Perceptual Encoding is Impervious to the Influence of Social Facilitation-Inhibition

topsalmonΤεχνίτη Νοημοσύνη και Ρομποτική

23 Φεβ 2014 (πριν από 3 χρόνια και 8 μήνες)

154 εμφανίσεις






The Inflexibility of Experts’
Perceptual Encoding is Impervious
to the Influence of Social
Facilitation
-
Inhibition




















Christopher Street



School of Psychology

University of Dundee

2009





This Dissertation is submitted in part fulfilmen
t of the requirements

For the Master of Arts Degree with Honours in Psychology.






2


































“Perception is where cognition and reality meet”

Neisser (1976, p9)



3





Table of Contents



i. Acknowledgements

................................
................................
................................
.....

4


1. Abstract

................................
................................
................................
.....................

5


2. Introduction

................................
................................
................................
................

6

1.1. Experts’ perception and performance

................................
................................
.

6

1.2. Physical Copresence

................................
................................
...........................

9

1.3. Beliefs H
eld About an Opponent

................................
................................
......

10

1.4. Present Study

................................
................................
................................
....

12


3. Methodology

................................
................................
................................
............

13

3.1. Participants

................................
................................
................................
........

13

3.2 Materials

................................
................................
................................
............

14

3.
3. Apparatus

................................
................................
................................
..........

14

3.3.1. Data recording

................................
................................
...........................

14

3.3.2. Flobopuyo

................................
................................
................................
..

15

3.3.2.1. Gameplay.

................................
................................
...........................

17

3.4. Procedure

................................
................................
................................
..........

17

3.5.
Design

................................
................................
................................
...............

18

3.5.1. Performance measures

................................
................................
...............

18

3.5.2. Self
-
reports

................................
................................
................................
.

19

3.5.3. Perceptual encoding measures

................................
................................
..

19

3.6. Statistical Analyses

................................
................................
...........................

20

3.6.1. Manipulation check: participant expertise

................................
................

20

3.6.3. Participants’ foveal allocation

................................
................................
..

21


4. Results

................................
................................
................................
......................

22

4.1. Manipulation Check: Partic
ipant Expertise

................................
......................

22

4.2. Beliefs regarding opponent ability

................................
................................
....

25

4.2.1. Self
-
report measures

................................
................................
..................

25

4.2.2. Performance measures

................................
................................
...............

27

4.3. Partici
pants’ Foveal Allocation

................................
................................
.........

29

4.4. Summary of Results

................................
................................
..........................

30


5. Discussion

................................
................................
................................
................

30

5.1. Experts’ Perceptual Encoding

................................
................................
...........

31

5.2. Beliefs Held About an

Opponent

................................
................................
......

33

5
.
3
. Conclusion

................................
................................
................................
........

38


6. References

................................
................................
................................
...............

39


7. Appendices

................................
................................
................................
..............

44



Word Count:


7,
479


4





i.

Acknowledgements


I would like to thank
Dr. Ben Tatler for his supervision throughout the past 2 years
and imparting his enthusiasm for the study of visual perception. This research could
not have been conducted without the voluntary participation of the novice and
experienced gameplayers who c
omprise the data set
, to whom I am also thankful
. I
am indebted to the patience and persistence of the confederate.
Finally,
I
would also
like to thank
all those who have provided me with access to their publications.



5





1. Abstract


This research aimed t
o discover whether the existance of inflexible perceptual
encoding mechanisms (e.g. Bilalić, McLeod & Gobet, 2008) has resulted from
studying experts in isolation of the social interactions in which they necessarily
engage whilst competing (French & McPher
son, 1999). The perceived ability of the
coactor was manipulated in an attempt to discover whether visual sampling strategies
could be rendered flexible when participants held a belief about their opponent (i.e.
social facilitation
-
inhibition, e.g. Muller

et al.
, 2004).


When n
ovice and experienced gameplayers (N = 12) competed against a a
human opponent
, they were given a cover story which led them to believe the
opponent w
as either (i) another naïve participant or (ii) a trained expert (between
subjects)
. Whilst the
cover story
was successful at modulating performance for
experienced participants

(novices performed at floor), overt perceptual encoding
strategies were not seen to vary. It was tentatively concluded that experts’ perceptual
encoding is res
istant to the influence of social facilitation
-
inhibition
, although no
conclusions were drawn about covert attention’s susceptibility to social facilitation
-
inhibition.





6





2. Introduction



Climbing the heights of expertise is considered an admirable attai
nment in our
society, and no more important is possessing expertise than when the aim is to
outperform others. Unsurprisingly then, there has been much interest in the
underlying differences between the novice and expert on competitive tasks,
particularly

in the domain of chess expertise.
Surprisingly, this field has been studied
separately under the sub
-
domains of social psychology and cognitive psychology.
This disection has resulted in the true sense of expertise (as a combination of
cognitive differe
nces and the social context in which it is engaged) being lost.

The present study will integrate expertise research from the these sub
-
domains
as well as outwith the field. By doing so, three key questions can be addressed: (i)
how do experts’ perception
and performance differ from their novice counterparts, (ii)
can the mere presence of a coactor cause differences in experts’ perceptual and
performance abilities, and (iii) does the belief held about the copresent other’s ability
affect an individual’s exp
ert perception and performance abilties
? Focus shall be cast
upon the last of these questions to demonstrate that even subtle manipulations of the
social environment can have a significant impact upon cognitive measures of
expertise.
This will highlight
the necessity for an interdisciplinary approach to
understanding human cognition in an ecological context.


1.1. Experts’ perception and performance

Years of research has demonstrated

that
experts
perceptually encode domain
-
specific information in fundamen
tally different ways

to novices
,
who fail

to encode
the cues that experts attend to (Klein & Hoffman, 199
2
). A smaller number of total


7





fixations are made by experts (Reingold, Charness, Pomplun & Stampe, 2001a)
demonstrating

that experts know what informa
tion is relevant to the task at hand and
can
filter out irrelevant cues.
Experts also produce

a shorter mean duration of a single
fixation (e.g. Chapman & Underwood, 1998; de Groot & Gobet, 1996, cited by Gobet,
1998)
1
, reflecting the time taken to cognit
ively process the incoming perceptual
stimulus (Rayner, 1998). Experts, then, efficiently process
perceptual stimuli (
Eccles,
2005;
see
also
Guo, Mahmoodi, Robert
son & Young, 2006; Irwin, 2004), as shown
in
a variety of domains from skilled
-
action domains

such as sports (Bahill & LaRitz,
1984, cited by Land, 2006;

Goulet, Bard & Fl
eury, 1989; Land & McLeod, 2000
) to
cognitively engaging
domains such as chess (Charness
,

Reingold, Pomplun &
Stampe,

2001; Chase & Simon, 1973; de Groot, 1946
/1978
) and video
-
ga
ming
(Jermann, Nüssli & Dillenbourg, 2008; Underwood, 2005).

E
xperts form long
-
term representations of domain
-
specific information
from
past experiences
. These long
-
term traces create
schemas

which are automatically
activated
when
similarities between the

current lay of the board

(for example, in
chess)

and past experiences
are perceived

by experts

(e.g. Chase & Simon, 1973;

Lane, Cheng & Gobet, 1999).

Over many experiences of gameplay, experts learn to
delineate schemas

with finer detail
.
That is, as pl
ayers learn to understand the finer
points of the game, they learn
also how to combine the finer points of

their

strategies
in
different
ways in order to
continue to
succeed.
Thus
with greater experience

players
will be more flexible (Ca
mpitelli & Gobet, 2004)
in the cues that they attend



1

Participants needn’t be compared on the extreme dimensions of ability to demonstrate perceptual
encoding differences. Studies examinin
g differences between novice and experienced (but not expert)
participants have shown similar findings (
Van Gog, Paas & Van Merriënboer, 2005
).



8





to and

will
become more
efficien
t

in processing them
(e.g. Chase & Simon, 1973; de
Groot, 1946
/1978
;
Erics
son & Lehman, 1996).

