Journal of Sport & Exercise Psychology, 2007, 29, 457-4
© 2007 Human Kinetics, Inc.
Mann and Janelle are with the University of Florida, Williams is with Liverpool John Moores University
and Florida State University, and Ward is with Florida State University.
Perceptual-Cognitive Expertise in Sport:
Derek T.Y. Mann,1 A. Mark Williams,2,3 Paul Ward,3
and Christopher M. Janelle1
1University of Florida; 2Liverpool John Moores University;
3Florida State University
Research focusing on perceptual-cognitive skill in sport is abundant. However, the
existing qualitative syntheses of this research lack the quantitative detail neces-
sary to determine the magnitude of differences between groups of varying levels
of skills, thereby limiting the theoretical and practical contribution of this body
of literature. We present a meta-analytic review focusing on perceptual-cognitive
skill in sport (N = 42 studies, 388 effect sizes) with the primary aim of quantifying
expertise differences. Effects were calculated for a variety of dependent measures
(i.e., response accuracy, response time, number of visual ﬁxations, visual ﬁxation
duration, and quiet eye period) using point-biserial correlation. Results indicated
that experts are better than nonexperts in picking up perceptual cues, as revealed
by measures of response accuracy and response time. Systematic differences in
visual search behaviors were also observed, with experts using fewer ﬁxations of
longer duration, including prolonged quiet eye periods, compared with nonex-
perts. Several factors (e.g., sport type, research paradigm employed, and stimulus
presentation modality) signiﬁcantly moderated the relationship between level of
expertise and perceptual-cognitive skill. Practical and theoretical implications are
presented and suggestions for empirical work are provided.
Key Words: expert, skill acquisition, anticipation, advance-cue usage, visual
Sport expertise has been deﬁned as the ability to consistently demonstrate
superior athletic performance (Janelle & Hillman, 2003; Starkes, 1993). Although
superior performance is readily apparent on observation, the perceptual-cognitive
mechanisms that contribute to the expert advantage are less evident. Perceptual-
cognitive skill refers to the ability to identify and acquire environmental informa-
tion for integration with existing knowledge such that appropriate responses can
be selected and executed (Marteniuk, 1976). Knowing where and when to look is
crucial for successful sport performance, yet the visual display is vast and often
458 Mann, Williams, Ward, and Janelle
saturated with information both relevant and irrelevant to the task. Sport perform-
ers must be able to identify the most information-rich areas of the display, direct
their attention appropriately, and extract meaning from these areas efﬁciently and
effectively (Williams, Davids, & Williams, 1999).
For nearly three decades, researchers have sought to better understand the
psychological factors that discriminate outstanding from less outstanding indi-
viduals in sport (■Starkes & Ericsson, 2003). Researchers have demonstrated
that experts possess extensive procedural and declarative knowledge that enables
them to extrapolate important information from the environment to anticipate
and predict future events (French & Thomas, 1987; French, Spurgeon, & Nevett,
1995; McPherson, 1999, 2000). Experts are typically more proﬁcient at making
decisions and possess an unparalleled ability to foreshadow or predict future events
and outcomes (Holyoak, 1991; Starkes & Allard, 1993; Williams et al., 1999).
Furthermore, expert performers possess enhanced perceptual-cognitive skills,
such as effective attention allocation and cue utilization, each of which have been
demonstrated across sporting and other domains. This has led to further inquiry
into the role of perceptual skill acquisition in the development of sport expertise
(Abernethy & Russell, 1987a, 1987b). Consequently, emphasis has been placed on
clarifying how experts learn to acquire perceptual cues, as well as understanding
the superior ability of experts to process task- and domain-speciﬁc information
Regardless of their individual attributes, all sport contexts require athletes to
focus attention on the most appropriate cues so as to perform effectively. It is not
surprising, therefore, that experts have been shown to differ from nonexperts on
sport-speciﬁc measures of attention allocation and information pickup. Despite
these empirical efforts, widely pervasive conceptual and methodological variability
has made it difﬁcult to extract information that can clearly advance the science of
expertise while offering practical recommendations for training perceptual-cognitive
skills. Several issues worthy of consideration are brieﬂy presented in the following
section and then revisited in the description of moderator variables given in the
Method section as they directly affect the ability to determine the magnitude of
the expert advantage.
Limitations of Extant Research
A multitude of research protocols (anticipation, decision making, recall, task per-
formance, spatial and temporal occlusion, and eye-movement registration) have
been used to elicit expertise differences in cognitive and perceptual skill. Although
valuable, such a rich and diverse research base has hindered the ability to compare
effects across different protocols. For example, although the occlusion paradigm
has been instrumental in identifying the importance of speciﬁc cues, research
employing this paradigm may not maintain ecological saliency on the perceptual
dimension (see Hoffman & Deffenbacher, 1993) or the essential characteristics of
the task to be captured in a holistic manner. As such, examination of the expertise
effects noticed in various paradigms is warranted.
A critical factor in the study of expert performance concerns the ability to create
experimental tasks and conditions that allow the expert advantage to emerge (Erics-
son & Smith, 1991). Detailed consideration of the experimental settings, whether
Visual Search and Expertise 459
laboratory-based or otherwise, is paramount to expertise researchers in attempting
to reproduce this advantage. However, researchers have relied on a wide range of
stimulus presentation and task performance modalities. For example, the use of
video (ﬁlm) and slide presentations is often employed in visual search investiga-
tions, potentially altering the perceptual and sensory experience (Isaacs & Finch,
1983). Although construct validity has repeatedly been demonstrated in dozens of
experiments, one could argue that two-dimensional stimulus presentations may not
adequately capture the dynamic nature of sport (Abernethy, Burgess-Limerick, &
Parks, 1994a). Few researchers have made explicit comparisons of presentation
modality in this regard.
Related to the mode of stimulus presentation, response characteristics have
often been insufﬁciently considered in the experimental design. The expert advan-
tage may be disguised or even masked by an inability to link stimulus characteristics
to response selection and execution in contrived settings. For instance, a baseball
player who watches a video segment of a pitch and then responds with a button
press (e.g., Radlo, Janelle, Barba, & Frehlich, 2001) may rely upon a different
perception-action coupling than when facing an actual pitcher and swinging a
bat. Therefore, it is relevant to compare the magnitude of the expert/nonexpert
performance difference across tasks, including those in the laboratory and those
in the actual sport setting.
The Current Project
As described, a number of techniques, protocols, and measurement tools have
been used to index differences in expert sport performance. The inability to extract
deﬁnitive conclusions regarding the magnitude of the overall effects warrants a
quantitative synthesis of the extant literature. Pursuant to this goal, the purpose of
this project was to conduct a meta-analysis of sport expertise to assess the most
prevalent outcome measures identiﬁed in the literature concerning perceptual-cog-
nitive differences between expert and nonexpert athletes (Rosenthal & DiMatteo,
2001). These measures included response accuracy, response time, number of
ﬁxations, ﬁxation duration, and quiet eye period. Response accuracy represents the
participant’s frequency of producing appropriate responses according to objective
standards and in accord with environmental constraints and task demands. Response
time is deﬁned as an objective measure of the elapsed time between stimulus onset
and the overt production of a response.
