THE INFLUENCE OF AFFECT-STIMULI
ON SUBSEQUENT PERFORMANCE
HERMAN IVAN LEVIN
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
The author wishes to thank the members of his supervisory
ca:mittee: Dr. James C. Dixon, Chairman; Dr. Rolland H. Waters;
Dr. Richard J. Andersonj Dr. Elmer D. Hinckley; and Dr. Albert M.
Barrett for the interest shown and the effort expended in the production
and development of this dissertation.
To my friends and fellow graduate students, Robert Procter,
Denis O'Donovan, Charlotte Ince, who in one ay or another have con-
tributed immeasurably to this work, also my sincerest thanks and
TABLE OF CONTENTS
ACKNOWIEDGEITS . . . . . . ........ ii
LIST OF TABLES.,L e ....... .. . .. iv
I. INTRODUCTION . . .. .. . . . . 1
II. PROSEDURE. .. . . . ........ . . 7
III. RESULTS. .. .. . . . . 16
IV. DISCUSSION . ... ... . ..... 23
V. SUMART ................. ...... 28
APPEDIXES. . . . . .... . . . 30
I. COMPARISON OF AVERAGE AFFECT RANKINGS OF STIMULI BI
JUDGES AND SUBJECTS. SALE OF AFFECT RANGED FROM 1
FOR VERT PLEASANT; THROUGH 3 for NEUTRAL TO 5 FOR
VERT UNPLEASAT......... ............ 31
II. STIMULUS MATERIAL: EXPERIMENTL TASKS PRECEDED BY
AFFICT-WiRDS ... . . . . . . 32
III. DIAGRAM OF CONTINUOUS VOICE KET. . . . . ... 34
IV. DIAGRAM OF WIRING OF SWITC7G ........... 35
BIBLIOGWAPTI . . . a . . . . . .. .. . 36
LIST OF TABLES
1. Results of the Complex Analysis of Variance for Speed . 17
2. Results of the Complex Analysis of Variance for Accuracy. 19
The presentation of a atbmlus to influence the responses of
a subject to susequent stiamlation is an epetrimetal situation fre-
quttly encountered in psychology. Experiments in such diversified
areas as larniag, transfer, motivation, and pereeptin have all used
this prooodure. The present experiment canferm to this general
appreskh by investigating the influence of affect-stiMul upon behavior
in a subsequent task situation. It attempts to find an answer to the
quastion-Doee the response of a subject to a standardized task differ
in either speed or accuracy when the task follows on ae eeeasion a pleas-
ant stimulus, on a second occasion a neutral stirmuia, and on a third
eceasies an unpleasant atialuam
The literature concerning the influence ef affect on future
performance can rather arbitrarily be divided into thoue studies which
are aat concerned with the pereeptaal variables of performnee and these
which are emst omeerned with the response variables of pferfirmae.,
Influenee of Aft ct en weoptumal Variables of Performane
Sines the work of tanford (10) en need in pereeptieo, it Lea
beems mere and mere apparent that peyebSloegeal falters ean and 4
influence the subject's perception of external stimuli. Stimuli may,
within the bounds placed by reality factors, be selected on the basis
of the subject's needs, desires, values, and fears (M). Pleasant,
valued, or need-related stimuli tend to be perceived more readily than
do neutral stimuli (4). Unpleasant, threatening, or taboo stimuli tend
in general to be perceived less readily-though there is considerable
evidence for wide individual variation in the perception of unpleasant
The affective nature of the stimulus not only influences the
perception of the affect-stimulus itself, but also tends to influence
the perception of other stimuli occurring in the same or in immediately
succeeding fields. Affect-stimuli seem to impose "severe limits . .
upon the breadth and character of perception" (1). Neutral stimuli in
the same or in immediately succeeding fields will be less adequately
perceived (1). A task to be performed in relation to these stimuli,
it may be assumed, would be hampered or left undone. Two studies
have been found in the literature that are concerned with the percep-
tion of neutral stimuli in affective fields.
McGinnies (7) tested the influence of unpleasant stimuli on the
thresholds of subsequently presented neutral stimuli in an attempt to
demonstrate the operation of motivational factors in perceptual defense.
His subjects were presented eight task words-four following the pre-
sentation of taboo words and four following the presentation of neutral
words. The task words were equated for frequency and neutrality. The
recognition tbreshe s for task words were sigtifieantly higher when
they folleeed taboo words than when they follend neutral words.
Seper (11) attempted to deamatrate Gebs hypothesis that the
preentatim of stidEml which a pleasant or unpleasant to the individual
result in a decreased ability to peeeive other details in the pereeptual
field. He presented his subjects with three eets of typed werdnlists
arraae soe that in ome ast of list. every fifth word w- pleasant in
a second, every fifth word was neutral; and in a third, every fifth
word was mpleeent. After oae-minute presentation of the lists, the
subject ware asked to recall as many words as peesible. Results in
general wore in the predicted direetias; that i., a significantly
greater number of field words were reported from the neutral than from
either the pLaesnt or unpleaant lists.
