Perception
& Psychophysics
2000, 62 (5), 960-968
Attending to an object's color entails
attending to its location: Support for
location-special
views
of visual
l~
le~
attention
~;~:~.
fects.Ts
YEHOSHUA
Tel Aviv
Van der Heijden,
Kurvink,
TSAL and DOMINIQUE
LAMY
University,
Tel Aviv, Israel
de Lange, de Leeuw,
and van der Geest (1996) argued
that the results
Experim
letterwa
Inare
(1996) (
sup-
porting the location-special
view obtained by Thal and Lavie (1988) were due to uncontrollable
shifts
of fixation,
rather than reflecting
the properties
of the attentional
system. In the present study, we present an improved
variation
of the Tsal and Lavie (1988) paradigm
and reassert our claim that location
is a special dimension.
Subjects were presented
with circular
arrays of six letters of different
colors.
Three of the letters were enclosed by (Experiment
1) or superimposed
on (Experiments
2, 3, and 4) different colored shapes. The subjects were instructed
to report the (target) shape with a given color (e.g.,
report whether the re~ shape was a square, ~ circ.le, or a trian~le)
and then either
freel~
report
letters
from the array (Expenments
1, 2, and 4) or Identify a prespecified
target letter (Expenment
3). In all
four experiments,
performance
was substantially
better for the letter that appeared in the location
of
the to-be-reported
shape (location
clude that attending
letter)
to the stimulus
color
than for the letter
entails
directing
that shared
attention
its color
(color
letter).
deed prG
results (!
f;v~te th
j
rep
leu;rs. J
der l(
~n
d tl
orce
. In the
We con-
t~ated tl
to its location.
dJen et I
onlyshf
then re~
The special
status
of stimulus
location
in visual
selec-
(e.g., Baylis
& Driver,
.1992, 1993; Duncan,
1984; Harms
exactly
tive attention has been extensively debated in recent years.
As van der Heijden, Kurvink, de Lange, de Leeuw, and
van der Geest (1996) have suggested, "a brief look at cur-
& Bundesen, 1983; Humphreys, 1981). For example, using
the Eriksen flanker paradigm, Baylis and Driver (1992)
reported that distant incompatible
distractors grouped
o~her ~~
s ape
to ~ppe~
rent ~eorizing c~ .make ~is cl~ar. There are two gro~p's of
theones: the 'posltlon-not-speclal'
ones and the 'posltIon-
w~th the t~et b~ ~olo~ or by good con~inuation.interfered
with target IdentIfIcatIon more than dId closer mcompat-
~i~::s ~
.
special' ones" (p. 1224).
Over the years, there has been a wide range of findings
that provided support for the unique role of location in
visual selection (e.g., Cave & Pashler, 1995; Hoffman &
ible distractors that were not otherwise grouped with the
target.
Recently, a controversy arose concerning the results
obtained under a particular paradigm introdl.1ced by Tsal
t
Nelson, 1981; Kim & Cave, 1995; Luck, Fan, & Hillyard,
1993; Posner, Snyder, & Davidson,
1980; Theeuwes,
1989; Tsal & Lavie, 1988, 1993;Vecera, 1994). For exampIe, Tsal and Lavie (1993, Experiment 4) presented a cuing
and Lavie (1988) that supported the location-special view.
The uniqueness of this paradigm was that it assessed the
effects of attending to location when location was completely irrelevant to the task. Van der Heijden et al. (1996) ~
:
I
display consisting of two peripheral disks, one black and
one colored (either pink or blue), followed by a probe
showed that these results were due to uncontrollable
fixation shifts and, as such, did not reflect the properties of
display containing a target letter. Subjects were told to
respond to the target letter only if the colored disk was
the attentional system. The present study is an improvement on our original paradigm. It shows that even under
pink. Although the location of the disk was entirely task
irrelevant, the probe was detected faster when it appeared
in the location previously occupied by the colored disk than
when it appeared in the location of the black disk.
In contrast to the above studies, a large number of findings have provided support for the notion that location is
not different in principle from other selection dimensions
and two anonymous revIewers for theIr helpful comments on an earlter
proper fixation conditions, attempting to attend to any aspect of a stimulus entails attending to its location.
Tsal and Lavie (1988, Experiment
I) presented a circular array containing three red, three green, and three
brown letters. The subjects were instructed to report first
one letter of a given color and then any other letters they
could identify. The letters reported additionally were more
often letters adjacent to the first reported letter (location
letters) than letters with the same color as the first reported
letter (color letters). Tsal and Lavie (1988) concluded
that the selective processing of a target specified by its
1 .
I.
draft.
c.o
This researchwas supportedby Grant 06 10710551from the israel
ScienceFoundation.We.aregrateful t«:,Kyle Cave,Lex van derHeijd~n,
Correspondence
concerning
this
article
should
be
addressed
h
to
Y. Tsal, Departmentof Psychology,Tel Aviv University, RamatAviv,
Tel Aviv 69978,Israel (e-mail: [email protected]).
Copyright
2000 Psychonomic
Society, Inc.
or
IS
accomp
IS
d
e
b
d
.y
atten
.
.mg
h
to
t
.
IS
,
target
I
,
1
loca-
s
d
tIon. They noted that this expenment may have suffere
from a methodological difficulty: "Since three of the letters
960
~;;5;,1
~
c,~
SUPPORTFOR LOCATION-SPECIAL VIEWS OF ATTENTION
\';i:~::':
961
i :J~;1',
..~
c.~"
"' ;::
,:;{
,1
t
:',
:;.",'
I ::'
"~
H
;:
~:
portin~ additi?nal red letters may ha~e~eflectedeffects
oft.he i?struCtionsratherthan the p~rSiStinge.ffectsof the
?ctivationof redness.In co~trast,smcelocation wastask
irrelevant, effects of location could not have stemmed
f~omtop-down ef!ectsof i~structio~sand c.oul~only in~icatethat atte~dmgto an item entaIledacti~at!nga sp~tial representation.In the presentstudy,we el1mmated
this
Experiments 2.a~d 3, b?t in these experiments, t~e target
letter was specif~ed by itS s~ape, rather than b~.its color.
In a recent senes of expenments, van der HeiJden et al.
proble.m, since the color letter used.to ~ssessthe.effects of
attendmg to color was not a potential Item for first-target
report.
;~C
;?I~
84. Harm ;', :
:1 ie, usins,, i"
v~r (199211
s groupedilc,
interfered,l[:,,
(1996) de.monstratedthat improper fixation. could indeedprovidethe bestaccountfor Tsal and Lavie's(1988)
results(Experiment I). They showedthat whenforced to
fixatethe centerof the display,the additional letterss~bjectsreportedweremore often color lettersthan location
letters.An advantagefor location letters wasfound only
underlow-contrast conditions when subjects were not
forcedto fi~ate the fix~tion point.