In addition to performance metrics, several indices of attentional allocation
differences between experts and nonexperts have been used by expertise research-
ers. During eye movement registration, both the number of ﬁxations and ﬁxation
duration index an individual’s point of interest and relative attention allocation. The
longer the eye remains ﬁxated on a given target, the more information is thought
to be extracted from the display (albeit not necessarily from the locus of ﬁxation),
permitting detailed information processing. Additionally, the number of visual ﬁxa-
tions during a given period of time provides an index of the search characteristics
representative of the most pertinent cues extracted from the environment to facili-
tate the decision-making process. It should be noted, however, that corresponding
movements of 5° or less are often considered noise and statistically removed from
the calculation of ﬁxation duration, which typically ranges from 150 ms up to
460 Mann, Williams, Ward, and Janelle
600 ms (Irwin, 1992). Sport scientists have recorded ﬁxations as short as 100 ms
and as long as 1,500 ms with corresponding movements of 1° or less (Williams et
al., 1999). Eye movements between successive ﬁxations, known as saccades, are
believed to suppress information processing. In sport, given the typically dynamic
context, researchers have typically interpreted visual search strategies involving
fewer ﬁxations of longer duration to be more representative of the expert than the
nonexpert performer, as this would allow more time for information extraction.
Finally, quiet eye is believed to be a period of time when task-relevant environmental
cues are processed and motor plans are coordinated for the successful completion
of an upcoming task (Vickers, 1996). Speciﬁcally, the quiet eye period represents
the elapsed time between the last visual ﬁxation on a target and the initiation of
the motor response (Vickers, 1996).
Given the diverse approaches for examining the expert/nonexpert difference
put forth in the literature, coupled with the numerous dependent measures for quan-
tifying the expert/nonexpert difference, our aims were threefold. First, our primary
aim was to determine the overall effect of perceptual cue usage and visual search
behaviors on performance. More speciﬁcally, we sought to determine the extent
to which perceptual cue usage discriminates between experts and nonexperts. A
second aim was to evaluate the relationship between visual search strategies and
expertise. We were speciﬁcally interested in whether experts require fewer ﬁxations
of longer duration in order to extract relevant information from the environment.
Furthermore, narrative reviews have failed to differentiate the impact of various
moderating variables. Therefore, our third aim was to assess the extent to which
the expert/nonexpert differences varied as a function of the research paradigm and
Experts were expected to demonstrate superior response accuracy coupled with
faster response times, while executing fewer visual ﬁxations of longer duration.
Furthermore, experts were hypothesized to exhibit a signiﬁcantly longer quiet
eye period than the nonexpert comparison group. We also predicted that across
all dependent measures (a) the research paradigm employed would signiﬁcantly
moderate the expert/nonexpert relationship, with more commensurate tasks on
both the perceptual and action dimensions evoking a greater expert advantage;
(b) a larger effect in favor of the experts would be evident for real-world tasks as
compared with ﬁlm and static slide presentations; and (c) sport type would mod-
erate the expertise relationship only for response time, ﬁxation duration, number
of ﬁxations, and quiet eye, but not for response accuracy. Regardless of the sport
type, performance accuracy was expected to be superior for the experts across
comparisons with nonexpert performers.
Visual Search and Expertise 461
An exhaustive search of the expertise literature was conducted in an effort to
locate all relevant studies, including the ancestry and descendancy approach and a
computer-generated key word search of Dissertation Abstracts Online (1861–2004),
PsychINFO (1967–2004), and SPORTDiscus (1830–2004). The key words included
anticipation, cue use, expertise, decision-making, eye movement, eye-tracking,
information processing, occlusion, quiet-eye, sport, visual attention, and visual
search. In accord with the ancestry approach, the reference lists of all obtained
review articles and research studies were perused, followed by a manual search
of the following peer-reviewed journals: Canadian Journal of Sports Sciences,
Human Movement Science, International Journal of Sport Psychology, Journal
of Applied Sport Psychology, Journal of Sport & Exercise Psychology, Perceptual
and Motor Skills, Quest, Research Quarterly for Exercise and Sport, and Sport
Science Review. In accord with the descendancy approach, reference to several
seminal works were entered into a database (e.g., Social SciSearch, Get Cited)
in an effort to locate those studies referencing the early work used to compile the
working database from which this meta-analysis was derived.
Studies were considered for inclusion in this meta-analysis if they assessed
performance differences (response accuracy and response time) or visual search
characteristics (ﬁxation duration, number of ﬁxations, and quiet eye), if they
employed an expert/nonexpert paradigm, and if data (means and SD, t value, exact
p value, or a simple effect F ratio) were available to compute an effect size (point-
biserial correlation; Rosenthal, 1984) expressed as rpb (Cooper & Hedges, 1994).
Additionally, studies were retained for inclusion if the author failed to include the
requisite data to compute an effect size but clearly stated the direction and sig-
niﬁcance of the expert/nonexpert relationship (i.e., no signiﬁcant difference). The
Results section provides an elaborate discussion of this inclusion procedure. The
multidimensional search process resulted in approximately 240 related abstracts,
and research and review papers. Of the 180 articles retrieved, 42 met the inclusion
criteria, generating 388 effect sizes from studies involving 1,288 participants, with
45.6% (n = 588) classiﬁed as expert and 54.35% (n = 700) classiﬁed as nonexpert
Independent study ratings were conducted for each study included in this meta-
analysis to assess potential coder drift and study quality. Interrater reliabilities were
computed for a number of study characteristics, including skill level, paradigm,
sport type, presentation modality, and study quality. Rater agreement ranged from
0.83 to 0.95 across categories and was therefore deemed acceptable. In the case
of an interrater discrepancy, a consensus was met before inclusion into the study.
Given that study quality was deemed consistent across those studies retained for
inclusion, study quality was not assessed a potential moderator variable.
462 Mann, Williams, Ward, and Janelle
The extent to which the magnitude and direction of the expert/nonexpert relation-
ship varied as a function of several moderator variables was examined. Based on
the limitations presented earlier, the (a) research paradigm employed, (b) mode
of stimulus presentation, and (c) type of sport, were identiﬁed and assessed as a
function of expertise.
Researchers have argued that perception and action are mutually interdependent,
cyclical processes that directly constrain and inﬂuence one another (Williams
et al., 1999). Although it has been well documented that the effective use of
relevant advance visual cues facilitates sport performance by means of anticipat-
ing opponents’ intentions (Williams & Davids, 1998; Williams et al., 1999), the
development of research protocols that permit relevant perception and action are
warranted. Furthermore, a comparison of the paradigms inherently restricting the
perception–action coupling (i.e., when individuals are asked to verbally or physi-
cally respond in a manner that is inconsistent with the way in which they would
typically perform the task) with those more representative paradigms (i.e., verbally
or physically performing the task in a manner that is consistent with the way in
which they would typically perform the task in the real world) may provide valuable
insight into the effects that the decoupling of perception and action may have on
performance, the visual search processes, and the corresponding expert/nonexpert
difference (Williams et al., 1999).
Researchers have made extensive use of the recall paradigm to assess the
degree to which the expert maintains a cognitive advantage over the lesser skilled
performer. The recall paradigm comprises both static and dynamic images, portray-
ing either a structured or unstructured task-speciﬁc display. In either case, upon
brief exposure to the image, the participant is required to recall the location of
each player present in the display. Performance is then ascertained as the level of
agreement between a priori–identiﬁed features in the actual display (e.g., player
positions) and the participant’s recall of those features (Williams & Davids, 1995).
Although expert/nonexpert differences have been reliably demonstrated across tasks,
the degree to which this task captures the essence of domain-speciﬁc performance
is questionable. Another concern in the task design is that it measures only the
accuracy of recall, neglecting the time taken to respond. Given the inherent time
constraints in sport, athletes must not only retrieve, encode, and respond accurately,
but also must respond under severe time pressure. Furthermore, the two-dimensional
representation of the sport context coupled with the frequent use of static images
may not truly capture expertise differences in sport given that movement may be
an integral component of the pattern recognition process (Williams et al., 1999).
As such, including the recall paradigm as a distinct level of a moderating variable
will help identify its utility on parsing the expert/nonexpert differences.
