S-R COMPATIBILITY: SPATIAL CHARACTERISTICS OF

STIMULUS AND RESPONSE CODES 1

PAUL M. FITTS
The Ohio State University
AND CHARLES M. SEEGER
Aero Medical Laboratory

The present paper reports the results extended to cover relations between
of two experiments designed to dem- concurrent stimulus activities, such as
onstrate the utility of the concept of take place during simultaneous listen-
stimulus-response compatibility2 in ing and looking, as well as to relations
the development of a theory of percep- between concurrent motor responses.
tual-motor behavior. However, the present paper will be
A task involves compatible S-R limited to a consideration of stimulus-
relations to the extent that the en- response compatibility effects in which
semble of stimulus and response com- the relevant information in the stim-
binations comprising the task results ulus source is that generated by
in a high rate of information transfer. changes in its spatial characteristics,
Admittedly, degree of compatibility and the relevant aspect of a response
can be defined in terms of operations is its direction of movement.
other than those used to secure a One of the earliest studies of the
measure of information, for example, behavioral effects of changes in the
it could be specified in terms of meas- spatial correspondence of S-R rela-
ures of speed or accuracy. However, tions is the well-known experiment of
the present writers prefer the preced- Stratton (12) on vision without inver-
ing definition because of the theoreti- sion of the retinal image. Recently
can interpretation that they wish to the majority of studies of such effects
give to compatibility effects. This have dealt with the spatial relations
interpretation makes use of the idea of between machine controls and remote
a hypothetical process of information visual displays that are connected to
transformation or recoding in the them by mechanical or electrical means
course of a perceptual-motor activity, (1, 3, 5, 9, 13). These relations are
and assumes that the degree of com- important for human engineering.
patibility is at a maximum when re- They have also become a matter of
coding processes are at a minimum. considerable interest for learning the-
The concept of compatibility can be ory, in part because of Gagne, Foster,
1
and Baker's (2) proposal that a rever-
This research was supported in part by the sal of S-R relations is the only condi-
United States Air Force under Contract No. AF
33(O38>1OS28 with the Ohio State University- tion leading to negative transfer effects
Research Foundation, monitored by the Human in perceptual-motor learning. How-
Resources Research Center, and was reported at ever, studies in both of these areas
the 1952 meeting of the Midwestern Psychologi- have consistently dealt with only one
cal Association. Permission is granted for repro-
duction, publication, use and disposal in whole or aspect of the compatibility problem.
in part by or for the United States Government. Investigations in the human-engineer-
2
The authors wish to credit Dr. A. M. Small ing area usually have compared -dif-
for suggesting the use of the term compatibility ferent ways of displaying information
in an unpublished paper presented before the
Ergonomics Research Society in 1951. when the S responds with a single type
199
200 PAUL M. FITTS AND CHARLES M. SEEGER