.
In the senesof expenmentspresentedbelow,we el1minatedthe fixation problem investigatedby van der Heidjenet al. (1996). Subjectshad to report the form of the
only shapein a circular array that had a given color, and
thenreport letters from the array. One of the letters had
exactlythe samecolor asthe shape(color letter), andanother letter.occupied exactly the same location a~ the
shape(Ioc~tionletter):The targetshapewasequallyI1kely
to appearm any location. The letters,:",ereall o~differe~t
colorsand were randomly presentedm the vanousPOSi-
The secondpro~l~mwith thi~ paradigmwasthat.it may
not havebeensufficiently sensitivefor demonstrating10cationeffects,becauseit wasclearly dimensionallyasyrnmetrical.Indeed,the lettersincludedin.the color category
wereexactlyof the samecolor asthe first reportedletter,
whereasletters included in the location categorydid not
occupy precisely the samelocation as the first letter reporte.d,but only positions adjacentt? .it. Thus, ?btainin.g
location responsesunder thesecondItionsrequIresadditional assumptionsaboutthe shapeand size of the attentional spotlight. In the presentstudy,we eliminated this
bias,s~ncethe loc.ationlett~rsused.toassessthe effect~of
attendmgto locationoccupiedpreciselythe samelocation
as the target item.
-,
Notethattheabov.ei~provementsonlyf~cilitateth~as- '-"
sessment~f the pers~stingeffects0: attendm.gto locat~on.
~he paradig~ remamsstrongly biasedagamstlocation,
smcethe subjecthasto attendto the color of the shapeand
f ;c'
incompat"
:d with thec~'
tions. T~?s, there ,:-,asno ben~fit in shifting :ixatio~ to
any position or regio.1}?fthe circular a~ray pn?r to stim-
the locations o!the shapesand letters are totally irrelevant
to the tas~. Still, t.he n~w results ~resented here ~upport
I-
S
~n
,.
f-
"rfi"
',i'"
were alwaysrele,,:~nt,su?jects~ay haveshifted.fixation
to any arra~ ~osition p~ior to.stimulu~ onset, smce any
randomposition w~s hIghly li~ely to i~clude a le~t~rof
the relevant color m that or m an adjacent pOSition"
.17).Thus,.theresults m.ayhavereflected sen~orydifWrencesin visual ~rocessmgr~t~er.than a.ttentionale!fects.Tsal and LaVie (1988) miD1mizedthISproblem m
~
; 1'"'
:"
,.
.,
~
~
f
~-
~:
~
~
j,
'"
ulusonset.Moreover,m the last expenment,m order to
the location-special view and remforce our claim that
the results'
ed by Tsal ~
,ecialview.'f
furtherensurethat the subjectswould not shift their eyes, whenattendingto the color of a given item, subjectscanwe followed the methodology used by v.ander Heij~en not avoid attendingto its location.
~tal. to control for e.yemovements.We m.clu~eda sign
'sessedthe 'J"
m thecenterof the display,the form of which mstructed
EXPERIMENT 1
, was com-; :;,subject.s whetherto r~~pondor withhold.their responses.
..
tal. (1996}.;,1."
>liablefix- :::
0 ertiesof :
n~m rove-;
evenPunder'
:i to anyas"" l'
tion.
':"
'nted
a
cir-",
, and three' ~
Followingvan d~r HeiJdenet al.'s~easoD1ng,
we assu~ed
,thatthe ce~tra.1item would re~uu.efoveal processmg,
thusnecessitatingcentral eye fixation. .
. In addition to the ~act.th~t the present.study.was designedto overcomethis fixation problem,It alsoIncluded
a numberof improvementsover Tsal and Lavie's (1988)
parad!g~. The origin?1 rationale was that the loc~t~on
supenonty
the targetwas
specified
by
its color effect
and itsobtamed
locationwhen
was completely
irrelevant
to
.In the.present~xpenme";t,t~e subjectswerepr~sented
with a cucular display of SiXdifferent lettersof diff~rent
colors. Threeof the letterswere enclosedby threedifferent colored shapes(seeFigure I )'. The subjectswere instructedto report the shapeof a given color (e.g., report
whetherthe red shapewasa square,a circle, or a triangle)
and then asmany letters as they could. One of the le!ters
had
the
same
colorThe
as the
target
shape
defmed
asthe
color
letter.
letter
enclosed
byand
thewas
target
shape
the task demonstrated that a.tt~n?ingto location is a
I" m~ndator,>!
proces~.However,it IS I~portant to ~~te.that
theY:'
.le~ers
weremore
thIS~ar~dlgmalsomvol.vedseve!albiasesthat mmi~ized
~r(location r
!heI1k~l1hood
of observmglo,cation~esponses
e~en1fsubrst reported",; , Jectsdid attendto the tar~et~ location. S~chbiasesmay
concluded,ic acc.~unt
for the color supenontyeffectobtamedby vander
ified by its'
HelJdenet al. (1996).
r et's loca-' "
The first bias concernedthe fact that the color letters
v~ suffered :,: ~erepotential targetsfor first-letter report. For instance,
)fthe letters;;';, ~mcethere.were three re~ letters in the array,when sub; i;"c,,;Jects
were mstructedto fust report a red letter, all three
, ::,;;red letterswere potential targetsand were thus likely to
wasdefinedasthe location letter. Thecritical comparison
conc~rnedthe fre9ue~cyof repo~ of c?lor letters and of
location I~tters.Fmdmg that subjectsmdeedtend to report location letters more fre9uently than color letters
wou!d dem~nstrate~hatattendmg.to the colo: of a.shape
entails that itS loc~tion, altho~gh ~rrelevant,is activated,
and more so than itS color, which IStask relevant.Sucha
finding would strongly support the idea that location is
a specialdimension.As in the studiesof Tsal and Lavie
(.1988) andVander H~ijden et al. (1996),the subjectsinitially attendedto a given color and then freely reported
any letters they could. Unlike thesetwo former studies,
~e ort first!
~~:ave undergonesomelevel.ofprocessing.As a result,re-
~
~
the location letter-namely, the letter usedto assessthe
.
-_.~
'
~~~~~=~
1
962
TSAL AND LAMY
&
~G
X
G
0
r;;-,
had to report correctly any letters of the array, and as many as possible. Thus: the letters reported could be, but d!d not need to be, the
;'~j"
lettersshanngthetargetshape'scolor or locatIon.