The Influeee ef Affeet en Rpese Variables in Perfornmee
Two area of research ceoerning the influeuee of afeet on
reepense variable here converge. One of them areas is the infliuaee
of reports of previous "weeNese or "fAilure" en perforneme; the other
is the influence of affect-stimli en asseelation to future stimulus
words on the Word Association Tfet.
Studies on Ieueees" and "failure" tend to show a deterieretion
in pfrfomanee after fatlue experience; ieirmeint in performed
after suees experienees Tse effects eeur beth in the learning of
verbal material and in the perfor eem e ef payshlmtu tracks. The sK-
peoted effeets, however, are not always the obtained results. SNveal
studies even show improvement under failure conditions. Lazarus (6),
in suwming up the review from which this material was obtained, states
that the nature of the task, the kinds and amount of stress, individual
differences, and the measures of efficiency must all be taken into
account in predicting the influence of affect on performance.
Both Jxng (5) and Rapaport (9) in their experiences with the
clinical application of the Word Association Test have observed a
perseverative influence of emotionally toned words on the associations
of subjects to subsequent stimulus words. Perseverative influences are
stated as being observed in the inappropriate repetition of the associa-
tion word and/or in aa increase in association time. Hull (3) attempted
as a minor part of his experiment on the diagnostic ability of complex
signs to test perseveration as one of his complex signs. He found that
neither of the clinically observed effects of perseveration occurred
in his subjects. The experimental design, not being specifically set
up to test perseverative influences, did not, however, permit an adequate
test of this phenmenon. The clinical observations of Jung and Rapaport
in this area remain untested.
Studies in both the areas of perception and performance tend to
indicate that unpleasant stimuli will negatively influence performance
on a subsequent task. Disagrearent occurs between studies in the two areas
on the influence of pleasant stimuli on subsequent performance. Since,
however, both measures in the present experiment are predominately
perfermsnce mamures, it is expected that the effects of pleasant
timali. will be to facilitate subsequiet perforaeee. The hypothesa
will, however, be east in the null form.
etst sois 1.-There wll be no relation betwea the affective
taen of a stimulus, whether pleasant, unpleasant, or neutral, and the
speed of the subject's response to a staanrdised task which immediately
follows upon the presentation of the stimulus. That is to say, when the
speed of reaction to a standardized task which follws an unpleasant
stimuis is compared with the speed of reaction to tir same standardized
task whme it follows ether a neutral ar a pleasant stimulus, no
statistically sigifieaat differeaee are expected to oecur.
korrtpig 2-There will be no relation between the affective
teoe of a stial s, whether pleasant, unpleasant, or neutral, and the
accuracy of the subjet's responses to a standardized task which
immediately follies upeo the presentation of the stimulus. That is to
say, whea the accuracy of reaction te a standardized task which follows
an uIpleasant stimulus is compared with the accuracy of reaction te the
ame standardized task when it follows either a neutral or a p] asant
stimulus, ne statistically sigtifieat differences are expected to occur.
To tet the aboer hypethewss, subjects were asked to spell
aled a series of three-consenant nonsense syllables following pi asant,
unpleasant and neutral words. Measures of speed and accureay of this
respense wre obtained.
Definition of a Term
At this point note should be taken of the terms used to denote
the subject's speed of reaction to the task following the affect word.
Both the terms speed of reaction and reaction time have previously been
used by experimenters to denote the temporal interval between a stimulus
and the response. In this experiment these terms are used to denote
the time it takes the subject to perform the task following the affect-
To test the hypothese, it was ne--msry to g through the
following steps in apparatus canstructier and preparation: (a) the
affect-stimul had to be selected; (b) the nceaneM-syllable tasks had to
be litrwise eeleetedj (c) slide containing one affeet-stimulu and one
meense-syllable task had to be eceatrueted; (d) the slides had to be
arranged J3 a particular eder for presentation to the subject; (e) an
inttraent to present the slides to the subjects had to be obtained
(f) appaatus that would record the speed and accuracy of respese to
the nsenmae-syl3able tasks had to be olnstrmated; and (g) the apparatus
had to be wired so the experimnter eaMco control from Moe sidtch the
presentation of the material and the apration of the instrument that
would record the speed of the subject's response to the task.
Selection of AMftet-worde
Affeet- srds had to be selected which weld shew a consistent
tendency to fall into either pehaant, neutral or unpleasant categories
when ramked by the judges. There was no need to equate the words in the
three affet-categories as to length or frequmeny sine it is the speed
and accuracy of nomaense-syllabl3 verbal report, rather thgn part of
the affect-stimulus, which is the subject's task.
The affect-words were obtained by presenting five graduate stu-
dents in psychology with a list of 394 words. These words were chosen
from the Thorndike Word List (12) with the intention of obtaining words
which were likely to have pleasant, unpleasant, or neutral connotations.
