Content-Type: text/html

The Power of Humorous Context to Affect Perception of
Commercials, Programs, and Products
 
 
 
Stephen D. Perry
University of Alabama
College of Communication
 
Stefan Jenzowsky
Free University of Berlin
 
Jeanne Gartenschlaeger, Joe Bob Hester
Cynthia King Jablonski, & Huiuk Yi
University of Alabama
College of Communication
 
 
 
 
 
Address correspondence to:
Stephen D. Perry
College of Communication
University of Alabama
Box 870172
Tuscaloosa, AL 35487-0172
Internet:  <a href="/cgi-bin/wa?LOGON=A3%3Dind9408D%26L%3DAEJMC%26E%3D0%26P%3D257035%26B%3D--%26T%3Dtext%252Fplain" >[log in to unmask]</a>
(205) 348-8604
 
Running Head:  CONTEXT EFFECTS OF HUMOR
 
 The Power of Humorous Context to Affect Perception of
Commercials, Programs, and Products
 
Abstract
        Levels of humor were manipulated in programs and commercials and ramifications
for advertisers examined.  Results support using higher levels of humor in the
surrounding context and in the commercial to create increased product appeal.
Recall is also improved by a more humorous commercial but reduced by more
humorous context.  Commercials are liked less as contextual humor increases, but
the rating of programs improves as the level of humor in the commercials
increases.
 
The situation comedy is the most popular genre of television programming,
accounting for over 45 percent of the 100 highest rated television series of all
time (Zillmann & Bryant, 1991).  According to network and advertising agency
estimates of prices for fall 1993 prime-time 30 second commercials, comedies are
also among the most expensive programs in which to advertise.  Eleven of the
thirteen most costly programs are situation comedies, and prices for 30 second
spots on these shows range from $200,000 to $325,000 (Mandese, 1993).  Other
program genres also feature humor including stand-up comedy, a common genre on
some cable networks, and television talk shows.  U.S. advertising executives
tend to believe that the broadcast media are well suited for the use of humor
(Madden & Weinberger, 1984),  yet little is known about the effects of humorous
programming on the commercials carried within the show.
        Research has provided evidence of an interaction between a program and the
commercials within that program (e.g., Schumann & Thorson, 1990).  Arousal has
been the focus of many context studies, with varying results (Singh & Churchill,
1987).  Program arousal has been shown to positively affect viewers' evaluation
of commercial pleasantness and effectiveness (Broach, Page, & Wilson, 1991;
Mattes & Cantor, 1982).  Mundorf, Zillmann & Drew (1991) compared information
acquisition from commercials following an emotionally disturbing news program
with a control condition, and for a period of two and one-half minutes after
exposure, information acquisition was significantly poorer following the
disturbing news story.  Pavelchak, Antil, and Munch (1988) investigated
commercial recall for viewers of Super Bowl XX.  They found no differences in
recall between respondents in the home cities of the winning and losing teams,
but respondents in a neutral city did have significantly higher recall scores.
        Goldberg and Gorn (1987) examined the interaction between happy or sad
television programs and emotional or informative commercials. Greater perceived
commercial effectiveness and slightly better recall was found for viewers of the
happy program compared to viewers of the sad program.  Kamins, Marks, and
Skinner (1991) also used happy and sad television programs to induce mood, and
they investigated the effects of the induced mood on happy and sad commercials.
Their findings tended to support a consistency effect (i.e., facilitation by
hedonically compatible viewing context).  For measures of liking the commercial
and purchase intention, happy ads were viewed more favorably after exposure to a
happy program and sad commercials were viewed more favorably after exposure to a
sad program.  Murry, Lastovicka, and Singh (1992) manipulated subjects' feelings
with exposure to a positive, negative, or neutral emotion-eliciting program.
They found that the viewers' liking of the programs had a positive influence on
both attitude toward the ad and attitude toward the brand, while the viewers'
feelings elicited by the programs had no effect.
Context Effects for Humor
        Few topics in advertising have received as much attention, discussion, and
debate as the impact of humor.  Research has demonstrated a positive link
between humor and attention to commercials (Madden & Weinberger, 1982) and a
negative link between humor and recall of commercials (Lammers, Leibowitz,
Seymour, & Hennessey, 1983) as well as gender differences.  The findings that
humor has opposite effects on attention and recall seems suspect, however, since
attention is necessary for recall.  (For a comprehensive review of the
literature on the impact of humor in advertising, see Weinberger & Gulas, 1992.)
        Little research has focused on humor as a mediating variable to explain the
impact of programming on commercials.  In a study of radio advertising, Cantor
and Venus (1980) found no significant effects between humorous and non-humorous
contexts for either recall or product evaluation.  In his doctoral dissertation,
Madden (cited in Wienberger & Gulas, 1992) found no effect of serious or
humorous context in radio.  Murphy, Cunningham, and Wilcox (1979) found limited
context effects in television.  For aided recall, humorous ads were more often
recalled in a non-humorous context, but for unaided recall, no context effects
were found.
        In their comprehensive review of the marketing, education, communication, and
psychology literature, Weinberger and Gulas (1992) argued that there is too
little evidence to reach a conclusion for or against the context effects of
humor, and they specifically called for more research in the area.  The present
investigation answers this call.  Its purpose is to explore the effects of
high-humor and low-humor contexts on high humor and low humor commercials and
the products advertised.
Hypotheses
        The following hypotheses are advanced:
H1:     The level of humor in the surrounding program will negatively influence
recall, and will positively influence product appeal and purchase intention.
H2:     More humorous commercials result in greater product appeal, improved
recall, and greater purchase intention than less humorous commercials.
H3:     Both product appeal and purchase intention will be affected by gender.
H4:     The more humorous the context in which a humorous commercial is placed, the
more positively the commercial will be perceived.
H5:     An increased level of humor in the commercials aired during a program will
result in higher program appeal.
 