~ii~
0
,;,;:""
,
Results and Discussion
~
~
£
X
~
E
T
~lf;c
Trialsin whichthe shapeof thedesignated
colorwas
R
fT'
n~medin.correctlywereexcludedfrom analysis(6% of the
tnals). Smcethe color letter and the .locationletter were
~
enclosed
byshapes,
weanalyzed
theresponses
onlyfor
Figure 1. Examplesof the stimuli usedin Experiment 1.
i!~:~
f"i;
'
-
the neutralletterthat wasalsoenclosedby a shape,in
~
order to make the three relevant response categories di-
"!:.'~'.'
p,., .'
rectlycomparable.
Thus,in eachdisplaytherewereone
color
letter,
one
location
letter,
and
.
one
relevant
~g;;;!k,'.'
neutral
.
;:~:.'~
",",c,'
.
"
letter. The latter servedas a baselmefor measunngany;':""'"
e.ffectsof attendmg.tolocatlon-occupled the sameloca~ facilitation resultingfrom sharingthe locationor the color
(JlI
t~onas~e shape.with the target color, ratherthan a loca of the target shape.Table I showsthe mean number of
i!¥:,:.'
tlon adjacent to It.
letters reported per trial in each category. On average, the
Method
subj,ectsreported 1.29relevantlettersper trial (excluding
the Irrelevantneutral
lettersnot enclosedby a shape).An
..
.
'
t
8 d
d t fi
TIA .
Su bJect 5, The sub~ec
s were un ergra ua es rom e VIV
University,Threewerepaid subjects,and5 participatedaspart of a
courserequirement.
All hadnormalor corrected-to-normal
vision.
Stimuli and Procedure. Stimuluspresentation
anddatacollection werecontrolledby an IBM PC/ATwith a VGA graphicdisplay.
A h'
d
b' l '
.
'
d'
43
c
mrest
was
use
to
sta
I
Ize
vIewIng
Istance
at
:
andth
ob,iJ ec1
overa~l analysIs of vanance (ANOYA) of letter cat~gory
(location, color, or neutral) X color (red, blue, or white) X
subjects revealed a highly significant effect for letter category [F(2,14) = 45.77,p < .0001]. Furthercomparisons
thesu'
result;
ing th
" .
h
cm
so
t
at
elml!
indicated
that
location
letters
were
reported
more
fre-
,
~ cmonthedisplayco~res~ondedto
1.33~ofvisualangl~.Thesubjects
werepresented
WIthcIrcularletter
dIsplays,EachdIsplaycon-. quentlythancolor letters[F(l ' 7).'= 37.51 p < .0005],
sistedof six different letters of different colors.The letterswere This ~as.true for each of the 8 subjects (see Table 2). The
randomlysampledfrom the entire alphabet.The colors were red supenonty of color letters over neutral letters was mar(RGB values,255/80/80;luminance,~I cd/m), blue (80/80/255; ginally significa?t [('(1,7) = 5.23,p = .056].
21cd/m),green(0/160/0;32 cd/m),whIte(255/255/255;III cd/m),
There was no significant effect of color [F(2,14) = 2.72],
magenta
(2,55/80/255;
37 cd/m),~d yellow(255/2.55/80,;
?8 cd/m). but there was a highly significant interaction effect beFor eachdIsplay,the letters,theIr colors,andtheIr posItIonswere tween color and letter category [F( 4 28) = 9 79 p >
randomlypaired,Eachletter subtended1.33°in heightand0.93°in
. . .
.
'
.,
.
width,andthecenter-to-center
interletterdistance
was5,32°of visual .000 ~], mdlcatmg that the difference between reporting
angle. The entire array subtended 9.31° of visual angle in diameter
location letters an? color letters ~as smaller when the tar-
white target shape, location letters were reported more:;
frequently than color letters [F(I,7) = 7.07,p < .05] and:!:
color letters were reported more often than neutral letters ~
",
[F(1,7) = 10.23,p < .05]. The above .mt~r~ctl°n.between::;;
stramtthata shape
neverenclosed
a letterof thesamecolor.The color and letter c.ategorywas also signIficant m all the ;~
shapeswerealwaysseparatedby one interveningletter,In half of
subsequent expenments reported below. We will address-;,
thedisplays,
th~y,occupied,the
topcentralpositionandt~erightand this issueby presentingadditional analysesin the Gen-J'!;
left bottomposItIons,and m the otherhalf, theyoccupIedthe bott~m centerpositionand the ri¥ht an~left top positions.T.h~se
two
dIsplaytypeswererandomlyIntermIxed.Thus,eachpositIonwas
equallylikely to containthetargetshape,andtherewasthereforeno
gain in shifting fixation to any position prior to stimulus onset,
Eachstimuluswas presentedfor 100msecand was immediately
preceded
by a graycentralfixation crosspresented
for I sec,Thesubjectswereinstructedto reporttheshape(thetargetshape)C?f
a given
colr;>r
(e,g" reportwhetherthered shapewasa square,a t~angle,or
a cIrcle)andthenasmany lettersasthey could,The subjectswere
presentedwith three45-trial blocks,eachprecededby 15practice
trials, Dependingon block, the designatedcolor for the targetwas
red,blue,or white.Theorderof blockpresentation
wasrandomized
acrosssubjects,
Thelocationletterwasdefinedastheletterenclosed
by the targetshape,The color letter wasdefinedasthe lettersharing the color of the targetshape,It wasalwaysseparated
from the
locationletterby oneinterveningletterandwasalwaysenclosedby
a nontargetshape.All the other lettersweredefinedas neutral,It
wasemphasized
to the subjectsbeforeeachblockof trials thatthey
eral Discussion section.
;;,
The present findings provide strong support for thet!'
.
..
.
.
., I '
locatl~n-speclal view of'!lsual attention. Not surp?smg y,~
attendmg to the shape With the relevant color entailed that;;
color letters were reported more often than neutralletters.~
More interestingly, it also entailed that location letters)
were reported more often than neutral letters and even,:
more often than color letters.
11bl 1
Mean Numberof Le~e~ Reportedper Trial
by Letter Categoryand Color Block in Experiment1
,
M
LetterCategory
Red
Blue
WhIte
e
Location
0.75
0.81
0.59
0.7
Color
0.24
0.31
0.36
0.3
Neutral
0.30
0.26
0.25
0,2
m.atl
slight!
posed
in the
shape
Onthl
letter..
'
h
Int e
entanJ
cepto
tifica1
,
andwascentered
onthefixationpoint,In eachdisplay,
threeof the get shapewas white than when It was red O[ blue. How- j1
letterswereenclosed
by threedifferentcoloredshapes:
a square ever, additional comparisonsshowedthat even for the;:
(s!de=-2,53°2,a circle(diameter= 2.66°~,a~danequilate~1triangle
(sIde- 3.01 ). The letter wascenteredI~sldethe e~closmgshape,
Oneofth,e~eshapeswasred,blue,or whIte(dependmg
on block, as
was
specifIed
below).
