Acta Psychologlca North-Holland
73 (1990) 159-170
EFFECT OF CONFLICTING CUES ON INFORMATION PROCESSING: THE ‘STROOP EFFECT’ VS. THE ‘SIMON J. Richard
SIMON
The Unrversrt~
of Iowa, USA
Accepted
February
EFFECT’
*
and Kevin BERBAUM
1989
This study exammed the relatlonshp between two sources of mterference m human mformatlon processmg the Stroop effect and the Slmon effect. Forty SubJects pressed a left- or right--hand key m response to a Stroop color word located on the left or right side of a screen For one group, mk color was the relevant cue and, for another group, word meamng was the relevant cue Independent variables were congruence, 1 e , agreement or lack thereof between the mk color and meamng of the Stroop word; spatial correspondence, 1.e) agreement or lack thereof between the locatlon of the Stroop word and the locatlon of the key used to make the response, and stimulus duration, 1 e ,400 or 100 ms Each of these vanables had a slgmflcant effect on RT, and there were no slgmflcant mteractlons According to Stemberg’s additive-factor logx, these fmdmgs suggest that the Stroop effect (congruence) and the Slmon effect (spatial correspondence) mvolve separate stages of processmg. If one assumes that mampulatlon of stimulus duration affects the encodmg stage, then results also suggest that neither the Stroop effect nor the Slmon effect mvolves the stimulus encoding stage
This research employed the Sternberg additive-factor method to determme the relationship between two potent sources of interference in human information processing, the Stroop effect and the Simon effect. There is an extensive literature concerned with information processing in situations where the stimulus provides relevant as well as irrelevant cues. The most widely used stimulus of this sort is the Stroop color word (Dyer 1973; Jensen and Rohwer 1966). In a typical Stroop * This research was supported by funds from the Graduate College of the Umverslty The authors gratefully acknowledge the assistance of Todd Bever and Renet Snuth Requests for repnnts should be sent to J.R. Simon, Dept of Psychology, University Iowa City, IA 52242, USA
OOOl-6918/90/$3
50 0 1990, Elsevler Science Publishers
B V (North-Holland)
of Iowa of Iowa,
160
J R Smon,
K Berbuum
/ Effect ofconfhrrng
cues
task (Stroop 1935), the stimuli consist of names of colors printed in other colors. Subjects experience interference when they attempt to name the ink colors and ignore the words. Despite literally hundreds of studies, there is still a controversy regarding the source of the Stroop interference effect (Dyer 1973; Glaser and Glaser 1982). The traditional explanation is in terms of response competition (e.g., Dalrymple-Alford and Azkoul 1972; Keele 1972; Klein 1964; Morton 1969; Warren 1972). It asserts that the Stroop effect is due to the lack of correspondence between the irrelevant attribute of the stimulus and the response. An alternative explanation is m terms of encoding. It argues that the Stroop effect is due to the lack of congruence between the relevant and irrelevant cues in the stimulus. Hock and Egeth (1970) propose that the interference occurs during perceptual encoding; i.e., the color word distracts from the identification of the ink color by divertmg attention from it. Seymour (1977) argues that the interference occurs during conceptual encoding which is a stage between perceptual encoding and response activation when the color information comes in contact with semantic memory. Simon and his associates (e.g., Craft and Simon 1970; Simon 1969; Simon 1970; Simon and Rude11 1967) have used a paradigm that is similar to the Stroop paradigm to investigate the effect of conflicting cues on information processing. In a typical task, subjects might press a left- or right-hand key, depending on the color of a stimulus light which appears on the right or left side of a display panel. Results indicate that the location of the light provides an irrelevant directional cue that interferes with processing the relevant symbolic cue which is the color of the light. In other words, reactions are faster on trials in which the location of the stimulus and response correspond than on trials in which they do not correspond. Hedge and Marsh (1975) have used the term ‘Simon effect’ to refer to this typical finding, and we use that label here as a convenient short-hand designation for the phenomenon. Simon et al. have suggested that their effect reflects a stereotypic tendency to respond toward the source of stimulation and, in a series of studies they have provided evidence that the interference is located in the response selection stage (Acosta and Simon 1976; Mewaldt et al. 1980; Simon 1982; Simon et al. 1975; 1976). Since the Stroop effect and the Simon effect are so similar, superficially at least, and since there is some evidence that both Stroop and Simon effects might involve the same processing stage, it is of interest
J R Swnon, K Berbaum / Effect
of confktmg cues
161
to determine the relationship between these two potent interference effects. Sternberg’s additive-factor-method (1969) provides a powerful tool for studying the stages of information processing and defining the processing which is accomplished by certain stages. The method assumes that the interval between the presentation of a stimulus and the subject’s response is filled with a sequence of independent stages or cognitive processes. Each stage receives an input from a previous stage, performs a transformatton on that input, and passes it along to the next stage. Total reaction time (RT) is simply the sum of the stage durations. When an experimental manipulation affects RT for a task, it does so by influencing the durations of one or more of the stages. Thus, tf two experimental variables affect the same processing stage, both variables will be contributors to the duration of that stage, and their expected effect on mean RT 1s an interaction. On the other hand, if two experimental variables affect two different stages, they will produce independent (additive) effects on total RT. In other words, the experimental variables should not interact in a statistical sense. The additivefactor method, then, involves conducting factorial experiments to determine which variables produce significant interactions and which do not. The variables which interact are assumed to affect a stage in common, whereas the variables which do not interact are assumed to affect different stages. Specifically, m the present study, a significant interaction between the Stroop effect and the Simon effect would be interpreted as evidence that they affect a common processmg stage whereas no interaction would suggest that these two phenomena involve separate stages. Stimulus duration was also manipulated in the present study in an attempt to locate the source of the Stroop and/or Simon effect.
Method SubJects The
subjects
mtroductory English
speakers
visual acuity. mk-relevant
were 40 undergraduates,
psychology
course
and reported
20 males
at the University normal
and 20
of Iowa.
females,
enrolled
All subjects
color vtston and normal
or corrected-to-normal
Half of the males and half of the females were assigned at random condition
and the other half to a word-relevant
m an
were nattve
conditton.
