|
page
page
Complexity and Blame Focus in Anti-Smoking Television Commercials:
The Role of Complexity and Individual vs. Industry Blame
on Smokers and Non-Smokers
One of the primary methods used to battle smoking in the United States has
been, and continues to be, televised anti-smoking commercials. Because of the
current tobacco settlements, states now have access to more funds than ever
before to produce televised anti-smoking campaigns to (1) foster the elimination
of tobacco use and (2) to denormalize smoking ("Master Settlement"). Current
state campaigns, however, have produced anti-smoking ads that use dramatically
different and diverse strategies. Indeed, at present, there is considerable
disagreement about which characteristics make anti-smoking ads more effective
from a public health perspective. The common belief is that to effectively
discourage smoking, these televised anti-smoking ads (a) must be
attention-getting, provocative, engaging, memorable, and (b) must illustrate
convincing reasons not to smoke. Indeed, considerations of how ads look
(structure), what they say (content), and the interactions between structure and
content are critical to the development of successful anti-smoking ads.
Like all ads, anti-smoking ads vary in their structure - it is the combination
of shot framing, editing, pacing, music, and the use of color or black-and-white
that create the "look and feel" of an ad. Understanding the effects of
production features in anti-smoking ads, although a new avenue of research, is
critical to ad development because these attributes have been shown to influence
attention, arousal, and evaluation of ads for commercial products (Yoon, Bolls &
Lang, 1998) and for political candidates (Geiger & Reeves, 1991).
In the realm of content, Dorfman and Wallack (1993) have developed a
distinction between traditional individual blame anti-smoking ads and what they
refer to as "counter-advertising." Counter-advertising is defined as attempts
to contextualize health problems and connect them to current social and
political conditions. Such advertising, then, shifts the blame from the
individual to society or institutions, and as such can be considered
structurally-focused industry blame anti-smoking ads. Empirical research on
this distinction, however, is contradictory with regard to information that
could guide decisions about effective message design (Antecol, 1998; Goldman &
Glantz, 1998, Pechmann, 1998; see also Azar, 1999).
The proposed study, then, addresses two important research questions: (1) how
does the structure and (2) the blame focus of anti-smoking ads affect
ad-specific responses (e.g., attention, arousal, evaluation, and importance) of
smokers and non-smokers. Structure will be examined by varying a construct
called "structural complexity." Structural complexity is defined in this study
as an ad's global complexity scores. Comparisons will be made on dependent
measures obtained after participants view ads that are high, medium or low in
terms of their complexity in the first experiment and ads that are high and low
in terms of their complexity in the second experiment. Blame focus will be
examined by comparing ads that focus on the behavior of individual smokers
(e.g., individual blame: by illustrating negative consequences of smoking and
second-hand smoke) and ads that focus on the behavior of the tobacco industry
(e.g., industry blame: by highlighting the deception and marketing tactics of
tobacco manufacturers).
Because smoking among 18-to-24 year olds remains high (Johnston, O'Malley, &
Bachman, 1992), this group was chosen as the participant population.
LITERATURE REVIEW
Ad complexity
Apart from content (i.e., whether an ad has an individual or industry blame
focus), one important attribute that may influence the effectiveness of an ad is
the complexity of its production features (e.g., Geiger & Reeves, 1993; Hockberg
& Brooks, 1978; Reeves, Thorson & Schleuder, 1986; Thorson, Reeves & Schleuder,
1985). Production features are characteristics of a message that are separate
from the message's factual content; the more of these features in a message, the
greater its complexity (Reeves, Thorson & Schleuder, 1986). Production features
include such factors as pacing, facial closeups, camera movement, music, and
multiple camera angles. In particular:
y Pacing concerns the number of cuts or edits that take place within a message.
Edits occur when one message scene shifts to another. These edits can be
between related content or non-related content. High-paced ads can often be
found in the new breed of music videos that run on MTV.
y Facial closeups occur when the face of a person within a message fills the
majority of the frame. Such closeups, for example, are often used during
interviews on 60 Minutes.
y Camera movement is defined by both camera pans and camera zooms. A pan occurs
when the shot moves across the scene such as from left to right or from top to
bottom; a zoom occurs when the camera closes in on a specific element within the
frame.
y Music here is more than simply its presence or absence. Instead, it is the
presence of music in the foreground rather than the background of the message.