The occlusion paradigm, popularized by Jones and Miles (1978), was tradi-
tionally espoused as the paradigm of choice to probe the perceptual behaviors of
athletes. Both temporal and spatial occlusion techniques have been employed to
systematically demonstrate expert/nonexpert differences in the use of information
Visual Search and Expertise 463
presented early in the visual display across a variety of sports, including tennis,
badminton, squash, cricket, baseball, and volleyball (Abernethy & Russell, 1987a,
1987b; Buckolz, Prapavessis, & Fairs, 1988; Starkes, Edwards, Dissanayake, &
Dunn, 1995). A summary of these experiments suggests that (1) experts are better
able to predict the direction and force of an opponent’s stroke based on kinematic
information that maintain subtle clues (such as the dominant arm of a tennis player)
(Abernethy, 1990b; Wright, Pleasants, & Gomez-Mesa, 1990) and (2) experts are
more adept than nonexperts at using early ﬂight cues to predict the ball’s end loca-
tion. These ﬁndings have been relatively consistent, signifying the attunement of
expert-level performers to advance cues otherwise neglected by nonexpert perform-
ers (Abernethy & Russell, 1987a; Buckolz et al., 1988; Jones & Miles, 1978).
Although the utility of occlusion paradigms has been clearly conﬁrmed, the
inherent limitations of this approach should be mentioned. First, both temporal and
spatial occlusion paradigms capture only a speciﬁc aspect of the task. When these
paradigms have omitted a physical or real-world response (e.g., Singer, Cauraugh,
Chen, Steinberg, & Frehlich, 1996; Williams & Burwitz, 1993), they may negate
the expert advantage, and may only partially capture speciﬁc elements of the deci-
sions made (Abernethy, Thomas, & Thomas, 1993). Second, the use of occlusion
techniques prohibits the connections of perceptual information (either temporal or
spatial) by restricting the sequential processing of subsequent perceptual cues and
therefore promotes the use of alternative cognitive strategies for decision making.
That is, rarely in sports are the athletes unable to view their opponents in their
entirety, yet occlusion paradigms inherently restrict the presentation of information.
From an information-processing approach, this may yoke very different connections
between perceptual stimuli and the declarative knowledge necessary to reach an
accurate problem solution (Abernethy et al., 1993; Williams et al., 1999).
A major point of contention thus far has been the lack of an ecologically
valid means for evaluating the expert/nonexpert difference. Therefore, studies
implementing sport-relevant tasks including the observation of actual performance
were isolated to in order to construct a “task performance classiﬁcation” for sub-
sequent moderator analyses. As such, those investigations preserving the tendency
to contrive the environment by means of occlusion, static slide, video, and or other
artiﬁcial means of manipulation (Williams, Singer, & Frehlich, 2002) were excluded
from the task performance classiﬁcation.
Several researchers (e.g., Bard & Fleury, 1987; Abernethy & Russell, 1987b)
have made extensive use of the frequency and duration of visual ﬁxations in the
absence of other performance measures or dependent variables in an effort to
unveil expert/nonexpert differences (Petrakis, 1986). These studies were classiﬁed
Static slide presentations inherently fail to present the participant with the dynamic
attributes of the visual environment consistent within most sporting domains (Aber-
nethy et al., 1994a). The use of dynamic ﬁlm or video may offer a more natural
perception of the scene when compared with static slides. However, both slides and
ﬁlm or video presentations reduce a three-dimensional world into a two-dimensional
image, potentially changing the perceptual and sensory experience. Abernethy et al.
464 Mann, Williams, Ward, and Janelle
(1993) suggested that tasks that take place in the real world should further discern
expert/nonexpert differences by exposing the participant to additional sources of
information not available in two-dimensional media, such as stereoscopic depth
information. Few explicit comparisons of these media have been made within a
single study. Therefore, a comparison of the effect sizes associated with ﬁlm, slide,
and real-world stimulus presentations were examined as potential moderators.
The current status of the perceptual-cognitive expertise literature suggests that the
perceptual strategies and corresponding decision-making processes of experts and
nonexperts is task dependent (Williams, Davids, Burwitz, & Williams, 1993; ■Wil-
liams et al., 1994; Williams & Davids, 1995). As such, the visual search behaviors
of expert and nonexpert players from one sport may be inconsistent with those
from another. For example, the contextual demands of anticipating a passing shot in
tennis may require different information-processing strategies when compared with
the underlying processes associated with anticipating a pass destination in a 3-on-3
soccer task. Therefore, sports were classiﬁed as interceptive (or coactive), strategic
(or interactive or invasive), and other (or independent or propulsive) to determine
the effect of sport type on expert/nonexpert comparisons. An interceptive sport
was deﬁned as any sport that requires coordination between a participant’s body,
parts of the body or a held implement, and an object in the environment (Davids,
Savelsbergh, Bennett, & Van der Kamp, 2002; e.g., squash, badminton, tennis);
a strategic sport was operationalized as a sport that involves multiple teammates,
often resulting in tactical formations during offensive and defensive series, and
emphasizing the importance of allocating attention to both the projectile involved
and the diverse array of participants (i.e., ﬁeld hockey, soccer); ﬁnally, a sport clas-
siﬁed as other included such characteristics as being closed, self-paced, and aiming
at a target (e.g., billiards, golf, target shooting). As a result of the varied contextual
demands of sport, it is not altogether surprising to suspect mixed perceptual-cog-
nitive strategies across sport. Therefore, conducting a moderator analysis on the
expert/nonexpert difference across sport types is necessary to further our current
understanding of the role of task speciﬁcity on expertise.
Calculation of Effect Sizes and Statistical Analyses
Meta-analytic procedures and statistical techniques outlined and advocated by
Hedges and Olkin (1985), Cooper and Hedges (1994), Rosenthal (1984, 1995), and
Rosenthal and DiMatteo (2001) were used to conduct a ﬁxed effects meta-analysis.
To clarify, a ﬁxed effects analysis restricts signiﬁcance testing to the total number
participants and not to the total number of studies. As such, a ﬁxed effects approach
results in greater statistical power (Rosenthal, 1995). ■Effect size estimates, rpb,
and overall mean rpb were calculated for each dependent variable. Many studies in
the expertise literature have assessed multiple dependent measures relevant to this
research synthesis, including response time, response accuracy, number of visual
ﬁxations, total ﬁxation duration, and quiet eye duration. Although each dependent
measure provides important information furthering our understanding of expert and
nonexpert differences, including multiple dependent measures in one quantitative
Visual Search and Expertise 465
synthesis inﬂates the sample size beyond the number of independent studies,
rendering it difﬁcult to estimate the true error associated with the overall effect
size, while also inﬂating the Type I error rate (Wolf, 1986). Grouping the various
dependent variables into one quantitative synthesis perpetuates the “apples and
oranges” criticism of meta-analytic reviews. To avoid this pitfall, and in accordance
with Rosenthal (1984), each dependent variable was analyzed separately.
Estimates of effect size are subject to positive bias in small samples and there-
fore should be adjusted to account for the within-study sample size variability. Each
effect size was therefore weighted by the reciprocal of its variance by using Fisher’s
(1925) variance stabilizing z-transform. An overall weighted mean effect size and
an estimate of the associated variance was obtained. Subsequent analyses included
the calculation of the mean rpb, 95% conﬁdence intervals (CI) around the mean to
determine whether effects were signiﬁcantly different from zero, and comparisons
of the mean rpb between levels of moderator variables (Cooper & Hedges, 1994, pp.
265-268). Additionally, the omnibus test statistics Q, QBET, and Qw, were computed
to determine within-group and between-group sources of variation (Hedges & Olkin,
1985). Heterogeneity was calculated and indicated whether Q (the weighted total
sum of squares about the grand mean; Cooper & Hedges, 1994) exceeded the upper
tail critical value of χ2 at k − 1 degrees of freedom (Cooper & Hedges, 1994, p.