of control, or else have examined the cussed the application of information

effectiveness of several kinds of con- theory to psychology have considered
trols or control motions when used coding problems, such as the possi-
with a single display. Investigations bility of recoding as an aid in remem-
of transfer effects have typically con- bering. However, interest has been
sidered only reversals of the stimulus- limited to stimulus coding (6).
response relations within one S—R Since the coding problem is central
ensemble, such as changes in the direc- to the topic of compatibility, it will be
tion of motion of the control on a par- worth while to consider briefly the
ticular psychomotor test (8), without available ways of choosing the sets
specifying the location of the original of stimuli and the sets of responses
task along a dimension of S-R com- employed in a visual-motor task. We
patibility. Only two studies (7, 11) shall consider only the case of a finite
have systematically investigated the series of discrete code symbols. A
effects of varying both stimulus sets stimulus code must utilize one or more
and response sets. Such an approach stimulus dimensions such as the inten-
is indicated by the concept of sity, the wave length, the duration, or
compatibility. the extent of visible light. A uni-
Information theory provides a con- dimensional code is one in which the
venient formulation of compatibility required number of unique code sym-
in terms of information coding. Shan- bols is provided by the selection of
non (10) and others have pointed out, points along a single stimulus dimen-
in the case of physical communication sion. A multidimensional code is one
channels, that information can be in which points are selected along two
transmitted at a rate approaching or more stimulus dimensions and each
channel capacity only if messages are code symbol is defined by specifying a
optimally encoded for the particular unique combination of several stim-
channel being used. In such systems ulus characteristics. In either case,
optimum encoding requires that mes- several code symbols can be combined
sages be expressed in a suitable form to form a larger multisymbol series.
and that their probability constraints A response code can be devised by
be matched to the physical constraints similar procedures. A unidimensional
of the channel. By analogy, it seems response code utilizes a single effector
reasonable to hypothesize that man's member and a specified number of
performance of a perceptual-motor points along one of the physical dimen-
task should be most efficient when the sions of a particular response con-
task necessitates a minimum amount tinuum, such as the direction, the
of information transformation (encod- force, or the duration of a movement.
ing and/or decoding), in other words, A multidimensional response code util-
when the information generated by izes the complex response of a single
successive stimulus events is appro- effector member, such as the motion of
priate to the set of responses that must the arm in positioning of a three-
be made in the task, or conversely, dimensional control, or the responses
when the set of responses is appropri- of several different effectors acting
ately matched to the stimulus source.8 concurrently, such as the simultane-
A few of the writers who have dis- ous responses of hands and feet in
moving a set of conventional aircraft
3
If an incompatible S-R relation exists, then controls. Both the set of stimuli and
a considerable amount of recoding of informa- the set of responses required in a par-
tion (such as the looking for meaningful associ-
ations, etc.) may be highly desirable. ticular task can thus be specified as
 S-R COMPATIBILITY 201
points lying in the multidimensional judged would be expected by most Ss,
space comprising the set of all pos- i.e., would agree with population stere-
sible discriminable stimuli and motor otypes. These sets of stimuli and
responses. responses were chosen as abstractions
Garner and Hake (4) have pointed of commonly used spatial patterns of
out that in order to attain maximum stimuli and responses, i.e., as familiar
average information transmission per ways of representing points on a two-
symbol, stimulus points should be dimensional surface.
selected with reference to their posi- A further consideration in selecting
tion along a scale of equal discrimina- the sets of stimuli and responses was
bility. A similar principle undoubt- that three of the nine S-R ensembles
edly holds for the selection of a set of (Sa-Ra, Sb-Rt,, and Sc-Rc) should rep-
responses. Discriminability, defined resent "corresponding" permutations
without reference to response rate, and combinations. Correspondence,
establishes an upper limit to the maxi- in this case, was judged by the Es on
mum average information that can be the basis of the direct physical simi-
generated by the selection of partic- larity of the two patterns. The experi-
ular stimulus or response points from ment therefore provides an incidental
a set of points lying along a specified test of how accurately S-R compatibil-
dimension. The problem of S-R com- ity can be predicted from a considera-
patibility remains, however, even when tion of the correspondence, with
sets of stimuli and responses are op- respect to the number of coding
timally chosen with respect to dis- dimensions employed, between the
criminability. It concerns the ques-
two sets of points employed in stim-
tion of how the correspondence of the
ulus and response coding.
sets of points in the stimulus and the
response symbol spaces affects infor-
mation transfer. Method
Apparatus.—The response required of Ss was
to move a stylus quickly in the direction indi-
EXPERIMENT I cated by a stimulus light or lights. The three
stimulus and the three response panels used in
The first experiment was planned the study are shown in Fig. 1 and a drawing of
to test the hypothesis that, in tasks the S's position, with the Sa-Rc combination in
requiring directional motor responses place, is shown in Fig. 2.
to spatial stimulus patterns, effective Stimulus Set A and Response Set A each con-
performance depends to a large extent sisted of eight permutations of direction from a
central reference point. The stimulus panel
upon the unique characteristics of contained eight lights forming the outline of a
S-R ensembles rather than on specific circle. The response panel contained eight path-
aspects of particular stimulus or re- ways radiating from a central point like the
sponse sets. Three sets of stimuli spokes of a wheel. Each light and each pathway
were separated from their neighbors by an angle
(S o , S(,, S c ), each comprising eight of 45°. These stimulus and response patterns
easily discriminable light patterns, and are characteristic of those provided by a pictorial
three sets of eight responses (Ra, Rb, azimuth or bearing display and an aircraft-type
Rc) were studied in the nine combina- control stick.
Stimulus Panel B consisted of four lights sep-
tions formed by combining each stim- arated by 90°. It provided four single-light posi-
ulus set with each response set. tions, plus thefourtwo-light combinations formed
Within each of these nine S-R en- by adjacent pairs of lights. Response Panel B
sembles, the pairings of stimuli and consisted of four pathways originating at a cen-
tral point and separated by 90° intervals. Each
responses were those that the Es path, as seen in Fig. 1, branched in a T and per-
202 PAUL M. FITTS AND CHARLES M. SEEGER