 
Method
 
Overview
 
Ninety-nine undergraduates, 66 females and 33 males, of a major university
participated in this research project.  All were enrolled in introductory
communications courses and received class credit for their participation in the
experiment.  The participants were randomly assigned to four experimental
conditions and were tested in eight combined gender sessions of up to 20
students each.
In each session, participants saw either (a) a high-humor stand-up comedy
program or (b) a low-humor talk program.  Each program was interrupted by four
commercial breaks, each containing one commercial.  Across each program
condition, participants saw either (a) humorous unedited commercials or (b) the
same commercials with punch lines removed.  The commercials were new to
participants, and featured unfamiliar products not available in the region of
the country where the study was conducted.  All unedited commercials included
humor as part of the advertising strategy, utilizing short, joke-like stories to
promote products.  The removal of  punch lines reduced the level of humor in
each of the commercials.
After exposure, the participants rated the commercials, the program, and the
product advertised on semantic differential scales. Additional tests for recall
and purchase intention were included.
Procedure
All sessions of the experiment were conducted in mixed gender sessions.  Each
experimental condition was tested in two independent exposure sessions.  All
experimental sessions were conducted in the afternoons of two consecutive days
between 4 p.m. and 8 p.m., and in no case were the two sessions of a condition
conducted at the same time of day.  Participants were supervised when entering
and leaving the test site in order to prevent the exchange of information about
testing procedures among participants.  All participants gave prior written
consent to participate in an experiment involving exposure to humorous TV items
and questionnaire testing.
A male experimenter indicated that after the exposure to a humorous program the
participants would be asked for evaluations of the program.  Then participants
were asked to make themselves comfortable and to watch the show as they would at
home or in a movie theater.  Participants giggled and laughed freely during
exposure, which lasted about 40 minutes.  Commercials interrupted the program at
intervals of about eight minutes.  The experimenters did not draw special
attention to the commercials, avoiding mention of them until the end of the
presentation.
After exposure, participants were given a six page questionnaire to complete.
Included in the questionnaire were scales evaluating the program, product-brand
and product-category unaided recall, aided recall, the advertised products, and
the commercials.
Manipulation of Commercial Funniness
The four commercials featured different products and services:  (a) British
Telecom, (b) Schwarzkopf News Hairstyling, (c) Bank of Boston, and (d) John
Smith Yorkshire Bitter (Beer). Both the commercials and the featured
products/services were unknown and not available in the area where the study was
conducted.  To accomplish this, the commercials were partly taken from foreign
television (British Telecom & John Smith Yorkshire Bitter from Great Britain,
Schwarzkopf News Hairstyling from Germany but synchronized in English for the
Cannes Commercial Film Festival, Bank of Boston from local television in
Massachusetts).
The manipulation of commercials used in this study is similar to that employed
by Lammers et al. (1983) in a study of humor in radio commercials.  In that
study, humorous quips by comedians were spliced out of the humorous version of
an advertisement to create a serious version.
In this study, all commercials told a short, humorous story about the use of the
product/service or the product/service itself.  Each ended with a punch-line and
then showed the product logo on the full screen.  By removing this punch-line, a
less humorous version of the same commercial spot was created.  Whenever
possible, only the video channel was removed during the punch-line sequence and
then replaced by a repetition of material from the same commercial.  Otherwise,
both the audio and video of the punch-line section were carefully removed, while
maintaining minimal differences in commercial length. All time differences were
less than three seconds.  Using this punch-line removal technique, a less funny
version of each commercial was created.  Each featured a still slightly humorous
and still meaningful story, that appeared to be complete despite the
manipulation.
The commercial featuring Schwarzkopf Hairstyling represents a typical example of
this punch-line removal.  In this commercial, a woman in a job interview
situation was shown.  Camera shots cut back and forth between her and the
prospective employer.  Each time the camera focused on her, her hairstyle
changed, becoming more and more  radical  as the commercial progressed.  Each
time her hairstyle changed, the employer stressed that this was  a very
conservative company,  implying that her hairstyles would not project the right
image.  In the final shot of the woman, she ended up with a conservative
hairstyle.  The punch-line then came as the camera cut once more to the employer
who, for the first time, had brightly colored spiked hair.  He was shaking her
hand and saying,  See you Monday, baby.   The logo and announcers voice closed
out the commercial.  In the low-humor version, the final cut of the employer was
removed.
Manipulation of Program Funniness
Two humorous television shows served as context program material.  In the
stand-up comedy show, A&E s an Evening at the Improv, the program segments
featured comedians performing monologues consisting almost exclusively of
humorous  one-liners  or jokes.  These speeches contained minimally harsh
language and a variety of humor types including self-disparaging and
 other-disparaging  humor, physical humor, exaggerations, and sexual humor.
This program served as the high-humor condition and featured more frequent jokes
and stories that were more jocular than those in the low-humor condition.
Segments taken from The Tonight Show with Jay Leno served as the low-humor
condition, which included two interviews and a comedian style speech by Leno.
Humor was woven into the program with jokes and short stories the host told and
from the interaction between the host and guests in the interviews.  The humor
included exaggerations, sexual humor, self-disparaging and  other-disparaging
humor, but was not typically of a crass or harsh nature.
Each program was shown only in part with segments taken from aired shows. The
edited versions had similar length of between 35 and 45 minutes including the
commercial breaks.
Measures
Immediately after exposure, respondents answered (a) a set of program evaluation
scales and (b) a product/service brand and product/service category recall test
for the products advertised during the program.  After reinstatement of brand
and product/service, respondents answered tests of (c) aided recall, (d)
purchase intention, (e) product evaluation, and (f) the evaluation of the
commercials for the four products advertised.
Program evaluation.  Respondents answered the following questions about the
television program they viewed:   Do you think the program was interesting?   Do
you think the program was funny?  and  Do you think the television program was a
good television program?   All answers were given on an eleven-point scale,
ranging from 0 (not at all interesting/funny/good) to 10 (extremely
interesting/funny/good). Furthermore, the program was rated on unipolar scales
for the characteristics amusing, likable, depressing, humorous, imaginative,
dull, informative, and entertaining.
Recall test of brand and product/service category.  Respondents were asked to
list as many of the four advertised product/service categories and brands as
they remembered.  An example was given:  "If you saw a commercial for Texaco
gasoline, the product category is gasoline and the brand is Texaco".  All
answers that had at least one word in common with the correct brand name or that
described a possible category of the product/service were given a score of one
(1) while those that did not meet these criteria were scored as zero (0).  The
scores for both category and brand were then summed to create an overall
category recall score and brand recall score ranging from zero to four.
The next four pages asked questions about specific commercials.  Questions about
aided recall, purchase intention, product evaluation, and commercial evaluation
were asked for each commercial in the order of exposure.
Aided recall test.  After the brand and product category were reinstated, the
respondents were asked how well they then remembered the commercial.  Scores on
this variable were given on an eleven point scale ranging from 0 (not at all) to
10 (extremely well).  A composite overall recall score was formed by averaging
the scores on this variable across all four commercials.
Purchase intention.  The next two questions were, "If this product/service was
available in your area, would you use it?" and "If available, would you
recommend this product/service to your family and/or friends?"  Both questions
were answered on an eleven point scale ranging from 0 (not at all likely) to 10
(extremely likely).
Product evaluation.  Respondents were then asked to give their impressions of
how much each product/service possessed the characteristics expensive, bad,
satisfying, friendly, interesting, dependable, exiting, harmful, likable, and
appealing.  All characteristics were evaluated on an eleven point scale ranging
from 0 (not at all) to 10 (extremely).
Commercial evaluation.  Participants were also asked to give evaluations of how
much each of the commercials possessed the characteristics amusing, helpful,
depressing, funny, imaginative, dull, informative, entertaining.  All were rated
on an eleven point scale ranging from 0 (not at all) to 10 (extremely).
 
