journal of Motor Behavior. 23 . 207-224. Spijkers, W.A.C., & Steyvers, F. (1984). Specification of direction and duration during programming of discrete sliding movements. Psychological Research. 46 . 59-71.

PAGE 175

166 Starkes, J.L. (1987). Skill in field hockey: The nature of the cognitive advantage. Journal of Sport Psychology. 9 . 146-160. Starkes, J.L., Allard, F., Lindley, S., & O'Reilly, K. (1994). Abilities and s kill in basketball International Journal of Sport Psychology. 25 . 249-265. Starkes, J.L., & Deakin, J. (1984). Perception in sport: A cognitive approach to skilled performance. In W.F. Straub & J.M. Williams (Eds.), Cognitive sport psychology (pp. 115-128). Lansing, NY: Sport Science Associates. Sternberg, S., Monsell, S., Knoll, R.L., & Wright, C.E. (1978). The latency and duration of rapid movement sequences: Comparisons of speech and typewriting. In G.E. Stelmach (Ed.), Information processing in motor control and learning (pp. 1 17-152). New York: Academic Press. Summers, J.J. (1977). The relationship between the sequencing and timing components of a skill Journal of Motor Behavior. 9 . 49-59. Temprado, J.J., & Spijkers, W.A.C. (1996). Effects of temporal patterns of the movement on reaction time: A review. International Journal of Sport Psychology, 27 . 337354. Terzuolo, C.A., & Viviani, P. (1979). The central representation of learned motor programs. In R.E. Talbot & D.R. Humphrey (Eds.), Posture and movement (pp. 113121). New York: Raven. Theeuwes, J. (1993). Visual selective attention: A theoretical analysis. Acta Psychological. 83 . 93-154. Turvey, M.T. (1990). Coordination. American Psychologist. 45 . 938-953. Turvey, M.T. (1992). Ecological foundations of cognition: Invariants of perception and action. In H. Pick, P. Van Den, & D. Knill (Eds.), Cognition: Conceptual and methodological issues (pp. 85-117). Washington, DC: American Psychological Association. Turvey, M.T., & Carello, C. (1986). The ecological approach to perceiving-acting: A pictorial essay. Acta Psvchologica. 63 . 133-155.

PAGE 176

167 Tyldesley, D.A., Bootsma, R.J., & Bomhoff, G.T. (1982). Skill level and eye movement patterns in a sport oriented reaction time task. In H. Reider, K. Bos, & K. Reischle (Eds.), Motor learning and movement behavior: Contribution to learning and knowledge (pp. 290-296). Cologne: Hofmann. van Wieringen, P.C.W. (1986). Motor coordination: Constraints and cognition. A faction to K.M. Newell In M.G. Wade & H.T.A. Whiting (Eds.), Motor development in children: Aspects of coordination and control (pp. 361-371). Dordrecht: Martinus Nijhoff. van Wieringen, P.C.W. (1988). Kinds and levels of explanation: Implications for the motor systems versus action systems controversy. In O.G. Meijer & K. Roth (Eds.), Complex movement behavior: The motor-action controversy (pp. 87-119). Amsterdam: North-Holland. Vickers, J.N. (1988). Knowledge structures of elite-novice gymnasts. Journal of Human Movement Science. 7 . 47-72. Vickers, J.N. (1992). Gaze control in putting. Perception. 21 . 117-132. Vickers, J.N. (1996). Visual control when aiming at a far target. Journal of Experimental Psychology: Human Perception and Performance. 22 . 342-354. Vickers, J.N. (1997). Control of visual attention during the basketball free throw. The American Journal of Sports Medicine. 24 . 93-97. Vickers, J.N., & Adolphe, R.M. (1997). Gaze behavior during a ball tracking and aiming skill. International Journal of Sport Vision. 4 . 18-27. Viviani, P. (1990). Eye movements in visual search: Cognitive, perceptual, and motor control aspects. In E. Knowler (Ed.), Eve movements and their role in visual and cognitive processes (pp. 353-393). Amsterdam: Elsevier Science. Whiting, H.T.A. (1972). Acquiring hall skill. London: Bell. Whiting, H.T.A, Savelsbergh, G.J.P., & Faber, C.M. (1988). "Catch" questions and incomplete answers. In A.M. Colley & J. Beech (Eds.), Cognition and action in skilled behaviors (pp. 257-271). Amsterdam: North-Holland.

PAGE 177

168 Williams, AM., Davids, K., Burwitz, L., & Williams, J.G. (1992). Perception and action in sport. Journal of Human Movement Studies. 22 . 147-204. Williams, A.M., Davids, K., Burwitz, L., & Williams, J.G. (1994). Visual search strategies in experienced and inexperienced soccer players. Research Quarterly for Exercise and Sport. 65 . 127-135. Wooden, J.R. (1988). Practical modem basketball . New York: Wiley. Wright, D.L., Pleasants, F., & Gomez-Mesa, M. (1990). Use of advanced visual cue sources in volleyball. Journal of Sport and Exercise Psychology. 12 . 406-414. Wright, R.D., & Ward, L.M. (1994). Shifts of visual attention: An historical and methodological overview. Canadian Journal of Experimental Psychology. 48 . 151-166. Yarbus, A.L. (1967). Eve movements and vision . New York: Plenum Press. Zangemeister, W.H., & Stark, L. (1982). Gaze latency: Variable interaction of head and eye latency. Experimental Neurology. 75 . 389-406. Zelaznik, H., Hawkins, B„ & Kisselburgh, L. (1983). Rapid visual feedback processing in single-aiming movements. Journal of Motor Behavior. 15 . 217-236.

PAGE 178

169 BIOGRAPHICAL SKETCH Shane Gary Frehlich was bom on December 14, 1968 in Macklin, Saskatchewan, Canada. His family moved to the province of Alberta shortly thereafter, and he graduated from Canmore Collegiate High School in 1986. In the fall of that year, he entered the University of Calgary and graduated with a Bachelor of Arts degree in psychology in the spring of 1990. In 1992 he entered the graduate program of the College of Health and Human Performance at the University of Florida, specializing in motor behavior and sport psychology. Shane earned his Doctor of Philosophy degree in the fall of 1997. Prior to completing his studies at the University of Florida, he was employed as a Lecturer in the Department of Physical Education at the State University of New York at Cortland, where he taught undergraduate and graduate classes in motor behavior, sport psychology, research methods, and the social psychology of sport and exercise. Shane is currently employed in his capacity as a teacher and researcher at SUNY Cortland.

PAGE 179

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Robert Smger, Chair Professor of Exercise and Sport Sciences I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. James Cauraugh Associate Professor of Exercise and Sport Sciences I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. 'Milledge Murphey ' & Associate Professor of Exercise and Sport Sciences I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Ira Fischler Professor of Psychology

PAGE 180

This dissertation was submitted to the Graduate Faculty of the College of Healt and Human Performance and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy December 1997 Dean, College of Health and Human Performance Dean, Graduate School