Each of the five graduate students ranked the 394 words on a five-point
scale. The scale points were: (1) very pleasant, (2) mildly pleasant,
(3) neutral, (4) mildly unpleasant, and (5) very unpleasant. The average
rank for each word was obtained. Any word which had an average rank of
at least mildly pleasant affect-tone and which was not ranked by any
judge as being unpleasant was considered for use as a pleasant affect-
word. Any word which had an average rank of at least mildly unpleasant
affect-tone and which was ranked by no judge as pleasant was considered
for use as an unpleasant affect-word. Only those words ranked by all
five judges as neutral were considered for use as neutral words. Twenty-
five pleasant, 25 neutral, and 25 unpleasant affect-words were chosen
randomly from thoe meeting the above criteria and were used in the
experimental slides. In addition, three pleasant, three unpleasant,
and three neutral words were selected to be used as part of the practice
slides. Practice slides were presented prior to the experimental
slides in an attempt to reduce practice effects which are likely to
occur in this type of experiment, and which would affect the dependent
variables of speed and accuracy.
Judges' ranks of the affect-words used in the experimental slides
eaa be see in Appendix I.
N nseme-syllable Tasks
The affeet-etimuli having been cheese, it was then napsesary to
shoose a sataWdardld task upon which e test their effects. A diffi-
eulty than areee in task election. If the m task were repeated
after each affeet-wurd, the practice effects would likely be ae great
that they wemld overshade the influsee of the affect- mrds themelvee.
If, on the other hand, different taaks were to follow each affest-werd,
differences in tak difficulty would like be so great that, again,
the influeP e of the affect-wrds weld be overshadowed. A cmpronse
was settled upen, using 25 different tasks, so that eash task would
felles one of the 25 pleasant werd, Me of the 25 neutral werds, and
me of the 30 pleasant werds. The task difficulty was therefore
equated by having the sam 25 tasks esur in seek of the three offeet-
grops. The effects of practice, which WLd have been great with the
repeed pressetation of the same task, were mininised by the fact that
there nre 25 different tasks to which the subject was to respond. Each
bjeet, then, received a total of 75 slede and made 75 ueasurable
To eonitutuute the taks nonsense syllables were choose. They
wen heksee because their degree effadliarity eeald be strolled;
because estimates of accuracy and peed of their prodetion could be
fairly easily measured; and because they wre a convenient material
with which to work. The fact that their degree of familiarity could
be controlled was most important to this experiment; it served to narrow
the range of difficulty among tasks without increasing the effects which
might result from practice. Under these conditions, the influence of
affect alone was more likely to be observable.
Eighty-four nonsense syllables consisting of three consonants
were chosen from Witmer's (13) list of association values of consonant
nonsense syllables. The range of association values from which the
nonsense syllables were chosen was from 0 per cent to 25 per cent.
Three nonsense syllables constituted a task. Nine of the nonsense
syllables were used to construct the three tasks for the practice
slides; 75 of the nonsense syllables were used to construct the 25
tasks for the experimental slides.
Preparation of Slides
Affect--words and nonsense syllable tasks were randomly assigned
to one another. Again, it should be noted the stipulation was that each
task would follow one unpleasant stimulus, one neutral stimulus, and
one pleasant stimulus. The physical preparation of the slides consisted
of typing the affect-word and the three nonsense syllables in a single
column on 3-1/2 x 4 inch sheets of thin drawing paper. Each sheet was
then placed between two thin pieces of glass to make a slide. Appendix
II shows the 25 experimental tasks preceded by their affect-words.
Arrangement of the Slides for Presentation to the Subjects
Since subjects tend to become more efficient in handling a
repeated similar situation, the slides were arranged in a stratified
random crder. Stratification was based on the assumption that in a
list in which three variables are to be distributed amoag 75 positions,
one run of four in a row, three runs of three in a row, eight runs of
two in a row, and 46 runs of one in a row ean be expected to occur.
Pleasant, unpleasant and neutral stimuli were assigned randeoly to the
various rans. The one stipulation ade was that 25 of the stimuli be
neutral, 25 pleasant, and 25 unpleasant. The result of assigning
affect-stimuli to runs provided one run of four, one run of three, two
runs of two, and 14 runs of ane for the neutral stimuli, and one run of
three, three runs of twe, and 16 runs of one for both the pleasant and
Runs were assigned to each of the eight deeiles and to each of
the units of each of the deciles from a table of random numbers. For
instance, the experimenter would begin with the run of four neutral
slides and look in a table of random numbers to determine the appropriate
decile. The number seven occurred. The run was, therefore, to be
located in the seventh decile. The experimenter would then look again
into the table of random numbers to determine the unit position in the
docile of the run. The number two occurred and the run was, therefore,
to begin in the second unit of the seventh demile, or the number 62.
Two restrictions were placed on the ordering: (1) no run could succeed
or precede another run of the same affective toae; and (2) in each third
of the 75 slides, eight words of one affect, eight words of a sacnd
affect, and nine words of the third affect were to occur.