However, another line of research
claims that

experts’
automatic activation of
schemas

results

in

inflexibility
. The Einstellung effect (Luchin, 1942, cited by Bilalić,
McLeod & Gobet, 2008a) occurs when individuals are unable to choose a more
optimal solution over a (non
-
optimal) familiar solution,

which has been
taken as
evidence of
an inability to be flexible
(even when it is
required by the task
) to
find
novel solutions

(see Bilalić
et al.
,
2008a;
Bilalić, McLeod & Gobet,

2008b;

Saariluoma, 1990).

According to schema theories of expert
ise (e.g. Gobet & Simon, 1996
)
,
superior performance ability

results from
the
perceptual encoding

of stimuli
. Bilalić
et al
. (2008b) used eye movement data to demonstrate that the Einstellung effect is
reflected by experts attending to
cues indicative of
a frequently occuring (and
sometimes suboptimal)
solution rather than
attending to the

full range of potential
cues available
to
deduce the
optimal solution. Thus experts demonstrate inflexibility
in
the selection of visual stimuli and
their perceptual en
coding
.

It is perhaps due to this

proposed inflexibility that
expertise
research thus far
has comprised of studying participants in isolation without recourse to acknowledging
a role for the opponent (French & McPherson, 1999; although see Marko
vitch &
Reger, 200
5
). And yet it is commonly believed that less task
-
relevant cues are an
important factor in determining one’s own performance (see Campitelli & Gobet,
2004 for an account of expert performance as becoming more sensitive to more
periphera
l cues). For example,
holding a belief that an opponent’s skill level is poor
can actually result in the opponent performing poorly
, behaviourally confirming the
belief
held by the individual
(see Snyder,

Tanke & Berscheid,

1977; Snyder & Swann,


9





1978). I
ndeed the same is
believed to be
true of ‘off
-
line’ games such as chess (Golz
& Keres, 1989, cited by Saariluoma, 1995).

Such accounts show how performance demonstrations are open to flexibility.
It also appears that perceptual processes can

be modulated
by such cues as the
presence of an
other person

(e.g. Bangerter, 2004; Bayliss,

Frischen, Fenske & Tipper,
2007
;

Brennan, Chen, Dickinson, Neider & Zelinsky, 2008
;

Crosby, Monin &
Richardson, 2008;

Frischen, Bayliss & Tipper, 2007;
Kingstone, Smilek, Friese
n &
Eastwood, 2003;

Land & Tatler, 2009;
Malcolmson, Reynolds & Smilek, 2007
;
Richardson, Dale, Tomlinson Jr. & Clark, 2008
). However this is yet to be
incorporated into expertise research to

examine
the flexibility of perceptual
processing. This has bee
n the motivation of the current research.


1.
2
. Physical Copresence

Many domains of expertise require individuals to collaborate with or compete
against another in each other’s presence, but this has been neglected
by research in
cognitive psychology
. The

more natural situation of physical collocation has been
examined in the human
-
computer interaction literature. It has been demonstrated that
when
coacting with
a human interactant there is a greater awareness of that person
(Gerhard, Moore & Hobbs, 2005)

and his/her
decision processes as arising from
conscious choices
(Gallagher, Jack, Roepstorff & Frith, 2002) along with stronger
physiological arousal and attention (
Ravaja
et al.
, 2006
)
in comparison to competing
against another who is not physically pre
sent (a virtual opponent). This has been

corroborated by functional imaging brain scans (
Rilling, Sanfey, Aronson, Nystrom &
Cohen, 2004
). Thus there are real differences in competing under copresent
conditions.



10





Copresent conditions have also been examin
ed in social psychological
research.
Zajonc (1965)
proposed
that
dominant
responses
(i.e. the response that have
the highest probability of being produced)
are facilitated by the presence of another
person
whilst learning is inhibited

(
compared to perform
ing alone
). When an
individual engages in a novel task, the generated responses tend to result in poor
performance. Therefore, novices will perform poorly under copresent conditions due
to facilitation of the dominant (incorrect) response. On the other h
and, experts who
have practiced in the domain have a well
-
learnt dominant response that is optimal for
the task.
Thus being experienced in a complex task
-
domain will result in the
practiced (and correct)
dominant
response(s) when coacting, facilitating pe
rformance.

This effect of facilitation of performance for experts and inhibition for novices has
been dubbed
social facilitation
-
inhibition

(SFI).

Because
expert
performance relies upon an individual’s perceptual encoding
strategy (e.g.
Chapman & Underwood, 1998; de Groot & Gobet, 1996, cited by
Charness
et al
., 2001
; Reingold
et al
., 2001a), one may suspect that

SFI

effects will be
roote
d in the duration

and distribution

of visual
-
attentional focus to task
-
relevant cues.

Thus
there is the potential for
perceptual encoding strategies
to
show a degree of
flexibility in as much as they may be modulated by mere presence

effects
demonstrated

in performance situations
.


1.
3
.
Beliefs Held About an Opponent

Cottrell (1972)
, however,

has argued that mere presence is not sufficient to
explain SFI effects. For example, Henchy & Glass (1968, cited by Amabile, Goldfarb
& Brackfield, 1984) found tha
t being observed by an audience who were believed to
be experts led to different levels of competent performance compared to being


11





observed by a non
-
expert audience
, showing either greater or poorer performance
depending upon the individual’s ability
. The
se differences
have been found time and
again
in reviews (Geen, 1991) and meta
-
analyses (Bond & Titus, 1983; Guerin, 1986,
cited by Geen, 1991).

Cottrell proposed an
evaluation apprehension

(EA)
theory to explain

SFI
effects
. He argue
d that being observed

by another who has the ability to evaluate
one’s performance (i.e. an expert)

leads to fear of being eval
uated and appearing
incompetent. This then somehow

results in the observed SFI performance effects
.

Two accounts have been provided for the intermed
iary step through which EA results
in SFI effects.

Zajonc’s (1965) response
-
dominance account has remained active as a
p
otential mechanism
. Evaluation by an expert opponent results in dominant response
facilitation, which then results in the observed SFI
effects. However, a different
mechanism has also been proposed.
As noted earlier, perceptual encoding
mechanisms
may be

flexible (cf. Campitelli & Gobet, 2004) and
could
potentially be
influenc
ed by the social context
. Because performance is associated with perceptual
encoding, EA may modulate performance by initially impacting upon perceptual
encoding

mechanisms
. Support for this notion has been provided by Baron (1986,
cited
by Geen, 1991)

i
n
a review of SFI effects
, who concluded that EA narrows
visual attention to central task
-
relevant stimuli and inhibits attention to more
peripheral stimuli
, resulting in SFI performance effects
(cf. Bruning
et al.
, 1968, cited
by Huguet, Galvaing, Monteil

& Dumas, 1999;

Easterbrook, 1959; Geen, 1980).

Similar to Zajonc’s response
-
dominance account, Baron’s
distraction
-
conflict

theory
claims that EA’s effect on

SFI
is modulated by
task
difficulty
. The theory
defines a
complex task
as
one that requires atte
nding to a greater number of cues in


12





order to perform well. Thus on complex tasks, EA inhibits attention to the greater
number of cues that are required for successful performance and results in relatively
poor performance. Provided support for such an a
ccount, one would conclude that
perceptual encoding
is flexible in that it can be
modulated by the

beliefs held about
another
.

Muller, Atzeni and Butera (2
004) have demonstrated that EA
can indeed have
an impact upon visual perception. Participants were g
iven 70 ms to process an array
of stimuli that included a set of distractors (S’s, vertical lines and horizontal lines) and
a single target ($). A response consisted of reporting whether the target was present
or absent.
An illusory perception

of a targe
t
-
present trial occurred as a result of
binding a set of visual primitives (S’s and vertical lines, see Muller
et al
., 2004 for
detail) even though the target ‘$’ was absent. When EA was induced by the presence
of
another who was believed to be more
accu
rate
than the participant
, attention was
reduced from peripheral (distractor) cues and delegated more st
rongly to central
(target) cues resulting in
more accurate
performance.

This study was taken to support the claim that EA operates via the
percep
tual
encoding
mechanism proposed by Baron (1986, cited by Geen, 1991). Note that

Muller
et al
. (2004)

did not test whether
the individual’s
expertise
could manipulate
the effect of EA. This is
because the
distraction
-
conflict theory
(Baron, 1986, cited b
y
Geen, 1991)
does not make any explicit claims that EA should have differential
effects as a function of
an individual’s own
ability.