266). To test the between-group differences for each moderator variable, the QBET
was calculated (Cooper & Hedges, 1994). Furthermore, preplanned linear contrasts
were performed on each moderator variable to test the difference among levels of a
given moderator variable. As such, the 95% conﬁdence interval and corresponding
χ2 value were calculated for each preplanned comparison (Cooper & Hedges, 1994).
In an effort to avoid the inﬂation of the Type I error rate, only the following linear
contrasts were computed for the moderator variable, paradigm: temporal – spatial,
anticipation – decision-making, task – anticipation, decision-making – task.
According to Rosenthal (1991), the probability of a meta-analyst accessing
all research, published and unpublished, is low, and furthermore the research is
unlikely to be a random sample of the existing research owing to publication bias.
To estimate the hypothetical effects of these limitations on the aggregated effect
size, a fail-safe n is necessary to estimate the number of studies averaging null
results needed to attenuate the observed effect and was thereby computed for each
dependent variable. Details for this calculation are provided by Rosenthal (1991).
Fail-safe n = [(ΣZ)/1.96]2 − k
where Z is the sum of the standard normal deviates for k studies.
All qualiﬁcations of the magnitude and effects of the estimated effect size
reported here are based on the recommendations of Cohen (1977) for correlational
effect sizes, such that the values of .10, .30, and .50, represent small, medium, and
large effect size estimates, respectively. Furthermore, to facilitate the interpretation
and practical signiﬁcance of the corresponding effect size, the results of a binomial
effect size display (BESD) will be presented for each dependent measure (Cooper
& Hedges, 1994). The BESD is a practical interpretation of the overall effect size
expressed as the difference in outcome rates between, in this case, the expert and
nonexpert groups for each of the dependent measures.
466 Mann, Williams, Ward, and Janelle
As mentioned previously, 388 effect sizes were calculated across the ﬁve dependent
measures: response accuracy, response time, ﬁxation duration, number of ﬁxations,
and quiet eye duration. Each dependent measure was analyzed separately (see
Rosenthal, 1984). A common ﬁnding across all but one dependent variable (i.e.,
quiet eye duration) was a signiﬁcant test of heterogeneity.
The source of total variation around the grand mean can be divided into within
and between sources of variability (Cooper & Hedges, 1994). Repeated calculation
of Qi (which is identical to Q with the ith effect removed) indicated that the source
of heterogeneity was explained by those studies that reported a lack of statistical
signiﬁcance and failed to provide sufﬁcient data to allow the actual effect size to
be calculated. These studies were assigned an effect size of rpb = 0.00 and a corre-
sponding one-tailed p value of 0.50 (Rosenthal, 1995). Inclusion of this procedure
is conservative; simply ignoring such null ﬁndings would result in the inﬂation of
the overall observed effect size for each dependent measure. Therefore, despite
the heterogeneity around the mean rpb, and in accord with the recommendations of
Rosenthal (1995), an overall estimate of the mean rpb was computed and moderator
analyses were conducted. Therefore, the results of each dependent measure will
include the overall Q statistic, in addition to the Qremoved statistic, to account for the
aforementioned source of variability (i.e., studies claiming “no effects”).
The analysis of 214 effect sizes in which response accuracy was assessed revealed
a medium mean effect size of 0.31 (95% CI 0.29–0.34), which was signiﬁcant (z =
13.83, p < .001). The fail-safe n was 1,386.9, indicating that approximately 1,400
studies averaging null results would be necessary to attenuate the signiﬁcance of the
current effect size at the .05 level. The distribution of effect sizes was heterogeneous,
Q(213) = 331.95, p < .001. When the effect sizes derived from missing data were
removed, the results approached signiﬁcance, Qremoved(195) = 226.60, p = .060. From
a practical perspective, it can be inferred that the experts were approximately 31%
more accurate across research studies as indexed by the BESD. Lastly, QBET was
calculated along with preplanned contrasts to test the difference between levels of
stimulus presentation, sport type, and study paradigm on the aggregated effect size
for response accuracy. A summary of these effect sizes is presented in Figure 1.
Sport Type. The type of sport performed was of primary interest as a potential
moderator to determine whether sport type (i.e., interceptive, strategic, and other)
inﬂuenced the skill-based performance difference. Although slight differences in
effect size magnitude were observed between the classiﬁcations of other (rpb = .37,
p < .001), interceptive (rpb = .32, p < .001), and strategic (rpb = .28, p < .001) sports,
these differences were not signiﬁcant, QBET(2) = 2.53, p = .28.
Research Paradigm. As Figure 1 indicates, the overall estimate of the between-
group difference is signiﬁcant, QBET(6) = 36.97, p < .001, suggesting that the
paradigm adopted to assess skill-based performance can yield variable effects.
Although signiﬁcant, no statistical differences were found for the preplanned
comparisons of interest.
Visual Search and Expertise 467
Stimulus Presentation. Researchers have questioned the degree to which various
stimulus presentation modalities adequately identify expert/nonexpert performance
differences in sport. A comparison of the presentation modalities yielded a sig-
niﬁcant effect, QBET(2) = 7.60, p = .02. The ﬁeld (rpb = .42, p < .001), video (rpb =
.31, p < .001), and static slide (rpb = .25, p < .001) stimulus presentations elicited
large-to-moderate effects with signiﬁcant increases in the magnitude of effects as
the mode of stimulus presentation became progressively more representative of a
real-world task (i.e., static, video, ﬁeld). Speciﬁcally, the preplanned comparison
between ﬁeld and static (χ2 = 16.99, p < .001), ﬁeld and video (χ2 = 5.19, p = .02),
and video and static were signiﬁcant (χ2 = 7.27, p < .001).
The analysis of 62 effect sizes in which response time was assessed revealed
that the aggregated effect was moderate, mean rpb = 0.35 (95% CI 0.30–0.40)
and signiﬁcant (z = 11.90, p < .001). The fail-safe n was 198.13, indicating that
approximately 200 studies averaging null results would be necessary to attenuate
the signiﬁcance of the current effect size at the .05 level. The distribution of effect
sizes was heterogeneous, Q(61) = 93.22, p < .001, and Qremoved(55) = 55.55, p =
.79). From a practical perspective, it can be inferred that the experts responded
approximately 35% faster across research studies as indexed by the BESD. Lastly,
QBET was calculated along with preplanned contrasts to test the difference among
levels of the aforementioned moderator variables. A summary of these ﬁndings is
presented in Figure 2.
Figure 1 — Summary of expertise difference for response accuracy.
468 Mann, Williams, Ward, and Janelle
Sport Type. Although the expert was reportedly more accurate than the nonex-
pert across the various sport classiﬁcations, the magnitude of this difference was
relatively consistent, suggesting that ■RA was not moderated by the nature of the
sport. Conversely, response time signiﬁcantly differed across sport type as a func-
tion of expertise, QBET(2) = 6.14, p = .05. Speciﬁcally, experts responded quicker
than their less skilled counterparts during interceptive sports (rpb = .37, p < .001),
strategic (rpb = .37, p < .001), and other (rpb = .15, p = .085) sports as evidenced by
the magnitude of the respective effect sizes across sport type, with notable sport
differences apparent between interceptive sports and those labeled other (e.g., bil-
liards; χ2 = 4.51, p = .033), and strategic sports and other (χ2 = 6.02, p = .014).
Research Paradigm. The paradigm adopted to assess response time signiﬁ-
cantly moderates the expert/nonexpert relationship, QBET(3) = 13.55, p = .01. The
anticipation paradigm evoked the largest performance difference (rpb = .43, p <
.001), followed by spatial occlusion (rpb = .37, p < .001), decision-making (rpb =
.31, p < .001), and recognition (rpb = .25, p < .001) paradigms. The preplanned
comparison of the decision-making and anticipation paradigms was signiﬁcant
(χ2 = 4.46, p = .034).