In other words, the music must be designed to be a vital part of the message,
not merely a background element.
y Multiple camera angles occur when the camera shoots a scene from a variety of
different perspectives such as looking down from the top and then up from the
bottom. For example, in the "Rain" PSA (California State Department of Health),
the viewer can see cigarettes falling from the sky from the vantage point of the
ground looking up and from the vantage point of the sky looking down.
Complexity hypotheses. The importance of considering the production features
found within an ad is twofold. First, for viewers to process a message they
must attend to it. Production features, such as those listed above, lead to
orienting responses (ORs), which have been described as "involuntary
physiological 'calls' for attention and processing" (Lang, 1991, p.223; see also
Geiger & Reeves, 1993; Lang, Geiger, Strickwerda & Sumner, 1993). Attention to
television is highest immediately following an OR and then diminishes over time
(Hochberg & Brooks, 1978; Thorson & Lang, 1992). Thus, complex ads elicit more
attention over time because they produce more ORs than structurally simple ads.
This attention translates into better memory for the message (Lang, Bolls,
Potter & Kawahara, in press). Put another way, ad-specific responses such as
attention and cognitive processing are heightened by these features because
increased complexity leads to increased physiologically stimulation and/or
arousal (Thorson, 1990). Although, no study has applied this research to
anti-smoking ads, the research suggests that increasing levels of complexity in
anti-smoking ads will lead to increased attention and physiological arousal in
viewers. Thus,
H1: Participants will exhibit higher levels of autonomic attention, as measured
by heart rate responses, to anti-smoking ads that are more complex.
H2: Participants will exhibit higher levels of autonomic arousal, as measured by
skin conductance responses, to anti-smoking ads that are more complex.
Because participants may or may not be aware of their physiological state, it is
not known whether these responses can be replicated on self-report measures.
Accordingly, two research questions are asked,
RQ1a: How will ad complexity affect the participants' self-reported levels of
attention and arousal?
RQ1b: Will participants' self-reported levels of attention and arousal be
consistent with their autonomic responses?
Only one study has purposefully manipulated the complexity of production
features to understand the effectiveness of televised PSAs on self-reported
ad-specific outcomes. In Jackowitz, Schooler and Flora (1997), the authors
found that PSAs with complex production features were rated higher than ads with
simple features. In particular, complex ads were rated as more informative,
more believable, and more motivating than simple ads. It is possible, then, to
hypothesize the following:
H3: Participants will rate complex ads as more likeable and more important than
simple ads.
Blame Focus of anti-smoking ads
Individual blame anti-smoking ads. Traditionally, the focus of anti-smoking
ads has been on the individual smoker. In other words, "if you smoke, it is
your fault" or "if you cannot stop, it is your fault." These "individual blame"
ads generally focus on the negative aspects of smoking, often rely on emotions
such as fear and humor, and sometimes give advice on how to quit smoking
(Pechmann, 1997). Because these ads are assumed to have direct effects, they
are usually targeted toward current smokers or youth who may be tempted to
smoke. One example of an individual blame ads is Thelon which was produced for
Health Canada. It opens with a shot of Thelon, a young man, leaning on an
outside wall. He talks directly into the camera. Periodically, the camera
zooms into his face. He says:
I started smoking when I was 8 and a half. First, I stole them from my dad. I
then I started smoking heavily up to about two packs a day by the time I was 15.
I started to get a pain in my chest. And the doctor only told me a few months
ago that it looks like it could be lung cancer. I'm very afraid. I don't want
to die. I'm not even 20 yet.
The ad ends with a purposely blurry shot of him lighting a cigarette. There is
an overlay on top of this: "We held an anti-smoking contest. Thelon won by
telling us what he lost." The desired outcomes from viewing individual blame
ads such as this one is to induce behavioral change in those who smoke, and
attitude change in those who may be tempted to smoke. From a public health
perspective, using such ads represents a status quo approach that lifts
responsibility for smoking from the tobacco companies and places it on the
individual (Dorfman & Wallack, 1993).