1.
4. Present Study

This study aims to answer three key questions. First,
do
experts perform better
than novices be
cause of the differential perceptual encoding strategies employed
?


13





The first experimental hypothesis states that experts will demonstrate shorter mean
fixation durations (Chapman & Underwood, 1998) with a smaller number of total
fixations (Reingold
et
al
., 2001a), although will attend to a greater number of task
-
relevant cues (Klein & Hoffman, 1992).

Second, it is aimed to show that expert performance is not merely a cognitive
issue, but a social one, and that the mere presence of another person can inf
luence
expert ability. Time limitations did not permit an examination of this. Focus shall
instead be cast upon the third aim of the present study. This final but primary aim
was to determine whether more subtle differences in an opponent’s presence can

result in different perceptual encoding strategies. The second
experimental hypothesis
predicts
a greater total fixation duration and/or a greater number of total fixations to a
smaller range of ROIs (i.e. an interaction) when participants believe the op
ponent to
be an expert compared to when they believe the opponent to be a naïve participant

(
Baron, 1986, cited by Geen, 1991).


3. Methodology

3.1. Participants

15 participants with normal vision were recruited
on an opportunistic basis
from the
Universit
y of Dundee who
voluntarily
took part.
The highest scoring participant

from

the novice and experienced groups received a £10 gift voucher. Participants were
excluded: (a) if he/she
provided a self
-
rating of

an
average ability

(N=2)
;

and

(b) if
he/she had

a prior relationship with the confederate (N =1). This left 12 participants
remaining for further analysis.



14





A male confederate was trained for
2 hours 50 minutes on the video
game
employed herein and competed against every participant
.

He
was blind to th
e aims of
the study, the

self
-
rated

ability of the participant
and
the
cover story
that participants
were given about the confederate’s ability.


3.2 Materials

A preliminary questionnaire assessed prior experience with games similar to Puyo
Pop (appendix 1
.
1
). Two follow
-
up questionnaires
measured
participants’ experience
of the
study (m
easures
of interest are listed in section 3.5.2
). A likert scale from 1

(total disagreement)

to 10

(total agreement) allowed participants to respond to
statements that ref
erred to the above measures
. These questionnaire
s were designed
for this study
and have not been tested for validity, reliability or internal consistency.
Prior to gameplay, participants were provided with
gameplay
instructions (appendix
2)
.


3.3. Appara
tus

3.3.1. Data recording


An SR Research Ltd. Eye
-
Link II eye tracker recorded eye movement data at 500Hz.
A calibration of eye positions was run using a 9
-
point grid

and

validated with another
9
-
point grid. Error in spa
tial accuracy of greater than 0.
5
o

resulted in a recalibration
of eye positioning. Recordings were taken from the eye which returned the greatest
spatial accuracy during calibration. The SR Research algorithm was used to define
saccades and fixations.



15





A video camcorder record
ed

compute
r
-
monitor activity during gameplay

to
allow subsequent analysis of performance.



3.3.2. Flobopuyo


Flobopuyo is
an easily
accessible

(due to a less complex aesthetical design),

zero
-
sum
competitive game where

the aim
is to cause
the opponent’s
third colum
n (from the
left) of a
6
-
column x 13
-
row screen to fill
to the top with puyos
(for a screenshot of
the game, see figure
1, overleaf
)
. Player
-
opponent competition was encouraged in the

instructions (and repeating verbally) by highlighting
the importance o
f forcing the
opponent to lose.

T
he computer opponent
was set to a medium difficulty.
A 21”
monitor displayed the game presented at a viewing distance of 60cm, subtending 30
o

x 40
o

of the visual field.

The computer keyboard was used as the input device
(see
appendix 2).


3.3.2.1. Gameplay.

Puyos, gelatinous spheres, fall from the top of the screen
in randomly coloured pairs until they contact the bottom of the screen
(
or previously
stacked

puyos
)

at a rate of
one row second
-
1

and can be
increased to
twe
nty

rows
second
-
1
.

There are 5 coloured puyos in total. Players must form groups of four or more
same
-
coloured puyos adjacently by rotating or translating the descending pair prior to
its contacting the ground. Upon doing so, the group disappears and gar
bage puyos
fall into the opponent’s screen. To win, players must prevent
the third column of
their
screen from filling entirely with puyos whilst simultaneously causing their opponents
screen to fill with (garbage) puyos. Thus central cues in this study
were
cautiously



16





Figure
1
. 14 regions of interest
. Quadrant divides of player and opponent areas were
employed as per Underwood (2005). Names of each player were not coded as a reg
ion of
interest.



(1) Area from which garbage is stacked prior to falling into opponent’s screen (
Opp
Garbage
)



(2) Upper
-
left (
Opp UL
), (3) upper
-
right (
Opp UR
)
, (4) lower
-
left (
Opp LL
) and (5)
lower
-
right (
Opp LR
) quadrants of opponent’s screen



(6) Opponen
t’s next puyo pair that will descend following the current puyo pair already in
play (
Opp Next
)



(7) Animation that pops up if either participant or opponent has stacked puyos to a height
of 3 puyos or less from the top of the screen in any column (
Animatio
n
)



(8) Scores of opponent and participant plus participants’ number of lives remaining
(
Score
)



(9) Participants’ next puyo information (
Own Next
)



(10) Participants’ garbage area (
Own Garbage
)



(11) Upper
-
left (
Own

UL
), (12) upper
-
right (
Own UR
)
, (13) lower
-
left (
Own LL
) and (14)
lower
-
right (
Own LR
) quadrants of participants’ screen


(
2
)


(
3
)


(
4
)


(
5
)


(
13
)


(
14)


(
12
)


(
11
)


(1)

(1
0
)

(
6
)

(
8
)

(
7
)

(
9
)



17





defined as those which pertained to the participants’ placement of puyo pairs in
his/her screen quadrants.

Creating
a group

of more than four puyos or a combo results in more garbage
dropping into the opponent’s screen
, as explained in the task instructions
. A combo
can be described as

creating tw
o or more puyo groups simultaneously (i.e. by
strategically placing a single puyo pair)
. These are more advanced strategies and
may
require attendance
to more peripheral cues, which may include

the
Own Next

region
(figure 1)
that
provides information to a
id decisions about future
potential actions.


3.4. Procedure

Participants were introduced to the confederate by name, ensuring no other
information exchanged between the two. The confederate was sat on
the
opposite
side of the room
to the participant
sepa
rated by a divide whilst
the
participant and
confederate read the information sheet, gave informed consent and completed the
preliminary questionnaire.
Participants’ self
-
rated ability
as above or below average
was used
to assign participants to the exper
ienced and novice conditions, respectively
.

Upon completion, both players were sat before the computer and given the
game instructions. For half of the novice (N = 3) and experienced (N = 3)
participants, the instructions highlighted that the human

oppone
nt was a trained
expert
. The remaining participants were led to believe that the opponent
was a fellow
naïve participant
. The confederate did not see the instructions. The
participants’
aim
was to “defeat your opponent in a battle by filling their grid
to the top with
garbage

and the instructions further stated that “this is the only way in which the game can be
won”.



18





After reading the instructions both players were given the opportunity to ask
questions before the eye
tracker

was

set up
. The game
was
played

for fifteen minu
tes
whilst the experimenter
sat behind the divide, out of view of the screen. A second
calibration corrected for drift at

the end of the 15
-
minute block after which
participants completed a questionnaire about their gaming experienc
e
. The second
block began with the setting up of the eye recording equipment followed by 15
minutes of competing against a computerised opponent set

to a medium difficulty
with

a calibration
at the start
and
end of
the block. The order of the human and
c
omputer opponent was counterbalanced between participants.


3.5. Design

A 2 (participant ability: experienced or novice; within subjects) x 2 (opponent’s
perceived ability:

naïve participant


or

trained expert

; between subject) mixed
design was used. T
he independent variables were participants’ ability and the
participants’ perceived ability of their opponent. The dependent varia
bles (DVs) were
of three sorts: performance measures, self
-
reports and perceptual encoding measures.