o
o o

SA

m-\ 5

RB Re
Scale , 6 in.
FIG. 1. The three stimulus panels (So, Ss, and Sc) and the three
response panels {£a, Kb, and £c) used in Exp. I. The Ss held a metal
stylus on the circular button in the center of the response panel.

mitted Ss to move from the center point to one ditions that it provided included^the four possi-
of eight terminal positions. The four corner bilities that a single light would come on, plus
points of the response panel could be reached by the four two-light combinations that could be
two alternative pathways; for example, the produced by the simultaneous presentation of
upper-right corner could be reached either by a one of each of the two pairs. Response Panel C
right-up sequence or an up-right sequence. The permitted a left or right response of the left hand,
Ss were told that these were equivalent responses. an up or down response of the right hand, plus
Response Panel B, therefore, involved the choice the four possible combined movements of the two
of a single directional movement, or of a sequence hands. Stimulus Panel C corresponds in general
of two successive movements, and permitted Ss to two separate single-scale instruments and
to terminate their responses in one of eight end Response Panel C provides a set of responses
states. similar to those present whenever two separate
Stimulus Panel C contained a pair of hori- hand controls are used.
zontally separated lights and a pair of vertically At the center of each response panel was a
separated lights. The set of eight stimulus con- §-in. diameter metal disc, surrounded by a thin
nonconductive ring. Reaction time was meas-
ured as the time taken by S to move the stylus
off this metal button.
The stimulus panels were mounted at a 60°
angle to the horizontal. Response panels were
30° to the horizontal. The Ss worked from a
seated position. The stimulus panels were 15°
downward and 28 in. from their eyes, and the
response panels were at a convenient location in
front of them. The Ss watched the stimulus
panels and seldom looked at their hands or the
response panels.
The JE'S station contained separate selector
switches for the eight stimulus combinations and
a single noiseless activation switch, which turned
on the selected light(s) and simultaneously
started a 1/100-sec. timer. The apparatus was
designed so that different stimulus and response
panels could be substituted at S's station without
FIG. 2. The S's station showing the correct any change at E's station. The E sat where he
two-handed response on Panel C to the lower could observe S's movements and could record
right-hand light of Stimulus Set A errors as well as reaction time.
 S-R COMPATIBILITY 203

Response Sets
Stimulus
Sets

o o

I u s cms (wean)

FIG. 3. Mean scores for the eight Ss in each of the experimental

groups. The datum for information lost is the theoretical informa-
tion in the stimulus (3 bits) minus the computed value for information
transmitted, uncorrected for sample size.