Industry blame anti-smoking ads. Increasingly, many states are attempting to
produce anti-smoking ads that shift the focus away from the individual and onto
the structural context in which smoking occurs (Cummings & Clarke, 1998). By
contextualizing smoking within current social and political conditions, these
"industry blame" ads challenge and delegitamize the institutions behind the
product. According to Dorfman and Wallack (1993, p.720), they "attempt to set
the agenda for an environmental perspective, conferring status on
policy-oriented strategies for addressing health problems. The primary purpose
of such ads is to challenge the dominant view that health problems reflect
personal health habits." Importantly, rather than being targeted only to
current smokers or at-risk youth, these ads also target non-smokers, in that
their purpose goes beyond smoking cessation and prevention to include
anti-smoking norm alteration.
One example of an industry blame ads is the 30-second spot titled Boardroom
(California State Department of Health). Boardroom opens in a dark,
smoke-filled room with men in dark suits listening to the chairman of the board.
He says:
Gentlemen, gentlemen. The tobacco industry has a serious multi-billion dollar
problem. We need more cigarette smokers, pure and simple. Every day 2,000
Americans stop smoking. And another 1,100 also quit. Actually, technically,
they die. That means that this business needs 3,000 fresh new volunteers every
day. So, forget about all that cancer, heart disease, emphysema, stroke stuff.
Gentlemen, we're not in this business for our health. (followed by loud,
cackling laughter)
The main expected outcomes of viewing such ads include creating strong
anti-smoking attitudes, norms, and intentions. In addition, there is some
overlap between the outcomes of viewing these anti-smoking ads and their
traditional individual blame counterparts, that is, both attempt to modify
behaviors of smokers and those tempted to smoke. However, because they are
perceived to be effective tools countering the tobacco marketing influences
(Dorfman & Wallack, 1993; Goldman & Glantz, 1998; Teinowitz, 1998), industry
blame ads have become an integral to anti-smoking campaigns in Arizona,
California, Florida, Massachusetts, Minnesota, and Oregon (Cummings & Clarke,
1998).
Effectiveness of individual and industry blame anti-smoking ads. There are
contradictory findings concerning the effectiveness of these two ad types. In a
secondary review of focus group findings, Goldman and Glantz (1998) found that
in terms of general effectiveness, for both youth and adults, industry blame
ads, compared to individual blame ads, achieved more desirable outcomes. The
two ad types, however, were effective for different reasons: adults found them
effective in overcoming guilt about not being able to quit, while youth found
them effective because they demonstrated that smokers were not "independent" but
were being manipulated. These results however, as recognized by the authors
themselves, are difficult to generalize because of the limitations inherent in
focus group research, namely, small sample size and group dynamics.
In contrast, Pechmann (1998), using an experimental methodology, found that
industry blame ads were not as effective in terms of ad-specific responses such
as cognitive and emotional involvement, or in terms of health-relevant responses
such as perceived importance of health risks, attitude toward smoking and
intentions to smoke in the upcoming year (see Azar, 1999). However, although
her involvement measures appear ad-specific in nature, it is unclear from the
report whether these responses were measured after viewing individual ads or
after viewing all the ads. If they were measured after viewing all the ads,
then they would be more akin to her health-relevant outcomes. Further, these
findings, however, may be limited to her participant population, namely 7th and
10th grade students from the very conservative and rather affluent Orange
County, California.
The only study that clearly used ad-specific measures that were measured after
the viewing of each ad was Antecol (1998). In his experimental dissertation
research, he found that, in terms of immediate impact, individual blame ads were
more attention-getting, arousing, interesting, involving and memorable than
industry blame ads. They also led to a more socially-positive attitude
regarding smoking. This was true for non-smokers, moderate smokers and regular
smokers. Although Antecol's (1998) results may be restricted to his participant
populations, namely undergraduate students (aged 18 to 24) from Kansas,
Missouri, and Indiana, the same populations are used here. Nevertheless, the
inconsistencies in the results of the above three studies, as well as the
inclusion of the complexity factor that will account for variance not previously
accounted for, makes it problematic to specify an hypothesis regarding ad type.
Accordingly, the following research question is asked:
RQ2: How will ad type affect the participants' ad-specific responses such as
attention, arousal, evaluation, and importance?