3.5.1. Performance meas
ures


Five

performance m
easures were taken as DVs . These measures were:

(i)

The proportion of levels won;

(ii)

The proportion of puyo pairs that fell
that created

a
puyo
group

(
proportion of
groups formed
)
;

(iii)

The proportion of groups formed that were combo
s

(see se
ction 3.3
.
proportion of combos formed
);



19





(iv)

The average size of a group formed
;

and

(v)

The proportion of garbage in the participants’ screen
that was cleared from the
playing field (
proportion of garbage cleared
)
.


3.5.2. Self
-
reports


The self
-
report measures
examined the participants’ perception of their
gaming experience. The four measures were:

(i)

Perception of opponent’s performance (irrespective of the participants’ own
performance);

(ii)

Perception of opponent’s performance in relation to the participants’
perfo
rmance;

(iii)

Perception of own performance in relation to how he/she had expected to
perform; and

(iv)

The level of participants’ effort during gameplay.


3.5.3. Perceptual encoding measures


Finally, three DVs measured perceptual encoding:

(i)

T
he mean fixation duratio
ns in each region of interest (ROI: see figure 1)
;

(ii)

The total fixation durations in each ROI; and

(iii)

T
he total number of fixations made in each ROI.




20





3.6. Statistical Analyses


3.6.1. Manipulation check
: participant expertise


A manipulation check was carried
out in order to assess (a) whether or not
participants could be categorised as novice or experienced throughout the duration of
the study, and (b) whether or not the self
-
ratings of ability matched their objective
performance.


A one
-
tailed 2 (participan
t’s self
-
rated ability: novice or experienced; between
subjects) x 3 (time of self
-
rating: preliminary, post
-
first block and post
-
second block;
within subjects) mixed analysis of variance (ANOVA) was conducted to carry out the
first assessment. The second

assessment was carried out by conducting a one
-
tailed
one
-
way between subjects ANOVA comparing self
-
rated novice and experienced
participants on the five measures of performance (section 3.5.1).


3.6.
2. Beliefs regarding opponent ability


T
he perceived a
bility of the opponent (as an expert or as another naïve
participant)
was manipulated in an attempt
to induce EA. To determine that the
instruction successfully manipulated the participants’ perceptions

of the opponent
, a
set of two
-
tailed 2 (participant
ability: novice and experienced) x 2 (opponent ability:
naïve and expert) between subjects ANOVA
s were

conducted on
self
-
report
measures
(section 3.5.2).

These perceptions of the opponent will be compared to performance
under these conditions to determine

whether
the different cover stories causes
differences in performance
.




21





3.6.
3
.
Participants’ foveal allocation


This research aim
ed

to discover whether differential beliefs about the opponent can
a
ffect
perceptual encodin
. The computer screen
was
divided

into 14 regions of
interest (ROI: figure 1) to aid analysis. A
two
-
tailed
2 (participant’s ability: novice
and experienced; be
tween subjects) x 2 (perceived opponent
ability:

naïve


and

expert

; between subjects) x 14 (ROI: figure 1; within subjects) m
ixed ANOVA
was
conducted to examine whether mean fixation durations show
ed

significant differences
as a result of the
participants’ belief

regarding their opponent
’s ability
. The same
analysis will be conducted on total fixation durations and total number

of fixations.

Zero values for mean fixation durations were substituted for the mean score
within the ROI being examined. This was to ensure that the mean fixation durations
were not skewed by a lack of fixating the region. The question that was of inter
est
here was whether the duration of the mean fixation was different between novices and
experts when participants fixated that region. Analysis of these results should be
interpreted with care, given that the mean fixation duration substitution will skew

the
results. However, it was believed that such a skew was more valid for analysis for the
above reason.


This transformation was not carried out on the other two measures. The total
fixation duration measure was analysed to give a measure of how long t
he ROI was
fixated over the course of the study, and
so
not fixating the region at all was an
equally valid recording. The same is true of the total number of fixations: the purpose
of this measure was to determine how often the ROI was fixated, and so a
recording
of zero fixations is a valid one for analysis.




22





4. Results

First, the manipulation check

was
carried out, followed by inferential tests to
disconfirm the null hypothesis. Homogeneity of variance will be assumed unless
otherwise stated.


4.1. Man
ipulation
C
heck
: Participant Expertise

Participants rated their ability on Puyo Pop (and/or Tetris, depending on which, if any,
they have played) on a scale of 0 to 10 with a higher score representing greater ability.
On this basis the experienced and nov
ice conditions were preliminarily defined
2
.
Perceptions of their own ability remained different between the two groups over the
course of the experiment, as shown in figure 2.





2

Note that, unlike chess, there is no objective standard of ability (such as
Elo

points in chess: Elo,
1978, ci
ted by Campitelli, Gobet, Williams & Parker, 2007). Thus there can be no claim that the
participants are true novices and experts; only that their abilities may differ significantly in the
directions of novice and expert ability. For ease of reading, tho
se tending towards novice ability shall
be referred to as novices, whilst those tending toward the other extreme shall be called experienced.



23









The above figure demonstrates that on all three occasions, novice partic
ipants
(mean = 2.87, SD = .49) rated themselves below 5 (midpoint) whilst experienced
players (mean = 7.50, SD = .49) rated themselves above 5 on all three occasions. A
one
-
tailed 2

(participant ability)

x 3
(time of rating) mixed ANOVA was conducted
3
.
N
ovices rated themselves as lower in ability compared to experienced players (F (1,
10) = 45.13, p < .001, η
p
2

= .777).


Performance measures tell a similar story. Means and standard deviations of
these measures are reported in the table below.











3

Detailed information about
the levels within each variable used for

inferential analyses can be found
in sectio
n 3.6

Figure
2
.

Novice and experienced

players’
self
-
ratings of ability over the course of the
experiment



24







Experienced players show higher mean scores than novices on all measures over the
course of the experiment. A one
-
tailed one
-
wa
y between
-
subjects ANOVA revealed

that the proportion of levels won

(F (1, 10) = 6.97, p = .013), groups formed (F (1, 10)
= 6.56, p = .014) and garbage cleared (F (1, 10) = 17.35, p = .001) were significantly
greater in the experienced condition. The proportion of combos formed approached
significance with experienced p
articipants creating more combos than novices (F (1,
10) = 3.07, p = .055), whilst the average size of a puyo group did not differ
significantly (F (1, 10) = 1.00, p = .341). It is important to note that novices
demonstrate floor performance, which could
potentially contaminate
any
interpretations
. Nonetheless, it appears that the expertise manipulation was
successful.


Participant’s ability


Novice

Experienced


Mean

SD

Mean

SD

Proportion of
levels won

.05

.09

.52

.42

Proportion of
groups formed

.26

.08

.35

.05

Proportion of
combos formed

.08

.05

.14

.07

Average size of
group formed

4.20

.07

4.25

.09

Proportion of
garbage cleared

.16

.03

.23

.
03

Table 1.

Mean performance measures for novices and experts and standard deviations



25





4.2. Beliefs regarding opponent ability

4.2.1. Self
-
report measures


It was necessary to exami
ne self reports to discover if
individuals w
ere of the opinion
that the opponent was more competent in the
opponent
-
as
-
expert condition

by
examining four self
-
report measures
.

The descriptive statistics can be found in table
2.

First,
novice participants believed the ‘expert’ opponent was more capa
ble
than the ‘naïve’ opponent (measure 1, table 2). The experienced participants show the
opposite trend: the ‘naïve’ opponent was perceived as
more
competent than the
‘expert’, demonstrating an interaction
(F (1, 8) = 9.39, p = .015, η
p
2

= .540).

(measure
2, table 2)
.

Second, novice participants perceived the ‘expert’ opponent as more
competent than themselves (compared to the ‘naïve’ opponent) whilst experienced
participants believed the ‘expert’ opponent

was less competent than themselves in
relation to the ‘naïve’ opponent (F (1, 8) = 6.92, p = .030, η
p
2

= .464).
s


Third, novices believed they performed worse than they had expected against
the perceived ‘expert’ opponent, more so than when they had compe
ted against the
perceived ‘naïve’ opponent (measure 3, table 2). Experienced participants show a
reverse pattern: these participants believed they performed worse than they expected
against the ‘naïve’ opponent, more so than against the ‘expert’ opponent.