Subjects.-—The Ss consisted of 72 airmen at move the stylus quickly to this position (indicate
Lockbourne Air Force Base, selected on the basis upper-right corner). If you start in the wrong
of two-choice reaction-time measures.4 Their direction, correct your movement as soon as
ages ranged from 18 to 29 years. All were right possible. Do not try to guess which light will
handed. They were excused from drill during come on as they will be presented in a random
the time spent as Ss and appeared to be well order. Work for both speed and accuracy
motivated. since both reaction time and errors will be
The 72 Ss were selected from a larger group recorded."
of 153 men so as to form eight equal-sized strata Each S was given 20 practice trials on his
which had homogeneous within-group average particular S-R combination, followed by 40 test
two-choice reaction-time scores. Each of the trials. The order of stimuli was randomized,
nine experimental groups was formed by drawing with the restriction of equal frequency for all
one S at random from each stratum. Each stimuli at the end of each series, and a further
group of eight Ss was then tested under one of the restriction against runs longer than two.
nine S-R combinations.
Procedure.—The instructions for each of the
groups were similar. Those for the Sa-Ra group Results
were as follows:
"Here is a stimulus panel of eight lights and a The experiment provides three meas-
response panel in which you can move this stylus ures of the effectiveness of each of the
to one of eight places. Hold the stylus in your S-R ensembles: (a) reaction time, (b)
right hand. When I say 'center' place it on this percentage of responses that were
center disc. I shall then say 'ready' and a few
seconds later one of the lights will come on. If errors, and (c) average information
this light (point) should come on, move the lost per stimulus. The means for all
stylus straight up. If this light should come on, three criterion measures are indicated
1
All Ss were tested by the second author. in Fig. 3.
204 PAUL M. FITTS AND CHARLES M. SEEGER

Reaction time.—In scoring the two- matched groups was then carried out
handed responses that were made on (see Table 1). The most important
Panel C, reaction time was taken as finding is the highly significant inter-
the average of the times for the two action effect. The variance that can
hands. This procedure is justified by be attributed to interaction is very
the finding that in two-handed re- much larger than the variance attrib-
sponses the times for the right and utable to the primary effects of either
left hands agree very closely. The stimulus or response sets alone.
average correlation of the two meas- The differences in the means for the
ures was .96 for the three groups that primary effects are interpreted as sig-
used Response Panel C, and the dif- nificantly different from chance, as
ference in the mean reaction time for indicated by the F ratios shown in
the two hands was only .004 sec. Table 1 for which the residual term is
However, for all one-handed responses used as the estimate of error. How-
on Panel C, the mean difference in ever, no reliable generalizations about
reaction time between the two hands these arbitrarily selected stimulus or
was .11 sec. in favor of the right hand, response codes can be made to situa-
and 23 out of 24 Ss using Panel C tions in which comparisons are made
were faster when using the right hand. among different sets of stimuli or dif-
In summary, when Ss had to move ferent sets of responses.
one hand alone, the right hand, which For every stimulus set there was a
moved away from or toward the body,
different best-response set, and for
was significantly faster (by about .108
every response set there was a differ-
sec.) than the left hand, which moved
ent best-stimulus set. For example,
to the right or the left, but when both
Response Set A led to the shortest
hands had to be moved together, they
had similar reaction times. The time mean reaction time in combination
for two-handed responses was approxi- with Stimulus Set A, but to the long-
mately the same as the mean for one- est reaction time in combination with
hand responses by the left hand. Stimulus Set C. The difference of
almost .4 sec. is 21 times as large as
The means shown in Fig. 3 are for the estimate of the standard error of
all stimuli in a set combined. The the difference. The three "best" com-
times for movements that were made binations are those that were pre-
in the wrong direction (errors) are dicted by Es on the basis of the corre-
combined with the data for correct spondence of the spatial codes. A
responses, since the mean reaction
times for erroneous and for correct
responses did not differ significantly. TABLE 1
ANALYSIS OF VARIANCE FOR THE
The mean reaction-time data for the
REACTION-TIME DATA
different Ss were tested for homoge-
neity of variance by Bartlett's test Mean
and no significant departure from Source df Square F