The specific affect-words were then placed randomly in the
appropriate positions. Here another stipulation was made. To assist
in controlling practice effects, no two similar sets of nonsense syllables
could occur in succession.
A projector that could handle slides of the appropriate dimn-
sions was obtained to project the slides on a projector screen.
A tape recorder was used to record the subject's report of the non-
sense syllables. Any errors, omissions or insertions in the spelling
of the nonsense syllables was considered an error.
A continuous voice key was constructed to measure the length of
time it took the subject to reproduce the nonsense syllables. Sounds
spoken into a microphone attached to the voice key closed the circuit,
and when the sounds ceased, the circuit was opened. By connecting this
apparatus to a signal magnet, by attaching a writing pen to the signal
magnet, and by placing the writing pen on a kymograph in motion, a
temporal pattern of speech sounds was obtained. Appendix III contains a
circuit diagram of the voice key.
The kymograph speed was regulated at 225 centimeters per minute.
When the speed was measured in millimeters, speed could be recorded to
27/1000 of a second. (Measurements of speed will be reported in milll-
meter units without converting them into seconds.)
The Wiring of the Apparatus
The apparatus was connected in such a way that the experimenter
could control the projector, kyegraph, and veio key by means of a
five-pole, two gang switch. It was arranged so that, in position 1,
the tape in the kyaograph was stationary, the pea was in position a"n,
the projector and the voice key were off; in positions 2 and 3, the tape
began te meoe, twh pea rose into position "b", the projector and the
voice key remained off; in position 4, the tape continued to move, the
pen fell back into position sa", the projector and voice key began to
work. While in petition I, each time the veLee key wma activated the
pen moved into position "b"; each time it was deactivated it returned
to petition "a". Thus in position 4, a graphic representation of the
subject's speech patterns was obtained. Position 5 as a cheek position
to test the workings of the voice key.
Appendix IV contains a circuit of the switch connections for
the two gangs.
Seventy-five male undergraduabe students from the University of
Florida were used as subjects. Meet, though not all, were volunteers
from the beginning psyehelogy courses.
The subjects were brought into the ream and the tape recorded
was started. Each subject was then heated facing the projector esreen
and handed the microphone. A few remarks were made by the experimenter
in an attempt to make the subject feel more cuofortable. The experi-
menter then read the following instructions to the subject:
"You are to watch the screen each time I say 'ready.' Soon
after, a slide will be projected on the screen. This slide will consist
of one familiar word and three nonsense syllables. You are to read the
familiar word to yourself and then spell out loud into the microphone
the nonsense syllables. Spell the nonsense syllables as fast as you
can. If you make a mistake, don't stop but go on to the next letter.
For instance, a slide is projected on the screen with the word THERE
appearing on it and underneath in a column appear the nonsense syllables
ZFT, QHN, MTC. You are to read the word THER to yourself and then
spell out loud into the microphone Z-F-T, Q-H-N, M-T-C. Do you under-
If any questions were asked, further explanations were given.
The 84 slides, including the nine practice slides and the 75
experimental slides, were then presented to the subject.
After the presentation of the slides, the subject was placed at
a desk with the list of 75 affect-words before him. The subject was
then given the following instructions:
"I want y~o to rank the words on this sheet of paper in one of
five categories. If the word has a very pleasant ring or association
connected with it, circle the letter P. If a mildly pe asant association,
circle the first dot. If the word has neither a pleasant nor unpleasant
association connected with it, but a neutral ome, circle the letter
NI If the word has a mildly unpleasant association, circle the second
dot. And finally, if the word has a very rmpeasant association,
circle the letters UP. Are there any questima*?
The subject then rated the affeet-vwrds in the five possible
This latter procedure, the rating of the experimental affect-
wards by the subject, was included to determine the similarity between
subjects' aad judges' ratings of the affeet-words. If the ratings are
markedly dimimilar, then the independent variable of affective tone
is cefeunded and the experimental results queetionable.
The Correlation between Judges' and Subjects' Ratings
The mean of the affect-ratings of the judges' and of the
subjects for each of the 75 experimental words can be seen in Appendix
I. A rank order correlation of .94 was obtained between the two sets
of data. Subjects' and judges' ratings of affect-words were, therefore,
highly similar in this study.
The Influence of Affect-stimuli on Speed of Reaction
Table 1 contains the results of a triple classification, complex
analysis of variance for task speed. Since 75 subjects were presented
with 25 tasks in each of three different affect-situations, the analysis
of variance has three categories in one variable (pleasant, unpleasant,
and neutral affect), 25 categories in another variable (each of the 25
sets of nonsense syllables) and 75 categories in a third variable (each
of the 75 subjects).
The 75 individuals differed significantly from one another in
speed of reaction to the tasks at less than the 1% level of confidence.
As expected, one individual reacted more swiftly to a situation of this
nature than did another.