Stimulus Presentation. Although the experts evoked a quicker response than the
nonexpert performers during video presentation (rpb = .37, p < .001) as compared
with static slide presentations (rpb = .25, p < .001), this difference was not statisti-
cally signiﬁcant, QBET(1) = 1.50, p = .22. An assessment of the ﬁeld condition as a
moderator was not included as a result of insufﬁcient data.
Figure 2 — Summary of expertise differences for response time.
Visual Search and Expertise 469
Number of Fixations
A total of 58 effect sizes were calculated for number of ﬁxations. The aggregated
effect was small to moderate, mean rpb = 0.26 (95% CI 0.20–0.32) and signiﬁcant
(z = 9.21, p < .001). The fail-safe n was 94.09, indicating that approximately 94
studies reporting null results would be necessary to attenuate the signiﬁcance of the
current effect size at the .05 level. The distribution of effect sizes was heterogeneous,
Q(57) = 117.11, p < .001, and Qremoved(49) = 63.72, p = .09. QBET was calculated
along with preplanned contrasts to test the difference between levels of stimulus
presentation, sport type, and study paradigm on the aggregated effect size for number
of ﬁxations. A summary of these ﬁndings is presented in Figure 3.
Sport Type. As indicated in Figure 3, sport type—speciﬁcally strategic sports
(rpb = .49, p = .011)—exacerbate the expert/nonexpert visual search differences.
Experts executed fewer ﬁxations as compared with the lesser skilled performers
when completing strategic tasks as compared with interceptive (rpb = .10, p = .435)
and other sports (rpb = .35, p = .197), QBET(2) = 37.66, p < .001. However, owing to
the presence of within-group heterogeneity, these differences are not signiﬁcant.
Paradigm. Temporal and spatial occlusion paradigms were removed from the
analysis of number of ﬁxations as a result of insufﬁcient data. The results dem-
onstrated signiﬁcant skill-based differences across research paradigms, QBET(3) =
29.01, p < .001. Accordingly, the preplanned comparison of the moderate effects
Figure 3 — Summary of expertise differences for number of ﬁxations.
470 Mann, Williams, Ward, and Janelle
for decision making (rpb = .44, p < .001) as compared with the small effects associ-
ated with anticipation (rpb = .12, p = .003) was signiﬁcant (χ2 = 21.47, p < .001).
No other preplanned comparisons were signiﬁcant; however, expertise differences
were evident, with the expert group demonstrating signiﬁcantly fewer ﬁxations
across paradigms, with the only exception being eye movement paradigm (rpb =
.05, p = .795).
Stimulus Presentation. Skill-based differences were observed across presenta-
tion modalities, QBET(2) = 10.99, p = .004: video (rpb = .19, p < .001) and static
slides (rpb = .41, p < .001), with the expert performers committing fewer ﬁxations
as compared with the nonexpert performers.
The analysis of 49 effect sizes in which response time was assessed revealed that
the aggregated effect was small to moderate, mean rpb = 0.23 (95% CI 0.16–0.30)
and signiﬁcant (z = 6.68, p < .001). The fail-safe n was 11.9, indicating that
approximately 12 studies averaging null results would be necessary to attenuate
the signiﬁcance of the current effect size at the .05 level. The distribution of effect
sizes was heterogeneous, Q(48) = 102.00, p < .001, and Qremoved(39) = 66.28, p <
.001. From a practical perspective, it can be inferred that the experts exhibited
ﬁxation durations lasting approximately 23% longer across research studies as
indexed by the BESD. QBET was calculated along with preplanned contrasts to
test the difference between levels of stimulus presentation, sport type, and study
paradigm on the aggregated effect size for ﬁxation duration. A summary of these
ﬁndings is presented in Figure 4.
Figure 4 — Summary of expertise differences for ﬁxation duration.
Visual Search and Expertise 471
Sport Type. Sport type signiﬁcantly moderated the expert/nonexpert ﬁxation dura-
tion relationship. The effect sizes for each sport type were as follows: interceptive
(rpb = .14, p = .016), strategic (rpb = .23, p = .003), and other (rpb = .32, p < .001);
QBET(2) = 36.09, p < .001. Only the preplanned comparison of interceptive tasks
to other tasks was signiﬁcant (χ2 = 5.06, p = .024).
Paradigm. Although slight differences were present among the varied paradigms
used to assess ﬁxation duration—including temporal occlusion (rpb = .22, p = .041),
anticipation (rpb = .24, p < .001), decision making (rpb = .15, p = .011), task perfor-
mance (rpb = .40, p < .001), and eye movement (rpb = −.11, p = .542) paradigms;
QBET(4) = 35.66, p < .001—only the preplanned comparison between decision
making and task performance was signiﬁcant (χ2 = 6.38, p = .011).
Stimulus Presentation. Not unlike the ﬁndings associated with the number of
ﬁxations, presentation modalities signiﬁcantly moderated the expert/nonexpert
relationship, with the expert performer committing longer ﬁxations as compared
with the nonexpert group: video (rpb = .30, p < .001), static slides (rpb = −.36, p <
.001), and ﬁeld (rpb = .32, p < .001); QBET(2) = 38.25, p < .001. However, contrary
to hypotheses, when viewing static slides, the nonexpert group engaged in longer
ﬁxations than did the expert group. Speciﬁcally, the preplanned comparison between
video and static (χ2 = 35.97, p < .001) and static and ﬁeld presentation modalities
were signiﬁcant (χ2 = 30.40, p < .001).
The analysis of ﬁve effect sizes derived from 150 participants across three separate
laboratories in which the quiet eye duration was assessed revealed a moderate-to-
large mean rpb of 0.62, (95% CI 0.40–0.82) and signiﬁcant (z = 5.76, p < .001)
aggregated effect. The fail-safe n was 0.38, indicating that 1 study reporting null
ﬁndings would be necessary to attenuate the signiﬁcance of the current effect size
at the .05 level. The distribution of effect sizes was homogeneous, Q(4) = 6.158, p
= .189. As a result of the small sample, subsequent analyses of potential modera-
tors were not conducted. The magnitude and direction of the reported effect size
supported the hypothesis that experts exhibit longer quiet eye periods coupled with
superior performance as compared with their less skilled counterparts. Experts
maintain a quiet eye period that is approximately 62% longer in duration across
research studies as indexed by the BESD.
The purpose of this investigation was to provide a quantitative synthesis of the
research on perceptual-cognitive expertise in sport and to assess the moderating
effects of a number of commonly employed research paradigms, participant charac-
teristics, and presentation modalities. Perceptual cue usage and visual gaze behaviors
were assessed using a number of dependent measures. These outcome measures
provided a natural framework from which this meta-analysis was constructed.
The analysis of performance measures conﬁrmed expectations that experts were
more accurate in their decision making relative to their lesser skilled counterparts
472 Mann, Williams, Ward, and Janelle
(mean rpb = .31, p < .001). Moreover, experts anticipated their opponents’ intentions
signiﬁcantly quicker (mean rpb = .35, p < .018) than less skilled participants. These
results are consistent with the notion that the use of advance perceptual cues has been
demonstrated to facilitate sport performance by means of aiding in the anticipation
of opponent’s actions and decreasing overall response time (e.g., Goulet, Bard, &
Fluery, 1989; Helsen & Starkes, 1999). As Abernethy (1991) contends, decision
making in sport is the product of a sequence of events occurring well before overt
movement is required. For example, during racquet sports, an ordered sequencing
of events occurs, commencing with a range of reliable kinematic cues preceding
ball ﬂight, which, when processed, can foretell the probability of a given outcome.