homogeneity was found, even though Matched Ss 1 42896

variance tended to increase somewhat Stimulus sets 2 86772 68.43*
as the mean reaction time increased Response sets 2 36352 28.67*
S-R interaction 4 170492 134.46*
(the range of SD's was from .036 sec. Residual 56 1268
for the Sa—Ra ensemble to .097 sec. for Total 71
the Sc~Ra combination). A double-
classification analysis of variance for * Significant beyond the 1 % level of confidence when
tested against the residual as an estimate of error.
 S-R COMPATIBILITY 205

direct test of the significance of the respect to the rankings assigned to the
difference between performance with different S—R ensembles. The three
corresponding and noncorresponding "best" combinations within each row
S-R ensembles is provided by a single- and column of Fig. 3 are the same for
classification analysis of variance, the two criteria. The difference be-
which is equivalent to a conventional tween the best and the worst combi-
t test between the means for matched nations appears to be relatively larger
Ss under the two conditions. Such an than was the case with the time scores.
analysis was carried out and the dif- It is interesting to note that Response
ference was found to be highly signifi- Set B, which in combination with its
cant (p<.001). The writers feel jus- corresponding stimulus set (Sb) led
tified in predicting that if other sets to the fewest errors, resulted in the
of stimuli and sets of responses should most errors when used in combination
be selected with due consideration with another stimulus set (S c ).
to the correspondence of the stimulus The data for responses to separate
and response codes, then it is highly stimuli revealed two important rela-
probable that interaction effects would tions. The first is that those S-R
again account for a large portion of the ensembles having lowest mean error
variance and spatially corresponding
scores also tend to have the most uni-
codes would again give superior
form time scores from one stimulus to
performance.
another. The second is that wherever
The reaction times for the Sa-Ra there is marked variability within the
and the Sb~Rb combinations did not responses to a set of stimuli, time and
differ significantly, but both were error scores tend to vary together, i.e.,
superior to that for the Sc—Rc combi- are positively correlated. For the
nation. It might be hypothesized four S-R ensembles with the greatest
that a bidimensional stimulus and number of errors the rank correlations
response coding scheme (two pairs of between time and error scores for the
lights and a two-handed response) is in- eight stimuli vary from .65 to .93.
efficient. It seems appropriate, how- Information lost.—The average in-
ever, to suggest an alternative hypoth- formation transmitted per stimulus
esis, that the inferiority may have was computed by Method 1 of Garner
been due to failure on the part of JEs to
and Hake (4). The data for infor-
observe some principle of response-
mation lost, shown in Fig. 3, are the
response compatibility in selecting the
differences between the theoretical
eight alternative movements consti-
information in each stimulus event (3
tuting Response Set C. It should be
mentioned that forward and back bits) and the average information
responses made with the right hand transmitted per stimulation.
only averaged .41 sec. for the Sc-Rc The results of the information anal-
combination, which is comparable to ysis agree closely with the total error
the mean reaction times for the Sa-Ra frequencies and, in fact, add little to
and Sb-Rb combinations, where the the grosser error analysis. The ranks
right hand was also used. of the nine groups are identical on the
two criteria except for a reversal of
Errors.—Approximately 10% of all
the two last (worst) groups. This
responses were in error, an error being
agreement is not surprising since the
defined as an initial movement in the
same type of error distribution, a pil-
wrong direction. These error data
agree in general with the time data in ing up of errors in certain cells of the
206 PAUL M. FITTS AND CHARLES M. SEEGER