to an
162
J R Svnon, K Berbaum / Effect of confktrng cues
Apparatus The apparatus measured keypress RT to the onset of a visual stimulus which appeared on a 60 cm square glare-resistant glass screen located 82 cm from the subject’s eyes and perpendicular to the line of sight. Nme Kodak Ektagrapmc prolectors with high-speed shutters attached (Urnbhtz Model 225 LOAXS) were located behind the screen and used to present the sttmulus materials Each projector contamed a single stimulus shde. Openmg the shutter m front of a projector resulted m the stimulus appeanng on the screen. The opening and closmg times for the shutters were 3 and 4 ms, respecttvely. A Digital Equipment Corporation (DEC) PDP 1 l/34 laboratory computer provided overall control of stimulus presentation and data acqutsttton The computer was interfaced with a specially-built shutter controller and was programmed to operate the shutters, thus determining the sequence of stimulus presentation, duration of exposure, and intervals between stirnull. The apparatus, then, formed a precise, highly flexible, multichannel tacmstoscope. The procedure elimmated stimulus rise time as a complicating factor smce prolector bulbs remained on throughout the session. Also, smce the stimulus slides did not move durmg the expertment, perfect sttmulus alignment was mamtamed. Two response keys, one on the right and one on the left, rested on a table m front of the seated subject Above each key was a white cardboard label with either the word Red printed m red ink, or the word Green printed m green mk Subjects operated the keys with their right and left index fingers. The computer recorded RT m ms and also recorded which key had been pressed. Smce this mformatton was also prmted tmmediately on a DEC Decwnter III, the expenmenter could monitor the subject’s performance durmg the session. Each tnal began with the presentatton of a small white fixation point with a luminance of 4.5 cd/m2 m the center of the screen at approximately eye level. The stimulus, i.e., the word Red or Green pnnted m red or green mk was then presented on the nght or left side of the screen at a distance of 8.9 cm from the fixation point or an arc of 6.2O at the eye. Henceforth, the subscripts R and G will be used to specify mk color; i.e., Red, denotes the word Red in red mk, Red, denotes the word Red m green mk, and so on. The first letter of each word was 4.3 cm high and the remaming letters were 2.85 cm htgh Stroke width for all letters was 0.79 cm The word Red measured 9 5 cm honzontally on the screen, and the word Green measured 15 25 cm. The lummance of the mdivtdual stimulus words (10.5 and 8.1 cd/m2) for red and green mk, respectively) was adjusted by means of neutral denstty filters and apertures m front of the prelectors so that the mk colors appeared to be equally intense and salient. The viewmg screen was steadily illummated at 3.6 cd/m* (incandescent light) m order to suppress colored after images. Procedure and experrment destgn Taped mstructions mformed subjects that the experiment would consist of a number of trials, each tnal begmnmg with a small white dot (ready signal) appearmg m the center of the screen. After a short Interval the word Red or Green prmted m red or
J R Smon, K Berbaum / Ejjecr ojconjhcrrng cues
163
green mk would appear to either the right or left of the dot. On some trials the word would appear very bnefly and on other trials it would appear for a longer duration. Half of the sublects (ink-relevant condttion) were told that their task was to respond to the color of the ink; i.e., ‘if the mk color is red, press the key labeled “Red” as quickly as posstble, and tf the mk color IS green, press the key labeled “Green” as quickly as posstble’. The other half of the subjects (word-relevant condition) responded to the identical sequence of trials but were told that then task was to respond to the meanmg of the word; i.e., ‘if the word is Red, press the key labeled “Red” as quickly as possible and tf the word is Green, press the key labeled “Green” as qutckly as possible’. For half of the males and half of the females m each conditton, the right key was labeled ‘Red’ and the left key was labeled ‘Green’. For the other half of each sex group, the key labels were reversed. Subjects placed their right index finger on the right key and their left index finger on the left key and were told to keep then fingers resting lightly on the two keys throughout the expenment Each subject performed on the same sequence of 192 test trials m which the 16 possible combmations of the four Stroop words (Reda, Redo, Green,, and Green,), two sttmulus locations (right and left of the ftxation point), and two sttmulus durations (400 and 100 ms) each appeared 12 times m a predetermined sequence. The sequence was arranged such that, in each successtve 48 trials, each of the 16 treatment combmations appeared three times m a random sequence. Pnor to the test trials, there were 16 practice trials and 16 unscored warm-up trials Both practice and warm-up trials included one of each treatment combmation. The durations of the various events m the trial sequence were as follows The ready signal remained on for 500 ms. It was followed by a 500 ms. blank interval after which the Stroop word appeared for either 400 or 100 ms. There was a 2-s interval between the subject’s response and the ready signal for the next trial. If the subject did not respond within 5 s, the program advanced automattcally to the next trial Subjects received no feedback regarding RT or the accuracy of their responses The experimental design consisted of three withm-subjects factors and one between-subjects factor. The within-sublects factors were congruence (i.e., the agreement or lack thereof between the meamng and mk color of the Stroop word), spatial correspondence (1 e., the agreement or lack thereof between the locatton of the Stroop word and the location of the key used to make the response), and stimulus duration (i.e., 400 or 100 ms). The between-sublects factor was the relevant dtmenston of the Stroop word (i.e., mk color or word meamng).
Results
Table 1 shows the mean RTs for the 16 treatment conditions, i.e., congruent (Red, and Green,) and mcongruent (Redo and Green.) Stroop words displayed m a location (nght or left) corresponding or not corresponding with the correct response key, and presented for 400 or 100 ms to the ink-relevant or the word-relevant group. Trials on which errors were made (2.6%) were excluded from the computations. The top third of the table combines the data from the group responding to ink color and the
Noncorrespondmg
Durations
and stimulus
Correspondmg
combmed
correspondence,
450 -435 442
428
437 -420
470 -436 453
454 -428 441
442 427 435
465 449 457
484 466 415
466
447
474 -444 459
475 457
Noncorrespondmg
454 430 442
492 -449 471
413 -440 456
Correspondmg
459 440 450
488 -457 473
461
474 -449
M
428
435 420
421 409 415
449 -423 436
425
441
471 438 454
435 416 -
Correspondmg
of Stroop
453 429 -
Noncorrespondmg
100 ms duration
time (m ms) to mk color and word meanmg
400 ms duration
on reactlon
459 436
M
duration
421
428 414
460 431 445
433
444 -423
M
color word stlmuh
_
Note Congruence refers to the agreement or lack of agreement between the meanmg of the Stroop word and the color of the mk m whrch 1t 1s prmted Correspondence refers to the agreement or lack of agreement between the location of the Stroop word (left or nght side of the screen) and the location of the key (left or nght) used to make the response
Mean
Incongruent Congruent
RT to word meanmg
Incongruent Congruent Mean
478 452 465
453
Mean
RT to mk color
464 443 -
Incongruent Congruent
RT to rnk color and word meanrng combrned
Congruence
Table 1 Effect of congruence,
h
9
3
B
J R Simon, K Berbaum / Effect of confltctmg cues
165
group responding to word meaning The mtddle thud of the table summarrzes the data from the group responding to mk color, and the bottom thud summarizes the data from the group responding to word meaning. An overall analysts of variance indicated that there was no srgmftcant difference between RT to mk color and RT to word meaning (459 vs. 435 ms), F(l, 38) = 0 53. The mam effects of congruence, spatial correspondence, and sttmulus duration were all statrstrcally significant Responses to incongruent strmuh were slower than to congruent strmuh (459 vs. 436 ms), F(1, 38) = 33 01, p < 0.001. Noncorrespondmg responses were slower than correspondmg responses (453 vs. 441 ms), F(l, 38) = 8 51, p c 0 01 Responses to long-duration strmuh were slower than to short duration stimuli (461 vs 433 ms), F(l, 38) = 17.13, p < 0.001 None of the mteracttons between congruence, spatial correspondence, and duration was stattstically stgruftcant which suggested that each factor affected a separate stage of mformatron processmg. (The F values [df 1, 381 for the Congruence X Correspondence, Congruence X Duratton, Correspondence X Duratron, and Congruence X Correspondence X Duration mteracttons were 0.94, 1.19, 1.07, and 3 24, respectrvely. Since the latter of these mteractrons approached srgruficance (p < 0.10) separate analyses were performed for the long and short duration condrtrons, but these, too, faded to show a srgmficant Congruence x Correspondence mteractron.) Separate data analyses were conducted for the group responding to mk color and the group respondmg to word meaning. For reactions to mk color (see middle third of table l), responses to incongruent sttmuh were slower than to congruent sttmuh (474 vs. 444 ms), F(1, 19) = 17.50, p -C 0.001, and responses to long-duration strmuh were slower than to short-duration strmuh (473 vs. 445 ms), F(l, 19) = 9.80, p < 0 01. The difference between noncorrespondmg and correspondmg responses was not srgmftcant (465 vs 453 ms), F(l, 19) = 2.23, p > 0.15. There were no significant mteracttons between any of these three independent vartables. (The F values [df 1, 191 for the Congruence x Correspondence, Congruence X Duratron, Correspondence x Duratron, and Congruence X Correspondence X Duration mteractrons were 0 80, 0.11, 1.89, and 2.88, respectrvely.) For reactions to word meaning (see lower thud of table l), responses to incongruent sttmuh were slower than to congruent strmuh (442 vs. 427 ms), F(l, 19) = 19 37, p < 0.001, noncorrespondmg responses were slower than correspondmg responses (442 vs. 428 ms), F(l, 19) = 11.77, p < 0.01, and responses to long-duration sttmuh were slower than to short-duration sttmuh (450 vs. 421 ms), F(1, 19) = 7 65, p < 0 01 None of the mteracttons between the three main effects was srgmftcant which again suggested that the expenmental mampulattons affected separate stages of mformatron processmg. (The F values [ df 1, 191 for the Congruence X Correspondence, Congruence X Duratron, Correspondence X Duratron, and Congruence X Correspondence X Duratron mteractrons were 0.17, 2.01, 0.19, and 0.56, respectively.)