The ability of expert performers to extract perceptual cues can alleviate the temporal
constraints imposed by reaction time alone (Buckolz et al., 1988). The presumption
is that the experts possess qualitatively different cognitive mechanisms and strate-
gies that facilitate anticipation, permitting reduced response times and increased
response accuracy (e.g., Ericsson & Kintsch, 1995).
In addition to performance indices, we were able to quantify the functional
signiﬁcance of the expert performers’ eye movement behaviors relative to their
nonexpert counterparts. Experts were characterized by fewer ﬁxations (mean rpb =
.26, p < .001) of longer duration (mean rpb = .23, p < .001). These ﬁndings support
the interpretation that experts in sport extract more task-relevant information from
each ﬁxation than do lesser skilled performers. Conversely, nonexperts typically
require more ﬁxations of shorter duration to gather sufﬁcient information to respond.
Because the ability to extract information from the display is reduced during sac-
cadic eye movements (Duchowski, 2002), one could argue that a strategy involving
more ﬁxations of shorter duration is less efﬁcient and effective than one involving
fewer ﬁxations or longer duration (Williams et al., 1993). Without sufﬁcient time
to process task-relevant cues, oversights and incorrect decisions are inevitable as
indicated by the inferior performance outcomes displayed by the novices.
In addition to typical ﬁxations and ﬁxation durations—although only few effects
were included in the analysis—the quiet eye period resulted in a large positive mean
effect size (rpb = .62, p < .001). Researchers have reliably demonstrated relatively
prolonged quiet eye periods as an effective marker for differentiating skilled and
lesser skilled athletes. Moreover, these ﬁndings have shown consistency across
domains as diverse as riﬂe shooting (Janelle et al., 2000) and billiards (■Williams
et al., 2002), for tasks that require aiming at a target (e.g., billiards and shooting),
and those that require the individual to receive a projectile momentarily while
aiming and releasing it to a designated target (e.g., volleyball; Vickers & Adolphe,
Several possible moderating effects of these results were examined. Results
indicated that sport type is not a signiﬁcant moderator of the expertise relationship
for response accuracy. Regardless of the type of sport performed, experts maintain
a perceptual advantage over their less skilled counterparts, facilitating response
accuracy. Response time, however, was inﬂuenced by sport type, with the largest
expert/nonexpert differences evident for interceptive sports (rpb = .37, p < .001), and
strategic sports (rpb = .37, p < .001), followed by other (rpb = .15, p = .085) sports.
The inherent temporal constraints associated with interceptive sports (e.g., tennis,
squash) render this ﬁnding intuitively appealing. Strategic sports (e.g., soccer, ﬁeld
hockey), in contrast, typically consist of a more elaborate sequencing of events,
Visual Search and Expertise 473
which may reduce the impending temporal pressures necessary to perform at a
superior level. However, the source of the greatest difference lies between inter-
ceptive sports and with those tasks classiﬁed as other (e.g., billiards, golf), which
are rarely faced with temporal constraints. Thus, although experts’ responses were
quicker, the speed with which this response occurs is at least partly constrained by
the nature of the task.
Similarly, the number of ﬁxations employed varied across sport type, with the
smallest margin of expert/nonexpert difference evident across interceptive sports
(rpb = .10). Clearly, this ﬁnding is attributable to the temporal constraints of task
duration. For example, an anticipation task in tennis, in which the service duration
from ball toss to racket contact may take no more than 300 ms (Abernethy, 1991),
is substantially shorter than a similar anticipation task in soccer, which may take
upwards of 9,000 ms (Williams et al., 1994b). In reality, however, the latter reﬂects
the time taken from the onset of a trial, whereas the anticipation response to some
critical event within that trial is likely to be more equivalent to that observed in
racket sports. Therefore, task duration alone will permit more ﬁxations in the soccer
task than is possible in a tennis task designed to assess the same ability. However,
the duration of each corresponding ﬁxation did not differ across sports, support-
ing the contention that experts seek the most information-dense areas of a display
while extracting task-relevant cues (Williams et al., 1993).
As predicted, the expert’s superior attunement to perceptual cues was moder-
ated by the research paradigm employed. For example, the effect sizes for response
time and response accuracy differed across paradigms, with the smallest expertise
difference noted with recall and recognition paradigms (Figures 1 and 2), that is,
tasks associated with the simple encoding and retrieval of sport speciﬁc informa-
tion. This ﬁnding questions whether performance on these tasks is predictive of
skilled performance (see also Ward & Williams, 2003). More likely, the primary
differentiation between expertise levels occurs when confronted with more complex
operations that occur rapidly, lack regularity, and are unpredictable. The remain-
ing protocols require the participant to not only encode and retrieve perceptual
information, but also to apply that information to a task that is skill dependent (i.e.,
anticipation and decision making).
The manner by which the testing stimulus was delivered to participants (i.e.,
video, static slide, and ﬁeld presentations) revealed a difference for response
accuracy and ﬁxation duration. The largest effect was reported in the ﬁeld stud-
ies, followed by video, and static slides (Figures 1 and 4, respectively), suggest-
ing that there is a greater likelihood of ﬁnding an expert advantage when skilled
participants are asked to perform in ecologically valid environments. Although a
number of the video-based paradigms have appropriately captured the essence of
expertise during task performance, other researchers have asked participants to
respond in an alternate manner or have changed the nature of the task such that
the expert advantage is diminished. Expert decision making was facilitated under
ﬁeld conditions, suggesting that the more realistic the paradigm, the greater the
Our quantitative ﬁndings support the early intuition of Jones and Miles (1978),
who discuss the inherent sterility of the laboratory and the inability of the laboratory
setting and task to accurately elicit comparable performance states. Such limitations
may confound the empirical estimates of perceived expert/nonexpert differences
474 Mann, Williams, Ward, and Janelle
garnered from such paradigms (see Abernethy et al., 1993). Although the argument
proposed by Abernethy and colleagues would appear to have merit, when ﬁeld-based
approaches are not permissible, video is a superior means of stimulus presentation
than static slides. Without question, ﬁeld-based approaches have elicited the great-
est differences; however, the real crux of the matter is whether the paradigm or
mode of presentation used accurately captures the superior performance of experts.
Moreover, one has to pay particular attention to the level of experimental control
achieved when testing in the naturalistic environment. Although effects sizes are
largest in the ﬁeld, it is difﬁcult to ascertain whether participants are responding to
different stimuli, rendering reliable comparison highly problematic.
To summarize, in this meta-analysis we have synthesized and quantiﬁed a
conceptually intricate body of expertise research. The locus of expert/nonexpert
difference has been a difﬁcult phenomenon to capture, given the diverse research
paradigms and varied experimental control, coupled with the wide-ranging opera-
tional deﬁnitions, techniques, and sampling characteristics. However, our quantita-
tive analysis has provided a means to objectively evaluate commonly held beliefs
concerning expertise in sport, conﬁrming that sports experts are typically more
accurate and quicker in their responses and generally employ fewer ﬁxations of
longer duration. More importantly, however, several factors (i.e., sport type, research
paradigm, and stimulus presentation modality) have been found to signiﬁcantly
moderate the various relationships between level of sport expertise and perceptual-
cognitive skill that should be used to guide expertise research in future years.
References preceded by an asterisk were included in the meta-analysis.
■Abernethy, B. (1988). Visual search in sport and ergonomics: Its relationship to selective
attention and performance expertise. Human Performance, 4, 205-235.
*Abernethy, B. (1989). Expert-novice differences in perception: How expert does the expert
have to be? Canadian Journal of Sports Sciences, 14(1), 27-30.
■*Abernethy, B. (1990a). Anticipation in squash: differences in advance cue utilization
between expert and novice players. Journal of Sports Sciences 8, 17-34.
*Abernethy, B. (1990b). Expertise, visual search, and information pick-up in squash. Per-
ception 19, 63-77.