matrix of transition probabilities, was data that are reported). In order to maintain
found for all nine S-R combinations. motivation Ss were scheduled in pairs (except for
the last 12 sessions of the odd S). One S was
It is not possible to compute a given a series of 16 trials on one of the three sets
meaningful estimate of average rate of S-R combinations in an unbroken sequence,
of information transfer in this experi- while the other S observed. The Ss then ex-
ment because of the relatively long changed places. This procedure was continued
time delay between successive re- until each S had been tested on each of the three
stimulus sets. The sequence of work and rest
sponses. However, short response and the order of trials on the three conditions
times were associated with small loss were balanced. Each stimulus appeared twice
of information both within the eight in a random order within each run of 16 trials.
conditions of each S—R ensemble and Each S made a total of 48 responses per session.
between the means for the nine differ- Initial instructions and other procedures were
the same as in the preceding experiment. At
ent S-R ensembles. Thus the results periodic intervals Ss were cautioned to try not to
of the information or error analysis, make errors. In addition to reaction time and
considered in relation to the reaction- errors, E recorded movement time, the time to
time analysis, provide empirical sup- traverse the selected pathway of Panel A. After
port for defining compatibility in each response Ss were told their reaction time
and if they had made an error, this was pointed
terms of the average rate of informa- out.
tion transmission. On Sessions 27 to 30 inclusive, a secondary
task, mental arithmetic, was carried on by Ss
concurrently with the perceptual-motor task.
EXPERIMENT II The E read aloud a series of numbers at a pre-
An important question concerning determined rate, and S gave the successive differ-
ences between the last two numbers read by E.
the results of the preceding experi- This secondary task was introduced to test the
ment is whether the differences be- hypothesis that the least compatible S-R combi-
tween various S—R combinations are nation would show the most deterioration under
transitory or permanent. An ex- conditions of additional load or stress. Standard
conditions were resumed on Session 31. The
tended learning study was carried out experiment was terminated at the end of Session
in an effort to answer this question. 32.

Method Results
The most desirable way to have conducted The mean reaction-time and move-
this study would have been to practice all of the
72 Ss used in Exp. I over an extended period. ment-time data for the five Ss are
However, this was not practical. Instead, a shown graphically in Fig. 4. Through-
single new group was studied for 32 training ses- out all of the standard sessions per-
sions covering approximately 2J months. The formance was consistently best when
Ss practiced making a single set of eight responses
(Response Set A) to each of the three different Ss responded to Stimulus Set A.
sets of eight stimuli employed in the previous Stimulus Set B gave almost equally
study. As an alternative Ss could have been good results. Stimulus Set C was
asked to make three sets of responses to a much the worst of the three. At no
single set of stimuli, but this was not done
because less initial habit interference is to be time during the 32 days did any of the
expected if the same responses are learned to Ss consistently respond as quickly to
three sets of stimuli, than if three sets of responses Stimulus Set C as they did to the
are learned to a single set of stimuli. Response
Set A was chosen because there is no ambiguity other two sets of eight stimuli. The
in the scoring of these responses. mean times for the three stimulus sets
Six male students at Ohio State University on Days 17 through 26 inclusive were
were started on the training series but one drop- as follows: Sa, -272 sec.; Si,, .286 sec.;
ped out after 20 sessions and his data are not
reported (although they are comparable to the Sc, .355 sec.
 S-R COMPATIBILITY 207

0.90
\
—'—S&~ RA
O80 SB-RA

070 1
SECONDS

\ DESTRACTlON '
(-TRIAUS

TIM
\ / \ REACTION ES^
— 0*5
TIME

^ ^ ^ ^ ^ , - 1

020

OIO A HOVEHBIT TIMES» ^

000
25 30
FIG. 4. Learning curve for five Ss during the 32 training sessions
of E i p . II. All Ss had 16 practice trials on each of the three S—R
combinations each session.