Discussion
The major purpose of this study was to determine whether the Stroop effect and the Simon effect involved a common stage of
166
J R Smon, K Berbaum
/ Effect
of conflmng
cues
information processing. A second purpose was to attempt to locate the Stroop- and/or Simon-type interference m a particular processing stage. At a superficial level, the interference produced by the mcongruity of ink color and word meaning m the Stroop task is similar to that observed in the Simon paradigm when an irrelevant location cue affects the time required to respond to a relevant symbolic cue (e.g., Simon 1970). Our results, however, suggest that the Stroop and Simon effects involve separate stages of processing. This conclusion is based on the absence of an interaction between congruence (Stroop effect) and spatial correspondence (Simon effect). According to Sternberg’s additive-factor logic, variables that do not interact are assumed to affect different stages. However, because there are possible alternative explanations for the absence of an interaction (Pachella 1974; Sanders 1980; Taylor 1976) this conclusion must be regarded as tentative. The stimulus duration manipulation was introduced in an attempt to specify more exactly the source of the Stroop and/or Simon interference effect. If one assumes that sttmulus duration affects stimulus encoding (the stage in which a stimulus representation is formed), then an interaction of the Stroop and/or Simon effect with stimulus duration would tend to locate the effect(s) m the encoding stage. ’ We ’ In order to mampulate the perceptual encodmg stage, mvestlgators have used various means of degradmg the quahty of the visual stimulus (Sanders 1980), e g , supenmposmg visual noise (Stemberg 1969; Van der Molen et al 1987) or reducmg visual contrast (Acosta and Simon 1976, Slmon and Pouraghabagher 1982, Stanovlch and Pachella 1977) To vary stimulus quahty drrectlj m the present expenment would have reqmred eight addItIona proJectors, Le , a degraded Image of the four Stroop words would have had to be proJected to both the left and right visual fields. We chose, Instead, to mampulate stimulus quahty rndrrectly by presentmg the Stroop words for a very brief duration Psychophysical evidence provided by Kuhkowskl and Tolhurst (1973) and Tolhurst (1973), and neurophyslologxal evidence provided by Enroth-Cugell and Robson (1966) suggest that there are two separate visual mformatlon channels The ‘sustamed’ channel 1s attuned to lugher spatial frequency than IS the ‘transient’ channel whereas the transient channel IS attuned to higher temporal frequency Therefore, the sustained system ~111 regster high detail so long as It does not move or change (turn on and off rapidly) The sustained channel IS the one medlatmg regstratlon of mformatlon (or lcomcally stonng mformatlon) required for reading The transient system ~111 regster moving or changing stlmuh, but only the low spatial frequencies Therefore, a moving or changmg pattern ~111 appear blurry because the fme details are not registered by either system. A short duration stimulus ~111 evoke less sustamed and more transient visual system response than a longer duration stimulus (Breltmeyer and Julesz 1975, Breltmeyer and Ganz 1976, 1977) This IS because the ratlo of steady-state to dynamic stlmulatlon IS lower It has even been suggested that the slope of the onset and offset of a stimulus determmes the relative response of transient and sustamed visual mechanisms (Breltmeyer and Julesz 1975) Thus, wlthm the visual system itself, stimulus duration affects stimulus quality relative to a readmg task which m turn affects the encoding stage
J R Smon, K Berbaum / Effect of confl~crmg cue-s
167
found no interaction of either the Stroop or Simon effect with stimulus duration which suggests that neither of these interference effects involve encoding. Thus, our results agree with previous findmgs indicating that the Simon effect does not involve the encoding stage (Acosta and Simon 1976; Simon 1982; Simon and Pouraghabagher 1978). Indeed, there is considerable evidence that its locus is in the response selection stage (Acosta and Simon 1976; Mewaldt et al. 1980; Simon 1982; Simon et al. 1975; 1976; Van der Molen and Keuss 1981). But what about the locus of the Stroop effect if, as the Sternberg logic suggests, it is in neither the encoding nor the response selection stage? Seymour (1977) suggests that Stroop interference occurs during conceptual encoding, a stage between perceptual encoding and response activation when color information contacts semantic memory. Simon and Berbaum (1988) conclude, based on finding a significant Stroop effect in a retrieval task, that the most likely locus of interference is the stage between perceptual encoding and response selection; i.e., the process of decoding or retrieving information from short-term memory. In the original Stroop study (1935), subjects asked to name the mk color experienced great difficulty from incongruent word meaning, but subjects asked to read the words encountered little interference from mcongruent ink colors. It is worth noting that we found not only the ‘standard’ Stroop effect but also the ‘reverse’ Stroop effect. Our results, then, support the more recent research which indicates that incongruent ink colors can also interfere with responses to word meaning (Flowers 1975; Simon and Sudalaimuthu 1979; Simon et al. 1985; Uleman and Reeves 1971). Another interesting aspect of our results was the finding that reactions were significantly faster to the 100 ms stimulus than to the 400 ms stimulus. Welford (1961) has proposed an interesting explanation for this effect. He suggested that subjects inspect a stimulus until ‘sufficient’ data are gathered to make a response. Reactions are faster to a short duration stimulus because the intake of data is stopped by the cessation of the stimulus, and, therefore, subjects gain little or nothing by delaying their reaction. Studies of the effect of stimulus duration on RT have produced contradictory findings. Some report a tendency for RT to decrease with an increase in duration (e.g., Froeberg 1907). Others, like the present study, report a difference in the opposite direction (e.g., Wells 1913; Botwmick et al. 1958; Gregg and Brogden, 1950). Still other studies have found no significant effects (e.g., Deup-
168
J R Simon, K Berbaum
/ Effect
of confktrng cues
ree and Simon 1963; Tolin and Simon 1968). While the contradictory findings seem to be related, at least in part, to the specific durations studied and to the complexity of the task (Welford 1980), the conditions under which stimulus duration affects RT remain unclear. There is a wide variety of tasks in which irrelevant cues intrude to affect information processing. Several recent studies have focused on the effect of stimulus-response compatibility in these Stroop or ‘Stroop-type’ tasks (Green and Barber 1983; McClain 1983; Simon and Sudalatmuthu 1979; Zakay and Glicksohn 1985). Harvey (1984) suggests that use of the term ‘Stroop-type’ should be avoided since it implies that the interference may have a different etiology from that giving rise to the color-word effect. He suggests that, until contrary evidence is provided, rt is more parsimonious to credit Stroop (1935) with discovering a general effect. Results of the present study strongly suggest that it would be a mistake to lump together indiscriminately all interference phenomena under the heading ‘Stroop effect’ since the locus of the interference; e.g., encoding or response selection may differ for different ‘Stroop-type’ tasks.