The
colorsof five
the other
shapes
rando~ly
sampled
from the
remaining
lettertwo
colors,
withwere
the con-
cOm
~",.,
,.teth
Sub
Univer
course
Stin
asthe1
totheI
perimp
R c'i
esn
As
the d..
clude,
respo;
Posed
with t
cTnean
On
te a..
rs
;~i
.
. c='"
-c-
"-
--"~-
SUP
FO
LO
VI
O
A
9
)s1It:~;
Tab
2
pe
loc
let
w
r
m
f
e
r
subje
Loca
Colo
Neu
twe
the
fre
of
re
co
le
a
n
h
l
.
k
h
d
f
.
I
.
:utrali
jc"~~
fec
of
co
[F(
=
6.
<
.0
T
i
b
n
g
any~
,;'~:"~
twe
an
let
ca
w
s
[F
=
i
colo,i
)c\::c
EXP
2
~~
c.
.,
.
,
l
.
g
t
th
d
n
h
32
II
P
<
ber
o\:~f'
.
...
rep
l.e
fro
th
c
w
~e,
thez~
One
migh
argu
that
m
the
first
exp
the
lett
sm
for
the
wh
ta
sh
fo
re
a
luding
and
the
surro
shap
form
a
coh
per
tar
sh
Ind
ad
an
o
Ie).
An~
};:
objec
(Trei
Kah
&
Bur
198
He
res
to
re
s
ltegor
;;:"
supe
perfo
for
loca
lette
ma
hav
can
dif
be
re
le
fr
th
ore
fre-i:i
m~tio
o~
a
per~
obje
m
the
sec
exp
by
for
the
wh
s~
~o
w
w
b
a
in
.OO0
displ
the
shap
so
that
~he
wer
sup
Ge
DI
se
th
re
o
t
s
h
Has
m~f,~
first
expe
the.
spa
~rra
of
the
per
sh
~n
ev
~t
lo
~),:
2.7t#
'c
On
the
contr
It
e?ha
t~e
salie
of
the
enc
lett
On
ma
th
co
th
lo
effect
~;~:
~ette
In
with
the
spat
arra
pre
sup
eff
wa
m
by
sp
a
n
).
79,
p..~lc
m
next
subj
had
to
effo
dls
by
ob
pr
repor
entan
shap
from
lette
bec
glob
per
\en
the~1
c.ept
pote
obje
they
form
wou
imp
ide
EX
3
}lue,
H.~'~
ufica
perfo
of
b?th
the
sha
and
the
lett
,en
for:!J
Th
thi
ex
wa
de
to
fu
c
orted
~~,l':
Meth
.
.
the
rel
co
dim
an
th
ir
l
o
<
.05)
~!l
~Ubj
.The
subj
we~
9
und
from
Tel
AvI
dim
un
mo
co
co
O
m
IocatlO
.It
le~?
UnIv
~Ive
were
paid
subj
and
4
part
as
par
of
a
'.
.
b
f
~utr~
co~~
All
had
no~
or
~orr
visi
arg
tha
m
the
firs
tw
ex
s
~
o
\on.
all"'t',
Stimu
and
Proc
Thl~
expe
was
exa
t~e
sam
som
~e
ha
ad
a
s.
st
th
~
tnt,
In
di~~i:
i!Sthe
~rst
one,
with
one
exce
All
the
sha
wer
disp
fav
of
th
l~
~
d
e
~Ill
adQ;f
~e~~
nght
by
1.33
As
a
re.su
I.eft
~ide
of
the
sha
was
sumg
the
eff
at
pn
o
t
c
l
h
t
c
c.
.
i
wit
ide
pr
ta
le
Th
.
h
d
.
d
.
..
b
.
surp~
fIrst
expe
trIal
~n
whi
of
a
p~
as
of
st
ch
o
t
a
h
IT
entall
colo
~as
nam
mc?
wer
exten
sy
W
u
th
sa
d
a
m
s
)
.
We
ana
y
z
ed
\':,oP
y
for
the
neut
lette
that
was
sup
:tters
~jfJ
';.'i~,W
in
?rde
to
mak
them
com
/\
:~~
1;g;~
c~of
repo
per
trIal
m
eac
cat
G
D
,r
~
r~.36';
1ria~
,,;c,,!
Cion
m
fi
orm
re
I
atlv
.
to
?e
eri~~
,~:1~
~~n
prob
be
attri
to
the
fac
tha
'~;;iC
if~~fi
Aga
the
ANO
reve
a
hig
T
0
'ii:'*~
ca~e
[F(2
16)
=
24.
<
Ind
t
d
.
h
h
f
.
y
s
es
'~k3!
'!fI:&.
i
"
7,ft:~r:
:he
.
i("f"~,
£'~;~;
Mean Number of Letters Reported per. Trial by
Subject and Letter Category in Experiment 1
-
first experiment, however,there was no difference be-
i;.,}c
LetterCategory
than color letters [F(I,8) = 26.28,p < .001].This wastrue
for each of the 9 subjects (see Table 4). In contrast with the
NaS!
c;,;;~i
.thef~,
verec;
';:cJI:
;~,!~~tjftf:
2
3
0.38
0.09
0.03
tralletters[F(I,8)=0.30].Inordertoaccountforthedis-
0.85
0.72
0.55
0.00
0.53
0.0 I
appearance
of the color letter effect,
it is reasonable
to
assume that the increased
difficulty
of the task reduced
ti
¥"!i{",c c
' ~r::I\! Ftlr,j~,Jic.:
4
5
0.71
0.85
0.48
0.43
0.38
0.46
t. e I e I 00
e, lRc; ,i~:?2\;c
6
0.60
0.09
0.06
tlon letter. As a result, floor effects may have masked the
s di-:,'~j'i;;;§~~\'c\!;;l~i;;
7
0.83
0.62
0.54
possiblesmalldifferencebetweencolor lettersandneutral
0.79
0.16
0.14
letters. The ANOVA
: one;~
:tit~t~~;b~~
etters m add Itlon
..
to t h e Ioca-
0 reporting
also revealed a significant
main ef-
000 1] in dl.cat
hite) 1~ j;;c resultedfrom.o~ject-based pro~esse.s,rat~er than reflecttter cat~;;'c;;~..~.oing
the superIorIty of the location dimension. In order to
categories [F(I,8) = 1.69, 1.96, and 3.03 for the differences between location and color, color and neutral, and
)ariso~!\'\c.oeliminatethis possibility,.we.disruptedthe pote~tial for-
locationandneutrallettercategories,respectively].Except
e 2). Th~,:icvo,posed
onthe lettersratherthan en~losmgthem. Note that
experiment,in which the shapesand the letters were su-
\J
']c\l;,c:c shape
andthelett~rwasnot likely to .dlsruptperformance. effectthan m ExpenmentI, m which the shapesenclosed
;,L."