Results
 
Preliminary Treatment Checks
        Initially, treatment materials were tested to confirm that the conditions
designed to be more humorous were indeed more humorous.  Responses to program
funniness were analyzed by a one-tailed t-test.  The stand-up comedy program was
rated significantly more humorous [t (97) = 1.70, p < .05] (M = 7.52) than the
alternative program (M = 6.88).  Analysis of program dullness also confirmed
that the chosen treatment programs were significantly different.  The Tonight
Show with Jay Leno was rated significantly more dull [t (97) = 2.13, p < .05]
than the stand up comedy condition (M = 1.52 for stand-up comedy and M = 2.49
for The Tonight Show).
        Additional treatment checks were performed by gender to help explain some of
the later findings.  Though means of program funniness and program dullness were
in the right direction for females (M = 7.62 for stand-up comedy and M = 7.25
for the talk program on program funniness, and M = 1.41 for stand-up comedy and
M = 1.82 for the talk program on program dullness) t-tests showed the difference
was not significant.  For males, however, program funniness approached
significance, t (36) = 1.54, p < .10, with stand-up comedy being rated funnier
(M = 7.33) than The Tonight Show with Jay Leno (M = 6.40).  Males also rated
program dullness significantly lower, t (36) = 2.13, p < .05, for the stand-up
condition (M = 3.45) than for the talk comedy condition (M = 1.72).
        Ratings of commercial funniness, averaged across all four commercials in each
treatment condition, also confirmed that the commercials that contained punch
lines were significantly funnier than the commercials from which punch lines had
been deleted: by one-tailed t-test, t (96) = 1.81, p < .05 (M = 4.85 for
commercials with punch lines and M = 4.21 for those without punch lines).
Commercial dullness was greater for the commercials from which a punch line had
been deleted, t (95) = 2.6, p < .01 (M = 2.67 for commercials with punch lines
and M = 3.68 for commercials without punch lines).
 
