AN INVESTIGATION OF THE
RELATIONSHIP BETWEEN CREATIVE
HUMOR AND AUTHORITARIANISM
ROBERT SCOTT CLELAND
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
To Dr. Justin E. Hatrlow, Jr., chairmana of the su-
pervisory committee, andk~ to the other members, Dr. E. D.
Hinckley, Dr. J. C. Dixon, Dr. R. J. Anderson, and Dr.
C. W. Fristoe, the author wishes to express his appreeia-
tion for unfailing support and eneourginagement
To the comm~ittee members who served as judges,
together with Dr. D. W. Soper, Dr. N., W. Coppinger, and
Dr. R. WJ. Bortner, the author extends sincere thanks for
performing an ard~uous t~ask
To Dean E. M. Miller and members of the Arts and
Sciences ffaculty of the University of IMiami at Coral
Gables, Florida, sincere appreciation for generous assist-
ance in obtaining subjects.
TABLE OF COI:TIENTSj
AClKrYNOrLEDGME~nTS . . . .
LIST OF TABELES .. . .. ..
LIST OF ILLUSTRlATIONIS .....
I. INTRODUCTION ... . .
I.REVIEW~ OF THE LITERATURE
III. PROBLEM . ,.. . .
IV. METHOD . .. .. .
V. RESULTS . .. . .
VI. DISCUSSION . . . .
VII. SUMMARFY . . .
BIBLIOGRAPiY . .. . .
A~PPENDIXSj . .. .. . .
. . . . .
.~ ... i
. ........ V
. . . . ... 65
:SPONSES .. ... 78
INSTRUCTIONS FOR TEST ADMI
INSTRUCTIONS FORi SUBE~CTS
INSTRUCTIONS FOR~ JUDGEtS
SPECIES OF HUM~OR. TEST RE
. . .
LIST OF TABLES
1. Seale Seore and ) Score Limits of the
Three F Scale Groups ......... 43
2. Total Hu]mor Points Received by Subjcts
in the Low, MTedial, and High F Scale
Groups ............... 50
3. Total Humor Points R~eeived by the Three
F Scale Groups Tabulated According to
Cartoons . ..... 52
4. F Seale Group Classifications of the Ten
"Most HumPorousn and the Ten nLeast
Humtorouas Subjects. .......... 55
LIST OF ILLUSTRAkTIONS
Distribution of F Scale Scores......
Distribution of Avera~ge Humoar 9artings.
Background of the Problem
Scholars representing m~any disciplines, from Plato
to the present, have written about humor. T~he word "humor"
is sometimes used in a limited sense, as in Freud's (23)
distinction between wit, humoI~Qr, and the comic. But except
where otherwise indicated, the more general meaning of the
word is intended in this paper. Humlor has had many defini-
tions, but most writers would agree that they are attempt-
ing to understand that quality or aspect of behavior and ex-
periene4 which mediates the amusing, the laughable, or the
funny. Mo~lst investigators have also noted that humor usual-
ly involves, either singly or in comnbination, the playful,
the surprising or unexpected, and the expression of emotion-
al or instinctive te~ndencies. Determinants of hukmor appre-
elation and the humor response have been recognized as being
related to an individual's intimate needs and feelings, yet,
at the same time, continuously influenced by the social en-
Since this quality called humor is so pervasive in
humarn behavior, it would ssee to offer a strategic base
from which to explore the, relationships among our emotional,
intellectual, and social selves. Yet relatively little re-
search on humor has been attemapted by psychologists. Philos-
ophers have provide-d insightful descriptions and speclla-
tions but scientific measuremennt has lagged. Except for the
efforts of a few pioneers such as M~artin (40), who devised
a rating scale for jokes in 1905, most of the~ experimental
work on humor has been done within the last twenty years.
Even the research that has been done has not been assimi-
lated into psychology as a whole. With a few exesptions,
textbooks on social psychology, clinical psychology, or even
personality thord~Y limit their disenasions on "sense of hu-
mor" to a sentence or two attesting to its import~ance.
There are several possible reasons for the limited
production of research on humor and for the failure of
textbook writers to discuss the suibject. One is that the
very quantity of allegedly humo~rous material to which we are
exposed daily may well stimulate an avoidance reaction in
persons intereestd in scientific inquiry. It is also true
t~hat hum3Vdor is a rather nebulous concept, the relationship
of which to personality and to behavior is both subtle and
com~plex. M~easurin65 devices are difficult to construct and
standards of excellence are hard to define. As in studies
of music, the prolonged exposure necessary for analysis of-
ten destroys the unique but elusive qualities one seeks to
Several advances in humor research have been made
in recent years. Factor analysis has been utilized in an
attem~spt to bring order and psychological meaning to the
commzon sense classifications of humor. Recent investiga-
tors have tended to avoid narrow theories and have shown a
greater awareness of the breadth of the problem and an ap-
preciation of the fact that not one, but many solutions
will eventually be needed.
Theoretical Orientation of the ,Present Study~
One of the limitations in some of the previous re-
search has been that the study of humoer has been subordi-
nated to some other goal. EFxamples are a study of the ef-
feetiveness of humork) in persuasive speech (37), and a humor
test which contribu~tes to a score in Wsocial intelligence"
(41). One nhumor" test (9) me1rely usesi~ jokse preferences as
a device for securing measures of other aspects of person-
alityr, such as se~x interest or hostility. Although "applied
humorM studies are of interest, they do little to cast light
on the nature of humor itself.
Probably the most striking deficiency in studies of
humo~sar has been their almost exclusive concern with appreci-
ation, rather than the cr-eation, of humor. Mos~t experiment-
ers have had skubjects; indicate their preference~ for jlokesl,
limeioricks, and the like, by rankring or rating. Another
technique requires subjects to choose the funniest" of
five endings to a joke (41). Eysenok (20, p. 225) is one
of the few writers who has recognized that there sar clea~nr-
ly two factors involved in "sense of humor," namely, "ap-
preciation and production." No states that his studies
with the production of huamnor have not yet yielded definite
enaough results to warrant publication.
ALnother weakness of previous investigations of hu-
mor seems to be an overemp~hasis of humor as an inddex of
possible psychopatholo~gy. This is especially true in psy-
choanalytically oriented studies, although Freud (23, p.
802~) himself recognized "harmlessM as well as "tent~denc~y"
wit and described humor as "the loftiest of the defen~s
functions." Vieving humorl as a useful defense function,
rath-er~ than as a symptom of conflict, seemsa a step in the
right direction. But humor's positive, creative, function
as an agent in superior adaptation" has received little
Some general problem~asl concerning taehniques and em-
phases in humor research have been discussed. Thea present
study does not attempt to suggest what should be the goals
in future research. However, it is thought best to point
out that th~e aims and the plan of the present stu~dy devel-
oped, to a large extent, in reaction to some of the previ-
ous emphases. First, interest is centered on the creation
of the humorous responses themselves, rather than on what
hidden meanings they mar~y contain or how others will react
to them. This orientation leads, then, to the question of
individual differences--nrot in appreciation of humor--but
in the production of h~umor. Why are some people success-
ful usjeras of humPor and others not? The present study als~o
attempts to avoid the bias of vie-wing humior primarily as
an expression of psychological conflict. Freud (23, p.
760) stated that the importance of unconscious tendencies
might explain nwhy the subjective conditions of wit are so
frequently fulfilled in the case of neurotic persons." It
is hoped to present evidence in this study suggesting that
these conditions may just as frequently, or more frequent-
ly, be fulfilled in the case of the psychologically healthy
Overall Pla.n for the Exrperiment
The problems pertaining to humor research discussed
thus far are broad and comsplex. They will be documented
and discussed in more detail in the review of the litera-
ture. The reactions on methods will present the complete
experimental procedure. However, the general approach will
be briefly described at this point.
It was desired to relate the ability to be humaorous
to individual differences in personality. First, it was de-
eided to limit the subjects to male, college undergraduates
since aest of the related research had utilized similar
groups. For the me8i~asure of humor "production,n a test con-
sisting of eight incomplete earteons was constptruted. (The
humo~ir test is reproduced in Appendix I.) The cartoo~ns de-
picted familiar campus situations in which one student la
saying something to another. The subjct is asked to fill
in a humsorous reply. The personality measure selected was
the California F (fascist or anti-democratic) acale, or
'"authoritarianismn' scale (1, p. 255). This scale was cho-
son because of the many studies suegesting that the F scPal
measured several personality variables suspected of being
cruacial to the production of humor.r
A total of 188 male subjects, college undergraduates,
were given the humor test to complete, followed by the F
scale. Sixty subjects were then selected on the basis of
their F scle scores:= the 20 scoring highest (mos~t a~uthor-
itarian), the 20 scoring lowest, rand the 20 H"mid-most" in
the distribution. The eartoosn responses of these 60 sub-
jects were then rated for funniness by a group of seven
Judges. The results of the humorPr test, as related to the
F acale classifications, constitute the principal findings.
REVI~EW OF THE LITER1TUR.E
The studies on humor to beg reviewed will be c~las-
sified under four general headings: the philosophers and
humbor, phylogenetic theories, psychoanalytic theories, and
experimental studies. The major emphasis will be placed
on the experimental studies. This group is felt to have
made the most significant contributions, not only because
of their methods, but because their studies have been the
mos~t recent and they have been in a position to bring to-
gether and test some of the earlier theories.
The Philosophers and Humor
Plato, in the Philebus (4~6, p. 339) noted that,
". .when we laugh at the ridiculous qualities of our
friends, we mix pleasure with pain, since we mix it with
envy. ."The emphasbis is on laughter as a subtle way
of a~ttacking those whomni we feel mray be suaiperior to us.
Aristotle (5, p. 1449a) however, believed that, "The Ridic-
alous may be defined as a mistake or deformity not produc-
tire of pain or hiarm to others.H This definition is prob-
ably the first exposition of the idea that humr in its
purest form is not hostile. Spinoza (55, p. 217) also ree-
ognized this difference. He stated, nI make a great dis-
tinction between moacckery .. and laughter; for laughter
and merriment are nothing but joy ,. ." Kant (33, p.
223) felt that humor could be understood in terms of phys-
iological functions. Hes stated, "In the case of jokes (the
art of which, just like music, should rather be reckoned as
pleasant than beautiful) the play begins with the thoughts
which together occupy the body, .. and as the understand-
ing stops suddenly short at this presentment, in which it
does n~ot find what is expected, we feel the effect of this
slackening in the b~ody by the oscillation of the organs,
which promotes the restorattion of quililbrliumIB and hats a fa~-
vourable influence upon health."
A more~r sophisticated theory of humor was set forth
by Scihopenhauer (52, p. 279). He suggested that laughter
results from the apprehension of an incongru~ity in which
perception triumphs over thought. n(perception~ .. is
the medium of the present, of enjoyment and Salety, more-
over it is attended with no exertion. But Cthoug~ht]. .
is the vehicle of our fears, our repentatnce, and all our
care. It must theroefre be diverting to see this strict,
untiring troublesome~ governess, the reason, for once con-
ricted of insufficiency."