This
interaction was significant (F (1, 8) = 4.90, p = .058, η
p
2
= .380
).

A final significant interaction was observed (F (1, 8) = 14.22, p = .005):
novice participants competed more seriously against the ‘naïve’ opponent than against
the ‘expert’ oppone
nt, whilst experienced players did not compete any more or less
strongly as a result of the cover story (table 2, measure 4).


26






Table
2
.
Questionnaire
measures (means and standard deviations) for each participant ability as a function of perceived opponent ability
.
(1 = totally disagree, 5 =
neutral,

10 = totally agree).

Novice Participant

Experienced Participant


Naïve Opponent

Expert opponent

Naïve opponent

Expert opponent


Mean

SD

Mean

SD

Mea
n

SD

Mean

SD

1.
I felt that my opponent
performed poorly

2.33

.58

1.00

.00

1.67

1.16

4.67

2.08

2.

I felt that my opponent
performed better than me

8.33

.58

9.67

.58

7.67

2.52

3.67

2.31

3.
I performed worse than I
expected

3.67

1.58

7.00

1.58

7.33

1.
58

3.67

1.58

4.
I competed very seriously

6.33

.71

2.00

.71

6.67

.71

7.67

.71



27





4.2.2. Performance measures

A two
-
tailed

2 (participant ability
) x 2 (opponent ability
)

x 5

(performance measures)

mixed ANOVA was conducted. The three
-
way interaction reached significance (F (4,
32) = 3.46, p = .019, η
p
2

= .302). A series of 2 (participant ability) x 2 (opponent
ability)
between
-
subjects
ANOVAs will examine each measure of performance in
turn to di
scover whether any performance measures produced significant interactions.

As can be seen from table 3 (overleaf), experienced participants won a greater
proportion of levels against the

expert


opponent than against the

naïve


opponent (F
(1, 8) = 6.91,

p = .030, η
p
2

= .464). Novices, meanwhile, demonstrated floor
performance against
the opponent irrespective of the cover story provided
. This
interaction approached significance (F (1, 8) = 3.79, p = .087). Secondly, the average
size of a group that wa
s created by experienced players was larger against the

expert


than the

naïve


opponent, a difference that was marginally significant (F (1, 8) =
5.13, p = .053, η
p
2

= .391), whilst again the novices demonstrated no significant
difference. There was no

significant interaction (F (1, 8) = .239, p = .638). Finally,
the proportion of garbage cleared by
both
the experienced player
s and the novice
players

was not statistically significant
when comparing the
opponent cover stories

(F
(1, 8) = 3.76, p = .089,

η
p
2

= .320).

From the significant main effects of the perceived
opponent’s
ability on
participants’
performance, it was concluded that
participants’ beliefs did impact upon
performance
. However, it was not possible to draw any conclusions about how this
i
nteracted with the participants’ ability
because novices demonstrated

floor
performance which contaminates any potential
interpretations
.




28





Table 3.
Performance measures (means and standard deviations) for each participant ability as a function of

perceived opponent ability
.


Novice Participant

Experienced Participant


Naïve Opponent

Expert opponent

Naïve opponent

Expert opponent


Mean

S
D

Mean

SD

Mean

SD

Mean

SD

1. Proportion of levels won

.07

.14

.04

.14

.25

.14

.78

.14

2. Proportion of groups
formed

.25

.04

.26

.04

.35

.04

.36

.04

3. Average group size

4.16

.04

4.24

.04

4.19

.04

4.31

.04

4. Proportion of garbage
cleared

.17

.01

.16

.01

.25

.01

.21

.01

5. Proportion of combos
formed

.08

.04

.09

.04

.13

.04

.16

.04



29





4.
3
. Participants’ Foveal Al
location

To determine whether novice and experienced participants allocated their attention
differently when competing against differently perceived opponents, a two
-
tailed 2
(participant ability) x 2 (opponent ability) x 14 (
ROI
: figure 1) ANOVA was
perfo
rmed on:

(i)

The mean fixation duration within each ROI (with missing cases
substituted for th
e series mean; section 3.6.3)

(ii)

The total fixation duration within each ROI
;

and

(iii)

The total number of fixations within each ROI.

Analyses
showed
no three
-
way interaction
s for
all three
measures. ROI
demonstrated a significant main effect for each measure

[
mean fixation duration (F
(13, 104) =

21.48, p < .001, η
p
2

= .772); number of fixations (F (13, 91) = 28.61, p <
.001, η
p
2

= .851); total fixation duration (F (13, 91) = 23.75, p <.001, η
p
2

= .82], and
the mean fixation duration within
ROIs interacted with
participant ability (F (13, 78)
= 1.
89, p = .044, η
p
2

= .149). However,
there were no
significant
effects as a result of
th
e cover story
.


The mean duration of

novices’ fixation
s

on
Own UL

(see figure 1) was
significantly longer compared to experienced participants (F (1, 6) = 11.02, p = .0
16,
η
p
2

= .647). Similarly, the mean duration of a fixation to
Own

Garbage

was longer
for novice compared to experienced players, approaching significance (F (1, 6) =
4.81, p = .071, η
p
2

= .445).




30





4.
4
. Summary

of Results

N
ovice participants show equally p
oor performance on all measures

(section
4.2.2).


T
his is reflected in their perception of their own performance

(section 4.2.1),
although t
hese participants
do
report competing less strongly against the
the
opponent

when the cover story claims the opponen
t is an expert compared to a cover story of
the opponent as a naïve participant
.


Experienced players demonstrate greater performance against the

expert


opponent

(section 4.2.2)
, with their perceptions of
the
opponent as easier to compete
against than th
ey had initially expected

when given this cover story (section 4.2.1)
.
Experienced

participants competed equally seriously against
the opponent irrespective
of the cover story
.

Thus the potential confound of withdrawing effort against the
‘expert’ oppone
nt was not present.


For all participants, the belief held about the opponent was not sufficient to
modulate their fixation durations or number of refixations to ROIs (section 4.3).


5. Discussion

This research aimed to
replicate past findings demonstratin
g that novice and
experienced participants (as defined by their performance) deploy their visual
attention differently. It was further aimed to determine whether or not
game
-
players
would demonstrate different patterns of visual fixation
(i.e. show flexib
ility in
perceptual encoding)
when
given a cover story about the skill level of their opponent.




31





5.1. Experts’ Perceptual Encoding

The present study predicted that experts would demonstrate shorter mean
fixation durations (e.g. Chapman & Underwood, 1998),
reflecting efficiency of
perceptual processing (Irwin, 2004) and was found in this study. Interestingly, such
differences have not always been found (Charness
et al
., 2001).

An
attempt to
reconcile these differences

will be provided in relation to the pr
esent study
by
claiming that experts’ perceptual encoding is flexible
.

E
xperienced participants only produced
relatively
shorter fixat
ion durations
when fixating
Own Garbage
and
Own UL

areas.
On Flobopuyo,
a pair of randomly
coloured
puyo
s

appear in the
O
wn Garbage

area and descend into
Own UL
. It is not
possible to see both puyos until
at least
one of the
puyos

has entered
Own UL
, even if
the pair is rotated 90
o

to obtain this information more quickly
(see Kirsh & Maglio,
1994). These ROIs therefore oft
en contain
a cue of central importance

for successful
task completion,
namely the puyo pair in play
.

Colour information about the pair must be remembered whilst the participant
scans the playing field to decide where the puyos will be placed, or vice versa
. The
decision must made by integrating the incoming stimuli with information held in
working memory. Shorter fixation durations may reflect a faster or more efficient
integration of the two in order to make a decision of placement, a task that is
perfor
med by the central executive (CE) (Robbins
et al
., 1996, study 2).