>Ll
:.~~~,ij",nt""tl~r and location responses: T~ble 3 shows the
X
"LJ
:f:~. ;~,:'.'
i,{~;~!ii~~'r~he
~ubJects
~eported
only 0.64 relevantletcc~,~1al.ThI
S redu t
.
!.x~
;~';
5 .
0.2 "
f,;' ;'l1i~~~P~ltlon
c.""
,,";".'C"~r.anal
'!('!;;);'/~:in$:';:
of the shapes and the letters made the
Ica e
at, as m t e Irst ex-
~
0
0
E
R
Figure2.Examples
of thestimuliusedin Experiment
2.
'.-+0'
,j
e
res:,ci'"rqm the analysIs (7% of the trIals
neutral,S,;'
. ie 'J:':;
jCr c...
ponses I
d
d
provi
~n
las~s
strategic
minimize
us
t
0
met
In thepresentexperIment,the freereporttaskwasreplaced
IS
.
..
I~n
;As'tn'
.i~~\:;i!
m the:o\W~~ic",: ,:,:~~§e~on the letter, as ISshown In Figure 2.
::
:i~j R:es
rt fo~j1"i;;""!'
u ts and DISCUSS'
ppo
;
..
TSAL AND LAMY
3
gories
LetterCategory
Red
Biue
White
Mean
Location
Color
Neutral
0.51
0.09
0.09
0.52
0.09
0,09
0.22
0,17
0,13
0.42
0,12
0,10
performance
letter
(I.e.,
with
~hen
appeared
that
when
it
e
er
was
the
tar~et
a color
letter
the
(i.e.,
y
d
'
I
was
of
letter
'
compare
location
was
equa
d
vve
the
m
was
ge
"'1
f
'
entl
a
re
,
'",
Jects
and
t '
a Ive
I
0
er
t
oca
fi
'
.
Ion
m
I
e co
or
to
to
the
the
d
or
tt
shared
the
Id
e
ers
t
er
,',
location
color
rt
th
0 repo
t
sugges
of the
dimension
,
I
e co
b
a
designated .
or shift
I
-
su
color
to an-
tt
or
e
e
that
t
other
Results
th
an
d
up
e
k
any
,&4",
"
.
position
outside
the
.
.
en
(
Ion.
"
the
4
.
location
supenonty
effect
was
".,
not
.
contamInated
by
eye
movements.
Note1988)
agaInthat,
unlikem
our prevIous
stu
y
sa
avle
m t h e th ree
expenmen
ts
d
(T
I &
.
L
'
'
. "
reported
'
h .
above,
the
subjects
,
could
' h
I
not
benefit
' ,
.
,
from
shlft-
,
mg t elr eyesto anypenp era posItionprIor to stlmuIus onset,sincee~c~positionwa~equallylikely t~ in-
er.
Method
Subjects,The subjectswere 10 undergraduates
from Tel Aviv
University,who participatedaspart of a courserequirement,Five
subjectsalso participated in Experiment 2, All had normal or
corrected-to-normal
vision,
Sti~uli and Procedure.The ~splayswer~identicalto thosein
~xpenment2, .exceptthat ea~hdIsplaycontalne~a targetletter,H
In half of thedIsplaysandU In the otherThhalf, ThIs
targetletter
was,
b
Idf
k I to appearIn
. anyposItIon.
..
equaIIy Iley
e su ~ectswereto Irst
to reportthe shapeof a givencolor andthento indicatewhichtar'
'
a
.
m
EXPERIMENT
target
t th
wou
tt
0
I
,
appearIng
f
ocus
;
I
th
0
.'
attend
switch
InitIally
need
to
th
t
\%!;~ Ne
shape)
.
I
ers
",c
1-
shape'scolor). Supenorperformance for the location lett
target
a location
target
,;~
'
IS
.
position.
,
II
'
clude the target. Stll,l, I,tmay be poss~blethat some su?Je~ts
adopted an eye-shifting strat~gy m order to m~xI~lze
performance on part of the trIals, In order to elImInate
this possibility, in the present experiment we adopted van
der Heijden et al.'s (1996) method for minimizing eye
movements by having subjects respond to a sign appear-
g men th
'
m
ce
t
of
th
er
.
cessmg.
d
e
'
spl
I
d
ayan
'
requi
r
.
g fiove
-
I pro
m
Results and Discussion
UnIversItywhopartIcIpated
aspa.Tt,ofa
courserequIrement.
All had,
the
first
two
.
experIments,
'
'0
we
excluded
.
trIals
m
ported n
p < .00]
normal
orcorrected-to-normal
:;c
Sti.mu
I.I and Procedure, ThVISIon,
e st.ImuII' and procedure were,den-\!!~
F(4,40,
orl
' ty of
'
~
of the trIals) and Included only responses for neutral let-
jects were presented with 60 trials in each block, Second, a gray
ters that were sup~rimposed on a shape. ~able~ ~ an? 6
show the proportions of correc,t target Identification
when the target letter was a location letter, a color letter,
and a neutral letter.
sjgn~as now presented.in
the centerof the display..It waseithera~;
~Iussign (~O%of the tnals) or an X (20% ofth~ tnals), ~he sub-;~
Jectswer~Instructedt? respond?nly on thosetnals that !n.clude.d~;
the plu~sIgn~ndto.wlthhol?the~rresponses
on the remainingt~-":
The
AN.oV.A,
gory
was
,
significant
that
[F(2,
18)
the
=
effect
19,40,
of
P
<
ther tests revealed that, as in Experiment
was
completely
accounted
locati~nresponses:
The
when
It was
a location
for
target
letter
by
was
than
the
letter
cate-
,0001],
Fur-
2, this effect
superiority
ide~tified
when
It was
of
color
1,453].identifying
The ANaVA
often
a color
letter
targets
and
targets
[F( 1 ,9) =
revealed
noneutral
signi.ric~~t
eff~ct
of
den
et al. (1996),
to process
the
We
central
followed
symbol
their
prior
logic
in assuming
to response
that
would
the
prevent
Mean
Number
of Letters
4
Reported
per
Trial
0,53
0.5\
0,09
0,11
0,\2
0.i2
P
3
0,\6
0,12
0,07
4
0.22
0.08
0.\3
5
0.63
0.20
0,\4
6
0.52
0.00
0.01
7
8
9
0,38
0.17
0.56
0,\4
0.\0
0,18
0,\0
0.13
0.\4
the
target
h
t
letter
h'
e w
Ite
for
the
various
h
I
t
arget
etter
t
an
in
categories
fi
or
h
t
ere
identifying
were
d
smaller
d bl
an
for
I
ue
target
et-
ters, Indeed, additional comparisons between responses
for the white target letters revealed no significant differpn~p~ hptw~~n r~n()rtinll letters from the various cate-
,
" --~--
--~~~-
;',
by Subjectand Letter Categoryin Experiment2~:
I
2
differences
anadditi
and indi
:,
action between color and letter category [F(4,36) =4,48,
the
sub"c-
Inoveme
Table
color [F(2,18)= 0.93],but therewasa significantmterthat
,I 0)
spective
~
Location
indicating
<.1
needc~!
jects from shifting attention to a peripheral position.