Kline, a psychologist who contributed to the
philosophyr of humosr, revealed a still broader understanding
of humor in noting that humor is a social process and that
it is also characterized by the elemeaant of freedom, He
states (34, p. 438), "Perhaps its [humeor's] largest fune-
tion is to detach us from our world of good and evil, of
loss an~d gain, and enable us to see it in proper perspec-
tive. It frees us from Yanity, on the one hand, and from
pesimismQ on the other by keeping us larger than what we
do, and greater than what can happen to us,"
Ph.Ylogenetic Theories of Humor
R~app (47, p. 21) states that wit and humor can be
traced to the primitive scerne where the victor uttered a
"roar of triumph" over the vanquished. He assumes that
laughter at persons preceded laughter from riddles or jokes,
and that laughter of ridicule preceded humane, genial laugh-
ter. He suggests that social training has caused laughter
or humor to be substituted for overt aggression. McCom~asc
(38) has also su86ested that laughter may have developed
before language, as a reaction to a find of good food or a
fresh spring. These authors have "explainedM laughter by
considering it an instinctive reaction. Fr~eud (23, p. 733)
allrso utilizes this approach when he suggests that laughter
maay develop from the contortions of the corners of the
mouth of the "satiated nu~raling5 when he drowsily quite the
EastmaP~n (16) has recognized tha&t theoies wQhich re-
duce laughter to an instinct have difficulty in accounting
for the wide variety of situations that can lead to humon~r-
ous responses., Eastmann explains adult hunimor as developing,
through learmning from childhood play. He refers to McDou-
gall (39) for support in his opinion that play is a form
of instinctive behavior. EaistmaEn also recognized Freud~'s
Wtendncy" wit, or nfurtivew wit, as ESartstman calls it. It
should also be noted that Frevud (23, p. 634) accepted the
definition of wit as a "playful judgment."
Psychoanalytic Theories of H~umor
Frekud's observations on wit, the com~ic, and humor
were insightful and comprehensive. He is probably best
lcitknon for his emphasis on ntendentiou" wri~t, or wit which
serves to release repressed unconscious im~pulsear, usually
sexual or hostile. He pointed out that somre jokes are sla-
11ar to dreams or lipss of the tong~ue" in this respect.
Freud also recognized nnon-tendentious" or harmless" wit
but had little to say about it.
Freud's attempt to explain the unique qualities of
humor resulted in the formulation of his principles of
"eonomy." He stated that the pleasure of wit ori61nates
from an economy of expenditure in inhibition, the pleasnsure
of comic from an economy of expenditure in thought, and the
pleasure of humor from an economy of expenditure in feeling.
(23, p. 802). These~ three divisions are suggestive, re-
spectively, of the sonative, cognitive, and affective divi-
siens later applied to humtor by Eysenekr (20, p. 227). How-
ever, Freud's terminology of Cceconomy of expendituree' has
not bee~n widely accepted, subsequent writers preferring to
attributed the pleasure to the expression of repressed ten-
dencies and the release of the energy used in repression.
Freud's frequently quoted distinctions between wit,
the codie, and humor do not seem entirely satisfactory.
For instance, he gives the following example as a combinap-
tion of the comic and wit (23, p. 763). A little girl who
was suffering from a severe cold and sneezing profusely,
pointed to her chest and said, "Dadday, Gesundheit hurtal".
Since our sympayLthy for the child is both mobilizead and re-
leased, this rma~rk could also be interpreted as humor,
according to Freud's doctrineh of neconomized feeling."
In 1928, Freud (24) published his theory of humo~r
making use of the superego, a concept which had not been
formullated at the timer of his book on wit. This concept
states that the ego adopts the point of view of the super-
ego and la able to look down on its difficulties with more~k
mature unrderstanding and detachmer~nt.
Freud's contribu~tionsa to an understanding of hum~or
have markedly influenced subsequent studies. However,
Flugel (22, p. 721) has noted that, "On the whole,...
the contributions of the 'metapsychologgy' of humor by pary-
choanalytie writers other than Freud himself have~ not been
as convincing, enlightening, or muttually consite~nt~ as might
have beebdn hoped for in view of -the considerable number of
papers detvote~d to the subject." An example of this neo-
Freudian approach to humo~r might be taken from Dooley (14),
who states, "Hu~mor thus becomes one of the ways in which
the ego .. 'wangles' restoration of the parents' love
by claiming t~he love of the parents' successor--the super-
er~fgo--in playful fantasy. (p. 44).
Expetrimental Studies of Humor
Tfhe experimental investigations will be discussed
in approximate chronological order. A atudy by Lantdia and
Ross (35) illustrates one of the earlier eIxperimental ap-
proaches to humor research in which a large number of jokes
were classified into categories by judges. In this case a
100 item test was constructed, jokes being classified into
the following seven categories (p. 157):
1. Humsor of Quaantity (exaggerations, under-statement,
2. Humzor of Incongruity (assoolation of incompat-
3. Humor of Unexpected (occurrence of some sur-
prising thought, etc.).
4. Humor of Truath (exposure of one's unrevealed
5. Humor of Superiority (difficulties of others
which seemn simple to us).
6. Humor of Repressions (release of tensions,
such as fear, sex, etc.).
7. Humor of Ridiculous (nonsensical use of logie,
play or words).
Jokes in these categories were further rated by writers
and by editors of a college humor publication as "Yery
good, n "good," "poor," and "very poor." Eiqual numbers from
each of these levels were placed in the test.
Subjects, in the portion of the experiment to be
discussed, were 124 male undergraduates at a me~n's univer-
sity and 154 female undergraduates at a women's college.
Subjects were also given an introve~rsion-extrov~ersion scale
and intelligence was mea~sured by means of a standard col-
lege ability test used for entering freshmen. Subjects
were asked to rate each joke on the humor test according to
the four-point scale of Hgoodness,w with no limit as to the
total r~Pnuber of humorr points" to be awarded. The subjects
were also asked to classify each joke according to the seven
humrr~ categories. The~ tests were scored for total "hlumor
points' and for total points assigned to eanch of the humor
categories. Results indicated n~o relationship between ei-
ther overall humor points or classification of itmsr with
the other ma~sures employed, that is, intelligence, or
introversrion-extrvroerion. The authors noted marked individ-
ual variations in humor preferene~ and concluded that per-
formance on the humorF test behaves as an independent per-
sonality factor. It was also noted that the men and women
differed significantly in humor preferences; evaluating the
categories from best to poorest in the following order (35,
Men: Repression, Unexpected, Ridiculous, Truth,
Incongruity, Quntity, Superiority.
Women: Rildiculousa, Incongruaity, Unexpected, Supe-
riority, Truth, Quantity, Repression.
The greatest difference is seen in the Repression classi-
fication which the men1 valued first and the woe~pn last.
This is the claselfication which includes mos~t of the jokes
related to sex.
In oriticism of the type of appro~ah used in thisa
study, Cattell (10) has stated, HValid humtJor scales cannot
be built up by items merely believed frome 'inspection' to
have a 'homooeneous' eentesnt, or even on psychiatric in-
sights of common dynamic content. ("Two psychiatrists agree
on the dynamic content and item ass~ociation of a joke scatrce-
lyr beyond chance.) The jokess that are put on a single fac-
tor sc.ale have to be proved by experime~nt to 'go together'
in the choices of the typical subject and to be at the saeEBC
time substantially independent of th other factor scales"
Anl example of the inconsistenciesa that c~an result
from classification by inapection occurs in the ~Landis and
Ross (35) study. The authors noted that some of the jokes
that they used were uased in a similar study by Kambouropou-
lou (32). They compared results and found that the relative
hum~or values assigned corresponded closely. However, the
classifications as to the type of humor involved differed
A study by Sears (54) has been widely quoted, prin-
oipally with reference to his useful distinction between
the schemastic and thematic aspects of humsor. TIhe former
refers to the structure of the joke, including such organi-
sational factors as timing, brevity, or the sudden shift of
elosure. Thed latter refers to the meaningful humatn content,
primarily asocial or repressed needs such as aggression,
superiority, or sex. Sears also classified his jokes on an
a priori basis. His choice of such classifications as
"overt anal"1 antd "covrert analn suggests the susceptibility
of this approach to theoretical bias,
Go~irdon A11port's (17) contribution to the litera-
ture on hum~or is more speculative than expperimental. How-
ever, chronologically, he should be considered at this
point, and he does begin his remarks on humor by referring
to a study in which ratings of persons for Hsense of humorM
were found to be correlated +.88 with their ratings for
"insightn (p. 220). Details of the experiment were not giv-
Allport states that "psychologically it is not prof-
itable to distin~guish between insight and sense of humsor"
(P. 422). He includes both traits under hself-objec~tifica-
tion." He rules out of his definition ab~aurdities, puns,
release of suppressed material, and the laughter of goo~d
spirits or play. What remr~ains is a highly intellectualized
distilled from the broader concept of humor. One might even
suspect that A11port has missed the essential ingredients of
humor when he6 states that, HTo view~ one's problems hum~or-
ously is to see th~!em as trivial and of no clonsebquatenc . .
(p. 225). This would appear to be self-deception rskather
than insight, and it is hardly an adequate evaluation of hu-
mor. Wat Allport, and other writers, find difficult to
explain is the special quality of humor which enables a per-
son to laugh when fully aware that a problem is far from
trivial. eAlthough A11port's .theory appears to be too narr-
row, it is an important contribution because of his point
of View.( His emphasis on the positive aspects of humeor
ctfa met a time when Hdischarge" and "camouflage" were being
A nelw approach to hum~bor research is represented by
the~B work of H. J. Eysenck, whose major contributions to
humor are contained in two articles (18, 19) and in his
book, Dimensions of Personality (20). Eyisenck is recog-
nized both for his "cogjnitive, conative, and affective"
theory of humor and for his factor analytic approach to the
factors in humo~r. In one~ study (18), 189 jokes were given
to 16 subjects to be ranked. Subjects were also given an
introversion-extroversion questionnaire. The humor rankings
were intereorrelated and subjected to factor analysis. The
main outcome appeared to be a bipolar factor contrasting
persons who preferred rather simaple, "funny," sex and ag-
gressive jokes with persons who preferred mBore complex,
clever" jokes not dealing with sexual matter. Eysenck
identifies the first group as oreetic (conastive and affee-
tive) jokes and the second group as cognitive jokes. Cor-
relations showed that extraverts tended to prefer the oree-
tic type of humor while introverts preferred the cognitive
Eiysenck believes that it is possible to classify
all humor un~der the three ma~in headings of cognitive (in-
congruity, contrast between ideas, and deceived expecta-
tions), conative (satisfaction of the desire for superior-
ity), and affective (emotional aspects). He calls this his
"~electic Theory of Humor," and illustrates it by means of
a triangular' diagram (20, Fig. 21, p. 228). The top corner
is labeled "Cognition" and the bottom corners "Affecction"
and nConation." The two bottom corners are drawn closer
together to represent the apparent closer interaction of
these aspects of the mind. ~Any joke can be located within
the triangle. As the joke approaches "&Affection" it can be
described as humoror" as it approaches HCognition" it is
adesribed as "comic,M and as it approaches HConation" it
takes on more of the characteristics of "wit." sear's (54)
Mscematt~ie" and nthematie" divisions are similar to Ey-
sanck's "cognitive" and "orectic," while Allport's (2) the-
ory is predominately cognitive.