References Acosta, E Jr. and J.R Bmon, 1976 The effect of Irrelevant mformatton on the stages of processmg. Journal of Motor Behawor 8, 181-187 Botwmck, J., J R. Brmley and J S Robbm, 1958 The mteractlon effects of perceptual dlfflculty and stimulus exposure time on age differences m speed and accuracy of response Gerontologa 2, l-10 Breltmeyer, B G and L Ganz, 1976 Imphcatlons of sustamed and transient channels for theones of wsual pattern maskmg, saccadlc suppresslon, and mformatlon processmg Psychologxal Review 83, l-36 Breltmeyer, B.G. and L Ganz, 1977 Temporal studies with flashed gratings Inferences about human transient and sustamed channels Vlslon Research 17, 861-865 Breltmeyer, B and B Julesz, 1975 The role of on and off transients m determmmg the psychophysical spatial frequency response Vlslon Research 15, 411-415. Craft, J L and J.R. Slmon, 1970 Processmg symbohc mformatlon from a wsual display: Interference from an Irrelevant dlrectlonal cue Journal of Expenmental Psychology 83, 415-420 Dahymple-Alford, EC and J Azkoul, 1972 The locus of interference m the Stroop and related tasks. PerceptIon and Psychophysics 11, 385-388 Deupree, R H and J.R Simon, 1963. Reactlon time and movement time as a function of age, stimulus duration, and task dlfflculty Ergononucs 6, 403-411 Dyer, F N , 1973 The Stroop phenomenon and Its use m the study of perceptual, cogmtlve, and response processes Memory and Cognition 1, 106-120
J R. Swnon, K Berbnum
/ Effect
of conflrctrng
cues
169
Enroth-Cugell, C and J G. Robson, 1966 The contrast sensltlvlty of retinal ganghon cells of the cat Journal of Physiology 187, 517-552 Flowers, J.H , 1975 ‘Sensory’ Interference m a word-color matchmg task PerceptIon and Psychophyslcs 18, 37-43 Froeberg, S., 1907 The relation between the magmtude of stimulus and the tnne of reactlon Archlves of Psychology 8, 1-38 Glaser, M 0 and W R Glaser, 1982. Time course analysis of the Stroop phenomenon Journal of Expenmental Psychology Human Perception and Performance 8, 875-894 Green, E.J and P J Barber, 1983 Interference effects m an audltory Stroop task Congruence and correspondence Acta Psychologca 53, 183-194 Gregg, L W. and W J Brogden, 1950 The relation between reactlon time and the duration of the audltory stimulus Journal of Comparative and Physlologlcal Psychology 43, 389-395 Harvey, N , 1984 The Stroop effect. Fadure to focus attention or failure to mamtam focusing? Quarterly Journal of Expenmental Psychology 36A, 89-115 Hedge, A and N W A Marsh, 1975 The effect of irrelevant spatial correspondences on two-choice response-time. Acta Psychologca 39, 427-439 Hock, H.S. and H. Egeth, 1970 Verbal Interference with encodmg m a perceptual classlficatlon task Journal of Expenmental Psychology 83, 299-303 Jensen, AR and W D Rohwer, Jr, 1966 The Stroop color-word test A review Acta Psychologca 25, 36-93. Keele, SW., 1972 Attention demands of memory retneval. Journal of Expenmental Psychology 93, 245-248. Klein, G S , 1964. Semantic power measured through the Interference of words with color-naming Amencan Journal of Psychology 77, 576-588 Kulikowslu, J J and D J. Tolhurst, 1973 Psychophyslcal evidence for sustamed and transient detectors m human vlslon Journal of Physiology 232, 149-163 McClam, L , 1983. Stimulus-response compatlblhty affects audltory Stroop Interference Perceptlon and Psychophyslcs 33, 266-270 Mewaldt, S P., C L. Connelly and J R Slmon, 1980 Response selectlon m choice reaction time Test of a buffer model. Memory and Cogmtlon 8, 606-611. Morton, J , 1969 Categories of Interference Verbal medlatlon and confhct m card sorting Bntlsh Journal of Psychology 60, 329-346. Pachella, R G , 1974 ‘The mterpretatlon of reactlon time m mformatlon processing research’ In B H Kantowltz (ed ), Human mformatlon processmg. Tutorials m performance and cogmtlon Hillsdale, NJ Erlbaum Sanders, A F., 1980. ‘Stage analysis of reaction processes’ In G.E Stelmach and J Requm (eds ), Tutonals m motor behavior. Amsterdam. North-Holland Seymour, P H K., 1977. Conceptual encoding and locus of the Stroop effect Quarterly Journal of Expenmental Psychology 29, 245-265. Slmon, J R , 1969 ReactIons toward the source of stlmulatlon Journal of Experimental Psychology 81, 174-176. Simon, J R , 1970 ‘StereotypIc reactlons m mformatlon processmg’. In L E Smith (ed ). Psychology of motor learnmg. Chlcago, IL The Athletic Institute Slmon, J R., 1982 Effect of an audltory stimulus on the processing of a visual stimulus under smgle- and dual-tasks condltlons Acta Psychologca 51, 61-73 Simon, J R , E Acosta, Jr and S.P. Mewaldt, 1975 Effect of locus of wammg tone on audltory choice reactlon time Memory and Cognmon 3, 167-170 Slmon, J R, E Acosta, Jr., S.P Mewaldt and C R Speldel, 1976 The effect of an Irrelevant dlrectlonal cue on choice reactlon time Duration of the phenomenon and Its relation to stages of processmg PerceptIon and Psychophysics 19, 16-22
170
J R Smon,
K Berbaum
/ Effect ojconjhctrng
cues
Simon, J.R and K Berbaum, 1988 Effect of irrelevant mformatton on retrieval ttme for relevant mformation Acta Psychologma 67, 33-57 Simon, J R, C Paulhn, SP Overmyer and K Berbaum, 1985 Reaction ttme to word meanmg and mk color of laterally-presented Stroop stimuli: Effects of handedness and sex International Journal of Neuroscience 28, 21-33. Stmon, J R and A R Pouraghabagher, 1978. The effect of aging on the stages of processmg m a choice reactton ttme task Journal of Gerontology 33, 553-561 Simon, J R. and A P Rudell, 1967. Auditory S-R compattbthty The effect of an Irrelevant cue on mformatton processmg. Journal of Applied Psychology 51, 300-304. Simon, JR and P. Sudalatmuthu, 1979 Effects of S-R mapping and response modahty on performance m a Stroop task. Journal of Expertmental Psychology Human Perceptton and Performance 5, 176-187 Stanovtch, K E and R G Pachella, 1977 Encoding, stimulus-response compattbihty, and stages of processmg Journal of Expenmental Psychology Human Perception and Performance 3, 411-421 Stemberg, S, 1969. The dtscovery of processmg stages Extenstons of Donders’ method. Acta Psychologtca 30, 276-315 Stroop, J R., 1935 Studies of Interference m senal verbal reacttons Journal of Expenmental Psychology 18, 643-662 Taylor, D A, 1976. Stage analysts of reactton time Psychologrcal Bulletm 83, 161-191. Tolhurst, D.J, 1973 Separate channels for the analysts of the shape and the movement of a movmg vtsual sttmulus Journal of Phystology 231, 385-402 Tohn, P and J R Stmon, 1968. Effect of task complexrty and stimulus duration on perceptual-motor performance of two dtsparate age groups. Ergonomics 11, 283-290 Uleman, J S. and J Reeves, 1971. A reversal of the Stroop interference effect, through scannmg Perceptton and Psychophystcs 9, 293-295 Van der Molen, M W and P.J.G Keuss, 1981. Response selectton and the processmg of auditory mtenstty Quarterly Journal of Expenmental Psychology 33A, 177-184 Van der Molen, M W , R.J M Somsen, J.R Jennings, R.T Nteuwboer and J.F. Orlebeke, 1987 A psychophystologtcal mvestigatton of cogmttve-energettc relattons m human mformatton processmg A heart rate/addttwe factors approach Acta Psychologtca 66, 251-289 Warren, R E, 1972 Stimulus encodmg and memory Journal of Expenmental Psychology 94, 90- 100 Welford, A T., 1961 Age changes m the times taken by chotce, discnmmatton and the control of movement Gerontologta 5, 129-145 Welford, A T (ed ), 1980. Reaction times. London Academtc Press Wells, G R , 1913 The Influence of sttmulus duratton on reaction time Psychological Monographs 15, No 5 (whole no. 66) Zakay, D and J Ghcksohn, 1985. Sttmulus congrutty and S-R compattbtlity as determmants of Interference m a Stroop-hke task Canadtan Journal of Psychology 39, 414-423.