Treatment Characteristics Analysis
        Products.  In order to determine correlations between characteristics of
products or services, ratings on each scale were first averaged for each
participant across all four commercials viewed.  Factor analysis (principal
components, orthogonal rotation) was then used to determine factors.  Three
factors emerged accounting for 76% of the variance.  Exciting, friendly,
interesting, likable, dependable, satisfying, and appealing loaded highly on the
first factor, with all variables loading at .69 or above.  The factor explained
54.1% of the variance.  The ratings on these seven variables were averaged to
obtain the variable "positive evaluation of product."  Its interitem consistency
was high:  Cronbach's alpha = .93.
        The second factor was comprised of the responses to two questions about the
product , ". . . would you use it?" and ". . . would you recommend this
product/service . . . ?"  Responses to both of these questions loaded at .89 or
above.  The second factor explained 12.8% of the overall variance.  The average
of these ratings was called "purchase intention."  Interitem consistency was
high:  Cronbach's alpha = .94.
        The third factor was comprised of responses to the adjectives bad and harmful.
Both of these variables loaded at .80 or higher and explained 9.1% of the
variance.  The variable ratings were averaged and termed "negative evaluation of
product."  The factor was retained for later analysis although the interitem
consistency of the variables was marginal:  Cronbach's alpha = .59.
        Commercials.  The characteristics of commercials were rated on eight scales
that were again averaged across the four commercials viewed.  The resulting
ratings were then the subject of a factor analysis (principal components,
orthogonal rotation).  Two usable factors emerged explaining 70.3% of the
variance.  The first factor included the four variables entertaining, amusing,
imaginative, and funny.  All variables loaded at .88 or above explaining 54.7%
of the variance.  The average of the ratings from these four variables was
termed "positive evaluation of commercial."  Interitem consistency was high:
Cronbach's alpha = .95.
        The second factor contained the variables helpful and informative, both of
which loaded at .88 or higher.  This factor accounted for 15.6% of the variance.
The ratings were combined and averaged to form the variable "information level
of commercial."  The interitem consistency was high:  Cronbach's alpha = .87.
        A third factor emerged with the factors depressing and dull loading highly.  It
was disregarded, however, due to a Cronbach's alpha of only .34 for the scale
combination.
        Program.  Factor analysis (principal components, orthogonal rotation) was
utilized to determine what variables loaded highly together from the eight
program characteristics scales.  The adjectives humorous, likable, amusing,
entertaining, and imaginative were positively loaded with each other and
negatively loaded with program dullness in the only usable factor to emerge from
the analysis.  This factor explained 57% of the variance with all factors
loading at .85 or higher with the exception of imaginative (.77) and dull
(-.77). The ratings of the five positively correlated variables and the inverse
of the negatively correlated variable were averaged to obtain a composite
measure of "program entertainment value."  Interitem consistency was high:
Cronbach's alpha = .92.
        A second factor emerged loading highly on the variables depressing and
informative.   This factor was disregarded, however, due to a Cronbach's alpha
of only .17 for the scale combinations.
Influence on Product Perception
        Recall.  The brand recall scores and product category recall scores were each
analyzed using analysis of variance procedures for the three independent
variables program condition, commercial condition, and gender of respondent.
Brand recall was significantly better when commercials were aired during the low
humor program, F (1, 90) = 4.79, p = .031 (for low humor program, M = 1.6; for
high humor program, M = 1.16).  A difference in the ability to recall brand
names was almost significant, F (1, 90) = 3.07, p = .083, with low humor
commercials (M = 1.21) not aiding recall as well as high humor commercials (M =
1.54).  Product category recall, however, was not affected significantly by the
context of the program, but was improved instead by a higher humor level in the
commercials, F (1, 90) = 6.553, p = .012 (for low humor commercials, M = 2.38;
for high humor commercials, M = 2.82).  Gender was not a relevant factor.
        Aided recall scores for how well respondents remembered the commercial were
tested across the same independent variables used in the analysis of unaided
recall .  Once again a significant difference was found for commercial
condition, F (1, 90) = 4.89, p = .030, with the low humor commercials (M = 6.19)
not being recalled as well as high humor commercials (M = 6.91).  A two-way
interaction between the humor level of the program and gender of respondent was
also significant, F (1,90) = 4.53, p = .036 as reported in Table 1.  Simple main
effects were found across program condition within male respondents with the low
humor program resulting in better recall than the high humor program, F (1, 90)
= 2.78, p < .10, and across gender within the low humor program with men
recalling the commercials better than women, F (1, 90) = 2.78, p < .10.
        Product evaluation.  Each of the product factors derived from the factor
analysis of product evaluation ratings were tested, using analysis of variance
for the three independent variables.  The factor "negative evaluation of the
product" is reported first.  Respondents were significantly more likely, F (1,
86) = 11.31, p = .001, to evaluate a product negatively depending on the
interaction between respondent's gender and the program treatment condition (see
Table 2).  Simple main effects were found across program condition within males
with products receiving a much more negative evaluation after the low humor
program than after the high humor program, F (1, 86) = 12.09, p < .001.  Another
simple main effect for negative evaluation of the product was found across
gender within the low humor program, F (1,86) = 11.14, p < .01.  Products
received higher negative evaluations from men than women.  When means were
compared across gender within the high humor program, on the other hand,
evaluations of products approached being significantly more negative for women
than for men, F (1,86) = 3.29, p < .10.
        The "positive evaluation of product" factor was not significantly affected by
the program context or gender of respondent.  However, commercials with a low
humor level (M = 4.49) approached a significantly lower positive evaluation of
the product, F (1, 84) = 3.72, p = .057, than did commercials with a high humor
level (M = 5.04).
        A three-way interaction, significant at F (1, 90) = 8.23, p = .005, emerged
from the ANOVA for the factor "purchase intention."  The associated means are
shown in Table 3.  Simple main effects across commercial condition within males
viewing the high humor program showed that low humor commercials created
significantly lower ratings of purchase intention, F (1, 90) = 12.79, p < .001,
than the high humor commercials.  Simple main effects were also found across
commercial condition within females viewing the low humor program.  The less
humorous commercials resulted in lower ratings of purchase intention than the
more humorous commercials, F (1, 90) = 6.25, p < .05.  A simple main effect
across gender within the high humor program and high humor commercials shows
significantly higher ratings of purchase intention from men then from women, F
(1, 90) = 4.06, p < .05.  The main effect across commercial condition was also
significant, F (1, 90) = 12.44, p < .001.   Purchase intention was lower overall
for the low humor commercials (M = 2.91) than for funnier commercials (M =
4.03).
Influence on Commercial Evaluations
        The ANOVA for the dependent variable "positive evaluation of commercials,"
reported in Table 4, yielded a main effect for humor level of commercials.  Low
humor commercials were rated significantly lower (M = 4.63) on the positive
scale than were high humor commercials (M = 5.36), F (1, 88) = 4.73, p = .032.
A three-way interaction between program condition, commercial condition, and
gender of respondent approached significance at F (1, 88) = 3.78, p = .055.
Simple main effects across commercial condition within females viewing the low
humor program gave a significantly lower positive evaluation of the low humor
commercials than the high humor commercials, F (1, 88) = 4.82, p < .05.  Simple
main effects approached significance across program condition within males
viewing the low humor commercials.  Low humor programs resulted in higher
positive evaluations of the commercials than high humor programs, F (1, 88) =
3.8, p < .10.  Simple main effects also approached significance across
commercial condition within males viewing the high humor program, F (1, 88) =
3.68, p < .10.  Low humor commercials resulted in a lower positive evaluation of
the commercial than did high humor commercials.         A main effect for the
"information level of commercials" factor was found to be significant for humor
level of program, F (1, 89) = 5.48, p = .021.  Commercials aired during the low
humor program were rated to be more informative (M = 3.76) than commercials
aired during the high humor program (M = 2.98).
Influence on Program Evaluation
        An ANOVA was performed on each of four dependent variables involving program
evaluation.  The "program entertainment value" factor varied significantly, F
(1, 88) = 4.528, p = .036, as a function of the humor level of the commercials
shown during the program.  Respondents rated programs more entertaining when the
humor level of the commercials was high (for low humor commercials, M = 6.89;
for high humor commercials, M = 7.65).  Programs also approached being
significantly more funny when high humor commercials were used, F (1, 90) =
3.57, p = .062, (for low humor, M = 6.83; for high humor, M = 7.58).
        Programs were also rated significantly lower on the variables good, F (1, 90) =
8.52, p = .004, and interesting, F (1, 90) = 5.931, p = .017, when low humor
commercials were included in the program (M = 6.04 for good and M = 6.50 for
interesting), compared to programs interspersed with the high humor commercials
(M = 7.30 for good and M = 7.42 for interesting).
 