Abernethy, B. (1991). Visual search strategies and decision-making in sport. International
Journal of Sport Psychology, 22, 189-210.
Abernethy, B. (1999). The 1997 Coleman Grifﬁth address: A juncture between psychological
theory and practice. Journal of Applied Sport Psychology, 11(1), 126-141.
Abernethy, B., Burgess-Limerick, R., & Parks, S. (1994a). Contrasting approaches to the
study of motor expertise. Quest, 46, 186-198.
■*Abernethy, B., Neal, R.J., & Koning, P. (1994b). Visual-perceptual and cognitive dif-
ferences between expert, intermediate, and novice snooker players. Applied Cognitive
Psychology, 18, 185-211.
*Abernethy, B., & Russell, D.G. (1987a). Expert-novice differences in an applied selective
attention task. Journal of Sport Psychology, 9, 326-345.
*Abernethy, B., & Russell, D.G. (1987b). The relationship between expertise and visual
search strategy in a racquet sport. Human Movement Science, 6, 283-319.
Abernethy, B., Thomas, K.T., & Thomas, J.T. (1993). Strategies for improving understand-
ing of motor expertise (or mistakes we have made and things we have learned!) In J.L.
Visual Search and Expertise 475
Starkes & F. Allard (Eds.), Cognitive issues in motor expertise. Amsterdam: Elsevier
■Allard, F., & Starkes, J.L. (1991). Motor skills experts in sports, dance, and other domains.
In K.A. Ericcson & J.Smith (Eds.), Toward a general theory of expertise: Prospects
and limits (pp. 126-152). Cambridge: Cambridge University Press.
■*Bard, C., & Fleury, M. (1976). Analysis of visual search activity during sport problem
situations. Journal of Human Movement Studies, 3, 214-227.
Bard, C., & Fleury, M. (1987). Considering eye movement as a predictor of attainment. In
I.M. Cockerill & W.W. MacGillivary (Eds.), Vision and sport. Cheltenham: Stanley
■Broadbent, D.E. (1958). Perception and communication. London: Pergamon Press.
Buckolz, E., Prapavessis, H., & Fairs, J. (1988). Advance cues and their use in predicting
tennis passing shots. Canadian Journal of Sport Science, 13(1), 20-30.
■*Cauraugh, J., Singer, R.N., & Chen, D. (1993). Visual scanning and anticipation of expert
and beginner tennis players. In S. Serpa (Ed.), Proceedings : VIII World Congress of
Sport Psychology. Lisbon, International Society of Sport Psychology, 336-340.
■Chase, W.G., & Simon, H.A. (1973). Perception in chess. Cognitive Psychology, 4, 55-
Cohen, J. (1977). Statistical power analysis for the behavioral sciences (rev. ed.). New
York: Academic Press.
Cooper, H., & Hedges, L.V. (Eds.). (1994). The handbook of research synthesis. New York:
Russell Sage Foundation.
■*Davids, K.W., Rex Pe Palmer, D., & Savelsbergh, G.J.P. (1989). Skill level, peripheral
vision and tennis volleying performance. Journal of Human Movement Studies, 16,
Davids, K., Savelsbergh, G., Bennett, S.J., & Van der Kamp, J. (2002). Interceptive actions in
sport: Theoretical perspectives and practical applications. In K. Davids, G Savelsbergh,
S.J. Bennett, & J. Van der Kamp (Eds.), Interceptive actions in sport: Information and
movement. New York: Routledge.
Duchowski, A.T. (2002). Eye tracking methodology: Theory and Practice, London:
■Ennis, C.D. (1994). Knowledge and beliefs underlying curricular expertise. Quest, 46,
Ericsson, K.A., & Kintsch, W. (1995). Long-term working memory. Psychological Review,
■Ericsson, K.A., Krampe, R.T., & Tesch-Römer, C. (1993). The role of deliberate practice
in the acquisition of expert performance. Psychological Review, 100, 363-406.
Ericsson, K.A., & Smith, J. (1991). Prospects and limits of the empirical study of expertise:
An introduction. In K. A. Ericsson & J. Smith (Eds.), Towards a general theory of
expertise: Prospects and limits (pp. 1–29). Cambridge: Cambridge University Press.
Fisher, R.A. (1925). Statistical methods for research workers. Edinburgh: Oliver & Boyd.
French, K.E., Spurgeon, J.H., & Nevett, M.E. (1995). Expert-novice differences in cognitive
and skill execution components of youth baseball performance. Research Quarterly
for Exercise and Sport, 66, 194-201.
French, K.E., & Thomas, J.R. (1987). The relation of knowledge development to children’s
basketball performance. Journal of Sport Psychology, 9, 15-32.
■Glass, G. (1976). Primary, secondary, and meta-analysis of research. Review of Research
in Education, 5, 3-8.
*Goulet, C., Bard, C., & Fleury, M. (1989). Expertise differences in preparing to return a
tennis serve: A visual information processing approach. Journal of Sport & Exercise
Psychology, 11, 382-398.
476 Mann, Williams, Ward, and Janelle
■*Goulet, C., Bard, C., & Fleury, M. (1992). Visual search strategies and information
processing in a racquet sport situation. In D. Brogan (Ed.), Visual search. London:
Taylor and Francis.
Hedges, L.V., & Olkin, I. (1985). Statistical methods for meta-analysis. Orlando, FL:
■*Helsen, W., & Pauwels, J.M. (1990). Analysis of visual search activity in solving tactical
game problems. In D. Brogan (Ed.), Visual Search: Proceedings of the First Interna-
tional Conference on Visual Search, University of Durham, England, September 5-9.
London: Taylor and Francis.
Helsen, W., & Starkes, J.L. (1999). A multidimensional approach to skilled perception and
performance in sport. Applied Cognitive Psychology, 13, 1-27.
■Hoffman, R.R. (1996). How can expertise be deﬁned? Implications of research from cogni-
tive psychology. In R. Williams, W. Faulkner, and J. Fleck (Eds.), Exploring Expertise
(pp. 81–100). Edinburgh: University of Edinburgh Press.
Hoffman, R.R., & Deffenbacher, K.A. (1993). An analysis of the relations of basic and
applied science. Ecological Psychology, 5, 315-352.
Holyoak, K. (1991). Symbolic connectionism: Toward third-generation theories of expertise.
In K.A. Ericsson & J.Smith (Eds.), Towards a general theory of expertise (pp. 301-336).
Cambridge: Cambridge University Press.
■*Houlston, D.R., & Lowes, R. (1993). Anticipatory cue-utilization process amongst expert
and non-expert wicketkeepers in cricket. International Journal of Sport Psychology,
Irwin, D.E. (1992). Visual memory within and across ﬁxations. In K. Rayner (Ed.), Eye
movements and visual cognition: Scene perception and reading (pp. 146-165). New
*Isaacs, L., & Finch, A. (1983). Anticipatory timing of beginning and intermediate tennis
players. Perceptual and Motor Skills, 57, 451-454.
Janelle, C.M., & Hillman, C.H. (2003). Expert performance in sport: Current perspective and
critical issues. In J.L. Starkes & K.A. Ericsson (Eds.), Expert Performance in sports:
Advances in research on sport expertise. Champaign, IL: Human Kinetics.
Janelle, C.M., Hillman, C.H., Apparies, R.J., Murray, N.P., Meili, L., Fallon, E.A., et al.
(2000). Expertise differences in cortical activation and gaze behavior during riﬂe shoot-
ing. Journal of Sport & Exercise Psychology, 22, 167-182.
*Jones, C., & Miles, J. (1978). Use of advance cues in predicting the ﬂight of a lawn tennis
ball. Journal of Human Movement Studies, 4, 231-235.