73 (1990) 159-170
EFFECT OF CONFLICTING CUES ON INFORMATION PROCESSING: THE ‘STROOP EFFECT’ VS. THE ‘SIMON J. Richard
SIMON
The Unrversrt~
of Iowa, USA
Accepted
February
EFFECT’
*
and Kevin BERBAUM
1989
This study exammed the relatlonshp between two sources of mterference m human mformatlon processmg the Stroop effect and the Slmon effect. Forty SubJects pressed a left- or right--hand key m response to a Stroop color word located on the left or right side of a screen For one group, mk color was the relevant cue and, for another group, word meamng was the relevant cue Independent variables were congruence, 1 e , agreement or lack thereof between the mk color and meamng of the Stroop word; spatial correspondence, 1.e) agreement or lack thereof between the locatlon of the Stroop word and the locatlon of the key used to make the response, and stimulus duration, 1 e ,400 or 100 ms Each of these vanables had a slgmflcant effect on RT, and there were no slgmflcant mteractlons According to Stemberg’s additive-factor logx, these fmdmgs suggest that the Stroop effect (congruence) and the Slmon effect (spatial correspondence) mvolve separate stages of processmg. If one assumes that mampulatlon of stimulus duration affects the encodmg stage, then results also suggest that neither the Stroop effect nor the Slmon effect mvolves the stimulus encoding stage
This research employed the Sternberg additive-factor method to determme the relationship between two potent sources of interference in human information processing, the Stroop effect and the Simon effect. There is an extensive literature concerned with information processing in situations where the stimulus provides relevant as well as irrelevant cues. The most widely used stimulus of this sort is the Stroop color word (Dyer 1973; Jensen and Rohwer 1966). In a typical Stroop * This research was supported by funds from the Graduate College of the Umverslty The authors gratefully acknowledge the assistance of Todd Bever and Renet Snuth Requests for repnnts should be sent to J.R. Simon, Dept of Psychology, University Iowa City, IA 52242, USA
OOOl-6918/90/$3
50 0 1990, Elsevler Science Publishers
B V (North-Holland)
of Iowa of Iowa,
160
J R Smon,
K Berbuum
/ Effect ofconfhrrng
cues
task (Stroop 1935), the stimuli consist of names of colors printed in other colors. Subjects experience interference when they attempt to name the ink colors and ignore the words. Despite literally hundreds of studies, there is still a controversy regarding the source of the Stroop interference effect (Dyer 1973; Glaser and Glaser 1982). The traditional explanation is in terms of response competition (e.g., Dalrymple-Alford and Azkoul 1972; Keele 1972; Klein 1964; Morton 1969; Warren 1972). It asserts that the Stroop effect is due to the lack of correspondence between the irrelevant attribute of the stimulus and the response. An alternative explanation is m terms of encoding. It argues that the Stroop effect is due to the lack of congruence between the relevant and irrelevant cues in the stimulus. Hock and Egeth (1970) propose that the interference occurs during perceptual encoding; i.e., the color word distracts from the identification of the ink color by divertmg attention from it. Seymour (1977) argues that the interference occurs during conceptual encoding which is a stage between perceptual encoding and response activation when the color information comes in contact with semantic memory. Simon and his associates (e.g., Craft and Simon 1970; Simon 1969; Simon 1970; Simon and Rude11 1967) have used a paradigm that is similar to the Stroop paradigm to investigate the effect of conflicting cues on information processing. In a typical task, subjects might press a left- or right-hand key, depending on the color of a stimulus light which appears on the right or left side of a display panel. Results indicate that the location of the light provides an irrelevant directional cue that interferes with processing the relevant symbolic cue which is the color of the light. In other words, reactions are faster on trials in which the location of the stimulus and response correspond than on trials in which they do not correspond. Hedge and Marsh (1975) have used the term ‘Simon effect’ to refer to this typical finding, and we use that label here as a convenient short-hand designation for the phenomenon. Simon et al. have suggested that their effect reflects a stereotypic tendency to respond toward the source of stimulation and, in a series of studies they have provided evidence that the interference is located in the response selection stage (Acosta and Simon 1976; Mewaldt et al. 1980; Simon 1982; Simon et al. 1975; 1976). Since the Stroop effect and the Simon effect are so similar, superficially at least, and since there is some evidence that both Stroop and Simon effects might involve the same processing stage, it is of interest
J R Swnon, K Berbaum / Effect
of confktmg cues
161
to determine the relationship between these two potent interference effects. Sternberg’s additive-factor-method (1969) provides a powerful tool for studying the stages of information processing and defining the processing which is accomplished by certain stages. The method assumes that the interval between the presentation of a stimulus and the subject’s response is filled with a sequence of independent stages or cognitive processes. Each stage receives an input from a previous stage, performs a transformatton on that input, and passes it along to the next stage. Total reaction time (RT) is simply the sum of the stage durations. When an experimental manipulation affects RT for a task, it does so by influencing the durations of one or more of the stages. Thus, tf two experimental variables affect the same processing stage, both variables will be contributors to the duration of that stage, and their expected effect on mean RT 1s an interaction. On the other hand, if two experimental variables affect two different stages, they will produce independent (additive) effects on total RT. In other words, the experimental variables should not interact in a statistical sense. The additivefactor method, then, involves conducting factorial experiments to determine which variables produce significant interactions and which do not. The variables which interact are assumed to affect a stage in common, whereas the variables which do not interact are assumed to affect different stages. Specifically, m the present study, a significant interaction between the Stroop effect and the Simon effect would be interpreted as evidence that they affect a common processmg stage whereas no interaction would suggest that these two phenomena involve separate stages. Stimulus duration was also manipulated in the present study in an attempt to locate the source of the Stroop and/or Simon effect.
Method SubJects The
subjects
mtroductory English
speakers
visual acuity. mk-relevant
were 40 undergraduates,
psychology
course
and reported
20 males
at the University normal
and 20
of Iowa.
females,
enrolled
All subjects
color vtston and normal
or corrected-to-normal
Half of the males and half of the females were assigned at random condition
and the other half to a word-relevant
m an
were nattve
conditton.