It is proposed that whilst experienced players do not ordinarily produce shorter
fixation durations than novices (as demonstrated in the participant
-
paced study by
Charness
et al
., 200
1
) th
ey are capable of doing so
when conditions become
unpredictable (i.e.

a lack of control

over game pace
)
. When the participant’s screen is
stacked high with puyos, the decision time of puyo
-
pair placement is relatively short.


32





Under these conditions, faste
r processing is necessary to match the colour of one of
the puyos in the pair with a colour in the top row of the stack before the decision of
puyo
-
placement can be made.

Faster processing is likely to incur some cost, cognitive and/or metabolic, and
so ra
pid processing engages in a trade
-
off. It would thus be interesting to compare the
durations of fixations when the participant’s screen is stacked high to when there are
relatively few puyos stacked in the screen to determine if fixation duration is a
fun
ction of alloted decision time
. The same argument can be applied to Chapman &
Underwood’s (1998) study which examined novice and expert drivers on roads with
many or few potential dangers

(rural and urban roads, respectively). The potential
dangers can b
e defined as those aspects of the task that are not self
-
paced.
Roads with
greater potential danger resulted in
shorter fixation durations, consistent with

more
rapid decisions (i.e. categorising the stimulus as dangerous or safe) when the situation
deman
ds
.

Thus it appears that there is some flexibility in the perceptual processing by
experts in as much as some (situation
-
paced) tasks result in shorter fixation durations
whilst other (self
-
paced) tasks do not. It may be fruitful for future research to
examine the duration of experts’ fixations as a function of
the locus of

control over
pace.

In the present study, it was also predicted that a greater number of fixations to
a wider range of ROIs would be demonstrated by experienced compared to novice
play
ers. The prediction was not supported. This may be a result of task complexity.
The task employed herein was complex, i.e. requiring attention to a number of cues to
perform well. On this basis, the prediction was made. However, Reingold
et al
.
(2001a
) have shown that the perceptual encoding advantages of experts are a result of


33





single fixations rather than their numerosity. The authors interpreted this as having a
larger perceptual span (i.e. the number of stimuli that can be processed during a singl
e
fixation). Thus experts may be attending to a greater number of stimuli, but this
would not be reflected in overt measures of attention.



5.2. Beliefs Held About an Opponent

Cottrell (1972) found that believing an observer is an expert can result in an

apprehension of being evaluated by him/her, which Cottrell termed evaluation
apprehension (EA). Henchy and Glass (1968, cited by Amabile
et al
., 1984) have
shown that experienced players who believe they are being observed by an expert
(EA) results in th
e player performing better than if he or she had believed the
observers to be novices. On the other hand, novices perform worse when they believe
they are in the presence of expert compared to novice observers.

Such modulation of
performance as a result
of EA has been termed social facilitation
-
inhibition (SFI).

Zajonc’s (1965) response
-
dominance approach has been used to explain SFI
effects. Believing an observer or coactor to be an expert results in initiating a
response with the highest probability of

being produced in that situation (known as a
dominant response), regardless of whether or not it is sufficient for good performance.
When individuals engage in novel tasks, the dominant response typically leads to poor
performance (Zajonc, 1965). Becaus
e the apprehension of appearing incompetent
(EA) results in a potentiation of the dominant response, novices will perform more
poorly when competing against a perceived expert opponent compared to a
supposedly novice opponent. Experienced participants’ do
minant response, on the
other hand, is the most practised response


i.e. the correct response for the situation.
Thus EA results in a potentiation of the correct response, thus explaining SFI effects


34





(the facilitation or inhibition of performance) as ari
sing from beliefs held about an
observer/coactor.

An alternative explanation of SFI effects has been provided by Baron (1986,
cited by Geen, 1991). His distraction
-
conflict theory claims that EA results in a
narrowing of attention to a relatively small se
t of central cues. Because complex tasks
require attention to a relatively large range of cues, this narrowing negatively impacts
upon performance. Muller
et al
. (2004) combined this approach with Festinger’s
(1954, cited by Muller
et al
., 2004) social c
omparison theory to account for the fact
that believing an opponent to be an expert results in social inhibition of performance
compared to a lack of social inhibition when the opponent is believed to be a novice
(Bond & Titus, 1983; Henchy & Glass, 1968,
c
ited by Amabile
et al.
,

1984
). That is,
on complex tasks, comparing oneself to an expert results in attention narrowing due to
EA.

In both accounts EA results in SFI effects, either by potentiating a response
(Zajonc, 1965) or via attentional focussing (
Muller
et
al
.
, 2004). A

difference
between the two accounts is that the former claims that EA results in performance
facilitation if the participant is experienced in the field but performance inhibition if
the participant is a novice. The latter a
ccount of SFI effects claims EA will result in
performance inhibition or a lack thereof (i.e. no performance facilitation), irrespective
of ability. Therefore the experimental hypothesis of the present study predicted that
believing the opponent to be an
expert would inhibit performance for both novice and
expert participants whilst believing the opponent to be a naïve participant would result
in all participants performing relatively better.

No conclusions could be reached about the
social inhibition of n
ovices’
performance
because the task was too complex

irrespective of the cover story
. Any


35





interpretation of novices’ self
-
reports and visual sampling strategies will be plagued
by this confound. Therefore, it was not possible to determine whether SFI eff
ects
were (Zajonc, 1965) or were not (Muller
et al
., 2004) a function of participant ability.

The same was not true of experienced players, who
demonstrated better
performance when competing against a
n


expert


opponent compared to
a ‘
naïve


opponent, as p
redicted by EA (Cottrell, 1972)
. This can be explained by both the
response
-
dominance approach (Zajonc, 1965) and the attention
-
narrowing approach
(Muller
et al
., 2004), but not as a result of withdrawing effort (a potential confound).
G
iven that the abi
lity to perform well relies heavily upon one’s deployment of visual
attention (
e.g. de Groot, 1946/1978
) as well the
beliefs held about the opponent
, it is
interesting to ask whether or not
the belief held about the opponent has its
influence
upon perceptu
al encoding

which then modulates performance
. Such a finding would
support the
claim that
perceptual encoding of

experts

is not inflexible (in
contradistinction to
Bilalić
et al.
, 2008b). Whilst there has been
some
support of this
proposal
from

reductionistic
studies
(Muller
et al.
, 2004), support was not found in
the present study which engaged participants in the situated and more complex
environment of competitiv
e videogaming.

The null hypothesis could not be rejected.
I
t appear
s

that the results are
in conflict with an attentional focusing account of EA
(e.g. Muller
et al
., 2004).

This null effect of visual sampling may be considered unsurprising. Visual
alloc
ation is a result of the demands of the task (e.g. Yarbus, 1967, cited by Findlay &
Gilchrist, 2003). The primary demand of the current task was always the same: to
compete (and win) against an opponent irrespective of the opponent’s ability.

However, du
e to limitations in the present study, explanations that pertain to the


36





distraction
-
conflict theory of SFI effects cannot be dismissed entirely.
T
hese come
under two headings: spatial detail and temporal detail.

For the former, there are issues concerning

methodological and hardware
solutions.
Concerning methodological issues
, ROIs
were not defined by the central
and peripheral cues described in section 3.3.2.1

because

they were dynamic
,
for
example, the continually changing height of the
up
per layer of the
stacked puyos in
the participants’ screen

(which allows for the puyo
-
pair plaecment decision to be
made).

Synchronising eye movement recordings with a camcorder recording of the
dynamic screen activity
4

was
too demanding given the time li
mitations. More
sensitive measures that pertained to
the dynamic
central and peripheral cues may have
been able to demonstrate that experts’ apparent perceptual inflexibilities (e.g.
Bilalić,
et al
.,

2008a, 2008b; Reingold,
et al.
, 2001b) are subject to influence from the beliefs
they hold about their opponent.

The eye recording hardware was limited in as much as an error of 0.5
o

in
spatial accuracy was inherent in the recordings of fi
xation location. Because
numerous ROIs bordered one another, fixations recorded as relatively near to any ROI
border could potentially be recorded as within the incorrect ROI. This may have
resulted in the insignificant differences found in fixations dir
ected to, say,
Own
Garbage

area (see figure 1
).