Discussion
        The results of the study confirm several of our hypotheses and fail to confirm
others.
        The first hypothesis was partially supported by the study.  As a program s
level of humor increased, the ability of viewers to recall brand names of
advertised products decreased as did the level of aided recall of commercials
for men.  Commercials aired during more humorous programs apparently have
trouble getting the attention of the viewer as easily as those aired during less
humorous programs.  Thus, humorous commercials seem to be more effective at
producing recall when aired during programs that don t contain high levels of
humor.  Recall of product category, however, was not affected by the level of
humor in the surrounding program.  Recall of product categories may be much
easier than brand names, enabling the viewer to overcome the distraction of the
more humorous program enough to remember product categories but not enough to
recall brand names.  Since many advertisers are pitting their product against
others in the same product category, recall of product category may not be
advantageous to the advertiser.  Thus, at least the more important area of brand
recall was affected by the level of humor in the program.
        The fact that increased aided recall ability was found only in men may be
explained by the types of programs used.  The difference in the level of comedy
between the stand-up comedy and the late-night talk format was much more
pronounced for men than for women as was seen by the treatment checks performed
separately on male and female respondents.  Therefore, the fact that men
achieved higher aided recall after viewing the low-humor condition than after
viewing the high-humor condition may still be the result of the perceived level
of comedy in the program.  The perceived level of humor, however, appears to be
affected by gender.
        The effect of level of humor in the surrounding program on purchase intention
did not support the first hypothesis since no significant effects were found.
The humor level of the program also had no overall effect on  positive
evaluation of the product  as predicted.  However, product appeal was affected
by context for males with the low humor programs leading to greater negative
evaluations than high humor programs, as hypothesized.  Again, women may have
responded differently because they perceived The Tonight Show as more humorous
than did men.  This would explain why men rated products more negatively than
women after viewing the low humor condition.  Women do not appear to have
perceived the stand-up comedy as significantly less humorous than did men,
however.  Instead, the crass and sexual humor often used in the stand-up comedy
condition may have been more offensive to women than men.  This would help
explain the different levels of negative evaluations of the products advertised
during the high-humor condition.
        The second hypothesis was supported.  Respondents perceived products as
something they were more likely to use and more likely to recommend to others,
the variables in "purchase intention," after viewing the more humorous
commercials.  Products also were perceived more positively when they were shown
in high humor commercials as opposed to the low humor commercials.
        High humor commercials also aided in recall of product category and seemed to
aid recall of brand more than low humor commercials.  High humor commercials
were also recalled better than low humor commercials for the aided recall test.
All of these results provide support for our assertion that increased recall
should result generally from higher humor levels in commercials as long as the
humor is related to the commercial message.
        Neither the evaluation of the product nor purchase intention were directly
affected by gender as predicted in hypothesis three.  Of more interest in this
study, however, gender did play a part in purchase intention as a function of
the stimulus material as discussed earlier.  During a stand-up comedy program,
men were much more likely to be inclined to use a product if the commercials
were of the more humorous nature.  During the more casual talk program, however,
women were more likely to be inclined to use and recommend a product after
viewing the more humorous commercials. Thus, the specific type of humorous
program seems to be important in its effect on purchase intention according to
gender.  This does not seem to be a function solely of the level of humor but
also may be explained by the types of humor used in each genre of program.  This
study, however, did not measure differences resulting from specific types of
humor.  Therefore, hypothesis three is not supported, but some interesting
findings are connected with gender nevertheless.
        The "positive evaluation of commercial" factor was used to examine the fourth
hypothesis.  No main effect on perception of commercial was found across program
condition as was predicted.  Thus, the hypothesis was not supported.  Since men
perceived a greater difference in the humor level of the programs, however,
there is some evidence that the benefits of a low humor commercial are
negatively affected by an increased level of humor in the surrounding context.
If the level of humor in the commercial is much below that of the program the
commercial actually seems to be rated less positively than when the level of
humor in the program is lower, closer to the level of humor in the commercial.
This is contrary to the hypothesis and indicates that as the level of humor in
the program increases, at some point it might produces a negative effect on the
evaluation of mildly humorous commercials aired within the program.  Further
study would be needed to clarify this point.
        The fifth hypothesis was strongly supported by several tests.  This is
surprising perhaps since most researchers have been interested in effects of
programs on commercials, results of interest to advertisers, and not in effects
of commercials on programs, results of interest to programmers or station
managers.  Ratings of the programs  entertainment value, program funniness, how
interesting, and how good the programs were were all rated higher when more
humorous commercials were used than when less humorous commercials were used.
This is strong support for the use of humorous commercials to boost program
enjoyment levels and thereby, one would assume, program ratings.
Future Research
        It seems that certain types of humor affect the genders differently.  The
stand-up comedy program is much less relational in nature and much more of a
competition to see who can get the most laughs than is the talk format.  It also
tends to use more brutal forms of comedy and much more overt sexual humor than
the talk format.  These differences appear to affect men and women differently.
More study should be done to determine precisely which elements in humor appeal
to women and which appeal to men.
        Future study of context effects on advertising could look at the context
effects of different levels of tragedy or seriousness in programs.  Also,
different types of comedy such as the situation comedy should be studied.  Types
of commercials other than those that are humorous could be manipulated in
similar studies to develop a broader understanding of the effects of context on
the commercials.  Also, the effect of commercials on the success of programs
should be looked at.  If syndicators or networks allow any paying commercial to
be aired, what might that do to a program s ratings?
Conclusion
        Results here indicate that programmers should be selective of the types of
commercials they allow to air during specific programs in order to improve the
programs  ratings.  By being more selective up front about commercials, ratings
may improve allowing programmers to charge more for advertising later on.  This
finding may also be significant for the local station that airs locally produced
commercials.  It may benefit the station to provide higher quality production
facilities to local advertisers in order to maintain or improve ratings on
programs with local availabilities.
        Overall, the more humorous a commercial is, the greater the benefit from both a
product standpoint and a programmer's standpoint.  The more humorous a program
is, however, the more dangerous it may be for advertisers to include their
humorous commercials in the program since it seems the commercials must be
funnier to be effective.  Specific types of comedy programs may also be more
appropriate for advertising to men or to women specifically because of the
appeal of the types of humor used in the program.
 Table 1
 