to an
162
J R Svnon, K Berbaum / Effect of confktrng cues
Apparatus The apparatus measured keypress RT to the onset of a visual stimulus which appeared on a 60 cm square glare-resistant glass screen located 82 cm from the subject’s eyes and perpendicular to the line of sight. Nme Kodak Ektagrapmc prolectors with high-speed shutters attached (Urnbhtz Model 225 LOAXS) were located behind the screen and used to present the sttmulus materials Each projector contamed a single stimulus shde. Openmg the shutter m front of a projector resulted m the stimulus appeanng on the screen. The opening and closmg times for the shutters were 3 and 4 ms, respecttvely. A Digital Equipment Corporation (DEC) PDP 1 l/34 laboratory computer provided overall control of stimulus presentation and data acqutsttton The computer was interfaced with a specially-built shutter controller and was programmed to operate the shutters, thus determining the sequence of stimulus presentation, duration of exposure, and intervals between stirnull. The apparatus, then, formed a precise, highly flexible, multichannel tacmstoscope. The procedure elimmated stimulus rise time as a complicating factor smce prolector bulbs remained on throughout the session. Also, smce the stimulus slides did not move durmg the expertment, perfect sttmulus alignment was mamtamed. Two response keys, one on the right and one on the left, rested on a table m front of the seated subject Above each key was a white cardboard label with either the word Red printed m red ink, or the word Green printed m green mk Subjects operated the keys with their right and left index fingers. The computer recorded RT m ms and also recorded which key had been pressed. Smce this mformatton was also prmted tmmediately on a DEC Decwnter III, the expenmenter could monitor the subject’s performance durmg the session. Each tnal began with the presentatton of a small white fixation point with a luminance of 4.5 cd/m2 m the center of the screen at approximately eye level. The stimulus, i.e., the word Red or Green pnnted m red or green mk was then presented on the nght or left side of the screen at a distance of 8.9 cm from the fixation point or an arc of 6.2O at the eye. Henceforth, the subscripts R and G will be used to specify mk color; i.e., Red, denotes the word Red in red mk, Red, denotes the word Red m green mk, and so on. The first letter of each word was 4.3 cm high and the remaming letters were 2.85 cm htgh Stroke width for all letters was 0.79 cm The word Red measured 9 5 cm honzontally on the screen, and the word Green measured 15 25 cm. The lummance of the mdivtdual stimulus words (10.5 and 8.1 cd/m2) for red and green mk, respectively) was adjusted by means of neutral denstty filters and apertures m front of the prelectors so that the mk colors appeared to be equally intense and salient. The viewmg screen was steadily illummated at 3.6 cd/m* (incandescent light) m order to suppress colored after images. Procedure and experrment destgn Taped mstructions mformed subjects that the experiment would consist of a number of trials, each tnal begmnmg with a small white dot (ready signal) appearmg m the center of the screen. After a short Interval the word Red or Green prmted m red or
J R Smon, K Berbaum / Ejjecr ojconjhcrrng cues
163
green mk would appear to either the right or left of the dot. On some trials the word would appear very bnefly and on other trials it would appear for a longer duration. Half of the sublects (ink-relevant condttion) were told that their task was to respond to the color of the ink; i.e., ‘if the mk color is red, press the key labeled “Red” as quickly as posstble, and tf the mk color IS green, press the key labeled “Green” as quickly as posstble’. The other half of the subjects (word-relevant condition) responded to the identical sequence of trials but were told that then task was to respond to the meanmg of the word; i.e., ‘if the word is Red, press the key labeled “Red” as quickly as possible and tf the word is Green, press the key labeled “Green” as qutckly as possible’. For half of the males and half of the females m each conditton, the right key was labeled ‘Red’ and the left key was labeled ‘Green’. For the other half of each sex group, the key labels were reversed. Subjects placed their right index finger on the right key and their left index finger on the left key and were told to keep then fingers resting lightly on the two keys throughout the expenment Each subject performed on the same sequence of 192 test trials m which the 16 possible combmations of the four Stroop words (Reda, Redo, Green,, and Green,), two sttmulus locations (right and left of the ftxation point), and two sttmulus durations (400 and 100 ms) each appeared 12 times m a predetermined sequence. The sequence was arranged such that, in each successtve 48 trials, each of the 16 treatment combmations appeared three times m a random sequence. Pnor to the test trials, there were 16 practice trials and 16 unscored warm-up trials Both practice and warm-up trials included one of each treatment combmation. The durations of the various events m the trial sequence were as follows The ready signal remained on for 500 ms. It was followed by a 500 ms. blank interval after which the Stroop word appeared for either 400 or 100 ms. There was a 2-s interval between the subject’s response and the ready signal for the next trial. If the subject did not respond within 5 s, the program advanced automattcally to the next trial Subjects received no feedback regarding RT or the accuracy of their responses The experimental design consisted of three withm-subjects factors and one between-subjects factor. The within-sublects factors were congruence (i.e., the agreement or lack thereof between the meamng and mk color of the Stroop word), spatial correspondence (1 e., the agreement or lack thereof between the locatton of the Stroop word and the location of the key used to make the response), and stimulus duration (i.e., 400 or 100 ms). The between-sublects factor was the relevant dtmenston of the Stroop word (i.e., mk color or word meamng).
Results
Table 1 shows the mean RTs for the 16 treatment conditions, i.e., congruent (Red, and Green,) and mcongruent (Redo and Green.) Stroop words displayed m a location (nght or left) corresponding or not corresponding with the correct response key, and presented for 400 or 100 ms to the ink-relevant or the word-relevant group. Trials on which errors were made (2.6%) were excluded from the computations. The top third of the table combines the data from the group responding to ink color and the
Noncorrespondmg
Durations
and stimulus
Correspondmg
combmed
correspondence,
450 -435 442
428
437 -420
470 -436 453
454 -428 441
442 427 435
465 449 457
484 466 415
466
447
474 -444 459
475 457
Noncorrespondmg
454 430 442
492 -449 471
413 -440 456
Correspondmg
459 440 450
488 -457 473
461
474 -449
M
428
435 420
421 409 415
449 -423 436
425
441
471 438 454
435 416 -
Correspondmg
of Stroop
453 429 -
Noncorrespondmg
100 ms duration
time (m ms) to mk color and word meanmg
400 ms duration
on reactlon
459 436
M
duration
421
428 414
460 431 445
433
444 -423
M
color word stlmuh
_
Note Congruence refers to the agreement or lack of agreement between the meanmg of the Stroop word and the color of the mk m whrch 1t 1s prmted Correspondence refers to the agreement or lack of agreement between the location of the Stroop word (left or nght side of the screen) and the location of the key (left or nght) used to make the response
Mean
Incongruent Congruent
RT to word meanmg
Incongruent Congruent Mean
478 452 465
453
Mean
RT to mk color
464 443 -
Incongruent Congruent
RT to rnk color and word meanrng combrned
Congruence
Table 1 Effect of congruence,
h
9
3
B
J R Simon, K Berbaum / Effect of confltctmg cues
165
group responding to word meaning The mtddle thud of the table summarrzes the data from the group responding to mk color, and the bottom thud summarizes the data from the group responding to word meaning. An overall analysts of variance indicated that there was no srgmftcant difference between RT to mk color and RT to word meaning (459 vs. 435 ms), F(l, 38) = 0 53. The mam effects of congruence, spatial correspondence, and sttmulus duration were all statrstrcally significant Responses to incongruent strmuh were slower than to congruent strmuh (459 vs. 436 ms), F(1, 38) = 33 01, p < 0.001. Noncorrespondmg responses were slower than correspondmg responses (453 vs. 441 ms), F(l, 38) = 8 51, p c 0 01 Responses to long-duration strmuh were slower than to short duration stimuli (461 vs 433 ms), F(l, 38) = 17.13, p < 0.001 None of the mteracttons between congruence, spatial correspondence, and duration was stattstically stgruftcant which suggested that each factor affected a separate stage of mformatron processmg. (The F values [df 1, 381 for the Congruence X Correspondence, Congruence X Duratton, Correspondence X Duratron, and Congruence X Correspondence X Duration mteracttons were 0.94, 1.19, 1.07, and 3 24, respectrvely. Since the latter of these mteractrons approached srgruficance (p < 0.10) separate analyses were performed for the long and short duration condrtrons, but these, too, faded to show a srgmficant Congruence x Correspondence mteractron.) Separate data analyses were conducted for the group responding to mk color and the group respondmg to word meaning. For reactions to mk color (see middle third of table l), responses to incongruent sttmuh were slower than to congruent sttmuh (474 vs. 444 ms), F(1, 19) = 17.50, p -C 0.001, and responses to long-duration strmuh were slower than to short-duration strmuh (473 vs. 445 ms), F(l, 19) = 9.80, p < 0 01. The difference between noncorrespondmg and correspondmg responses was not srgmftcant (465 vs 453 ms), F(l, 19) = 2.23, p > 0.15. There were no significant mteracttons between any of these three independent vartables. (The F values [df 1, 191 for the Congruence x Correspondence, Congruence X Duratron, Correspondence x Duratron, and Congruence X Correspondence X Duration mteractrons were 0 80, 0.11, 1.89, and 2.88, respectrvely.) For reactions to word meaning (see lower thud of table l), responses to incongruent sttmuh were slower than to congruent strmuh (442 vs. 427 ms), F(l, 19) = 19 37, p < 0.001, noncorrespondmg responses were slower than correspondmg responses (442 vs. 428 ms), F(l, 19) = 11.77, p < 0.01, and responses to long-duration sttmuh were slower than to short-duration sttmuh (450 vs. 421 ms), F(1, 19) = 7 65, p < 0 01 None of the mteracttons between the three main effects was srgmftcant which again suggested that the expenmental mampulattons affected separate stages of mformatron processmg. (The F values [ df 1, 191 for the Congruence X Correspondence, Congruence X Duratron, Correspondence X Duratron, and Congruence X Correspondence X Duratron mteractrons were 0.17, 2.01, 0.19, and 0.56, respectively.)