RESULTS OF THE
COMPLEX ANALYSIS OF VARIANCE FOR SPEED
Soere Ste of Derees of Variane Significance
Squares FreedeM Ektimate Level
Affect 22.64 2 11.32 Net significant
Tasks 5657.18 2k 235.72 Belor 1%
Subjects 41474.88 74 560.07 Bolow 1%
and objects 11191.50 1776 6.30 Blelw 1%
and affect 4j3.38 48 9.23 Belar 1%
and subject 473.72 148 3.20 Not significant
Triple interaetio 13357.90 3552 3.76
Taeal 72621.20 5624
am-nce-syllable tasks also differed significantly from me
another in their tendency to influenee the speed of the subjects'
reaetioms. This wm so, despite ti attempted to control task difficulty
on the basai of familiarity. Certain nowsese-eyllable groups, regardless
of te affeet-stImula wIhih proved them, were rested to ore aloirly
by the subjects tha were other aaense-syllable greape.
The results of the toet of tih first hypothesis are rather eem-
phex. Stim li of different affect-tone did not sigific atly influence
the speed of reaction to subsequent tapr wiMe tea s if d ai
significantly from one another. However, when affect-tone and the nature
of the task were taken jointly into consideration (the interaction of
task and affect), significant findings below the 1% level were obtained.
The experimental answer, then, is that p3e asant, unpleasant and neutral
stimuli influenced the task differentially when the nature of the task
was taken into consideration. When compared with neutral stimuli, un-
pleasant stimuli slowed the response and pleasant stimuli speeded the
Concerning the other interactions, the interaction between
tasks and subjects was significant; the interaction between affect and
subjects was not significant.
The Influence of Affect-stimuli on the Accuracy of Report
Table 2 contains the results of a triple classification, complex
analysis of variance for the accuracy of the subjects' reports. Since
the same number of measurements were made on the accuracy of the subjects'
responses as on the speed of the response, again the analysis of variance
has three categories in one variable (affect-stimuli), 25 categories
in another variable (nonsense-syllable groups), and 75 categories in a
third variable (subjects).
Individuals again differed significantly from one another
(below 1% level of confidence) in their reactions as measured by accuracy
of report; that is, some subjects tended to be more accurate than others.
Nonsense-syllable groups also again differed significantly from one another
in the accuracy of the reactions which they tended to elicit. Some
nemeemne-syllable groups tended to elicit more error than did others.
RESULTS OF THE
COMPLEX ANALYSIS OF VARIANCE FOR ACCURACY
Sms of Degrees of Varioane Significance
Square Freedom atimate Level
Affeet .96 2 .48 Not significant
Task 38.23 24 1.59 Below 1%
Subjects 61.22 74 .83 Below 1%
Interaction : tasks
and subjects 43j.12 1776 .2k Below 1%
and affect 17.7k4 S .37 Below 1%
and subjects 28.03 148 .19 Net significant
Triple intereetien 637.27 3552 .18
Tetal 1217.57 562k
The second hypethesie-that differing affeetive stimuli will net
result in varying accuracy on the taks--eannot be rejected, when the
tka sva ry sigifieaatl, mi t pmalves. HeMver, Aea affective tone
and the inherent task differences wre taken jointly into eernideration
(interaction et task and affect), signifioant differences beler the 1%
level of confidene were obtained. Both pleasant and unpleasant stimuli
elicited significantly fewer errors than did mutral stimuli.
Concerning the other interactions, the interaction between tasks
and subjects was significant; the interaction between affect and subjects
Comparison of the Two Complex Analyses of Variance
If we regard fast and accurate report as task efficiency, then
efficiency varied with the response measure used (when task differences
also were taken into consideration). When the speed of reaction to the
nonsense syllables is taken as the measure of influence, unpleasant
stimuli instigate the longest reaction times (average 15.02 mm. per
task per subject); neutral stimuli the next longest (average 15.00 nm
per task per subject); and pleasant stimuli the shortest (14.88 mm
per task per subject). When accuracy of report of the nonsense syllables
is taken as the measure of influence, more errors were made on nonsense
syllables following neutral words (average .187 per task per subject);
than on nonsense syllables following unpleasant words (average .161 per
task per subject) or pleasant words (average .157 per task per subject).
The results indicate that while unpleasant stimuli, when compared with
neutral stimuli, resulted in slower reactions, they also produced more
accurate reactions. Pleasant stimuli resulted in faster reactions and
more accurate reports when compared with both neutral and unpleasant
Nature of the Task that Interact with Affect to
Pro de Siaif iat rmaml
It has been noted above that the interaction of task differences
and affective tone was significant. The following data are presented
in an attempt to test the assumption that the significant interaction
between task and affect is a function of task difficulty. If task
difficulty is the important property of the task in the interaction
effects, the responses to the ten hardest tasks would be more likely
influenced in the direction expected from the obtained results than
would the responses to the ten easiest tasks. This would mean (a)
that, for speed, responses following pleasant stimuli would be the
shortest responses folloLng unpleasant stimili the longest; (b)
that, for accuracy, reepnases following pleasant stimuli would contain
the least errors; responses following neutral stimuli would contain the
.Fth sign tests of sigaifianese and t tests of significance
were obtained for the ten hardest and the ten easiest tasks caoparing
the speed and accuracy of respeonse following each of the three different
affects (for any one sign test or any one t test, of course, only tasks
following two different affects were compared). The results were ambiguous,
at least as far as the interpretation is coieerned. Trends were of such
a nature that the more difficult tasks tended to fall in the expected
order more frequently than did the easier tasks for both measures of
speed and accuracy. The sign tests only tended to confirm the trends 4
for speed (the comparison between neutral and pleasant and between
pleasant and unpleasant were significant for difficult tasks below the
5% level of confidence). The t tests showed no significance for either
measure of response either for difficult or easy tasks. Any interpreta-
tion of task difficulty as being the property of the task contributing
to the significant interaction between affect and task would have to be
made with extreme caution.