Subject
tween
,005],
[F,eto r
the
more
[F(I,9) = 21.11,p< ,005],andtherewasnodifference
be-
<
when th
shapett
color jet
between
c'tiv
als, This manIpulatIon was IdentIcal to that used by van der HeIJ'~
Indicated
Letter
Category
Color-
Neu~"~
th
29
[ ' 5], ~
~
which t~e shape ~fthe color was named Incorrectly (5>10 tical to thosein Experiment2, with two exceptions.First,the sub-{
,
of
c..cr
le
th""IDall '
'
~ubj~cts, The s~b~ectswere 11 undergraduate~from Tel AvIV
,.'
m
Ity
"
1'
Method
.
j,
a,
get letter (an H or a.U) ~ey saw. In a~1other ~espects,the stimuli
andProcedure were IdentIcal to those In Ex penment
2,
As
by]
11j"" Co
As
ensure
a.
any
. .ty
I
.
dls-
er
or
m
~~~~
furth
cation
appear
"~
fit!'
receiving
0
an
'
to
be-
location.
epic
the
the
0
name
tt
t
me
I
I
its
a region
b
'
d
con
k
t
ley
to
whether
tar
to
in
t
.
ay
'
ley
report
'
d
l
Th
differences
eslgne
p
to
U
the
directed
more
required
then
H
for
appearing
l .k
t
were
and
d
was
target
was
subjects
color
'
spotlight
a location
prIOrI
The
given
ta
attentional
was
a
'
2,
of
I
the
a result,
Th
shape
2,95
'
.
Expenment
f .
Irst
'
and
iment minimized strategic factors, the most parsimonious
explanation for the location superiority effect obtained
h
h
I
f
h
d d
here IS
' h
t at w en t e co or 0 a given s ape was atten e
attentlona
.
3.35,
Because by eliminating free report, the present exper-
to,
,
.05,
'
.
=
[F(I,9)
tween location and color letters, color and neutral letters,
and location and neutral letters, respectively].
expenmen
Table
Mean Number of Letters Reported per Trial by
Letter Category and Color Block in Experiment 2
present
964
;1
CC,
::'
"
i
,
SUPPORT
el;
FOR
LOCATION-SPECIAL
VIEWS
OF
ATTENTION
965
it;
~
I
rences be, ~
ralletters, ~;
~
-
ent
experiJ
.
.
Category Red
~r.
t
Color
of,
Neutral
s a~ended'
'catlo?.As
~recelvmg
wd ~p than
lutslde
!,
~
:1,
the!
Mean
.
.37
-.
80
.34
.57
.43
.39
.39
.36
.38
.
56
nd in whlch
the
~ncorrectly
shape
(4%
of the
of the
and
target
6%
deslgnated
trials).
in order
color
We a~alyzed
that
the
these
conditions,
the . sub-
to process
that
narned
mergmg
object
of
would
have
the formation
of response-relevant
rep~ence,
the results
could
not reflect
the opobject-based
processes
but,
rather,
reflected
space-b~sed
processes..
evIdence
under
very
location
constrain~d
supenonty.
conditions,
minimized
.penme~tsal!
from
S.hlft..1:~'
perimposed
viously
due
r to
to responding
Sti~U'~,
reported
on a shape).
to the need
to
only
irre.ievant
the
0.25
and
8.
relevant
neu~ralletter~,
Thls
substantial
t.o process
the
penpheral
On
n?t
reducti.on
lS ~bce~tral
Slg~ pnor
targ~t.
Still,
desplte
color.
This
experime~t
ellmmate~
posslble
tha!
the
receiving
attentional
likely
to be picked
~he
other
position
target
free
strateglc
loc~tio.n
outside
the
lt occupl~d
report
effects.
target
pnonty
than was
up
whe~
co~ld
suggest
subjects
7
a predeslgnated
It showed
subj.ects
suits
(excluding
trial
Tables
detect
3 provided
fore,
the
ters per
m
easily
for
Experi~en.t
su-
the average,
.0ur.prevI',::
Itammate.d
presented
more
Any
perceptual
let-
l:
are
to
mdependently.
cohesive
with
stronger
that
them
a single
the location of the targe! I~tter than when lt sharedltS
results.
on a s~ape
of
even
responses
into
make thesecomparablewlth the c.olorand location reThe
supenmposed
re-
alarms)
was
only
fact
:f
sponses.
w~s
was
false
the
for the neutral
~
that
central
to
:,f
~nsure
letter
the
mis~es
be attributed
ratherhad
than
jects
to enclosin
segregate the.superimposed letter and ~lgure
interfered
resentations.
eration
of
(3%
not
g it. Under
74
Results
and
Disc.ussi.on
.
We excluded
tnals
m WhlCh
t.o incorrectly
could
suIt whenthe shapewas supenmposedon the letter
the two
f
her
White
results
target shape and the enclosed. location le~ter fo~ed a
perceptual object, since Expenm~nt 2 repllcated thlsre-
--
sponded
!
Blue
-e
.
86
LocatIon
Slmon.lolls
t obtamed
These
5
Mean Proportionsoi~:t;ers Reportedper Trial
by Letter Categoryand Color Block in Experiment3
and,
~here-
He?ce,
appeared
lts.re-
m a reglon
and was
there~ore.more
a target
app~anng
m a~y
focus
of
attenti?n.
Ex~en-
'-'
kely t~ mol:
very low level of correct letter reporting, a.clear locat~on
ment 4 eliminated the possibility that the subjects shlfted
rle su?Je~ts.'!.
superiority
their
'
~
m~xl~lze
ellmmate
effect
~
eye
from
these
tables..
T~lS
~bservation
wasconfirmed by an overall ANOVA, 1Odlcatingthat the
~
d~p.tedvan ~
nlzmg
emerges
of letter
category
was
significant
[~(2,20)
':"
22.25,
eyes
to a peripheral
position
prior
to displ.ay
onset.