In a second study by Eysenck (19), 100 anormal su~b-
jects ranked8 250 humorous items grouped into five different
tests. lAverager interoorrelations between subec~ts were
very low, as were the interoorrrlations between ecp~h sub-
ject's perforrma8~nce on all five tests. In other words, there
was a surprisingly small amount of agreement as to the rel-
ative goodness of the items. It was f~ound~, however that
individuals tended to be consistent in the~ amount of "fun"
that they got out of the materials. A person who found
many jokes funny te~nde~d to find ma6~ny cartoo~ns and verses
funny. Similarly, a person who did not find the jokes fun-
ny was equally rejectinS of theg other hum~aorous ma~zterials.
Eysenckr otes that this affective reaction seems to conbsti-
tute a genuine personality trait.
Anadrewsr (4) has reported an interesting study which
was designed to determine whether or not humoar should be
considered as a single unit quality of a genmeral nature.
Several hulnzdred jokesa, puns, limerickts, and cartoons were
sorted by college students into nine piles, in order of fun-
niness. Twenty-four representative items, covering the arn-
tire range of humr values, were intercorrelated and~ factor
analysis carried out until six factors had been extracted.
No~ general or unjiversal factor was found. The six factors
were desribed as follows:
Factor (1) Feelings of superiority over persons
seen~ as inferior.
Factor (2) Escape from, or sym~Bpathy~ for, debauch-
Factor (3) Subtlety (puazzlers, or hidden meaninS)*
Factor (4) The pun, or play on words.
Factor (5) sexual jokes.
Factor (6) Ridieulous, nonsensical, wise-eracks.
It is interesting to note that these factors can
also be grouped according to the classifications of Sears
(54) and Eysnck (20). Factor 3, 4, and 6 could be de-
scribed as scheaitie or cognitive, and Facters 1, 2, and 5
as themaBtic or orectic. However, Abndrew's factors, like
those of other investigators, esem far from satisfactory.
There is a great deal of overlapping. For example, a play
on words might also be ridiculous and, in addition, carry
sexual connotations. Tracing out such connections still
leaves us, in Freud's words, with Hdisjointed fragments
which we should like to see welded into an organic whole
(23, P. 637).
R. B. Cattell is another investigator who has done
extens~ive resara~ch on hut~ser. In collaboration with Lubopr-
sky (11), 100 jokes were rated by 100 young men and women.
Factor analysis yielded five general personality factors:
(1) good natured assurance, (2) rebellious dominance-, (3)
sex repression, (4) passive derision, and (5) sophistication.
The last four factors seem to be closely related to factor
obtainged by previous invresti~atozrs. The first factor,
good natured assurance," is of interest because it sug~-
gests the waraI yet stimwulating, affective qualities we of-
tebn~ ~~assoiater with a person who has a ''good sense of humoi~r."
Th~is aspect of humortr seems to have beent overlooked in other
eCttell has published a humo~r test (9) which is
based on an extensive series of factor analytic studies.
This test consists of 76 items (152 jokes, in pairs) for
which the srubject indicates his preference. Since each
pair of jokes has been mratehed for relative nfunniness" on
the basis of previous ratings, extraction for individual's
"total funniness" sc~ore is impossible. Th~ie subjects' choi-
ees contribute to scores on the following 12 bipolar fac-
tors: (pp. 11-13).
Factor 1. Debonair sexual and general uninhibited-
ness vs. anxious considerateness.
Factor 2. Good-naturead play vs. dry wit.
Factor 3. Tough self-com~pos~ure vs. reassurance in
Factor 4. Gruesomeness vs. Flirtatious playfulness.
Factor 5. Hostile derogation vs. urbane plesa~nt-
Factor 6. Resignation vs. impudent defiance of de-
Factor 7. Gold realismr vs. thea~tricalisml.
Factor 8. Ponderous humor vs. neat, li~jhthcac~rted
Factor 9. Whimsical retort vs. damaging retort.
Factor 10. Mistreatment humor ve. cheerful indepen-
Factor 11. Ev~asiont of responsibility and guilt vs.
Factor 12.Scorns of ineffectual male vs. rebound
against feminine aggression.
Tlhe preceding array of facters vividly illustrates the, com-
plexity and subtlety with which the "sense# or "quality"
called humo~r is woven into our personalities. However,
reading, this list hardly gives one a feeling of "closure,"
and the question la raised as to whether or not this ap-
proach to understanding~ hum~ior has reached a point of dimin-
ishing returns. Apparently a reaction has already set in
as, Ya~rnold and Berkelley (58) have rleanal3yZed some of Cat-
tell's data and have suggested reducing the number of fac-
tors to seven.
The four studies which will be reviewed next are
relatively recent. They seem to reflect a change in orien-
tation, from attempting: to understand and describe humor as
a whole, to the testing of hypotheses deriveda from limited
portions of humcor theory. Redlich, Levine, and Sohler (48~),
in a study of reactions to humorous materials of 59 psychi-
atrie patients and 24 normal subjects, made two interesting
contributions to technique. First, as stim~ulus miaterals,
they used 36 cairtoons depicting different objects of aggres-
sion and distortion. The authors found that cartoons were
orerb quickly and easily responded to than were verbal a~teB-
rials. Second, in addition to having subjects sort the ear-
toons according to their likes and dislikes, the authors
rated the subjects' overt reactions to the cartoons accord-
in8 to a Wmirth response scale." These ratings were taken
unobtrusively and noted the degree of apparent rairth from
Hno response" to Mlaugh~ter." The r~esults of this frankly
exploratory study were not conclusive with respect to dif-
ferences between groups. However, the depressed patients
raected2 as expected, with little or no overt response.
Schizophrenics tended to miss the point and exhibit Wgpara-
humosr. The me~thod was described as being a promising clin-
ioal tool because reactions to the cartoons were observed
to be closely related to the dynamics of individual patients.
A later study using the "mirth respon~s scale" was
carried out by Berkowits (6). Forty-four collqeo students
were given a group Rorschzach test antd their protocols
scored for hostility. Half of the su~bjects (praise group)
were then given prepared reports reporting favorable per-
formsane, while the other half were given~akr unfavorable re-~
pobrts (streBss group). Each subject was then~ asked to sort
twelve cartoons into three equail piles, like most," "Like
least, and "neutral," and to explain the ''point" of each
eartoon.~L Overt reactions to the cartoons were noted on the
birth response scale." Since jUdges had previously rated
the cartoons for aggresalve content, the data could then be
analyzed in several ways. Tfhe independent variables were
the measrtuess of hostility and the stress-praise var~iable.
The dependent variables were the mirth response score, the
cartoon preference (aggressive versus non-aggressive), and
the degree of understanding or distortion of the pointsn
of the eartoons--also rated by judges.
The principal findings were as follows:
1. Su~bjects in the "f:raise" group scored signifi-
cantly higher in 'total mirth response" than did
subjcts in the "stressr" group.
2. Average "mirth response" scores were not signifi-
cantly different for Hhigh hostility" and "low
hostility" groups. However, "high hostility" sub-
jects gave significantly more extreme responses
(frowns or laughter) while "low hostility" sub-
ects gave more neutralt" or "smile" responses.
3. igh hostility" subjcts nade s1 nificantly more
"distortionsM than "low hostility subjects on
the "high aggression" cartoons but not on the nlow
These results suggest that Htendency" theories of
humsor cannot assumeI a simple relationship between the de-
gree of a Htendency," such as hostility, and the degree to
which the humor response may be affeeted. Finding number
one appears to give experimental support to the frequently
observed depressinS~ effect of anxiety on the ability to
respond humorously. However, the findings with respect
to hostility suggest that theP relationship of anxiety to
humor may also be Im~ore complex than is apparent under the
conditions of the experiment.
Grziwok and Scodel (26), asked 140 male college
students to rate a series of 40r eartoone for "funniness.n
The cartoons were selected from a pool of 250 New Yorker
cartoons which had been placed by judges into the following
1. Humorous effect based on aggression, either
explicit or deliberately understated.
2. Humorous effect obtained by a parody on sex.
3. Humorous effect based on the exaggerated or
paradoxical use of social stereotypes.
4. Humrlorous effect based on obvious or striking
Ten cartoons were selected from each category. The
authors state that the first two categories (aggression and
sex) can be su)Pbsumed under "orectic humor" and the second
tvo categories (social gcmme~antaryJ humor and loIical incon-
grutity) under ncognitive hum~sor." (Ciategory nm~ber three
appears to be misplaced in view of the potentially highly
orectie nature of ca~rtoons which exaegerate social stereo~s-
The subjects were also given the Allport-Vernon-
Lindzey Study of Values~ (3) and were asked to write stories
for seven Thematic Apperception Test (TAT) cards (42). The
TAT stories were sEored according to sp~eelall constructed
scales for degrees of aggressive and sexual content, and,
in addition, intraception versus extraception.
The significant results were:
1. S's high in TABT aggression prefer aggressive
humor while thosee low in TAT a66resalon prefer
social commentary humor.P
2. With respect to value orientations:
a) Aesthetic scale. "Highs" prefer logically
incongruous cartoons while "lows" prefer ag-
b) Soelal scale. "HighsH prefer aggressive car-
c) Theo~retical scale. "Lowa" prefer sexual car-
The authors ~usummarize their results as follows: "In ma~ore
general terms~ a preference for orectic humBor, as opposed to
cognitive humor, seems to be characterized by more fantasy
aggression, more extraersrion or outgoingness, less prooc-
cupation with intellectual values, and less psychological
subtlety or comtplexity"~ (26, p. 42).
The above findings are in agreement with Eysenok's
(18) finding that introverts prefer cognitive humor while
extraverts prefer orectic humor. The results of the Garzivok
and Scodel study m~ust be interpreted with caution, however,
because only a few of the masny com~parisons possible are re-
ported as being significant, giving rise to possible sampling
The last humo~r reseaErch study to be reviewed is by
Epstein and ~Smitha (17). Their experiment attempted to test
the Freudian view that repression favors an appreciation of
humor in which the repressed material is represented. The
authors noted that previous studies have demonstrated a re-
Ilationship betw~eenr manifest hostility, for example, cand
preference for hostile humor, buet have not directly attacked
the Freudian hypothesis that repression of a tendency favors
its appreciation in humorous form.
The authors measured preference for hostile humor
in 32 members of a college fraternity by three techniques:
observing 8's expressive reactions to hnostle6 and control
cartoons, having 8 rate the cartoons, and having him eart
them on a Q sort of funniness. In order to obtain a meas-
ure of repression, self-ratings of hostility were compared
with average ratings assigned by fellow fraternity members.
~The 8's were thereby divided into insight and repressiojn
groups, according to the relative accuracy or inaccuracy
The principal findings were as follows:
1. No relationship was found between repres~~sin
and general preference for heatile cartoons.
2. A si~nificant positive relationship between
insight and sense of humor was found, sense of
humor being melklasur byr degree of correspondl-
eance~ of 8s ratings with group normsr.
The negative finding with resrpectet to the hyrpother-
sized relationship between repressed hostility and humor
is difficult to interpret. The author suggest that there
may be no general relationship. Their findings indicated
that the rteltionship maay hold true for cartoons which are
the most obviously hostile. The rauthror also suggest that
a relationship might be found if suppression (conselous in-
hibition) or external frustration of an impgulse wercr in-
restigated rather tha~n repression (unconscious inhi~bition).