Aided Recall of Commercials by Program Condition and Gender of Respondent
 
--------------------------------------------------------------------------------
----------------------------
Program condition
Respondents'
--------------------------------------------------------------------------------
gender                  Low humor                       High humor
--------------------------------------------------------------------------------
----------------------------
Male                        7.29a/A                        6.17b/A
Female                      6.32a/B                        6.53a/A
--------------------------------------------------------------------------------
----------------------------
 
Note.  Ratings could range from zero to 10 (averaged across four commercials
within the specified program condition).  Lower-case superscripts describe
comparisons across program condition (horizontal) within gender.  Upper-case
superscripts describe comparisons across gender (vertical) within program
condition.  Mean scores not sharing superscripts in the specified comparisons
differ at p < 0.10 by F test.
 
 
 
Table 2
Negative Evaluation of the Product by Gender of Respondent and Program Condition
 
--------------------------------------------------------------------------------
----------------------------
                                                  Respondents'  gender
Program condition
-------------------------------------------------
                                        Male                    Female
--------------------------------------------------------------------------------
----------------------------
Low humor                               3.42a/A                 2.18b/A
High humor                              1.96b/B                 2.65a/A
--------------------------------------------------------------------------------
----------------------------
 
Note. Ratings could range from zero to 10 (average of two max 10 scores,
averaged across four commercials within that condition).  Lower-case
superscripts describe comparisons across gender (horizontal) within humor
condition.  Upper case superscripts describe comparisons across humor condition
( vertical) within gender.  Mean scores not sharing upper-case superscript
letters in the specified comparisons differ at p < 0.001 by F test.  Mean scores
not sharing lower-case supercript letters in the specified comparisons in the
low humor condition differ at p < 0.01.  Mean scores not sharing lower-case
supercript letters in the specified comparisons in the high humor condition
differ at p < 0.10 by F test.
 Table 3
 
Purchase intentions by Program Condition, Commercial Condition and Gender of
Respondent
--------------------------------------------------------------------------------
----------------------------
                                                Commercial condition
                                        -------------------------------------------------
Program         Respondents'
condition       gender                     Low humor            High humor
--------------------------------------------------------------------------------
----------------------------
Low humor       Male                          3.44a/A               3.60a/A
                Female                        2.71b/A               4.28a/A
                ------------------------------------------------------------------------------
---------
High humor      Male                          1.75b/A               4.75a/A
                Female                        3.10a/A               3.42a/B
--------------------------------------------------------------------------------
----------------------------
 
Note.  Ratings could range from zero to 10 (average of two max 10 scores,
averaged across the four commercials within the specified program condition.)
Lower-case superscripts describe comparisons across commercial condition
(horizontal) within program condition and within respondent's gender.
Upper-case superscripts describe comparisons across only respondent's gender
(vertical) within commercial condition and within program condition.  The
upper-case letters do not signify differences across program condition (across
the dotted line).  No significant simple effects differences were found between
program conditions.  Mean scores not sharing supercript letters in the specified
comparisons differ at p < 0.05 by F test.
Table 4
 
Positive Evaluation of Commercial by Commercial Condition, Program Condition and
Gender of Respondent
 
--------------------------------------------------------------------------------
---------------------------
                                                    Commercial condition
                                                ----------------------------------------------------------
Respondents     Program
gender          condition               Low humor               High humor
--------------------------------------------------------------------------------
----------------------------
Male            Low Humor                  5.48a/A                  4.81a/A
                High Humor                 3.79a/B                  5.32b/A
                ------------------------------------------------------------------------------
---------
Female          Low Humor                 4.55b/A                   5.83a/A
                High Humor                 4.43a/A                  5.26a/A
--------------------------------------------------------------------------------
----------------------------
 