In
this experiment, the subjects were presented wlth a central
sign
and
were
required
to identify
i~.before
respondin~.
p < .OOOl].Furthertestsrevealedthat,as10Expenme~t2,
We followed the logic of van der HelJdenet al. (1996).10
this
assuming
effect
was
completely
accounted!or
by
the
supenor-
that
the processing
of this
s':(mbol.wo~ld
requlre
gn appear-~
ity of the location responses.Location letters were re-
foveal processing and thus necessltateflxation at the
foveal
ported
center
proof
t
color
.
Aviv~
~:t~~11had'f
iden-
;irst,
the sub-]
:cond,
~ gray
t was
tr
.,
i
[F(4,40)
j
ority
=
of
and
~
~',
color
letters
letters
ll.
79,p
the
<
.0001],
letters.
[F(l
[F(.i,l?).
,10)
= 23.~6,
= 0.05].
target
was
mdlcating
over
white.
color
Indeed,
was no difference
rep?rts
[F(I:lO)
between
report1Og
location
.
tlve to neutral
letters
were
i- ~c
Agal~
the
between
= 0.55],
letters
only
that.the
supen-
letters
dls~ppeared
for
I prevent
spectively].
sub-.,
The
an additional
and indicate
,
:,
rial
movement
results
of the
replication
that
this
white
the.
more
target
This
and .color
letters
margma
y slgm
II
.
rela-
.f lcan
'
t
present
09,.re-
expenment
of the location
effect
could
not
provlde
superiority
be due to
g::4
:::
DISCUSSION
0.13
-~:.!~--~~
findings
views
provide
of visual
strong
att~nti°.n'
support
::
!i"
first
shape
of a given
color
experiment
showed
that
than
the
between
may
of the
stantiall
r
y more
e duction
in
and
the
letter
that
had
the
They
th~n
show.
report
subjects
target
neutral
dlfflcult
ThlS
effe~t
most
shape
I and
by the fact that
the letters
mad~
the
other
.ex-
the supenmthe task sub-
10 the latter
expenments
(as the
P erformance
m these
expenments
"
suggests),thusreducingthelikellhood?fr~port1Og1etters
'-.l
. .
.
in addition
to the
location
letter.
Table
This
fmdmg
further
of Letten
Re~orted
per Trial
3
Color
Neutral
.62
.47
that
2
.90
:;g
:;~
~
:1~
.00
.00
5
.87
.68
.62
6
.88
.77
.24
letters.
7
.92
.52
.~~
often
8
.;;
:~
:04
:67
.29
.38
I~
em-
6
.92
s color.
,
com-
expenments.
Expenment
be explained
s~a~es
a~d
overall
letters.
three.subsequent
Location
reported the letter that was enclosed by the targe,t shape,
;P ;ii::rather
~itj;
the
than
I
In all four experiments, the subjects were mstructed
to name
::
!...c,:he
"J,
color
for
direct attention to its location...
: i:
:
Although
by Subject and Letter Category In Experiment
Letter Category
Jk~:: ~henselecti.nga t~rgeton .thebasIsof its color, subjects
0.01
0.10
present
)ocation-spe.cial
o.o~
on performance.
in all
difference
Subject
The
letters
disappeared
Mean Proportions
GENERAL
i~
0.12
shape
effect
an eye
~%
0.12
display.
target
color
periments
position
strategy.
~{
-~~~}
the
of the
pletely
location.
letter and
and the dlfferences
[F(I,10)=3.83,p<.08;andF(I,10)=3.5~,p<.
f
need
of
Another importantfinding is the minimal effectsof the
color
~
vanderHeij-~'
~
neutral
location
shape, there
color letter
~
emammg
that
letters
when
d ed
thatmclu
than
themain effect of colo~was not slg?lflcant [F(2,20) - specified the target and was thu.sthe o?ly ta~k-relevant
2.95], but this factor Interac.te~wl.th letter categor.y dimension,only in the first expenmentdld ~ubJects
report
~
~~hs~b~
lis).~
frequently
p < .00I], and therewasno differencebetweenreportl?g
i:
re were
more
.
966
.
TSAL AND LAMY
Table7
MeanNumberof LettersReported
perTrial
byLetterCategory
andColorBlockin Experiment
4
LetterCategory Red
Blue
White
Mean
Location
Color
Neutral
0.19
0.02
0.02
0.24
0.02
0.02
0.08
0.07
0.04
nificant only for the red andthe blue targetshapesin Experiment 1. Hence, once the differential color salience
wascontrolled for, the location superiority effect wasobtained for all three colors of target shapes.
0.17
0.04
0.04
The aboveanalysesmakeit possibleto distinguish between bottom-up and top-down influences on letter reports.The resultsof all the experimentsclearly showthat
phasizesthe superiority of location letters overcolor letters in showingthat attending to the location of the target shapeis a mandatoryprocess.
A general problem involved in comparing the efficiencyof selectionby location versusthat by color is the
differential discriminability of valuesalongthesedimensions.In principle, one can neverbe surethat the particular values selectedalong each stimulus dimension are
equallydiscriminable.For example,is the discrimination
betweenr~dandblueaseasyasthediscriminationbetween
locationsseparatedby 2° of visual angle?This problemis
commonto many studiesthat compareefficiency of selection on the basis of different dimensional cues(e.g.,
color and location). However, discriminability along a
particular dimensionmay affect performanceonly if the
task involvesa judgment along this dimension.Because
subjectshad to respondto colors and becausethe locations theseo~cupiedin spacewere completelyirrelevant
to taskdemands,whetherthe particularpositionscontain-
the perceptualsalienceof lettershada considerableeffect
on performance(bottom-up influence), since the salient
white letter was reported substantially more frequently
thanthe redandthe blueletters.The importantpoint, however,is whether there was an interaction betweencolor
discriminability and top-down factors. That is, was 10cation superiority overriddenby top-down factors (task
relevanceof the color sharedby the target shapeand the
color letter) when color discriminability was highest?
That is, for white letters,werecolor lettersreportedmore
often than location letters? Inspection of the rightmost
threecolumnsof Table9 clearly showsthat the superiority of location letter reportsover color letter reportswas
highly significant, evenfor the very discriminable
target. Note that in the last experiment,this superiority
wasevenstrongerfor the white targetthanfor the bluetarget. From this analysis,we may now concludethat when
selectingthe color of a given shape,the subjectsattended
to the location of this shapemorethan to its color, this eF
fect beinghighly significant whetherthe color washighly
or poorly discriminable.Therefore,althougha direct com-W
\
ing the letters were more or less discriminable than the
parison between the discriminability
particularcolor valuesusedin thepresentexperimentswas
not likely to haveany effect on performance.