This latter possibility is suggested by thet frequently noted
interest in sexua~tl jokes of sexually frustrated groups such
as adolescents (eulturaslly frustrated) and men in the armed
srvuiesr (geographically frustrated) who are not necessar-
With respect to the findings of a positive relation-
ship between insight and sense of humror, the authors' meas-
ure of the latter should be noted. The accuracy with which
the eartoons were evaluated, with respect to group norms,
may be largely a cognitive function. In this case, these
results would tend to support A~llport's (2) equating of in-
sight andb humor, at least as far as the cognitive aspects
of humor are concerned.
The California F Scale
Since the F scale from The Authoritarian Personal$-
Ag! by Adorno, et arl. (1) is the personality measure usebd
in this study, it is felt that a brief description of its
development, together with a summary of subsequent research
with the scale, is desirable.
Research in humtror appreciation has tended to be
concerned with specific personaLlity traits, such as aggree-
sivenesso or insight. However, since creation or production
of humor was to be measured in the present study, the ques-
tion was first formul~Pated as, nWha~t might be the persaonail-
ityf characteristics tending to mai~ke a person a 'humoa~rist,'
or a 'non-humorist'?" It was felt that mranay personality
traits might contribute: flexibility, insight, empathy,
intraceptiveness, oad the like. However, it is difficult
to find satisfactory measures of these~ traits. In addi-
tion, it was felt that a se91nse of humBiICor was more likely to
depend on a constellation of these traits than on any sin-
gle trait. Therefore, a global measure was sought which
would, in effect, combined measures of theics above traits.
The~ Ca-liornmia F scale wasl selected for several
reasons. The development of the test can be traced from
the initial attempts to get at personality characteristics
underlying racial prejudice and anti-democratic attitudes
through the severa~cl form changes needed to improve sceni-
tivity and bring the average reliability to .90. Abundant
clinical data from the original research, together with
many subsequent studies, enable one to judge with some con-
fidence as to what is included in the omP~nibas~ term "author-
Th~e authors of" the scale have su66ested that the F
scale measures the followxing nine traits (1, p. 228).
2. Authoritarian submission.
3. Authoritarian aggression.
5. Superstition and stereotypy.
6. Power and "toughness."
7. Destructiveness and cynicism.
9. Exaggerated concern with sex.
The authors point out (p. 262) that these groupings
are a priori aids to discussion only, not clusters in the
statistical se~nse. However, all individual items in the
test correlate significantly (average r = .33) with the
test as a whole. One factor analytic study of the F scale
by O'Neil and Levinson (44) reve~arled four factors, which
they named: religious conveentionaelism, authoritarian sub-
mission, mascouline strength facade, and moralistic control.
Possibly a more adequate conception of what the test meas-
ures can be obtained from the findings of subsequent re-
searchers who havde orrelated the F scale with other crite-
ria, Some of the personality characteristics which have
been found to be, related to Hauthoritarianism" are as fol-
1. Intolerance of ambiguity: Jones (31) found that
inability to tolerate a high rate of reversal in the Necker
Gube was positively related to F; scale authoritarianism.
Similar results, with various measres of intolerance of
ambiguity, were obtained by Block and Block (7), and O'Con-
2, Misanthropy: Campbell and M/cCandaless (8) found
the F acale to correlate highly with a seale meag~suring gen~-
eral dislike of oth~ers. Su~llivan and Adelson (56) obtained
similar results with a scale of ethnic prejudice rewritten
Iso~ thatterms such as Uall people" were substituted for mi~-
norities originally specified.
3. A~nti-intraception: Dorris, Lev8inson, and Hanf-
mann (15) found~ll that authoritarians tended to deny self-
reference to items after having taken a sentence completion
test. Jones (30) and Scodel and IAslsen (53) have found
authoritarians less capablea of judging the psychological
and personality characteristics of other.
4. Doymmatism: Ro~keach (49) constructed a "Dogmatic
Personality Scale" composed of statements concerning over-
identification with a caused, and the like, th~at correlated
.67? with the F scale.
5. Intelligncce: H~ollander (29) studying Naval Avri-
ation Cadets, found that intelligence as measured by the
A. C. E. test correlated -.21 with the F scale. Adorno,
et al. (1, p. 283) report an r of -.32 fo~r the F actale with
the Otie Higher Form A Intelligenice Test, in a group of
veterans applying to the U. 8. Employment Service.
6. Concern with aggression: Sanford and Rosenstock
(51), using responses to cartoons, found authoritarians
gave more extrapunitive and intrapunitive repliesB, hence
were overconcerned with aggression.
It is evidesnt that the personality dimeansion tapped
by the F scale la broad and coma~plex. Gr~egory (25, p. 642)
has obser-ved that the authoritariania personality" aight
be6tterh be thought of as, "personali~ty syndrome x which has
sometimes been called authoritarian." However, inspection
of the six traits discussed above does suggest that they
tend to fall into two groups, as follows:
M~isanthrlopy Intolerance of ambiguaity
Abnti-intrac ept ion Do6matism
Concern with Intelligence (negative
It will be ob-served that the first group could be classi-
fled as oreetic traits donationn, affection) while the
second group are cognitive traits. In view of these find-
ings it might be meaningful to think of F scale nauthori-
tarianism" as being characterized, effectively, by hostil-
ity and intellectually, by rigidity.
Additional evidence that the F scale measures a
meaningful personality dimension is contained in studies
that show a positive relationship between measured author-
itarianism and overt behavior. Sanford and Rosenstock (51)
found that authoritarians tended to reject cartoons which
were being used as projective devices. Block and Blook
(7) reported that authoritarians in an exrperiment on auto-
kinetic movement submitted more rea~dily to arbitra~pry de-
m8~ands by the extperiments.
Deskins (13) has recently com-pleted a factor ana-
lytic study of several sealed related to authoritarianism.
The F scale appeared with s~ignificant loadings on two fac-
tors. One factor included high positive loa~dingss on the
F scale an~d reli61ous conventionalism with a negative
loading on a vocabulary scale. The other factor had posi-
tire leadings on masculinity (as opposed to femininity),
ethnoscentrism, political-economic conservatism, and the F
scale. Theser findings tend to support previous studies
which su~gget that nauthorita~rianism" is mu~ltifactorial
and cannot be considered a simple, personality "type.w
Theoretical Backpround for the Problem
Before formur~lating the problem it seems advisable
to summa~8rize the various theoretical approaches and exper-
im~erntal contributions to hum~or. Freud (23) has probably
had the greatest influence on subsequent research, partly
because of the breadth of his observations, but also be-
eause his "tendency" theory of wit has been e-nthusiasti-
eally accepted by a generation of psychologists interested
in revealing the hidden. The~ work of Sears (54) and oth-
ers has demonstrated that humo~r often is used as a social-
ly acceptable device for releasing feelings of superiority,
hostility or sexual desire. However, later work by Berko-
witz (6) and EZpstein and Smnith (17) has demonstrated that
the relationship between "tendencies" and humor is not
simple and that HmBany factors, such as depth of repression,
~a~y be important. Ea~stmaan's (16) emphasis on humlor as a
form of play and as a means for getting pleasure suggest
moare positive aspects of the humor response, as does All-
port's (2) equating of humor with insight. Ahndrews (4),
Gattell (10) and others have verified, through factor
analysis, the importance of tendency humoI~8r together with
play and pleasure seeking factors. Eysenck (20) h~a shown
that humorrous material involves the cognitive, cenative
and affective aspects of the mind. He alsoa noted the dis-
tinction between the appreciation of humlor and the produea-
tion of humraor.
In delimiting the problema for the present study it
was decided to investigate the production? or creative~ as-
pects of hnum8or rather tha~n appreciation. The latter aspect
has~ been overemphasized. In addition, although apprecia-
tion tests of hukmor measure individual differnces in huorr
preferences they do not reveal a person's readiness and4
ability to be humosrous. In this study it is desired to
investigate some of the personrality factors involved in
Many issues and questions that need to be d~ei8dd
are apparent in the review of the humor literature. Hew-
ebve, the one general aspect of humor theory that it is
desired to investigate in this study is one of emphasis.
Both "positive"' and nnegative" sources of the h2umor re-
sponse have been described. That is, humor can be the ex-
pression of a constructive, insightful, process that is
pleasurable in its own right; or humto~r can serve as a
technique for camouflaging and releasing socially tabooed
conflictual maehterial. In the latter case, humor is the
me~an to an end and the pleasure comes primarily froma oth-
er sources. It is this "release" or Hfurtive" aspect of
h~morr which has been m~ost studied and mosat widely reco5-
nized. Freud, (23, p. 801) for example, has referred to
humor as "one of the psychic correlates of the flight re-
flex.H Kost writers have emphasized this escape function
and have given little attention to humor as a possible ex-
pression of superior adaptation, as an index of personal
adjustment rather than ma~ladjustment, or as a social eat-
alyst rather than a social weapon.
In order to obtain a more accurate evaluation as
to the relative importance of these two aspects of humor
it should be possible to compared, on a test of readiness
and ability to be humorous, subjects having personality
characteristic conducive to tendentious humoPr With sub-
jects having personality characteristics conducive to non-
tendentious humoar. Results indicating that potentially
"tendentious" subjects are the ~most humorous would tend to
support humor theories which emphasize discharge of re-
pressed feelings. Results in favor of the "non-tendentious"
group should tend to support "superior adaptation" or nin-
slightH theo~ries.B It should be noted that either type of
humor m~ay make use of cognitive, conative and affective
A humor test was construlcted, consisting of eight
incomplete oartoons which depict a variety of interperson-
al situations. One student is shown speakring to another,
for whom the subjct is asked to fill in a humorous reply.
The repl.ies area rated for "funninvess"d by a group of judges.
The situations are designed so as to stimulate both cog-
nitive and orectic humor.
The California F scale, or "authoritarianrism"
scale, was selected as the psersnality measure because it
has been shown to mea~dsaure the more important cog~nitivea and
orec~itie traits believed to be crucial in various huorar
threries. Flugel (22, P. 725) hias stated, MAuthoritarian-
ism in any formp seems inimical to humor--and here again
is a matter deserving further study." Flugel apparently
prefers to view~a humra ~~ as an expression of psychop~logical
health ratherr than path.ology. However, the pre~occupation
with sutperiority, sex, and ae~jression found in "authori-
tarians" should stimulate humshorous respon~ss, if humor is
The hyp~otheses caen now bse stated.
WFhen subjects classified as low, medial, and high
in authoritarianism, as measured by the California F scale,
are judged for ability to be humorous on a test involving
incomplete cartoons, their relative humo~ar scoires will be
1. Subjects low in authoritarianism will achieve
ai~nificantlyT higher hiumor seorest than will subjects higih
2. Subeects low in authoritariatnism wrill achieve
significsantly higher hum~or scores than will subjcts
medial in authoritarianism.
3. Subjects medial in authoritarianism will achieve
significantly higher humor scores than will subjects hefgh
The original pool of subjects for this experiment
consisted of 188 mable students at two large universities
in southeastern United States. The students were freshmenc
and sophomores enrolled in beginning classes in psychology,
English literature, and sociology. Testing was done dur-
ing class time ran all of the mahle students present in the
classes contaeted agreed to servea as subjects. Since the
experiment was conductedl during the summe~ar session whebn
the num~ber of classes was reduced, the samaple represents a
sizable proportion of the marle students enrolled in these
classes at that time. Subjects were selcte3d from the
particular classes mentioned because an insufficient nuim-
ber of psychology students were available and it wi~as de-
sired to obtain a relatively homogeneous group of arts and
+selencs students. This ntender-minded" sample was desired
for two reasons: first, the judges of humor were to be psy-
chologists, and second, it was felt that inclusion of sub-
Jects from the Htough-minded" disciplines might introduce
dissimilarities in interests and training that would tenr~d
to dichotomize the sample on a basis other than measured
personality traits. Freshmen and sophonores were used as
subjects since it was dseired that they be relatively na-
ive with respect to personality tests in general and the
F scale in particular. TPhe s~aple was restricted to males
because certain studies (35) have indicated marked sex
differences in humor preference which would be difficult
to control in an lexpe~riment utilizing incomplete cartoons.