Movement time in responding to five Ss for successive blocks of time,

Stimulus Set C was slightly but con- are given in Table 2. It can be seen
sistently slower than to Sets A and B. that the Ss never succeeded in elimi-
Averaged over Trials 17 through 26 nating all errors. In fact, errors
these times were as follows: Sa, -059 occurred more frequently as training
sec.; Si>, .061 sec.; Sc, .067 sec. progressed, indicating perhaps that S
The error data, averaged over all emphasized speed at the expense of
accuracy. Errors were made most
TABLE 2 frequently to Stimulus Set C through-
PERCENTAGES OF MOTOR RESPONSES THAT
out the training period. Thus both
W E R E IN ERROR DURING DIFFERENT time and error data indicate consis-
BLOCKS OF TRAINING SESSIONS* tently poorer performance on the
Sc-Ra condition.
S-R Combination On Days 27-30, when the load on
Sessions
Ss was increased by the addition of a
Sb-Ra
secondary task, reaction times were
i-6 2.5 3.7 11.9 substantially slower for all three S-R
7-16 5.6 8.7 12.5
combinations. All five Ss continued
17-26 6.9 10.0 11.9
27-30** 4.4 6.9 15.6 to give the slowest reaction to Stimulus
31-32 8.1 9.4 12.5 C; however, the relative differences
*SessionB consisted of 16 responses per S-R combi-
between groups were much smaller
nation for each of the five Ss. than they had been previously. The
** On Sessions 27, 28, 29, and 30 the Ss responded to
stimulus lights and solved mental arithmetic problems frequency of arithmetic errors was ap-
concurrently. The number of arithmetic problems
attempted during each session was 50 per S and the proximately the same for all three
percentage of arithmetic errors for the three sets of
stimulus lights was 4.1, 2.9, and 4.2, respectively. groups, but there was a large differ-
208 PAUL M. FITTS AND CHARLES M. SEEGER