Lasaru (6), as has been noted in the introduction, states that
the nature of the tack, individual differences, and measures of efficiency,
as wll as the kind and degree of affect influeme the results of a
subject's performance on a bask following affective stimulation. This
is what teded to happen in this study. The degree of affect is mild.
en this is eo ad task vary greatly &a g theselves, affect does
net seem to inflwuae performance. When, however, the nature of the
task is taken into consideration, even this mild degree of affect
become ipertant in the efficiency of produetiea of the release. When
different response masires of efficiency are used, a given affect, un-
pleasant in nature, may have at oe and the sae tim facilitating
effects ae one type of response measure and detrimental effects on
another type of mersre. The oely aae of the peesible determinant of
the task performaaee mentioned by LMearue that is net observed in this
study to differentiate reaetions is individual differences in the reaction
be affect-tomes (the interaction between affect and subjects is not
significant). This seold eeily be due to the mildnea of the affective
Interaction of Affect and the Nature of the Task
Mild affect, when considered alone, does not influence either the
speed or accuracy of the performance of a subject on a series of tasks
which differ widely among themselves. When, however, the nature of the
task is taken into consideration the affect does seem te influence the
performance. A similar situation has occurred previously in the
literature. This was encountered by Postman and Schneider (8) in their
study comparing the relative effects of value and frequency on word
threshold. These authors sum their results thus, "We find frequency
of word usage is significant whereas value in and of itself is not.
The interaction between frequency and value is, however, significant, i.
e., the effect of value rank on duration thresholds depends on the level
of word frequency." The obvious correlate of word frequency in the
present study is task difficulty. Task difficulty, however, did not
unambiguously allow for the rejection of the null hypotheses. Though
trends were in the appropriate direction (performance on the more
difficult tasks occurring more in line with the order predicted on the
basis of the obtained results), variations among tasks were too great
for these trends to reach significance.
Influence of Varying Degrees of Affect on Performance
Only one degree of affect was used in this study, a very mild
one. Previously in the literature muh greater degrees of affective
stimulation have been used-to the point of emotional arousal. To fit
the results of this study into this body of literature, Duffy's (2)
activation theory of emotion is used. Each of three emuerpts frao her
work is followed by an attempt to integrate the results of this study.
"The chief point in regard to arousal, . is that arousal
occurs on a continuum, from a low point during deep sleep to a high point
during extra effort or great excitement, with ne distinguishable break
for such conditions as sleep or 'emobiem. Our affective stimuli are
sew-here between the lewr ead and the middle of the continue. In
addition to the faster of the waking state, perforuanwe situation, etc.,
pleasant and unpleasant stimuli can aly be Slightly more activating than
"When performasne has been observed to vary nader certain con-
ditions . it is suggested that the variatioes ay be due, at least
in part, to the effect of varying degrees of arousal. The diserganise-
timn of responses frequently reported during 'oremetivatioa' or 'metioa'
. may be conselved of as resulting in part from too high a degree of
arousal. Such a eoaditdn would be repr4Eented at one end of the U-
shaped curve. A similar dis*ganiateon ef responses, found erastima
during drowiness or fatigue, would be represented at the other end of
the curve skewing the relationship between aresal and perforsiaee."
One woued uppees that mild affect, pleasant or unpleasant in nature,
would increase performance abeat equally--eine both would increase
activation approximately the esAe amount. Hewer, it was fomud thb%
pleasant and unpleasant stimli have different effects on perforaeasn
when speed is the response measure. To account for the results, another
of Duffy's terms "direction" of behavior needs to be defined.
"'Direction' in behavior refers to the fact that the individual
does 'this' rather than 'that', or responds positively to certain cues
and negatively to others." Pleasant stimuli produce positive responses,
unpleasant stimuli, negative ones. Probably in an activation theory
such as Duffy's a much greater amount of attention, than is at present
placed, should be paid to these directional factors.