Measure of Authoritarianism (California F Scale)
Theb Galifornia F" scale has already been dsesribed
in the review of the literature. The form used is the
final form of the F seale (form 45-40) as developed by
Adorno et al. (1, P. 225) except that itemn number 22 was
omitted, making a seale of 29 items ingsted of the origin-
al 30. Itaemi numrber 22, which refers to postwar~ Germrany,
has lost its tim~eliness and has been omnitte~d by other re-
The Likert method of sealing (36) was used in con-
struaction of the F seale. Subjects are asked to indicate
degree of ~~ragrement or disagreement with each item accord-
in8 to the following scale:
+1 slight support, agreement
+2 moderate support, agreement
+3 strong suppp~ort, agreement
-1 alight opposition, disagreement
-2 mboderat eopposition, disagreement
-3 strong opposition, disagreement
Since higher scores are intended to B9express increasing
authoritarianism, the responses are converted into scores
-3 = 1 point +1 = 5 points
-2 = 2 points +2 = 6 points
-1 = 3 points +3 = 7 points
It will be noted that the scoring skips from 3 to 5 points
between -1 and +1. Four points represents the hypotheti-
cal neutral response and was assigned when an item was
omitted. The authors state that this scheme~ was used main-
ly because there seemed to be a greater psychological gap
between -1 and +1 responses than between any other adja-
eent responses. A person's scale score is the sum of his
scores on the single items.,
Thea distribution of the F seale scores obtained
(N = 188) is shown in Figure 1. Curves for actual fre-
quencies and for frequencies smoothed once are plotted.
The mea~bn ecore was 3.519 with a standard deviation of .84r0.
The median was 3.583. The mean score of 3.549 corresponds
to a response between -1, slight disagreement, and zero,
the hypothetical neutral point. Skewness was calculated
as -.144, a value not quite significant at the .05 level.
Kiurtosis = +.012 (platykurtic), alse not significant.
Therefore, deviation of this ourve from a normal curve can
CY ~Oao 49 ~ QO ~ ~ pr(O
(8 8/= N) S~133~9~6
be attributed to sampling errors.
ILn border to mak~e jumdging feasible, ther cartoon re-
sponses of only 60 of th1e original 188 subjects were used
in the main part of the experiment. These 60 subjects were
selected according to their total F scale scores as fol-
It~he 20 sub+ects scoring highnest in authoritari-
anism (h_(gh group).
The 20 subjects scoring "mid-mostw in authoritar-
innism (medial group).
Thea 20 subjects scoring lowest in authoritarianism
It-was originally intended. to compare ony the two
extreme groups, WhighsH and Wlows." However, it was de-
cided to included a sample of medical scorers in order to
more adequately messure the relationship between humor and
authoritarianism over the entire range of F scale scores.
For example, it was desired to allow for the possibility
that subsects #average" in authoritarianism alght be more,
or less, humorous than scorers at either extreme.
Although it was noted that the distribution of F
scale scores could be treated as a normal ourve, there
was sufficient negative skewness to affect the relative
distances from the men of the three+ groups of 20 subjects.
Scale sco~re limits and corresponding scores are shown
in Table 1 for each of the three groups.
SCALE SCORE AND J SCORE LIMITS OF THE THREE GROUPS
OF SUBJECTS SCORING HIGH, MEDIAL, AND LOW ON
THE F SCALE. (ME;AN = 3.549, 8. D. = .840)
F Scale Group Scale Score Corresponding
Limits t Scores
High (N = 20) 5.897 +2.80
Medial (N = 20) 3.724 +0.21
Liown (N = 20) 2.448 -1.31
It can be seen from Table 1 that the inner limits of both
extremrre groups are greater than 1 S. D. from the mean.
Howeever, this value is -1.31 for the low group while it is
only 1.07 for the high group. This is not as likely to
affect our results as is the difference in relative dis-
tances of both extreme groups from the medial group. The
Score distance between the low and maedial groups is
1.23; for the high and medial groups it is .86. There-
fore, we would expect the medial group to be somewkhat
closer to the high group than to the low group with re-
spect to the characteristics measured by the F scale.
In the present study, the humorous productions were
obtained by means of a carto~on test devised by the ex~peri-
mernter,. (TPhe eartoon tes~t is reproduced in Appendix I.)
Rosenzweis (50), and Sanford and Rosenstock (51) have used
incomplete cartoons as projective~ devices in which the
faces of the various characters are drawn almost without
features and are purposely ambiguous as to facial expres-
sion. In p~retestinga with variousl cartoons for the present
study it was found that subjects need the stimulus of a
potentially hum~~o~ro situation, with characters drawn as
amilin8, in order to respond appropriately. Pretesting
also revealed that it is necessary to dir~et subjecatsr spe-
cifically to give humrorous response. (Instructions for
subjects are reproduced in Appendix III.) Spontaneous hu-
mor would be theoretically of greater interest. However,
it was found that in the ellsaroom situation, students are
set to treat all tests with deadly seriosetness and need to
be reassured thate they are supposed to give "funny" replies.
Humor preference tests, in which the suabject is presented
with a series of jokes, encounter less difficulty in this
The cartoons were drawn for a college undergraduate
sample. An attcemsrpt was made to include a representative
seleetion of campus scenes, both academic and reoraeational.
The "stimulus remairks"~ used in the cartoons were selected
to allow for both cognitive h~umo, such as puns or plays
on words, and the uual)P1 range of cam~pus oreetic humor,
including sex, superiority, and aggression. An attempt
was mnade to avoid frustrating situaitions of the intensity
depleted in the Rosensweig cartoons. Rather, the scenes
were designed with the intention of showing good-natured
banter of the sort that can stimulate either nharmnless"
or tendentious wit. The number of cartoons, eight, was
arprived at as a compromise beatweean what wasb felt to be
n~ecssary for reliability and what would constitute a rea-
sonaPPble task for subjects an~d judges.
Administration of the Tests
The tests were given to the students during the
last thirty minutes of one class period. The experimenter
administered the tests to eight of the classes and another
graduate student in psychology administered the tested to
the remaining three classes. Copies of the instructions
for the test admainistrator and for the subjects are in-
c~luded in Appendix II and Appendix III respectively.) Nel-
ther test was timed. Subjects were asked to complete the
humor test First before going on to the F scale. This
order of presentation w~as chosen because it was assutmed
that the eartooin test responses, being projective in
maature, might be markedly influenced by~ the structured and
highly orec~tic items containQed in the F ocale. It was
felt that any order effects of the oartoon test on the F
scale would be less marked, and in any~ capse, the order
would be the &ame for all subjects.
In selecting judges for studies in experimental~a
aesthetics there are usually recognized experts to choose
from, although it is still necessary to describe in detail
the characteristics of the judges, as, for example, an
artist's identification with a certain nachool.n "Experts "
in humoir, however, are likely to be highly specialized,
both in the type of humor that they disseminate and in the
type of audiences they attract. In presvious studies of
humor with college students (19) subjects' preferences
have been co~praresd with the~ results for the group as a
whole. The result was a significant lack ef agreement as
to what was "funny."
In the present study, it was decided to use pay-
chologists as judges for several reasons., Due to the nai-
ture of the experiment, the selection of a largee and
representatives" group of judges was not feasible--if it
ever la in the study of humor. Therefore, it was felt that
the judges rshuld be a "known" group, fairly homogeneous
with respect to backg~round, interest, and education. It
was also desired, due to the complexity of the humor vari-
able, that judges Be familiar with rating procedures and
skilled in the evaluation of verbal material. Assum~ing;
that individual judges demonstrate satisfactory reliabil-
ity in their ratings, and that significant interjudge
agreement is shown, it should be possible to interpret the
results in terms9 of the "known" group. Seven men with
Ph. D. degrees in psychology volunteered to serve as judfg-
es in the present experiment. Five of the judges were
university profesrsors and two were clinical psychologists
with the Veterans Admrinistration. The latter two judges
have had continuing contact with a campus environment.
The comp~lete instructions for judges is reproduced
in Appendix IV. Each judge was given eight packe;trs of
cartoon replies. Each packet contained, typed on lips of
paper, the 60 replies to one cartoon. The cartoon replies
had been given cogde numbers and had been shuffled. Judges
were instructed to shuffle the eight packets before judg-
ing in order to control for order effects between cartoons.
Shuffling, together with seaEtrate codinS for each cartoon,
controlled for order effects between subjects. Replies
were typed to avoid idetiztfication through handwriting
which might have led to n"halo" effects. Each judge worked
Judges were asked to place the 60 slips for each
cartoon into four categories: "not funny," Mslightly fun-
my," moderatelyy funny,w and "very funny." Each category
was te receive 15 slips. Judgest~ were instructed to, "Ac-
cept a broad definition of what is humorous and group the
replies according to the degree to which they amuse you."
Judges then recorded by code number their selections for
ea~Pch category. Thisl method of rating has been called by
Guilford (27, p. 264) a Wdefined-groupM scale, in which
". the ju~de is given instructions as to what propor-
tio~ns of the samples should be expected to fall in each
After decoding of the judges' ratings, scoring was
done as follows: "no~t funny = 0 points, Hslightly funny"
= 1 point, "m~oderatealy funny"n = 2 points, atnd "very funny"
= 3 points, Since each judge placed 15 responses in each
of these categories for eaich cartoon, he assigned a total
of 90 humosr points per cartoon. With eight cartoons, each
judge assigned a total of 720 humor points for 60 subjects.
If h~umr points were distributed aceerding to
chance (nall hypothesis), each of the three F scale groups
should receive 2410 humror points per judge, or a total of
1680 points for ~ech group.. The results actually obtained
are presented in Table 2.
With respect to the total humor scores for the
three F scarle groups, as shown in Table 2, Chi square was
usebd in order to determiPne whether or ncoit the obtained re-
sults, as a whole, differed significantly from a trichotomy
(Bi5, p. 431l). TPhe Chi square test ( X= 27.56, df = 2, P
=<.01) vindicated that the obtained distribu~tion of humor
points can be considered significantly different fromi~ a
chanceb, or trichotomous distribution. The testing of the
significanc eof thme obtained differecnes between the groups,
TOTA1L HUMOR POINjTS RECE~IVEbD BiY SUBJECTS INd THEr
LCW ME'DIAL, AND HIGH F SCALE GROUPS
Fa Scale Groups
Low Medial High
Judge (N = 20) (NJ = 20) (n =e 20)
1 262 238 220
2 268 24Q0 212
3 264 234 222
4 268 222 230
5 264 240 216
6 263 221 236
7 263 230 227
Total 1852 1625 1563
against the null hypotheses, was also carried out by Chi
squaPtre, as follows:
1. A Chi equetre test of the difference in total
humosr points received by the low and htgh F scale groups
wars significant at the .01 level of confidence (1 ?C= 24.46,
df = 1, p =<.01). TPherefore, the low_ group rtecived sig-
nificantly mocre haumor points than did the high group, a
result interpreted as supporting Hypothesis 1 (Above, p.