Temporal detail refers to the beliefs held over the duration of the game.
P
articipants were able to re
-
evaluate their initial
belief
about the opponent’s ability
because
the
opponent’s performance

could be m
onitored on screen
.
Participants may



4

This solution provided a technically undemanding alternative to writing a program of the videogame
into the eye movement
-
recording software. Had such an endeavour been attempted it may have been
possible to define ROIs as dynamic, as per Jermann
e
t al
. (2008).



37





have changed the belief they held, gradually or suddenly, about the opponent’s ability
from an ‘expert’ to a less difficult ability. This was indeed the case, as shown by the
self
-
report measure examining expectations

of performance
.
Because EA results from
being observed by an expert (which in turn modulates
perceptual encoding
: Muller
et
al
.,
2004), it is possible that cue utilisation did differ significantly at the beginning of
gameplay, but this effect was
diluted

by the shift in belief held about the opponent
.
Future studies should examine

perceptual encoding strategies
by employing

shorter
game durations
(
so participants would not have had sufficient time to re
-
evaluate their
opponent
) or by equally dividing the

15 minutes of gameplay
into a number of time
periods and determining whether
a shift in cue utilisation was observed

as a function
of the changing direction of social comparison.

Employing the latter methodology, it would

also

be possible to determine
whe
ther the overarching goal of competing to win can be divided into a number of
component subgoals. It is too rigid to assume that participants always aim their
behaviours towards solving
the overarching
goal (Kirsh & Maglio, 1994). For
example, when the p
articipant’s screen

becomes almost full with puyos

a switch in a
subgoal may be found: from attempting to knock out the opponent with advanced
strategies (appendix 2)
(an attacking subgoal)
to more basic strategies in an attempt to
merely survive

(a defens
ive subgoal)
.
These

may require utilising different perceptual
cues, such as the next puyo pair to fall

(
Own Next)

with an attacking

subgoal (which
may not be attended
with a deensive
subgoal).
Because

“additional time spent on
perception is… [used] to r
e
-
examine the most important

[task
-
relevant]

elements [of
the environment]” (Yarbus, 1967; cited by Findlay & Gilchrist, 2003, p.132), one
may

expect differences in
either
total fixation duration on different screen regions
or in
perceptual span (cf. Reing
old
et
al
.
, 2001)

as
a result of these changing subgoals.



38





Finally, it should be noted that the current study, u
nlike Muller
et al
.
’s

(2004)

research
, cannot determine whether experts’ perceptual encoding mechanisms do
show flexibility in terms of covert at
tention. Future studies should aim to examine
whether this is the case

in situations that are more ecologically valid than that
employed by Muller
et al
. (2004)
.



5.3
. Conclusion

In summary, the present research aimed to discover whether or
not
experts’
inflexible
perceptual encoding mechanisms were truly inflexible by bringing the study of
expertise into a
dyadic
competitive arena. Even under these ecological conditions,
experts appear to demonstrate a high degree of automaticity. This was inferred fro
m
the results exhibiting similar duration and number of fixations with

irrespective of the
belief held about the opponent
, even though
it was sufficient to modulate
performance. The present study thus detracts from explanations of
SFI effects

as a
result
of attentional focussing
(Baron, 1986, cited by Geen, 1991), even when taking
into account the social comparison factor that has been shown to modulate SFI effects
(Muller
et al
., 2004). However, because the methodology did not afford the
opportunity to
e
xamine

covert attentional mechanisms, neither conflict
-
distraction
theory nor the potential flexibility of perceptual encoding strategies can be dismissed.
Indeed, there appears to be
some albeit
indirect support for strategic and flexible
perceptual enco
ding as demonstrated by the mean fixation duration
s of experts

on this
study in comparison to other studies.

The way forward is to examine
experts’ perceptual encoding by first defining
the central and peripheral cues and investigating whether or not the t
otal duration of
fixations
and the mean duration of a single fixation
can be understood as a function of


39





these cues (Baron, 1986, cited by Geen, 1991). Until such studies are conducted
though, it seems that the perceptual encoding of domain
-
specific exper
tise must be
explained as largely automatic and highly resistant to influence from external

soures
of comparison
.


6. References


Amabile, T.M., Goldfrab, P. & Brackfield, S.C. (1984).
Social influences on
creativity: Evaluation, coaction and surveillance
.
Toronto, Canada:
the
Annual Convention of the American Psychological Association
.

(ERIC
Document Reproduction Service No. ED256095)


Bangerter, A. (2004).
Using pointing and describing to achieve joint focus of
attention in dialogue
, Psychological Scie
nce, 15
(6), 415
-
419


Bayliss, A.P., Frischen, A., Fenske, M.J., & Tipper, S.P. (200
7
).
Affective
evaluations of objects are influenced by observed gaze direction and
emotional direction
,
Cognition,
104
(3), 644
-
653


Bilalić, M., McLeod, P., & Gobet, F. (20
08
a
).
Inflexibility of experts


reality of
myth? Quantifying the Einstellung effect in chess masters
,
Cognitive
Psychology, 56
(2), 73
-
102


Bilalić, M., McLeod, P., & Gobet, F. (2008
b
).
Why good thoughts block better
ones: The mechanism of the perniciou
s Einstellung (set) effect
,
Cognition,
108
(3), 652
-
661



Bond, C.F., & Titus, L.J. (1983).
Social facilitation: A meta
-
analysis of 241
studies
,
Psychological Bulletin, 94
(2), 265
-
292


Brennan, S.E., Chen, X., Dickinson, C.A., Neider, M.B., & Zelinsky, G.J
. (2008).
Coordinating cognition: The costs and benefits of shared gaze during
collaborative search
,
Cognition, 106,

1465
-
1477


Campitelli, G., & Gobet, F. (2004).
Adaptive expert decision making: Skilled
chess players search more and deeper
,
Journal of
the International Computer
Games Association
,

27
(4),

209
-
216


Campitelli, P.R., Gobet, F., Williams, G., & Parker, A. (2007).
Integration of
perceptual input and visual imagery in chess players: Evidence from eye
movements
,
Swiss Journal of Psychology, 66
, 201
-
213


Chapman, P.R., & Underwood, G. (1998).
Visual search of driving situations:
Danger and experience
,
Perception, 27
(8), 951
-
964



40






Charness, N., Reingold, E.M., Pomplun, M., & Stampe, D.M. (2001).
The
perceptual aspect of skilled performance in ch
ess: Evidence from eye
movements
,
Memory & Cognition, 29
(8), 1146
-
1152


Chase
. W.G.,

& Simon
, H.A.

(
1973
).
Perception in chess
,
Cognitive Psychology, 4,
55
-
81


Cottrell
, N.B.

(1972)
.
Social facilitation
. In C. McClintock (ed.),
Experimental
Social Psych
ology

(pp. 185
-
236). NY: Holt, Rinehart & Winston


d
e Groot
, A.D.

(
19
78).
Thought and choice in chess

(2
nd

ed.). The Hague: Mouton
(Original work published in 1946)


Easterbrook, J.A. (1959).
The effect of emotion on cue utilisation and the
organisatio
n of behaviour
,
Psychological Review, 66
(3), 183
-
201


Eccles, D.W. (2005).
Thinking outside the box: The role of environmental
adaptation in the acquisition of skilled and expert performance
,
Journal of
Sport Sciences, 24
(10), 1103
-
1114



Ericsson, K.A.,
& Lehman, A.C. (1996).
Expert and exceptional performance:
Evidence of maximal adaptation to task constraints
,
Annual Review of
Psychology, 47
, 273
-
305


Findlay, J.M., & Gilchrist, I.D. (2003).
Active vision
. Oxford: Oxford University
Press


French
, K.E
.,

& McPherson
, S.L. (
1999
).
Adaptations in response selection
processes used during sport competition with increasing age and expertise
,
International Journal of Sport Psychology,


Frischen, A., Bayliss, A.P., & Tipper, S.P. (2007).
Gaze cueing of atte
ntion: Visual
attention, social cognition, and individual differences
,
Psychological Bulletin,
133
(4), 694
-
724


Gajadhar, B.J., de Kort, Y.A.W., & Ijsselsteijn, W.A., (submitted).
Rules of
engagement: Influence of co
-
player presence on player involvement
in
digital games
,
_________
-


Gallagher, H.L., Jack, A.I, Roepstorff, A., & Frith, C.D. (2002).
Imaging the
intentional stance in a competitive game
,
NeuroImage, 16
, 814
-
21


Geen, R.G. (1980).
Test anxiety and cue utilization
. In I.G. Sarason (Ed.),
Tes
t
Anxiety: Theory, Research, and Applications
(pp.43
-
61). Hillsdale: Erlbaum


Geen, R. G. (1991).
Social motivation
,
Annual Review of Psychology, 42
, 377
-
399


Gerhard, M., Moore, D., & Hobbs, D. (2005).
Close encounters of the virtual kind:
agents simul
ating copresence
,
Applied Artificial Intelligence, 19
, 393
-
412



41






Gobet
, F.