Differential Effects of Unpleasant and Neutral
Stimuli for Different Response Measures
When speed is the response measure, unpleasant stimuli have a
negative effect on performance when compared to neutral stimuli. When
accuracy is the response measure, unpleasant stimuli have a positive
effect on performance. This tends to go counter to the experimental
results reported by Lazarus (6) that in stress situations (assumed to
be of unpleasant affective nature) speed increases while accuracy
decreases. It does, however, fit in with his theorizing that depending
upon the variables involved in the situation, such as expectancy, stress
will either increase or decrease in particular performance measure. It
is suggested that, though speed and accuracy were both stressed in the
instructions to the subjects, accuracy was emphasized more by the
subjects. Accuracy, rather thcn speed, is what has previously been
stressed in similar situations. Of the two variables which might be
detrimentally affected by unpleasant situations, the less "important"
of the two is the ome hieh is hampered.
The purpose of this experiment was to observe the influence
of affect-stimuli on the performance of an immediately following task.
The stimulus material, presented on a projector screen to 75 subjects,
consisted of one affect-word (the affect-stimulus) followed by three-
consonant nonsense syllables (the task). Affect-words-25 pleasant,
25 neutral, and 25 unpleasant--were chosen on the basis of the judgments
made by five graduate students in psychology. Confirmation of the
affective nature of the stimulus was obtained through the judgments of
the 75 subjects. Each of the 25 tasks on one occasion followed an
unpleasant stimulus, n a second occasion followed a neutral stimulus
and on a third occasion a pleasant stimulus.
The instructions were for the subject to first read the affect-
stimulus word to himself, then to spell aloud as fast and accurately
as possible the nonsense syllables. The measure of the influence of
affect-stimuli on performance ws the speed and accuracy of the subject's
spelling of the nonsense syllables.
The results show that the influence of affect on the responses
to the task is a complex one. Both the nature of the response measure
and the nature of the task have to be taken into consideration.
Affeet-stimili considered alone did net influence the subject's
performance on the tasks. When, however, the affeet-stimulus and the
nature of the task fer the separate measures were takb jointly into
ceaaideration, significant result below the 1 level of confidence were
obtained. In comparison with the influeme of neutral still on the
task, unpleasant stimuli, when the response measure as speed, tended
te hinder performiae;j pleasant stimuli tended to facilitate performance.
When the response measure was accuracy, both pleasant and unpleasant
stimuli tended to facilitate performance.
Task difficulty as investigated to determine if this was the
task property that interacted differentially with the affect to produce
differential results. Trends in the predicted direction were obtained,
but they failed to reach significanee.
COMPARISON OF AVERAGE AFFECT RAWKIGS OF STnIULI BI JUDGES AND SUBJECTS.
SCAJ OF AFFECT RANWGJ FROM 1 FOR VERY PLEFLMT; THOUGH 3 FOR NEUTRALj
TO 5 FOR VERY IFNPLASANT.
Affeet Judge's Subject's Affeet Judge's Subject's
eWrd Bank Rank Word Rank Rank
1. Fun 1.80 1.24 39. Tree 3.00 2.92
2. Uestasy 1.60 1.10 40. Cable 3.00 2.89
3. Thrill 1.60 1.68 l1. Bavrlve 3.00 3.00
4. Favorite 1.80 1.88 42. Gibe 3.00 2.92
5. Luxury 1.50 1.71 43. Besh 3.00 2.99
6. Beautiful 1.4 1.39 14. Sinae 3.00 2.93
7. Popular 1.80 1.73 45. During 3.00 2.97
8. Delightful 1.80 1.45 16. Declare 3.00 2.88
9. Irresistible 1.80 2.23 47. Tapie 3.00 2.91
10. Pleasant 1.60 1.61 48. Circle 3.00 2.85
11. Party 1.80 1.81 49. Desen 3.00 2.92
12. Veluptuou 1.80 2.08 50. Saple 3.00 2.88
13. bealleat 1.80 1.55 51. Terent 4.20 4.31
14. Laughter 1.80 1.59 52. Quarrel 4.20 4.32
15. Nppy 1.20 1.51 53. Failure 4.60 4.51
16. Beliciows 1.60 1.63 54. Behead 4.20 4.47
17. HHwr 1.80 1.75 55. Slander 4.ao 4.1
18. oGreaul 1.80 1.77 56. Fear 4.IO 4.30