Discussion
The major purpose of this study was to determine whether the Stroop effect and the Simon effect involved a common stage of
166
J R Smon, K Berbaum
/ Effect
of conflmng
cues
information processing. A second purpose was to attempt to locate the Stroop- and/or Simon-type interference m a particular processing stage. At a superficial level, the interference produced by the mcongruity of ink color and word meaning m the Stroop task is similar to that observed in the Simon paradigm when an irrelevant location cue affects the time required to respond to a relevant symbolic cue (e.g., Simon 1970). Our results, however, suggest that the Stroop and Simon effects involve separate stages of processing. This conclusion is based on the absence of an interaction between congruence (Stroop effect) and spatial correspondence (Simon effect). According to Sternberg’s additive-factor logic, variables that do not interact are assumed to affect different stages. However, because there are possible alternative explanations for the absence of an interaction (Pachella 1974; Sanders 1980; Taylor 1976) this conclusion must be regarded as tentative. The stimulus duration manipulation was introduced in an attempt to specify more exactly the source of the Stroop and/or Simon interference effect. If one assumes that sttmulus duration affects stimulus encoding (the stage in which a stimulus representation is formed), then an interaction of the Stroop and/or Simon effect with stimulus duration would tend to locate the effect(s) m the encoding stage. ’ We ’ In order to mampulate the perceptual encodmg stage, mvestlgators have used various means of degradmg the quahty of the visual stimulus (Sanders 1980), e g , supenmposmg visual noise (Stemberg 1969; Van der Molen et al 1987) or reducmg visual contrast (Acosta and Simon 1976, Slmon and Pouraghabagher 1982, Stanovlch and Pachella 1977) To vary stimulus quahty drrectlj m the present expenment would have reqmred eight addItIona proJectors, Le , a degraded Image of the four Stroop words would have had to be proJected to both the left and right visual fields. We chose, Instead, to mampulate stimulus quahty rndrrectly by presentmg the Stroop words for a very brief duration Psychophysical evidence provided by Kuhkowskl and Tolhurst (1973) and Tolhurst (1973), and neurophyslologxal evidence provided by Enroth-Cugell and Robson (1966) suggest that there are two separate visual mformatlon channels The ‘sustamed’ channel 1s attuned to lugher spatial frequency than IS the ‘transient’ channel whereas the transient channel IS attuned to higher temporal frequency Therefore, the sustained system ~111 regster high detail so long as It does not move or change (turn on and off rapidly) The sustained channel IS the one medlatmg regstratlon of mformatlon (or lcomcally stonng mformatlon) required for reading The transient system ~111 regster moving or changing stlmuh, but only the low spatial frequencies Therefore, a moving or changmg pattern ~111 appear blurry because the fme details are not registered by either system. A short duration stimulus ~111 evoke less sustamed and more transient visual system response than a longer duration stimulus (Breltmeyer and Julesz 1975, Breltmeyer and Ganz 1976, 1977) This IS because the ratlo of steady-state to dynamic stlmulatlon IS lower It has even been suggested that the slope of the onset and offset of a stimulus determmes the relative response of transient and sustamed visual mechanisms (Breltmeyer and Julesz 1975) Thus, wlthm the visual system itself, stimulus duration affects stimulus quality relative to a readmg task which m turn affects the encoding stage
J R Smon, K Berbaum / Effect of confl~crmg cue-s
167
found no interaction of either the Stroop or Simon effect with stimulus duration which suggests that neither of these interference effects involve encoding. Thus, our results agree with previous findmgs indicating that the Simon effect does not involve the encoding stage (Acosta and Simon 1976; Simon 1982; Simon and Pouraghabagher 1978). Indeed, there is considerable evidence that its locus is in the response selection stage (Acosta and Simon 1976; Mewaldt et al. 1980; Simon 1982; Simon et al. 1975; 1976; Van der Molen and Keuss 1981). But what about the locus of the Stroop effect if, as the Sternberg logic suggests, it is in neither the encoding nor the response selection stage? Seymour (1977) suggests that Stroop interference occurs during conceptual encoding, a stage between perceptual encoding and response activation when color information contacts semantic memory. Simon and Berbaum (1988) conclude, based on finding a significant Stroop effect in a retrieval task, that the most likely locus of interference is the stage between perceptual encoding and response selection; i.e., the process of decoding or retrieving information from short-term memory. In the original Stroop study (1935), subjects asked to name the mk color experienced great difficulty from incongruent word meaning, but subjects asked to read the words encountered little interference from mcongruent ink colors. It is worth noting that we found not only the ‘standard’ Stroop effect but also the ‘reverse’ Stroop effect. Our results, then, support the more recent research which indicates that incongruent ink colors can also interfere with responses to word meaning (Flowers 1975; Simon and Sudalaimuthu 1979; Simon et al. 1985; Uleman and Reeves 1971). Another interesting aspect of our results was the finding that reactions were significantly faster to the 100 ms stimulus than to the 400 ms stimulus. Welford (1961) has proposed an interesting explanation for this effect. He suggested that subjects inspect a stimulus until ‘sufficient’ data are gathered to make a response. Reactions are faster to a short duration stimulus because the intake of data is stopped by the cessation of the stimulus, and, therefore, subjects gain little or nothing by delaying their reaction. Studies of the effect of stimulus duration on RT have produced contradictory findings. Some report a tendency for RT to decrease with an increase in duration (e.g., Froeberg 1907). Others, like the present study, report a difference in the opposite direction (e.g., Wells 1913; Botwmick et al. 1958; Gregg and Brogden, 1950). Still other studies have found no significant effects (e.g., Deup-
168
J R Simon, K Berbaum
/ Effect
of confktrng cues
ree and Simon 1963; Tolin and Simon 1968). While the contradictory findings seem to be related, at least in part, to the specific durations studied and to the complexity of the task (Welford 1980), the conditions under which stimulus duration affects RT remain unclear. There is a wide variety of tasks in which irrelevant cues intrude to affect information processing. Several recent studies have focused on the effect of stimulus-response compatibility in these Stroop or ‘Stroop-type’ tasks (Green and Barber 1983; McClain 1983; Simon and Sudalatmuthu 1979; Zakay and Glicksohn 1985). Harvey (1984) suggests that use of the term ‘Stroop-type’ should be avoided since it implies that the interference may have a different etiology from that giving rise to the color-word effect. He suggests that, until contrary evidence is provided, rt is more parsimonious to credit Stroop (1935) with discovering a general effect. Results of the present study strongly suggest that it would be a mistake to lump together indiscriminately all interference phenomena under the heading ‘Stroop effect’ since the locus of the interference; e.g., encoding or response selection may differ for different ‘Stroop-type’ tasks.