2. A Chi square test of the difference in total
humor points resceivd by the low and medial F scale groups
was significant at the .01 level of confidnene (X == 14.~82,
df = 1, p =<.01). Therefore, the low_ group received sig-
nificantly more humor points than did the medial group, a
result interpreted as supporting Hypothesis 2 (Adbove, p.
3. A Chi square test of the difference in total
humor points received by the medial and high F scale groups
was not significant (3 := 1.2, df = 1, p =>.20). There-
fore, the difference in humor points received by the medial
and high groups could easily have resulted from chance.
This result fails to support Hypothesis 3 (Above, p. 37).
The results of the humor ratings, tabulated accord-
ing to the different scores received by the three F scale
groups on each of the eight cartoons, is shown in Table 3.
Inspection of Table 3 reveaals that only in car-
toons B, E, and H are the ratings in the predicted order
of low, medial, and high. These three cartoons were the
only ones which, taken singly, discriminated significantly
beteenthe10gandhig~h F acale groups. The Chi squares
were, respectively, 6.40, p<.05; 8.84c, p*C.01, and 19.6,
p<.01. (The remaining Chi squares, none of which were
significant, were as follows: A, ?CL = 1.466 p>.10; C, 7C
= .114, p >.70; D, Xa= 3.12, P>.05; FIX = .186, p>.50;
G3, 7C = .80, p > 30.) However, the direction of the scor~e
difference for the low group over the h_1g~h group la con-
slatent for all cartoons except cartoon C.
TOTAL HUMOR POINTS RECEIVED BY THE THRiEE F SCALE
GROUPS TrABULATED ACCORlDING TO CARTOON\S
F Scale Glroups
Low Medial H'igh
A 234 189 207
B3 237 208 185
C 210 203 217
D 238 191 201
E 244 204 182
F 220 199 211
Gt 212 224 194
H 257 207 166
Tok~takl 1852 1625 1563
aCartoons as given to the subjects, were numbered
from 1 through 8. However, as an aid toe the judges, the
eartoons were later designated consecutively by letter from
A thnrough~ H.
The distribution ourv~e for the humor ratings is
shown in Figure 2. The possible range for each subject was
from 0 (0 points for each cartoon) to 24 (3 points for
each cartoon), for any one judge or an average of all
judges' ratings. liThe o~btained range for an individual
judge's ratings was~ from O to 22 points. The range of the
averaged judges' ratings was from 2.57 to 20.0 points.
In other words, the judges agreed almost unanimo~usly, in
the case of certain subjsect, that none of their cartoon
responses were funny. Also, there was almost unacnimous
agreement that some subjects succeeded in giving "~very
funny" responses to every cartoon.
obtain ed frequencie s
doo Ied" once
,o o Af=/. oo, cr- =4' /P' o o
Ja 4 6 8 to /2. / /b /P .Zo 12~ .24
AVERAGE //l./AOR rQAT I ~6
Fig. 2.--Distribution of avera~hge humorr ratin88s,
showing obtained frequencies and sm~oothed curve~.
In the distribution of humor ratings the mea8n of
12 was determined by the judging procedure. The standard
deviation was 4.18. There waas a su6Gestion of negative
skewnesns (Sk = -.26) and of kurtosis (Ku = +.027, platy-
kur-tic), but neither value was larger enough to indicate
significant deviation from the normal curve.
The self-reliability of each judge's ratin68 was
ealeulated by obtaining correlations of his ratings for
each subJect on cartoons Ai, D, F, and G with his ratings
for cartoons B3, C, E, and H. The test was split in this
way in order to control f'or both order and page position
effects. The reliability coefficients, corrected by the
Spearman-Brown formula, ranged from .50 to .70. havr
age self-rerl,1ability obtainedd by using Fisher's at fune-
tion) was .63. The overall reliability for the test wasl
calculated by the same method used for self-reliabilities,
except that the combined ratings of all judges were usecd.
The corrected reliability obtained wase .77.
The degree of agreementS~ between the various judgesrp'
humo~sr ratings was calculated by obtaining a c~orr~elation of
each judge's distribution with that of every other judge.
The twecnty-one i~nterJud~e reliabilities ranged from .65 to
.83. The average interjudge reliability (obtained by ue-
ing Fishor's z function) was .73.
In addition to coms~paring the performances of the
three F scale groups on the hum2or test, it is possible to
approach the results from the standpoint of individual
diffe~renes.e That is, one can determine a subject's posi-
tion on the humlor test and then note his F scale group mems-
bership. This has been done in Table 4 for the 10 sub-
jects having the highest, and for the 10 subjects having
the lowest, humor ratings.
F SCALE GROUP CLASSIFICATIONS OF THEa TEN W"MO6ST HUMORSOUS"
AND THE TEN LEASTT HUMOROUS" SUBJECTS
F Scale Group Classifiention
Low M~edial Hi gh
Ten "Most 6 2 2
Humorous" subjects 34
Before discussing the findings which relate direct-
ly to the hypotheses it ma~y be advisable to consider some
of the seondary findings, such as the reliability of the
judges, form of the humor distribution, and the like, which
will influence interpretation of the results. The fact
that an essentially normal distribution of humor ratings
(see Figiure 2) was obtaine-d is of intreret in itself, since
studies of humor preference cast no light on the distribu-
tion of sense of humor." Freud (23, p. 728) has stated,
"Wit maDking is not at the disposal of all, in generral,
there are but a few persons to whom one can point and say
that they are witty .. ." Such statements reflect our
admiration for the few "humtorists" at the "top* of the
distribution without casting any light on the shape of the
rest of the eurve. Neither the limited nature of the task,
the use of extreme~i groups for subjects, nor the rating
method employed suit the present experiment to test for a
general "curve" of ability to be humorous. However, in
the restricted range of stimulus situations presented in
this experiment, and with a homogeneous group of judges,
the ability to be humorous was found to be normadlly dis-
The range of "humo~r talent" appeared to be rather
limited in spite of the breadth of the sample. Most of
the judges were disappointed at the scarcity of really hu-
morous replies. The judges' ratings of "very funny" were,
therefore, relative to the sample and not to their usual
standards of humor. In order that the reader can better
appreciate the nature and variety of replies, the cartoon
responses of four of the subjects have been reproduced in
In the preliminary planning for this experiment the
question was frequently raised as to whether or not judges
would be able to show any agreement with them~selves, or
with each other as to judgments of Hfunniness." ]Previous
research, such as the study by Grzivok and Scodel (26),
has shown that judges can reliably differentiate humor con-
tent, as, for example, between aggressive and incongruous
humor. But ability of jud,-es to agree with each other as
to what is "funny" has not been demonstrated. In fact,
ysen~ck (19) has stressed the absence of conformityt" in
humor appreciation. Previous studies, however, have given
judges an extremely wide variety off jokes and other varie-
ties of humo3rous productions from which to select prefer-
ences. In the present study, with a more specifle task,
judges were able to achieve what would appear to be mosder-
ate intrajudge and interjudge reliability, in view of the
complexity of the judgments and the briefness of the humr~s~~
test. Wilson, Guilford, and Christensen (57), using an
approach similar to the present stuidy, asked subjects to
inve~nt titles for two brief stories. T~he titles were then
rated for "cleverne~ss on a six-point seale by three judg-
es. Reliabilities of the individual judges ranged from
.69g to .77 (range in the present study wras .50 to .70);
interjudge correlations of ratings ran~ged from .53 to .76
(range in the present study was .65 to .83); and the re-
liability computed from the composite ratings was .76
(composite reliability in the present study was .77). It
is interesting to note that the reliabilitica achieved in
rating humoro" compnare favorably with the reliabilities
found in rating "cleverness," although Hcleverness" would
appear to be mo~re of a cognitive and less of an affective
variable than is humor, and, presumably, more susceptible
to reliable measurement. It would also be of interest to
know, especiallyr with judges highly trained in verbal
skills, what the correlation might have been had the judg-
es rated the same responses for both "cleverness" ad~"hu-
From the breakdown of the ratings by cartoons,
shown in Table 3 (Above, p. 52), it is apparent that isom~e
of the cartoons diserisinate between the F scale groups
considerably better than do others. Cartoon H, in which
the first student is saying, "I'll bet she forgets you
when she goes home weekendal", produced the sharpest dis-
crimination. Cartoon C, in which the stimulus words are,
"How about lending me five bucks until next week?", was
the poorest discriminator. It is difficult to explain
these results, although an inspection of the replies of-
fers some clues. Cartoon C seems to stimulate rather
abrupt, negative reactions which tend to be stereotyped.
The situation, since it involves money, calls for action
and not reflection. In addition, the stimulus character
in this cartoon was inadvertently drawn with a mildly
threatening facial expression, further restricting the af-
fective range of the responses. Cartoon H, on the other
hand, presents a stimfulus which is muach richer in its ca-
pacity to elicit the subject's feelings about himself and
others. The nature of the responses suEgests that Mnon-
authoritarians" were ~seure enough to be able to use humor
as "superior adaptation," whereas the "authoritarians"
tended, to react with defe~nsive hostility to a degree which
inhibited humorous responses. Analysis of the qualitative
differences between the humorous replies of hig~h and low
F acale group subjects was beyond the scope of the present
study. However, some appreciation of the differences in
"tendlentiousnessa" can be gained from the sample responses
given in Ap~pendixr V.
The principle findings of the experiment were set
forth in Table 2 (Above, p. 50). It was noted that the
results tended to confirm the predicted superiority in hu-
mor scores of the low F scl ru vrtehigh_ and medi.-
al groups. However, the score for the medial group was
not significantly higher than that for the high group.
The failure of this difference to be sig5nificant may be
related to the negative skewness of the F scale distribu-
tion. Although the skwnenss~ was not si~gnificant, i.e.,
might result from sampling, variations, its effect was real
enough in the present experiment, causing the 20 cases
grouped around the median to be closer to the hLgh group
thaln to the 1005 group, along the F scale con~tinuum. The
results with the medial group are in the predicted diree-
tion and the absence of a significant difference between
the medial and which groups does not ma~terially influence
the findings as a whole. The medial group was included in
the experiment primarerily as a hed~e against t~he possibill-
ty that the person who is "'averageH in F scale authLoritar-
lanism might be the most humorous. The results obtained
make it possible to reject this possibility.
In formulating the hypotheses for this study, it
was suggested that Prsults showing superiority of the Iow
over the higrh F seale group in ability to be humorous could
be interpreted as supporting theories which see humo~cr as
an index of psychological health, as opposed to theories
which view humor as the by-product of repressed conflicts.
The question is largely one of emphasis because mcost a~uthor-
ities have recognzized both aspects of humor. Brut the bulk
of the literature has concerned itself solely with humor's
function in releasing repressed hostile or sexual impulses.
If this were hum~or's primary function, the high F scale
scorers should have had a significant advantage on the hu-
mor test. In spite of the impure factorial comp-osition of
the F scale, there is substantial evidence that it taps a
wide spectrum of tendentious material; that is, partially
repressed and socially unacceptable sexual, superiority,
and aggressive needs. But the overall results su86est
that these needs tend to be associated with below aver-
a~o, rather than above average, humor ability.