ence between groups in the frequency This view holds that S's behavior,
of movement errors. The introduc- notwithstanding long experience in a
tion of an additional task late in learn- particular situation, is never entirely
ing in this case apparently served to relevant to the specific constraints of
reduce the degree of readiness for the that particular situation. Instead, S
motor task, minimizing reaction-time responds as if additional possibilities
differences, but increasing differences were present.
with respect to errors in the motor This interpretation supports the
task. However, the data are not view that stimulus and response sets
conclusive on these points and the are optimally matched when the result-
test of the hypothesis regarding the
ing ensemble agrees closely with the
effects of increased load is considered
basic habits or expectancies of indi-
inconclusive.
viduals, i.e., with individual and with
population stereotypes. It must be
DISCUSSION
remembered, however, that the expec-
The results of the two experiments tancies referred to are those which
demonstrate clearly the importance of hold for the particular situation under
stimulus and response coding for the study, and that population stereotypes
maximum rate of information transfer should be determined with due regard
in a perceptual-motor task. It also to the total situation. For example,
is clear that some S-R compatibility in the present experiment the JBS deter-
effects are relatively unaffected by mined, by preliminary trial, that Ss
extended practice. "preferred" to move toward rather
An interpretation of the basis for than away from a stimulus light.
the relative permanence of these ef- However, it is known that in many
fects, which seems appropriate to the stimulus tasks where the S controls
concept of compatibility, is one stated the stimulus, the opposite motion rela-
in terms of the capacity to learn to tion is the expected one. When Ss
deal with sets of probabilities (proba- have to learn to deal with the proba-
bility learning). The response that a bilities inherent in a particular situa-
person makes to a particular stimulus tion, the correspondence of these
event can be considered to be a func- specific expectancies to the more gen-
tion of two sets of probabilities; (a) eral expectancies of the individual with
the probabilities (uncertainties) appro- respect to that kind of situation is an
priate to the situational constraints important aspect of the learning task.
established by E's instructions, by the Further support for considering S-R
reinforcements experienced in the ex- compatibility to be a function of stim-
perimental situation, and by other ulus and response matching comes
aspects of the immediate situation; from the positive correlation between
and (b) the more general and more the two criteria, time and errors, used
stable expectancies or habits based on in evaluating performance. This posi-
S's experiences in many other situa- tive correlation would be expected if
tions. It is suggested that extended additional information transforma-
training will nearly always lead to tions, or re-encoding steps, were added
changes in the former but often will to a communication system, since each
have relatively little effect on the transformation would be likely to add
latter. an additional time delay and to in-
 S-R COMPATIBILITY 209
crease the total probability of errors. three sets of stimuli persisted over the
In the human it is hypothesized that 32 days.
the fastest responses may be the most The results are interpreted in terms
accurate because they involve S-R of probability learning and the neces-
ensembles in which the transfer of sity for (hypothetical) information
information from stimulus to response transformation or re-encoding steps.
is most direct, i.e., involves the mini- It appears that it is very difficult for
mum number of recoding steps. The
concept of intervening information Ss to learn to deal effectively with the
transformations does not attempt to information (uncertainties) charac-
explain how such recoding occurs teristic of a specific situation, if these
within the nervous system, but it is in uncertainties are different from the
agreement with the subjective reports more general set of probabilities which
of Ss who maintained that it was diffi- have been learned in similar life
cult, in the case of the less compatible situations.
S-R ensembles used in the present
study, to be prepared to respond to all REFERENCES
of the eight stimulus possibilities. 1. FITTS, P. M., & SIMON, C. W. Some rela-
tions between stimulus patterns and per-
formance in a continuous dual-pursuit
SUMMARY task. / . exp. Psychoi, 1952,43, 428-436.
2. GAGNE, R. M., BAKER, K. E., & FOSTER, H.
Experiment I was planned to On the relation between similarity and
test the hypothesis that information transfer of training in the learning of dis-
transfer in a perceptual-motor task is criminative motor tasks. Psycho!. Rev.,
in large measure a function of the 1950, 57, 67-79.
matching of sets of stimuli and sets of 3. GARDNER, J. F. Direction of pointer motion
in relation to movement of flight controls.
responses. Nine S-R ensembles, in- Wright Air Development Center, Dayton,
volving variations of the spatial pat- Ohio. AF Technical Report No. 6016,
terns of stimuli and responses, were 1950.
4. GARNER, W. R., & HAKE, H. W. The
studied in an eight-choice situation, amount of information in absolute judg-
using groups of matched Ss. ments Psychoi. Rev., 1951, 58, 446-459.
The results, analyzed in terms of 5. JENKINS, W. L., & OLSON, M. W. The use
reaction time, errors, and information of levers in making settings on a linear
scale. J. appl. Psychoi., 1952, 36, 269-
lost, support the hypothesis. They 271.
indicate that it is not permissible to 6. KLEMMER/E. T., & FRICK, F. C. Assimila-
conclude that any particular set of tion of information from dot and matrix
stimuli, or set of responses, will pro- patterns.^/, exp. Psychoi., 1953, 45,
15-19.
vide a high rate of information trans- 7. KNOWLES, W. B., GAKVEY, W. D., & NEWLIN,
fer; it is the ensemble of S-R combi- E. P. The effect of "speed" and "load"
nations that must be considered. on display-control relationships. / . exp.
Experiment II was planned to test Psychoi, 1953, 46, 2, 65-76.
8. LEWIS, D., SMITH, P. N., & MCALLISTER,
the permanence of three selected S-R D. E. Retroactive facilitation and inter-
compatibility effects. Five Ss were ference in performance of the modified
trained for 32 days to make a par- two-hand coordinator. / . exp. Psychoi.,
ticular set of responses to each of three 1952, 44, 44-50.
9. MITCHELL, M. J. H., & VINCE, M. A. The
sets of stimuli. Differences in reac-
direction of movement of machine con-
tion time, movement time, and fre- trols. Quart. J. exp. Psychoi., 1951, 3,
quency of errors in responding to the 24-25.
210 PAUL M. FITTS AND CHARLES M. SEEGER

10. SHANNON, C. E. A mathematical theory of 1896, 3, 611-612; 1897, 4, 342-351,

communication. Bell Syst. Tech. /., 466-471.
1948, 27, 379-423; 623-656. 13. WARRICK, M. J. Effects of motion rela-
11. SIMON, C. W. Instrument-control configu- tionships on speed of positioning visual
rations affecting performance in a com- indicators by rotary control knobs.
pensatory pursuit task. Wright Air Wright Air Development Center, Dayton,
Development Center, Dayton, Ohio. AF Ohio. AF Technical Report No. 5812,
1949.
Technical Report No. 601S, 1952.
12. STRATTON, G. M. Vision without inversion (Received for early publication
of the retinal image. Psycho!. Rev., May 22, 1953)