References Acosta, E Jr. and J.R Bmon, 1976 The effect of Irrelevant mformatton on the stages of processmg. Journal of Motor Behawor 8, 181-187 Botwmck, J., J R. Brmley and J S Robbm, 1958 The mteractlon effects of perceptual dlfflculty and stimulus exposure time on age differences m speed and accuracy of response Gerontologa 2, l-10 Breltmeyer, B G and L Ganz, 1976 Imphcatlons of sustamed and transient channels for theones of wsual pattern maskmg, saccadlc suppresslon, and mformatlon processmg Psychologxal Review 83, l-36 Breltmeyer, B.G. and L Ganz, 1977 Temporal studies with flashed gratings Inferences about human transient and sustamed channels Vlslon Research 17, 861-865 Breltmeyer, B and B Julesz, 1975 The role of on and off transients m determmmg the psychophysical spatial frequency response Vlslon Research 15, 411-415. Craft, J L and J.R. Slmon, 1970 Processmg symbohc mformatlon from a wsual display: Interference from an Irrelevant dlrectlonal cue Journal of Expenmental Psychology 83, 415-420 Dahymple-Alford, EC and J Azkoul, 1972 The locus of interference m the Stroop and related tasks. PerceptIon and Psychophysics 11, 385-388 Deupree, R H and J.R Simon, 1963. Reactlon time and movement time as a function of age, stimulus duration, and task dlfflculty Ergononucs 6, 403-411 Dyer, F N , 1973 The Stroop phenomenon and Its use m the study of perceptual, cogmtlve, and response processes Memory and Cognition 1, 106-120
J R. Swnon, K Berbnum
/ Effect
of conflrctrng
cues
169
Enroth-Cugell, C and J G. Robson, 1966 The contrast sensltlvlty of retinal ganghon cells of the cat Journal of Physiology 187, 517-552 Flowers, J.H , 1975 ‘Sensory’ Interference m a word-color matchmg task PerceptIon and Psychophyslcs 18, 37-43 Froeberg, S., 1907 The relation between the magmtude of stimulus and the tnne of reactlon Archlves of Psychology 8, 1-38 Glaser, M 0 and W R Glaser, 1982. Time course analysis of the Stroop phenomenon Journal of Expenmental Psychology Human Perception and Performance 8, 875-894 Green, E.J and P J Barber, 1983 Interference effects m an audltory Stroop task Congruence and correspondence Acta Psychologca 53, 183-194 Gregg, L W. and W J Brogden, 1950 The relation between reactlon time and the duration of the audltory stimulus Journal of Comparative and Physlologlcal Psychology 43, 389-395 Harvey, N , 1984 The Stroop effect. Fadure to focus attention or failure to mamtam focusing? Quarterly Journal of Expenmental Psychology 36A, 89-115 Hedge, A and N W A Marsh, 1975 The effect of irrelevant spatial correspondences on two-choice response-time. Acta Psychologca 39, 427-439 Hock, H.S. and H. Egeth, 1970 Verbal Interference with encodmg m a perceptual classlficatlon task Journal of Expenmental Psychology 83, 299-303 Jensen, AR and W D Rohwer, Jr, 1966 The Stroop color-word test A review Acta Psychologca 25, 36-93. Keele, SW., 1972 Attention demands of memory retneval. Journal of Expenmental Psychology 93, 245-248. Klein, G S , 1964. Semantic power measured through the Interference of words with color-naming Amencan Journal of Psychology 77, 576-588 Kulikowslu, J J and D J. Tolhurst, 1973 Psychophyslcal evidence for sustamed and transient detectors m human vlslon Journal of Physiology 232, 149-163 McClam, L , 1983. Stimulus-response compatlblhty affects audltory Stroop Interference Perceptlon and Psychophyslcs 33, 266-270 Mewaldt, S P., C L. Connelly and J R Slmon, 1980 Response selectlon m choice reaction time Test of a buffer model. Memory and Cogmtlon 8, 606-611. Morton, J , 1969 Categories of Interference Verbal medlatlon and confhct m card sorting Bntlsh Journal of Psychology 60, 329-346. Pachella, R G , 1974 ‘The mterpretatlon of reactlon time m mformatlon processing research’ In B H Kantowltz (ed ), Human mformatlon processmg. Tutorials m performance and cogmtlon Hillsdale, NJ Erlbaum Sanders, A F., 1980. ‘Stage analysis of reaction processes’ In G.E Stelmach and J Requm (eds ), Tutonals m motor behavior. Amsterdam. North-Holland Seymour, P H K., 1977. Conceptual encoding and locus of the Stroop effect Quarterly Journal of Expenmental Psychology 29, 245-265. Slmon, J R , 1969 ReactIons toward the source of stlmulatlon Journal of Experimental Psychology 81, 174-176. Simon, J R , 1970 ‘StereotypIc reactlons m mformatlon processmg’. In L E Smith (ed ). Psychology of motor learnmg. Chlcago, IL The Athletic Institute Slmon, J R., 1982 Effect of an audltory stimulus on the processing of a visual stimulus under smgle- and dual-tasks condltlons Acta Psychologca 51, 61-73 Simon, J R , E Acosta, Jr and S.P. Mewaldt, 1975 Effect of locus of wammg tone on audltory choice reactlon time Memory and Cognmon 3, 167-170 Slmon, J R, E Acosta, Jr., S.P Mewaldt and C R Speldel, 1976 The effect of an Irrelevant dlrectlonal cue on choice reactlon time Duration of the phenomenon and Its relation to stages of processmg PerceptIon and Psychophysics 19, 16-22
170
J R Smon,
K Berbaum
/ Effect ojconjhctrng
cues
Simon, J.R and K Berbaum, 1988 Effect of irrelevant mformatton on retrieval ttme for relevant mformation Acta Psychologma 67, 33-57 Simon, J R, C Paulhn, SP Overmyer and K Berbaum, 1985 Reaction ttme to word meanmg and mk color of laterally-presented Stroop stimuli: Effects of handedness and sex International Journal of Neuroscience 28, 21-33. Stmon, J R and A R Pouraghabagher, 1978. The effect of aging on the stages of processmg m a choice reactton ttme task Journal of Gerontology 33, 553-561 Simon, J R. and A P Rudell, 1967. Auditory S-R compattbthty The effect of an Irrelevant cue on mformatton processmg. Journal of Applied Psychology 51, 300-304. Simon, JR and P. Sudalatmuthu, 1979 Effects of S-R mapping and response modahty on performance m a Stroop task. Journal of Expertmental Psychology Human Perceptton and Performance 5, 176-187 Stanovtch, K E and R G Pachella, 1977 Encoding, stimulus-response compattbihty, and stages of processmg Journal of Expenmental Psychology Human Perception and Performance 3, 411-421 Stemberg, S, 1969. The dtscovery of processmg stages Extenstons of Donders’ method. Acta Psychologtca 30, 276-315 Stroop, J R., 1935 Studies of Interference m senal verbal reacttons Journal of Expenmental Psychology 18, 643-662 Taylor, D A, 1976. Stage analysts of reactton time Psychologrcal Bulletm 83, 161-191. Tolhurst, D.J, 1973 Separate channels for the analysts of the shape and the movement of a movmg vtsual sttmulus Journal of Phystology 231, 385-402 Tohn, P and J R Stmon, 1968. Effect of task complexrty and stimulus duration on perceptual-motor performance of two dtsparate age groups. Ergonomics 11, 283-290 Uleman, J S. and J Reeves, 1971. A reversal of the Stroop interference effect, through scannmg Perceptton and Psychophystcs 9, 293-295 Van der Molen, M W and P.J.G Keuss, 1981. Response selectton and the processmg of auditory mtenstty Quarterly Journal of Expenmental Psychology 33A, 177-184 Van der Molen, M W , R.J M Somsen, J.R Jennings, R.T Nteuwboer and J.F. Orlebeke, 1987 A psychophystologtcal mvestigatton of cogmttve-energettc relattons m human mformatton processmg A heart rate/addttwe factors approach Acta Psychologtca 66, 251-289 Warren, R E, 1972 Stimulus encodmg and memory Journal of Expenmental Psychology 94, 90- 100 Welford, A T., 1961 Age changes m the times taken by chotce, discnmmatton and the control of movement Gerontologta 5, 129-145 Welford, A T (ed ), 1980. Reaction times. London Academtc Press Wells, G R , 1913 The Influence of sttmulus duratton on reaction time Psychological Monographs 15, No 5 (whole no. 66) Zakay, D and J Ghcksohn, 1985. Sttmulus congrutty and S-R compattbtlity as determmants of Interference m a Stroop-hke task Canadtan Journal of Psychology 39, 414-423.