One factor that should be considered in interpret-
ing the results of this study is the significant negative
correlation between the F scale and ea~s~ures of intelli-
gence. Correlations of -.21 and -.32 were previously not-
ed (Above, p. 30). The implications for the present study
are difficult to determine because the relationship be-
tween humor and intelligence has never been adequately in-
vestigated. Freud (23, p. 728) desscribed the asense of
humor as n. . a special ability, fairly independent of
intelligence .n . ndis andi Ross (35) found negli-
gible correlations between their humor test and measures
of intelligence. On the other hand, A11port (2, p. 224)
has~~ said that to achieve a sense of humor requires ".
a high level of intelligence.M One reason for this lack
of agreement may be due to the failure of investigators to
distinguish between the cognitive and the orectic aspects
of humor. Intelligence is certainly substantially in-
volved in cognition, but presumably So n lesser extent in
affeeiation and contion. In the present study an attem~ipt
was made to compensate for the frankly verbal nature of
the task by the use of eartoons and by selecting stimulus
situations designed to~ tap sources of "tendency humor."
However, it is difficult to estimate~ to what degree the
humor scores in this study reflect intellectual ability.
To the extent that humor is viewed as an index of health-
ful adjustment, at positive correlation with intelligence
would be exa~p~eted. However, in view of the fact that in-
telligence correlates only to a low or moderate degre
with either the F scale or with Hsense of humor," intel-
ligence is probably a si~gnificsant bu~t nott a determining
factor in the preser;n~t eixperimeabnt.
An additional question that arises in the present
study invrolves the characteristics of thEe judges with
respect to the various traits m~neasured by the F scale.
P~re sumably, p ayehelogis8t s tend to be "non-autho ri tari an "
and therefore might tend to favor humorous responses giv-
en by the Ilew F acale scorers. In a sense, this begs the
question since tendency theory would mai~~ntain that a re-
sponse would be amusing in proportion to its un~acceptabil-
ity. In order to control for such influences one would
have to obtain judges with "average" F scale scores; but
since the F scale is designed~ for naive subjects it would
not be possible to obtain comnpeten~t jsdges in this way3. A
study by Seodel and Murssn (53) bears on this point. They
foundC that "non-authoritarians" can correctly judge the
attitudes of "authoritarians" but that the reverse is not
true. Likewise, E~Spstein and Smaith (17), in their study
relating to insight and tsnse of haumor, found that ability
to judge cartoons accurately was directly related to accur-
acy of self-evaluation. It would appear that knowledge of
the ways in which humorsr ratings are influenced by the
characteristics of the judges will have to be gained froma
accumulate~dd studies with groups of competent judges who are
homogeneous with respect to sorme variable. It seemsH~ un-
likely that humo~tr ratings by a "random samplfle" of the pop-
ulation would be either reliable or meaningful.
Although signirficnt differences in humor scores
were obtained between the low and the highr, and the low and
the medial F scale groups, it weas noted that there was con-
siderable overlap. Thio la apparent in Table 4 (Above, p.
55) wh~ich' showsJ the F acale group classificat~ions for the
10 hi~heat and the 10 lowoot scorers on the humor test.
Taking; the 10 nmoasst humorous" individuals first, it is ob-
served that six camei from~ the low F scale group and two
each from the medial and hiph groups. With respect to
Flugel's stateenk~rt (22, p. 725i), "Authoritarian ism in any
form sooms inimical to humor . .", our results are in
gnegra~l grseement; buit the above- exemptions indicate that
there are wide individual variations. Also, on the basis
of the prceset study, one cannot maelss the converse of this
stateament, to the effect that nnon-authoriterianism is in-
imnical to the abse~nce of humo~r." Of the 10 n1cast humor-
our" individuals, three came from~r the low and medial groups
and four came from the ht~gh group, obviously a chances dis-
tribution. The ab ovelr findin60, of course, correspond with
common sence observation. Some of our acquai-ntances, who
are liberals in every sense of the word, seem completely
lackin6 in sense of humogr. Other people, who seefm to com-
bine lall of the various neaativistice traifts that led to
the concept of authoritarianism in the first place, may
h~ave quite a livealyr sense of humonr. The results of the pre-
sent study ougges that these cases are exceptions, but un~-
til we can explain such inconsistencies our theories of
humor are incomplete.
1. One hundred and eighty-eight university under-
graduate maale subjects were given the California F scale
(Authoritarianise), together with a cartoon test of humor.
Sixty experimental subjects were then selected on the basis
of their F scale scores: the 20 scoring highest (most au-
thoritarian), the 20 scoring lowest, and the 20 scoring
Hmid-nost" in the distribution. The responses of these 60
subjects were then rated on a four-point scale of relative
funniness by seven psychologists who acted as judges.
2. It was predicted that with respect to humor
scores: (1) the law F scale group would significantly excel
the high group, (2) the low F scale group would significant-
ly excel the medial group, anld that (3) the medial F scale
group would significantly excel the high group. The ob-
tained differences were tested by Chi-sq~uare. Differences
were significant and in the predicted direction with re-
spect to hypotheses (1) and (2). The difference for hypo-
thesis (3) was in the expected direction but was not sig-
3. The results were interpreted as tending to
support theories of humor which emphasize the adaptive, as
opposed to the pathological, aspects of humor. Possible
biasing factor noted were the negative correlation of the
FT scale with intelligence and tfhe~ use of Sudges who prob-
ably tend to be nonauthor~itarian. Wide individual differ-
ences in hum~sor ability for subjcts~ in all three experi-
me~tall groups were noted.
4. Corrected split-half intrajudge reliabilities
for the humor ratings ranged froma .50 to .70. Interjudge
reliabilities ranged from .65 to .83. Com~posite reliabil-
ity for the test was .77. These results were interpreted
as suggesting th8at humnorous productions are susceptible
to being rated with satisfactory reliability and inter-
judge agreement if a homogeneous group of judges is used.
1. Adorno, T. W., Frenkel-Braunavik, E., Levinson, D. J.,
and Sanford, R. N. The authoritarian personality.
New York: Harper, 1950.
2. Allport, G. W. Personality, a psychological interpreta-
tion. New York: Holt, 1937.
3. Allport, G. W., Vernon, P. E., & XLndtey, G. Study of
Values. Boston: Houghton M~ifflin, 1951.
4. Andrews, T. G. A factorial analysis of response to the
comic as a study in personality. J. ren. Psychol.
1943, 28, 209-224.
5. Aristotle. Des Poetica, Translated by By~water, I.
Oxford: Clarendon Press, 1924.
6. Berkrowitz, M. A study of somne relationships between
anxiety and humror. Unpublished doctor's disser-
tation, Pennsylvania State University, 1954.
7. Block, J. &c Block, J. An investigation of the relation-
ship between intolerance of ambiguity and ethno-
centrisml. J. Pers., 1951, 19, 303-311.
8. Cam~pbell, D. T., & MlcCandless, B. R. Eth~nocentrism,
xenophobia, and personality. Hum. Relat,, 1951,
9, cattell, R. B. IPAT Humor test of personality, Form A.
Champaign: Institute for Personality and Ability
10. Cattell, R. B. Handbook for the IPAT humor test of ner-
sonality, Forms A, B, and C. Champaign: Institute
for Personality and Ability Testing, 1955.
11. Gattell, R. B., & Liuboraky, L. B. Personality factors
in response to humor. J. abnorm. soc. Psychol.,
194~7, 42, 402-421.
12. Conkrlin, E. S., & Sutherland, J. W'. A comparison of the
scale of values method. with the order of merit
me~thod. J. exp. Psychol., 1923, 6, 44-57.
13. Deskins, A. J. MaI~glicl thinking, superstition and au-
thoritarian trends in personality. Unpublished
doctor's dissertation, University of Florida, 1957.
14. Dooley, L. The relation of humor to masochism. Psycho-
anal. Rev., 1941, 28, 37-468.
15. Dorris, R. J., Lev~ginson, D. J., & Hanfmann, E;. Author-
ita~rian peirsonality studied by a new variation of
the sentence completion technique. J. abporp.
soc. Psychol., 1954, 49, 99-108.
16. EbsastmaFn, M. The sense of humor. New York: Seribners,
17. E~pstein, S., & Smith, R. Repression and insight as re-
lated to reaction to cartoons. J. consullt. Payebhol.,
1956, 20, 391-395.
18. Eysenck, H. J. The appreciation of humor: an extpei-
mental and theoretically study. Brit. J. Psychol.,
1942, 32, 295-309.
19. Eysenek, H. J. An~J experimental analysis of five tests
of appreciation of humaor. Edue. psychol. M~easyt,,
1943, 3, 191-214.
20. E~ysenek, H. J. Dimensions of personality. London: Kegan
21. Feldman, S. A sutpp~d~Qlent to Freud's theory of wit.
Psychoanal. Rev., 1941, 28, 201-217.
22. Flugel, Jr. C. Humor and laughter. In Handbook of Social
Psychology, Vol. II. Cambrid~e, Mass.: Addison-
Wesley, 1954. Pp. 709-734.
23. Freud, S. Wit and its relation to the unconscious. In
The basic writings of Sigmund Freud. New York:
Random House, 1938. PJp. 633 803.
24. Freud, S. Humor. Int. J. Psychoanal., 1928, 9, 1-6.
25. Gregory, W. E. Authoritarianism and authority. L.
abnorm. soc. Psychol., 1955i, 51, 641-643.
26. Grziwok, R., &c Scodel, A. Some psychological corre-
lates of humor preferences. J. consult. Psychol.,
1956, 20, 42.
27. Guilferd, J. P. Psychometric Methods, (1st Ed.). New
York: M8cGraw-Hill, 1936.
28. Gfuilford, J. P. Psychometric Methods, (2nd Ed.). N'ew
York: McGraw-Hill, 1954.
29. Bollander, E. P. Authoritarianism and leadership
choice in a military settin~g. J. abnorm. soc.
Psychol., 1954, 49, 365-370.
30. Jones, E. E. ~Authori~t~raianis as a determinant of
first impression formration. J. Pers., 1954, 23r
31. Jones, M. B. Authoritarianism and intolerance of flue-
tuation. J. abnorm. soc. Psychol., 1955, 50, 125-
32. Kiambouropoulou, P. Individual differences in the sense
of humo~r and their relation to tem~peramental dif-
ferences. Arch. Psychol., 1930-31, 19, 1-83.
33. Kant, I. Critique of judgement, TraLnslated by Bernard,
J., L~ondon: Maicmillan, 1931.
34. Kline, L. W. The psycho~logy of humoar. Amer. J. of
Psychol., 18, 1907, 421-441.
35. Landis, c., & Ross, J. W. H. Humor and its relation
to other personatlity traits. J. soc. Psychol.,
1933, 4, 156-175.
36. Likert, R. A technique for the messurement of atti-
tudes. Arch. of Psychol., 1932, 140O, 55 pp.
37. Lull, P. E. The effectiveness ofd h~um~or in persuasive
speech. Speech Nb~nog,, 1940, 7, 26-40.
38. McComas, H. C. The origin of laughter. PSyobol. Rev.,
1923, 20, 45-55.
39. Mc~ougall, W". An introduction to social ~psychology.
(2nd~i Ed.) Boston: Luce, 1909.