Note.  Ratings could range from zero to 10 (average of 6 max 10 scores, averaged
across the four commercials within the specified program condition.)  Lower-case
superscripts describe comparisons across commercial condition (horizontal)
within respondent's gender and within program condition.  Upper-case
superscripts describe comparisons across only program condition (vertical)
within commercial condition and within respondent's gender.  The upper-case
letters do not signify differences across respondent s gender (across the dotted
line).  No significant differences across respondent's gender were found.  Mean
scores not sharing underlined superscript letters in the specified comparisons
differ at p < 0.10 by F test.  Mean scores not sharing plain text superscript
letters in the specified comparisons differ at p < 0.05 by F test.
 References
Broach, V. C., Page, T. J., & Wilson, R. D. (1991).  The context effect of
emotional response to TV programs on emotional response to commercials:  A
cognition-arousal perspective.  In R. Holman (Ed.), Proceedings of the 1991
Conference of the American Academy of Advertising (pp.88-90).  New York:  D'Arcy
Masius, Benton & Bowles, Inc.
Cantor, J. R., & Venus, P.  (1980).  The effect of humor on recall of a radio
advertisement.  Journal of Broadcasting, 24(1), 13-22.
Goldberg, M. E., & Gorn, G. J.  (1987).  Happy and sad TV programs:  How they
affect reactions to commercials.  Journal of Consumer Research, 14(3), 387-403.
Kamins, M. A., Marks, L. J., & Skinner, D. (1991).  Television commercial
evaluation in the context of program induced mood:  Congruency versus
consistency effects.  Journal of Advertising, 20(2), 1-14.
Lammers, H. B., Leibowitz, L., Seymour, G. E., & Hennessey, J. E.  (1983).
Humor and cognitive responses to advertising stimuli:  A trace consolidation
approach.  Journal of Business Research, 11, 173-185.
Madden, T. J., & Weinberger, M. G.  (1982).  The effects of humor on attention
in magazine advertising.  Journal of Advertising, 11(3), 8-14.
Mandese, J.  (1993, September 6).  'Home Improvement' wins $ race.  Advertising
Age, pp. 3, 26.
Mattes, J., & Cantor J. R.  (1982).  Enhancing responses to television
advertisements via the transfer of residual arousal from prior programming.
Journal of Broadcasting, 26(2), 553-566.
Mundorf, N., Zillmann, D. & Drew, D.  (1991)  Effects of disturbing televised
events on the acquisition of information from subsequently presented
commercials.  Journal of Advertising, 20(1), 46-53.
Murphy, J. H., Cunningham, I. C., & Wilcox, G.  (1979).  The impact of program
environment on recall of humorous television commercials.  Journal of
Advertising Research, 8(2), 17-21.
Murry, J. P., Lastovicka, J. L., & Singh, S. N.  (1992).  Feeling and liking
responses to television programs:  An examination of two explanations for
media-context effects.  Journal of Consumer Research, 18(4), 441-451.
Pavelchak, M. A., Antil, J. H., & Munch, J. M.  (1988).  The Super Bowl:  An
investigation into the relationship among program context, emotional experience,
and ad recall.
Singh, S. N., & Churchill, G. A.  (1987).  Arousal and advertising
effectiveness.  Journal of Advertising, 16(1), 4-10.
Schumann, D. W., & Thorson, E.  (1990).  The influence of viewing context on
commercial effectiveness:  A selection-processing model.  Current Issues and
Research in Advertising, 12(1/2), 1-24.
Weinberger, M. G. & Gulas, C. S.  (1992).  The impact of humor in advertising:
A review.  Journal of Advertising, 21(4), 35-59.
Zillmann, D., & Bryant, J.  (1991).  Responding to comedy:  The sense and
nonsense in humor.  In J. Bryant & D. Zillmann (Eds.), Responding to the screen:
Reception and reaction processes (pp. 261-279).  Hillsdale, NJ:  Erlbaum.
 
Context Effects of Humor  page  11
 
 
 
 v
 
 
            a   E   G   J   '   )   1   3           5           >   ?   X   Y
j   k   6   7   9   :   +   -   3   5   7   C   ^   i
  6&  \&   1  &1   2  {wsokgc_[WSOKGC=93/)%               }yuqmiea
 
                   |                   |                   |                   |
                   |
                                                             % 2  $2   4  /4  q6
  {6  -8  ?8  88  C8  d8  i8   9   9  o:  #:  y;   ;   ;   ;   =   >   >   >  }?
   ?  H@  Z@  (B  =B  `C  bC  kC  mC   C   D   D  'D  (D  {wsokgc_[WSOKGC?;73/+'
#         {wsmic
 
 
 
 
 
                                          &(D  1D  2D  YD  ZD  SE  qE   E   E  $
E  %E  QE  RE  rE  sE   E   F  XF  YF  yF  zF  1G  2G  RG  SG  xG  yG   H   H  :
H  ;H  IH  gH  iH  jH  {uqkgc_YUOKEA;73/)%       }wsmiea[
 
 
 
 
 
 
 
 
 
 
 
 
 
 
          "jH  ,H  -H  ;H  <H  `H  aH   I   I  -J  .J  <J  =J  EJ  FJ  sJ  tJ  {
J  |J   K
K   K   K  @K  AK  rK   K   K   K  `N  eN   P   P   Q   Q  {uqkga]WSMIC?95/+%!
     {wsmic_Y
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
          " Q   Q   Q  ~S   T  :U  ?U  GU  LU  vU  }U  vX  {X   Y   Y  DY  eY  f
Y  lY  +[  ,[  =[  >[  ^[  _[   [   [  U[  V[  g[  h[
\   \  L\  N\  {wsmic_YUQMGC=951-)#       {wqmgc]
 