(
1998
).
Expert memory: A comparison of four theories
,
Cognition,
66
(2), 115
-
152


Gobet, F., & Simon, H.A. (1996).
The roles of recognition processes and look
-
ahead search in time
-
c
onstrained expert problem solving: Evidence from
grandmaster level chess
,
Psychological Science, 7
, 52
-
55


Goulet
, C.,

Bard
, C.,

& Fleury
, M. (
1989
).
Expertise differences in preparing to
return a tennis serve: A visual information processing approach
,
Jo
urnal of
Sport and Exercise Psychology, 11
(4), 382
-
398


Guo
, K.,

Mahmoodi
, S.,

Robertson
, R.G.,

& Young
, M.P. (
2006
).
Longer fixation
duration while viewing face images
,
Experimental Brain Research, 171,

91
-
98


Huguet, P., Galvaing, M.P., Monteil, J.M., &

Dumas, F. (1999).
Social presence
effects in the Stroop task: Further evidence for an attentional view of social
facilitation
,
Journal of Personality and Social Psychology, 77
(5), 1011
-
1025


Irwin, D.E. (2004).
Fixation location and fixation duration as

indices of cognitive
processing
. In M. Henderson & F. Ferreira (Eds.),
The Interface of Language,
Vision and Action: Eye Movements and the Visual World

(pp.105
-
134).
London: Psychology Press


Jermann, P., Nüssli, M.A., & Dillenbourg, P. (2008).
The eff
ect of conflict,
expertise and conflict on eye fixations in collaborative tetris
,
ICLS 2008
Workshop: Four Eyes, Eye tracking methods in collaborative work/learning
research,
1
-
8


Kingstone, A., Smilek, D., Ristic, J., Friesen, C.K., & Eastwood, J.D. (2003
).
Attention, researchers! It is time to take a look at the real world
,
Current
Directions in Psychological Science
,
12
(8), 176
-
84


Kirs
h
, D.

& Maglio
, P.

(1994)
.
On distinguishing epistemic from pragmatic
action
,
Cognitive Science, 18
, 513
-
549


Klein
,
G.A.

& Hoffman
, R.R.

(
199
2).
Seeing the invisible: Perceptual
-
cognitive
aspects of expertise
. In M. Robinowitz (Ed.),
Cognitive Science Foundations
of Instruction

(pp.203
-
226). N.J.: Lawrence Erlbaum


Knoblich, G., & Jordan, J.S. (2006).
Action coordina
tion in groups and
individuals: Learning anticipatory control
,
Journal of Experimental
Psychology: Learning, Memory and Cognition, 29
(5), 1006
-
1016


Land, M.F. (2006).
Eye movements and control of actions in everyday life
,
Progress in Retinal and Eye Rese
arch,
25
(3), 296
-
324


Land
, M.F.,

& McLeod
, P.

(
2000
).
From eye movements to action: How batsmen
hit the ball,
Nature Neuroscience, 3
(12), 1340
-
1345




42





Land, M.F., & Tatler, B.W. (2009).

Looking & acting
. Oxford: Oxford University
Press


Lane
, P.C.R.,

Che
ng
, P.C.H.,
& Gobet
, F.

(
1999
).
Learning perceptual schemas to
avoid the utility problem
.
Proceedings of the Nineteenth SGES International
Conference on Knowledge Based Systems and Applied Artificial Intelligence
,
pp.72
-
82. Cambridge: Springer


Malcolms
on, K.A., Reynolds, M.G., & Smilek, D. (2007).
Collaboration during
visual search
,
Psychonomic Bulletin & Review, 14
(4), 704
-
09


Markovitch, S., & Reger, R. (200
5
).
Learning and exploiting relative weaknesses
of opponent agents
,
Autonomous Agents and Mul
ti
-
Agent Systems
,
10
(2), 103
-
130


Muller, D., Atzeni, T., & Butera, F. (2004).
Coaction and upward social
comparison reduce the illusory conjunction effect: Support for distraction
-
conflict theory
,
Journal of Experimental Social Psychology, 40,
659
-
665


N
eisser
, U.

(
1976
).
Cognition and reality: Principles and implcations of cognitive
psychology
. CA: Freeman


Ravaja, N., Saari, T., Turpeinen, M., Laarni, J., Salminen, M., & Kivikangas, M.
(200
6
).
Spatial presence and emotions during video game playing:
Does it
matter with whom you play?
,
Presence
: Teleoperatores and Virtual
Environments
,
15
(4)
,
381
-
392


Reingold
, E.M.,

Charness
, N.,
Pomplun
, M.,

& Stampe
, D.M.

(
2001a
).
Visual span
in expert chess players: Evidence from eye movements
,
Psychological
Scien
ce, 12
(1), 48
-
55


Reingold
, E.M.,

Charness
, N., S
chultetus
, R.S.,

& Stampe
, D.M.

(
2001b
).
Perceptual
automaticity in expert chess players: Parallel encoding of chess relations
,
Psychonomic Bulletin & Review, 8
(3), 504
-
510


Rilling, J.K., Sanfey, A.G., Aro
nson, J.A., Nystrom, L.E., & Cohen, J.D. (2004).
The
neural correlates of theory of mind within interpersonal interactions
,
NeuroImage, 22
, 1694
-
1703


Robbins, T.W., Anderson, E.J., Barker, D.R., Bradley, A.C., Fearnhough, C., Henson,
R., Hudson, S.R., &
Baddeley, A.D. (1996).
Working memory in chess
,
Memory & Cognition, 24
(1), 83
-
93


Saariluoma, P. (1990).
Appereption and restructering in chess players’ problem
solving
. In K.J. Gilhooly, M.T.G. Keene & G. Erdos (Eds.),
Lines of Throught:
Reflections on

the Psychology of Thinking

(pp.41
-
57). London: Wiley


Saariluoma
, P.

(
1995
).
Chess players’ thinking: A cognitive psycological
approach
. London: Routledge




43





Snyder, M., Tanke, E.D., & Berscheid, E. (1977).
Social perception and
interpersonal behaviour:

On the self
-
fulfilling nature of social stereotypes
,
Journal of Personality & Social Psychology, 35
, 656
-
666


Snyder, M., & Swann, W.B. (1978).
Hypothesis
-
testing processes in social
interaction
,
Journal of Personality & Social Psychology
,
35
(11), 1202
-
1
212


Underwood, J. (2005).
Novice and expert performance with a dynamic control
task: Scanpaths during a computer game.
In Underwood, G. (Ed.),
Cognitive
Processes in Eye Guidance

(pp. 303
-
324). Oxford: Oxford University Press


v
an Gog
, T.,
Paas
, F.,

&
v
an Merriënboer
, J.J.G.

(
2005
).
Uncovering the problem
-
solving process: Cued retrospective reporting versus concurrent and
,
Journal of Experimental Psychology: Applied, 11,
237
-
244


Zajonc
, R.B.

(
1965
).
Social Facilitation
,
Science, 149
(3681), 269
-
274



44





7. Appendices


1.
1
.

Preliminary questionnaire

1.
2
.

Game experience questionnaire after block 1

1.3
.

Game experience questionnaire after block 2


2.

Example of task instructions (for participant vs. expert opponent in block 1.
Presented via PowerPoint).


3.

Participants’ means & S.Ds for when responding to the statement: “I felt that
my opponent performed better than me”