40. Martin, L. J. Psychology of esthetics. ~bAmer. J.
Psychol., 1905, 16, 35-118.
41. M~oss, F., Hunt, T., &G Omwake, K. Social intelligence
test. W~ashington: Center for Psychological Ser-
42. Mrurray, H. A. Thematic Apporception Test. Cambridge:
IRarvard University ~Press, 1943.
43. O'Connor P., Ethnocentrism, "intolerance of am~bigul-
ty,& and abstract reasoning ability. J. abnorm.
soo, Psychol., 1952, 47, 526-530.
44~. O'Neil, H. M., & Levinson, D. cJ. AC factorial ex~pleras-
tion of authoritarianism and some of its ideolog-
ical1 concomitants. J. Pers.. 1954, 22, 449-463.
.45. Peatm~an, J. G. Descriptive and sampling statistics.
New York: Harper, 1947.
46., Plato Philebus Tri~eanslated~ by Fowrler, H. New York:
P~u;tln~am's s;-ons, 1925.
47. Rapp, A. The origins of wit and humor. Newn York: Dult-
48. Redlich, F. C., Levine, 6., & Sohler, T. P. A mirth
response test. Am. J. Orthopsychiatry, 1951, 21,
49. Rokeach, EM. Dogmatism and opinionation on the left and
on the right. Amer. Psycholopist, 1952, 7, 310-
50. Rosensweig, 8. T3he+ picture asscication me~thod anrd its
application in a study of reactions to frustra-
tion. J.Pes, 1945, 14, 3-23.
51. sanford, F'. H., & Rosentock. I. M. Projective tech-
niques on the doorstep. J. abnormo. soc. Psychol.,
1952, 47, 3-16.
52. schopenhauer, A. The world as will and idea. Trans-
lated by Haldane, R., and Kemp, J., Vol. II.
London: Kegan Paul, 1928.
53. scodel, A., & Mussen, P. Seeial perceptions of author-
itarians and nonauthoitarians. 3. abnorm. soc.
Psychol., 1953, 48, 181-184.
54. Sears, R. N. Dyn~same ic tors in the psychology of hur~-
mor. Unpublished doctor's dissertation, Harvard
University, 1934, eited by Miurray, H. Ad. Ex~plora~-
tions in personality. New York: Oxford University
Press, 1938, pp. 583-584.
55. Spinoza, B. Ethic. Translated by WIhite, W., (4th E~d.)
London: Oxford University Press, 1937.
56. ~sullivan, P. L., & Adelson, J7. tEthnocentrism and misan-
thropy. J. abnorm. soc. Psyvchol., 1954, 49, 246-
57. Wilson, R, C., Guilford, J. P. & C~hri~sense~n, P. RB.
The measurement of individual differences in ori6-
inaflity. Psychol., Bull., 1953. 50, 362-370.
58. Yar~nold, J. li., & Berkeley, M. H. An analysis of thea
Cattell-Luboraky humor~9 test into homogeneous
scalesr. J. abnorm. soc, Psyvchol,, 1954r 49g, 543-
r9 # T~E
APPENDIX I. HUMOR TESTa
INSTRUCTIONS: IN EACH OF THESE PICTURES TWO COLLEGE STUDENTS ARE
TALKING. THE WORDS SAID BY ONE STUDENT ARE ALWAYS SHOWN. IMAGINE
YOURSELF IN THE OTHER STUDENT'S PLACE AND THINK OF HOW YOU MIGHT
RESPON~D. WRITE YOUR REPLY IN THE EMPTY BOX.
W~AT HAPPEn/ FD
/ Y DoNl'r p o Yo u
) I e ON/ TH IS T~EsT r
OR ;TH E
WEC ? I3-
a~hi isafciieo h uorts.Teoii et
gient tesujet wr dplcte n tadrdB b 1 nc apr
I LL BET -SH~E
WH EN ,SHE
6 0ES 14 0 E
ArPPENDIX II. INSTRUCTIONS FOR TEST1 ADM'l~INISTRAFTORS
Subjects, are to be MAkLE undergraduates. Mo0st su~b-
jects complete the tests in 20 minu~tes; a few will take 25.
It is probably most convenient to use the last 30 minutes
of a class period, allowing 5 minutes for passing out the
tests and reading instructions.
First hand each subj+ect a copy of the F scale, a~sk-
ing subeects to lea~sve them face~ down. The~n hand him the
cartoon test, face up. (This is done rseadily13 if tests are
previously sorted ~baick to back.) As the tests are being
passed out, ask the sub9ects to write their names on both
Dues to the na~ture of the research, the distribu-
ting of the tests andi the reading of instructions should
preferablyg be handled as informally as possible. Fortu-
nteP~4ly, som~ie students help to set the mood by laughing
when they see the earteegns.
APPENDoIX III. INSTRUCTIONS FOR SUBJECTS
These tests that you have bee4n given are part of a
research study being conducted by a graduate student.
These tests are nobt a part of your regular course work and
most students find themt interesting. Please check to see
that your name is on both tests. All n~ame will later be
changed to code numbers and results will be used only for
hen yfou have BcompletedQ theg cartoon test, turn
ovrer the seond test, which is a list of statements to be
rated, and complete it. Neither test is timed, bu~t each
one should take only 10 minutes. Affter 10 minutes I will
remind you to go on to the second test.
Now, with respect to this cartoon test. In fill-
ing in the cartoons, the idea is to give humorous repliesi
If you can't think of a humorous reply, fill in an appropri-
ate response anrd go on to the next cartoon. Be natural,
but try to think of humorous replies. All right, go ahead.
APPENDIX IV. INSTRUCTIONS FOR JUDGES
You are being asked to rate for Hfunnineoss the re-
sponses of sixty subjects to a humoa~r test consisting of
eight incom8plete cartoons. Replies have been given~ code
numbers, shuffled, and assembled separately for each car-
teoon The sixty replies for each cartoon are judged and
the results recorded before proceeding to the next cartoon.
Ma~terials to be used include:
(1) Eight pacekets, lettered from A through H.
Packet A contains the sixty replies to cartoon
A; Packet B, the sixty replies to cartoon B3,
(2) A samiple copy of the eartoon test.
(3) A set of eight recording sheets, one for ea~ch
cartoon, for use in recording the judge's
Thoroughly shuffle the eight unopened packets of
cartoon responses, leaving them face down in a pile. As
judging proceeds, select one packet at a time from the top
of the pile. After finishing with each packet and record-
ing the results, replace packet face up at bottom of pile.
Eabch packet contains the 60 responsesa to one car-
toon. Remover slips and shuaffle thoroughly. Take slips
one at a time and place them, face Fup, into four categor-
les; from left to right--Not funny, Slightly funny, Moder-
ately funny, and Very funny. After slips have been distri-
buted according to your first impressions, glance over them
again and rearrange themr so as to have 15 slips in each
category. In rating the replies, accept a broad defini-
tion of what is humorrous and group the replies according
to the degree to which they amuse you.
After each group of 60 replies has been rated,
there should be four piles of 15 replies eaceh. Take the
"Not funny" pile first, turn it over, and record the code
num~~bers (omitting the letter) by marking short lines
through the appropriate numbers under the "Not funny" eel-
umn on the recording sheet. Repeat for the other three
columns and make sure that the letter designating the ear-
toon is entered at the top of the blank. Then recombine
the 60 slips, shuffle, and replace in the packet. Proceed
to the next cartoon.
After all eight packets have been judged, check
the recording sheet for completeness and place in the enve-T~
lope provided and seal. It requires from 20 to 25 minutes
to Judge and record each packet and total Judging time will
be~ from 3 to 34 hours. Therefore, it is su~ggestedf th~t
Judging be done in twor or more sessions.
APPENdDIX V. SPECIMENS OF UM]OR TEST RESPOIJSE~S
1. Cartoon responses of a subject from the low F scale group
who scored h_1gh (rank order 1.5) in the humor test.
A "My Cadillac broke dowvn and I doubt if my date
wo9~uld consentl to ride in a Chevrolet."c
B "M~y ankswers were so intelligent and abstract
that t~he professor was unable to grasp their
0 "I would .. except my father broke his leg
and is inc~apacitated, my mother was laid off
her job, and I lost money in a poker g~ame, and
as a result have $3.00 left for the next two
D "If it didn't require that I be blind or the
lights out, to enjoy her company."
E "Any person who appreciates ageing and mellow-
ness, plus a high degree of mechanical un~der-
standing, and not interested in superficial
F "Does this mean 12u are going to take over may
folkrs' ,job of supporting5 me, or do Bou need a
G "No, who wants to be B. M. O. C. and have all
those beautiful girls running after you, and
receive free, tuition, board, etc."
H HYes, she is probablyB very engrossed in her ctol-
lection of shrunk~en heads and her deciphering
of E:gyptian heiroglyphics,"
2, CalrtoonP~ respons8es of a subject from the low F scale
group who~ scored low (rankn order 56) in the humor test.
Ab No Esmonyl
B "I failed to study."
C "W~hen, next week?"
D "What's she like? Have you met her?"
E "I'm giving it away."
F "O. K. 'Prof,' baea.t it."
G "I wish I could play tha~st good but I don't
think I'd like to Bb on a school team."
H "She couldn't forget ag."
'3. Cartoon responses of a subject from the high_ F sealed
group who scored hZtgh (rank order 5) in the humor test.
A "Wrho in bell would have me? I couldn't get
a date on~i a Saturday night at the YWlCA with
~10 bills sticking out of my earis."
B MInstead of looking over m~y notes, I over-
C NIf I had five bucks I would stay up all
night looking~ at it."
D "Yes, but if she is a dog I'm going to
E *Oh, I'll find some blind guy."
F "I have to or I'll be digging ditches for
G MYeh, the team of iMa~rtin and Lewia--at
least I would have a few buc~ks."
H "Well, don't you think the mouse (me) will
play when the cat is away?"
4. Cartoon responsesr of a subject from the _Zigh F scale
group who scored low (rank order 58) in the humor test.
A "Because my wife does not want to go and she
won't let me go with anyone else."
B nfI'm not sureQ what h~appened. TLhey seemed
quite funny to me."
C nI'm sorry, but I do not hav~'Fe that antreh with
D UI'm a married mr~an, but th~:iank anyway."
E Anyone who is in need of cheap transporta~-
F "You bet, I have to because of my low grades
G MI sure do. I have always wanted to play
but never got the chan~e."'
Hi. "You abould not rsayv things like that unless
you k~now wh~nat you are talkinG~ about. En~-
courage, do not disoaura~ge.
Robert Scott Cleland was born Janulary 16, 1920, at
Pittburgh, pennsylvania. He received the Bachelor of Arts
degree from Monmouth College in 1941. After serving in
the navy he worked for several years in industry. During
this period he did graduate work at Northwestern University.
He entered the graduate school of the University of Florida
in 1949 and received the N~aster of Arts degree in psychol-
ogy in September, 1950. He is a member of the Amerlean
This dissertation was prepared under the direction
of the chairman of the candidate's supervrisory committee
and has bee~n alpproved byr all mem~dbersg of that committee. It
was submi~tted to the Dean of the College of Ar1ts and Sci-
~enes and to the Graduate Council, and was raproved as
partial fulfillment of the requirements for the degree of
Doctor of Pshilosophy.
June 3, 1957
Dean, College of Arts and Sciences
Dean, Graduate School