Marteniuk, R.G. (1976). Information processing in motor skills. New York: Holt, Rinehart,
McPherson, S.L. (1999). Tactical differences in problem representations and solutions in
collegiate varsity and beginner women tennis players. Research Quarterly for Exercise
and Sport, 70, 369-284.
McPherson, S.L. (2000). Expert-novice differences in planning strategies during collegiate
singles tennis competition. Journal of Sport & Exercise Psychology, 22, 39-62.
■*Moran, A., Byrne, A., & McGlade, N. (2002). The effects of anxiety and strategic plan-
ning on visual search behavior. Journal of Sports Sciences, 20, 225-237.
■*Paul, G., & Glencross, D. (1997). Expert perception and decision-making in baseball.
International Journal of Sport Psychology, 28, 35-56.
Petrakis, E. (1986). Visual observation patterns of tennis teachers. Research Quarterly for
Exercise and Sport, 57(3), 254-259.
Radlo, S.J., Janelle, C.M., Barba, D.A., & Frehlich, S.G. (2001). Perceptual decision-making
for baseball pitch recognition: Using P300 latency and amplitude to index attentional
processing. Research Quarterly for Exercise and Sport, 72(1), 22-31.
■*Renshaw, I. & Fairweather, M.M. (2000). Cricket bowling and the discrimination ability
of professional and amateur batters. Journal of Sports Sciences, 18, 951-957.
Visual Search and Expertise 477
■*Ripoll, H., Kerlirzin, Y., Stein, J.F., & Reine, B. (1995). Analysis of information process-
ing, decision-making, and visual search strategies in complex problem solving sport
situations. Human Movement Science, 14, 325-349.
Rosenthal, R. (1984). Meta-analytic procedures for social research. Beverly Hills, CA:
Rosenthal, R. (1991). Meta-analytic procedures for social research. Newbury Park, CA:
Rosenthal, R. (1995). Writing meta-analytic reviews. Psychological Bulletin, 118(2), 183-
Rosenthal, R. & DiMatteo, M.R. (2001). Meta-Analysis: Recent developments in quantitative
methods for literature reviews. Annual Review of Psychology, 52, 59-82.
■*Savelsbergh, G.J.P., Williams, A.M., Van Der Kamp, J., & Ward, P. (2002). Visual
search, anticipation, and expertise in soccer goalkeepers. Journal of Sport Sciences,
■*Shank, M.D. & Haywood, K.M. (1987). Eye movements while viewing a baseball pitch.
Perceptual and Motor Skills, 64, 1191-1197.
■Simon, H.A., & Chase, W.G. (1973). Skill in chess. American Scientist, 61, 394-403.
■Simonton, D.K. (1996). Creative expertise: A life-span developmental perspective. In
K.A. Ericsson (Ed.), The road to excellence: The acquisition of expert performance
in the arts and sciences, sports, and games (pp. 227-253). Mahwah, NJ: Lawrence
*Singer, R., Cauraugh, J., Chen, D., Steinberg, G.M., & Frehlich, S. G. (1996). Visual search,
anticipation, and reactive comparisons between highly-skilled and beginning tennis
players. Journal of Applied Sport Psychology, 8(1), 9-26.
■*Starkes, J.L., (1987). Skill in ﬁeld hockey. The nature of the cognitive advantage. Journal
of Sport Psychology, 9, 146-160.
Starkes, J.L. (1993). Motor experts: Opening thoughts. In J.L. Starkes & F. Allard (Eds.),
Cognitive issues in motor expertise (pp. 3-16). Amsterdam: Elsevier Science.
Starkes, J.L., & Allard, F. (Eds.). (1993). Cognitive issues in motor expertise. Amsterdam:
Starkes, J.L., Edwards, P., Dissanayake, P., & Dunn, T. (1995). A new technology and ﬁeld
test of advance cue usage in volleyball. Research Quarterly for Exercise and Sport,
■Starkes, J.L., Helsen, W., & Jack, R. (2001). Expert performance in sport and dance. In
R.N. Singer, H.A. Hausenblas, & C.M. Janelle (Eds.), Handbook of sport psychology
(2nd ed.) (pp. 174-201). New York: Wiley.
■Thomas, J.R., & French, K.E. (1986). The use of meta-analysis in exercise and sport: A
tutorial. Research Quarterly for Exercise and Sport, 57(3), 196-204.
Vickers, J.N. (1996). Control of visual attention during the basketball free throw. The
American Journal of Sports Medicine 24(6), S93-S96.
Vickers, J.N., & Adolphe, R.M. (1997). Gaze behaviour during a ball tracking and aiming
skill. International Journal of Sports Vision, 4, 18-27.
Ward, P., & Williams, A.M. (2003). Perceptual and cognitive skill development in soccer:
The multidimensional nature of expert performance. Journal of Sport and Exercise
Psychology, 25, 93-111.
■*Ward, P., Williams, A.M., & Bennett, S.J., (2002). Visual search and biological motion
perception in tennis. Research Quarterly for Exercise and Sport, 73(1), 107-112.
■Williams, A.M. (2000). Perceptual skills in soccer: Implications for talent identiﬁcation
and development. Journal of Sport Sciences, 18, 737-750.
Williams, A.M., & Burwitz, K. (1993). Advance cue utilization in soccer. In T. Reilly, J.
Clarys, & A. Stibbe (eds.). Science and Football, vol. III. London: E & FN Spon.
478 Mann, Williams, Ward, and Janelle
*Williams, A.M., & Davids, K. (1995). Declarative knowledge in sport: A by-product of
experience or a characteristic of expertise? Journal of Sport and Exercise Psychology,
■Williams, A.M., & Davids, K. (1997). Assessing cue usage in performance contexts: A
comparison between eye-movement and concurrent verbal report methods. Behavior
Research Methods, Instruments & Computers, 29(3), 364-375.
*Williams, A.M., & Davids, K. (1998). Visual search strategy, selective attentive, and exper-
tise in soccer. Research Quarterly for Exercise and Sport, 69, 111-128.
■Williams, A.M., Davids, K., & Burwtiz, L. (1994a). Ecological validity and visual search
research in sport. Journal of Sport and Exercise Psychology, S16, 22.
■Williams, A.M., 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. (1993). Visual search and sports
performance. The Australian Journal of Science and Medicine in Sport, 25(2), 147-
*Williams, A.M., Davids, K., Burwitz, L., & Williams, J.G. (1994b). Visual search strategies
in experienced and inexperienced soccer players. Research Quarterly for Exercise and
Sport, 6(2), 127-135.
Williams, A.M., Davids, K., & Williams, J.G. (1999). Visual Perception and Action in Sport.
London: E & FN Spon.
■*Williams, A.M., & Elliot, D. (1999). Anxiety, expertise and visual search strategy in
karate. Journal of Sport and Exercise Psychology, 21, 361-374
Williams, A.M., Singer, R.N., & Frehlich, S.G., (2002). Quiet eye duration, expertise, and task
complexity in near and far aiming tasks. Journal of Motor Behavior, 34(2), 197-207.
■Williams, A.M., Ward, P., Knowles, J.M., & Smeeton, N.J. (2002). Anticipation skill in a
real-world task: measurement, training, and transfer in tennis. Journal of Experimental
Psychology: Applied, 8(4), 259-270
■Williams, A.M., Ward, P., Smeeton, N.J., & Allen, D. (2004). Developing anticipation
skills in tennis using on-court instruction: Perception versus perception and action.
Journal of Applied Sport Psychology, 16(4), 350-360.
Wolf, F.M. (1986). Meta-analysis: Quantitative methods for research synthesis. Newbury
Park, CA: Sage.
*Wright, D.L., Pleasants, F., & Gomez-Meza, M. (1990). Use of advanced visual cue sources
in volleyball. Journal of Sport and Exercise Psychology, 12, 406-414.