19. Best 1.60 1.93 57. Unhappy 4.80 4.39
20. Swee 1.40 1.45 58. Violate 4.20 3.78
21. Friendly 1.80 1.i9 59. Murder 14.0 4.53
22. Spledid 1.80 1.61 60. Shame 4.60 4.19
23. fuprb 1.60 1.55. 61. Angry 140 4.05
24. Smile 1.80 1.68 62. Kill 4.0 4.29
25. Jelly 1.60 1.69 63. Argue 14.0 3.97
26. Varies 3.00 2.57 6k. Sueide 4.o 4.51
27. sert 3.00 2.92 65. Mutilate 44.o 4.51
28. Tjp 3.00 2.77 66. Molest 4.30 4.23
29. Number 3.00 2.97 67. uly 4.20 4.13
30. Garpwt 3.00 2.83 68. Supid 4.20 4.19
31. Bmame 3.00 2.97 69. Hate 4.60 4.51
32. Term 3.00 3.14 70. Lanely 40 4.28
33. Wagen 3.00 2.88 71. Blood 4.20 3.52
3-. Assemble 3.00 2.73 72. Diagr m 4U.0 .439
35. While 3.00 3.00 73. Freak 4.O 3.93
36. Riddle 3.00 2.96 74. Deop-e 4.80 4,27
37. Similar 3.00 2.81 75. 'Maseere 4.20 4.47
38. Material 3.00 2.95
EXPERIMENTAL TASKS PRECEDED BY AFFECT-WORDS
POPULAR FAILURE ASSEMBLE SUCC
ZXJ S~ SGI 3G0
CXJ XZL XZL XZL
XJF GXJ GXJ GXJ
LAUGHTER SIAllDER WAGON FAVI
PJZ QHX QHX QHX
FQJ BJH BJH BJH
GZK GQH CQH CQH
IRRESISTIBLE UNHAPPY SDICE SPI2
JHQ CSF CSF CSF
ZKG KXZ KXZ KXZ
XJQ CGJ CGJ CGJ
PARTY MURDER TERM SUPER
JGP BJS BJS BJS
KXB QZM QZM QZM
FJH GIC 0ox GXC
SimLE SH BENCH BEAU
ZC QXH QXH QXH
DJQ XKB YXi XKB
MGQ KMH KMH KMH
FU, KILL VARIOUS ECS1
RJG GI GKI GKl
BFM HFC HFC HFC
QJ XZG x XZG
LUXURY MUTIIATE DOZEN BES1
ZMF SFM SFM SFM
JCF DJX DJX DJX
ZQJ ZBJ ZBJ ZBJ
HAPPY MOLEST SORT EXCE
RBM TJH TJH TJH
GQC BQJ BQJ BQJ
FBM CFP CFP CFP
GRACEFUL STUPID MATERIAL HUMO
JSB FHJ FHJ FHJ
TZH ZJF ZJF ZJF
NPB XOD XQD XOD
W -- or)
^ <1 <-
co l Li
RELAY OF \
z TAPE TRANSPORT
-- - 2v BATTERY
DIAGRAM OF WIRING
1. Combs, A. F. Intelligence from a perceptual point of view.
J. abnorm. sec. Psychol., 1952, h7, 662-673.
2. Duffy, Elizabeth. The psychological significance of the concept
of "arousal" or "activation." Psychol. Rev., 1957, 64, 265-275.
3. Hull, C. L., and Lugoff, L. S. Complex signs in diagnostic free
association. J. exp. Psychol., 1921, 4, 111-136.
4. Jenkin, N. Affective processes in perception. Psychol. Bull.,
1957, 5i, 100-127.
5. Jung, C. G. Studies in word-association. London: William
6. Lazarus, R. S., Deese, J., and Osler, Sonia F. The effects of
psychological stress upon performance. Psychol. Bull., 1952,
7. McGinnies, E. Generalization of perceptual defense. J. abnorm.
soc. Psychol., 1952, 47, 81-85.
8. Postman, L., and 5chneider, B. H. Personal values, visual recogni-
tion and recall. Psychol. Rev., 1951, 58, 271-284.
9. Rapaport, D. Dignostic psychological testing. Vol. 2. Chicago:
Year Book Publishers, 196
10. Sanford, R. N. The effects of abstinence from food upon imaginal
processes. J. Psychol., 1936, 2, 129-136.
11. Soper, D. A study of changes in the perceptual field in responses
to the perception of pleasant and unpleasant stimuli. Un-
published doctoral dissertation, Syracuse Univ., 1952.
12. Thorndike, E. L. The teachers word book. New York: Teachers
College, Columbia Univ., 1921.
13. Witmer, L. R. The association value of three-place consonant
syllables. J. genet. Psychol., 1935, 47, 337-359.
Reran Ivan Levin was born in PenCmeola Florida, on April 23,
1932. RB attended the public schools aC Penscola and graduated from
PM elaal High School in Jae, 1949. In September of that yer he
wter1a d eary Unifverity. Two years later h transferred to lasiana
8tate Universityr, a in Jume, 1953, kh rewlwed his Baehlar of Arts
degre froa this university. In Septalbr, 1953, he matriewlated at
Teachers Collge, Celumbia University and reeivd his Master of Arts
degree in Jue, 195I. In September, 1955, he entered the [uiversitt
at Flerida. There he has remained; emept that ring the last year he
has bean in the Veterans Aminia tratie clLnical psye.ology training
proem' at ulfport, Missisippi.
This dissertation was prepared under the direction of the chairman
of the candidate's supervisory committee and has been approved by all
members of the committee. It was submitted to the Dean of the College of
Arts and Sciences and to the Graduate Council and was approved as partial
fulfillment of the requirements for the degree of Doctor of Philosophy.
August 9, 1958
Dean, College of Arts and Sciences
Dean, Graduate School
L 'K /7^ f
,1K2 -$,.kb -1)..