 
 
 
 
 
 
 
 
 
 
 
 
                          "N\  }\  ~\   ]   ]  5]  6]  []  \]  v^  w^   ^       ^  2
^  3^  r^  s^  j_  k_  {_  |_  @`  A`  R`  S`  @`  A`  R`  S`  ]`  o`   b   b
b   b  {uqkga]WSMIC?95/+%!      }ysokga]W
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
          " b  rc  sc   c   c   d       d   d   d   e   e  /e  0e  ce  de  2f  3f  D
f  Ef  |f  }f  2f  3f  Df  Ef  pg  qg   h   h   i   i  "i  #i  \i  {uqkga]WSMIC?
95/+%!      }ysoie_[
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
          !\i  ]i  ni  oi  :j  ;j  Mj  Nj   j   j  Ej  Fj  Qj  vj  Dk  Ek   l
l   l   l  #l  $l  +l  ,l  =l  >l   m   m  $m  %m   n   n   n   n  !o  zvplfb\XR
NHD@<62,("       zvplfb\X
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
"!o  "o  3o  4o  (p  )p  :p  ;p   p   p  Wp  Xp  cp   q   q   q  1q  2q   r   r
 *r  +r   s   s  's  (s  As  Bs  [s  \s  .s  /s  @s  As  [s  zvplfb\XRNHD@<62,("
       zvplfb\X
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
"[s  \s  ns  os  1t  2t  Gt  Ht  %t  &t  ;t  <t  Yt  et  0~  @~      %   u   w
     O   X   %   .   1   ;   >   S   V   `   c   W   [   ^   _       T   k   n
 u   zvplfb\XRNHD@<840,&"            z t n h
 
 
 
 
 
 
 
 
 
 
          (u   x                                   1   2   ?   @   C   D   d   =
  Q"  T"  ^"  a"  z"  }"   "   "   "  g"   #   #   #   #  -#  0#  :#  =#  1#  5#
   %   %  1%  2%  w&  x&   &   &  3(  5(  y sm ga [ U O I C = 7 1 + * $
      ~ x
 
 
 
 
 
 
 
 
 
 
                                                              /5(  6(  @(  C(  \
(  ](  ^(  _(  i(  j(  l(  r(  I)  e)  h)  r)  u)   )   )   )
)   *   *  o+  t+   +   +   +   +  h+  l+  W,  X,  J-  K-  _-  6.   .   /  3/  5
/  7/  4/  P/  R/  T/  y s mga [U T N H B < 6 5 4 3 2 , & "      ~zt
 
 
 
 
 
 
 
 
 
 
  -T/   0   0  !0  #0  S1  o1  q1  s1  h1  ~1   2   2  R2  a2   3   3   3   3  N
3  d3  f3  h3   4   4
4  !4  ]5  y5  {5  }5   6  #6  %6  {uqkga]WSMIC?95/+%!      }ysoie_[
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
          !%6  '6  :7  d7  f7  h7  Q7  g7  i7  k7  y8  48  _8  a8  z8  {8   9
9   9   9   9
9   9  zvplfb\XRNHD  B @ > <
 
 
                                                      B
 
 
 
 
 
 
             <   a   c   e   g   y       *   ,   >   Y   [       (   ?   Y   [
 ]   _   a   c           -   D   P   k               G   I       '   )   3
     (
      5                   =   W   i   5   8   )   +   -   5   7   A   C   M
  ^   i
  yyyyyssssssssssssssllllffffffff_______YYQKKKKKKKKBBBBB9 3   +++%
 `           P  `                           d} `
  0} `           `           P  `            `             `
     `                         ` @i
  {#  /%  6&  \&  P)  i*  \-   1  &1  }3  ?5  q6  {6  -8  o:  %=   >  }?  H@  (B
  `C  bC  kC  mC   C   F   I  pK  rK   K  nN   P   Q   T  HU  uU   Y  DY  eY   ]
  \`  9e        f  Qj  vj  cp   q  is  Wt  Yt  et  >u  4y  4|  B   5   k       i
  xxxrjjjjc[[[UMMMMMMMMG?88222 ,&&&&&&&&&&&&&&&&&&&                      `
                   `                       `
                 P  `            `           P  `              `           P  `
           `           P  `  <    %   #   w       M   N   X   Z   %   '       (
 
   1       B   g   U   W   u   w   y   {       V   X   F   k   W   l   Z   |
           X   b   d   =
  +!  G!   !   !  H!  6"  e"   "  i"   #  A#  /#  1#   &   &   &   &   '  n'   '
  h'   (  ,(   (  G(  p(  K)  z)  $)  xqkkkk    e^XXXXXXXXXXXXRLLLLLLLLLL FFF@::
:::::::::: 4..............             p           `           p           `
       p           p           `           p             p           p
 `             `               ` D$)   *   *  _-  l-  >.  D/  c/  /0   1   2  o2
  -3  u3  .4   5  :6  36  v7  x7  _8  a8   9   9
9   9  zz sjjjjjjjjjjjjjjjjd^^ X
 
 
 
                        `   s                   `              d} `
`           p                                        F           :  r          `
 @!   s          ` @!                        r s                          ^
7     9      '   )        9  ]      9  k =
  Tms Rmn       ` Symbol    Helv
0 Courier    Times New
Roman          "    7   !    "         P         " c  Cc vJc       ,   i4
 ]   'humor appendix study discussion section    Huiuk Yi$Institute for
Communication
Research