Despitethe argumentsabove,one can still arguethat
the location superiority demonstratedhere may be contingent on the choice of the particular colors usedin our
experiments.Indeed,in all four experiments,the superiority of location letters over color letters was either reduced or eliminated when the target shapewas white.
color is not possible,the resultsaboveshowthat location;
superiority remainshighly significant acrosssubstantial ~
variationsalong the discriminability of the color dimen- 1
sion. Obviously,we would expectthat despitethis supe-]:
riority, extremeconditions that impair location processing while facilitating color processing might override;:,
location superiority,but we would considerthat to be the::~
exceptionrather than the rule.
I;
of location and
items. It is thereforereasonableto assumethat the color
letter was reported more frequently in the white shape
conditionthanin the red or blue shapeconditionsbecause
it waswhite andthus more salientthan becauseit shared
ority demonstratedin the present experimentsand the 1
color superiorityobtainedin the study of van der Heijden!~
et al. (1996)canbe explainedbestby the fact that the lat- iJ
ter includedunnecessary
biasesagainstlocationselection,::
the attendedshape'scolor. This possibilityis strongly
!,
'E'
Whiteitemsweresubstantially
brighterthanredor blue
supportedby subsequentinspections of the data. They
indicatedthat in all the experiments,the white letter was
reportedmore often than the red and blue letters across
letter categories.In order to eliminate the effectsof differentialsensorys~lience,we conductedadditio~alanalysesin all the expenments.Wecomparedthe relatIvereport
frequencies
by lettercategory,separatelyfor thewhite,red,
andblue letters,so that the comparedlocation, color, and
neutral letterswere all of the samecolor. The resultsare
summarizedin Table9. As the table shows,in threeof the
four experiments(Experiments 1,2, and 4), location let-
Theevidentdiscrepancy
betweenthelocationsuperi-!
MeanNumberOf~a;~~
8Reported per Trial
bySubjectandLetterCategoryin Experiment
4
LetterCategory
Subject
Location
Color
Neutral
I
0 03
0 0I
0 02
2
0:17
0:00
0:06
3
0.25
0.08
0.04
4
0.21
0.03
0.02
5
0.28
0.01
0.01
ters were reported more frequently than color letters for
each of the three target shape colors, and this effect was
highly significant. Moreover, the difference between reporting color letters and reporting neutral letters was s.ig-
~
~.~~
8
9
10
iI
0:17
0.01
0.30
0.15
~.~~
0:09
0.02
0.01
0.02
~.~:
0:10
0.01
0.03
0.03
I
dimensi<
ThepI
viewsof
Berge&
1988,1
se)1t~ind
1111-,te
1~essen
tlonsin"
at reg.io
~ttentlo
tlcularfe
the.mast
catIon(1
,
Vand
thatnow
the~rie
dec
deuc
can1
oduc
etha
nce.l
stitu
nt iQ
arate
aI, l~
proc
ceptu
,
--~~~~-~~-~-
"
intr<
ho
-ses
a p<
Ie 10
Ieco
ous
he p
spec
I; Bl
enik
proc
hing
basis
one
cientbut
c"
3"
SUPPORTFOR LOCATION-SPECIAL VIEWS OF ATTENTION'
967
Table9
StatisticalSignificanceoftbe Differences
for Different Categories,Within EachColor
LocationVersusNeutral Color VersusNeutral
LocationVersusColor
Experiment Red Blue White Red Blue White Red Blue White
I
--
--
--
-
-
n.s.
--
--
--
2
3
4
-n.s.
-
---
---
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
n.s.
-n.s.
-
--
-n.s.
--
-approachedsignificance(p < .1). --highly significant (p < .0I).
icant.
n.s.,nonsignif-
tion of internal structuresrepresentingtheseselectionattributes,but by increasingthe sensitivity of the locations
the items occupy in space.
REFERENCES
BAYLIS,G. C., & DRIVER, J. (1992). Visual parsing and response competition: The effect of grouping factors. Perception & Psychophysics,
51, 145-162.
BAYLIS, G. C., & DRIVER, J. (1993). Visual attention and objects: Evidence for hierarchical coding of location. Journal of Experimental
Psychology: Human Perception & Performance, 19,451-470.
BROADBENT,D. E. (1971). Decision and Stress. London: Pergamon.
BUNDESEN,C. (1990). A theory of visual attention. Psychological Review, 97,523-547.
CAVE, K. R., & PASHLER,H. (1995). Visual selection mediated by location: Selecting successive visual objects. Perception & Psychophysics,
57,421-432.
CAVE, K. R., & WOLFE, J. M. (1990). Modeling the role of parallel processing in visual search. Cognitive Psychology, 22, 225-271.
DUNCAN,J. (1981). Directing attention in the visual field. Perception &
Psychophysics, 30, 90-93.
DUNCAN, J. (1984). Selective attention and the organization of visual
information. Journal of Experimental Psychology: General, 113,
501-517.
HARMS,L., & BUNDESEN,C. (1983). Color segregation and selective
attention in a nonsearch task. Perception & Psychophysics, 33,11-19.
HOFFMAN, J. E., & NELSON, B. (1981). Spatial selectivity in visual
search. Perception & Psychophysics, 30, 283-290.
HUMPHREYS,G. W. (1981). Flexibility of attention between stimulus dimensions. Perception & Psychophysics, 30, 291-302.
KAHNEMAN, D., & HENIK, A. (1981). Perceptual organization and attention. In M. Kubovy & J. M. Pomerantz (Eds.), Perceptual organization (pp. 181-211). Hillsdale, NJ: Erlbaum.
KIM, M., & CAVE, K. (1995). Spatial attention in visual-search for features and feature conjunctions. Psychological Science, 6, 376-380.
LABERGE,D" & BROWN,V. (1989). Theory of attentional operations in
shape identification. Psychological Review, 96,101-124.
LUCK, S. J., FAN, S., & HILLYARD, S. A. (1993). Attention-related modulation of sensory-evoked brain activity in a visual search task. Journal of Cognitive Neuroscience, 5,188-195.
NEISSER,
U. (1967). Cognitive psychology. New York: Appleton-Century-
Crofts.
POSNER.M.I., SNYDER,C. R. R., & DAVIDSON,B. J. (1980). Attention
and the detection of signals. Journal of Experimental Psychology:
General, 109,160-174.
THEEUWES,J. (1989). Effects of location and form cuing on the allocation of attention in the visual field. Acta Psychologica, 72,177-192.
TREISMAN,A. M. (1988). Features and objects: The fourteenth Bartlett
memorial lecture. Quarterly Journal of Experimental Psychology,
40A,201-237.
TREISMAN,A. M. (1990). Variations on the theme of feature integration:
A reply to Navon. Psychological Review, 97, 460-463.
TREISMAN,A. [M.], KAHNEMAN, D., & BURKELL, J. (1983). Perceptual