Interactions between complexity ad and blame
It is possible that complexity may interact with blame focus. Indeed, Geiger
and Reeves (1991, p.158) argued that "the influence of structure on evaluation
indicates that television viewers treat structure and content as interdependent
dimensions, relying on visual structure and content for the attribution of
meaning." Previous research related to political communication suggests that
interactions between complexity and content do, in fact, occur. For example,
one study found an interaction between content and complexity such that
political candidates were evaluated more positively when they were presented in
ads with complex production features rather than in ads with simple features
(Geiger & Reeves, 1991). The particular importance of this latter finding is
that the processing of political ads has been found to be similar to the
processing of other types of ads, including anti-smoking ads (Thorson, Christ &
Caywood, 1991).
Although these latter findings suggest that interactions will be seen between
an anti-smoking ad's complexity and blame focus, there is no direct research to
suggest exactly how these two factors will interact. It is not known, for
example, whether complexity will interact with blame focus to increase or
decrease an ad's effectiveness. This leads to the following the following:
RQ3: How will an ad's blame focus interact with its complexity level?
Smoking status
It is clear that smokers differ depending on the amount they smoke. Those who
smoke regularly differ from those who smoke on a more casual basis. Likewise,
both groups differ from those who do not smoke at all (Grube, McGee & Morgan,
1986). In this study, we differentiate between those who have smoked at least
one cigarette in the last month from those who have not smoked at all. These
two groups are labeled here as "smokers," and "non-smokers." Although we expect
smoking status to play a role, because of the lack of research on the questions
asked in this study, we offer no hypotheses or research questions based on
smoking status. Instead, we will report those instances where smoking status
had a significant effect.
EXPERIMENT 1
General design. This experiment is designed to test the main effect hypotheses
of complexity and to answer Research Question 1. It was designed as a mixed 2
(smoker: smoker vs. non-smoker) x 3 (complexity: high vs. medium vs. low) x 2
(message replication: 2 for high complexity, 2 for medium complexity, 2 for low
complexity) analysis of variance (ANOVA). Smoking was a between-subjects
variable. Complexity and message replication were within-subjects variables.
Participants (N = 65) representing the 18-24 year-old demographic were recruited
from journalism classes at a large Midwestern university.
Procedure. The experimenter arrived at least 15 minutes prior to the
participant to set up the lab and to run a series of necessary checks on the
data collection equipment to ensure the participants' safety. The experimenter
greeted the participants upon their arrival and explained the purpose of the
study and that skin conductance and heart rate would be recorded using small
sensors while viewing 14 video-taped ads. After obtaining informed consent, the
participant was seated in a comfortable chair about five feet from a 19 inch
television set. The participant's forearms and hands then were washed with
distilled water to control hydration and Beckman standard AG/AGCL electrodes
were applied to those areas. They participated individually.
The experimenter stressed that because of the sensitivity of the recording
instruments, it was imperative that the participant remain as still as possible
during the experiments. The experimenter then started the stimulus tape and the
participant viewed 14 ads; six of these ads were the actual commercials of
interest while the rest were filler commercials. The filler commercials
consisted of ads for general household products and were placed around the
anti-smoking ads so that participants saw commercials other than the ones under
study. To ensure there were no order effects, the order of the six anti-smoking
commercials was counterbalanced; six orders were constructed and the
participants were randomly assigned to one of these. The position of the filler
commercials remained constant. While viewing each ad, heart rate and skin
conductance data were collected. After viewing each ad, the stimulus tape was
stopped and the participant completed the pencil-and-paper questions regarding
the ad. Autonomic responses were not collected at this time. After viewing all
the ads, the participants completed the remaining self-report measures.
Apparatus. The videocassette recorder, experimenter and physiological
equipment were separated from the participant by an eight-foot wooded wall. The
lab was controlled by a 386 computer with a LabMaster AD/DA board installed.
Coulbourne physiological equipment was used in the collection of data. Heart
rate data were collected through three electrodes placed on the participants'
forearms. It was measured as the milliseconds between heart beats and was later
converted to heart rate per second for statistical analyses. Skin conductance
data were collected as an analog signal sampling 10 times per second through the
participant's non-dominant hand. Score XY, a subprogram of VPMANLOG (Cook,
1995), was used to score the spontaneous SCRs during each ad.
Independent variables. Complexity. The overall complexity of an ad can be
considered in both the visual and auditory modalities. There are several
operational definitions of visual complexity. Basil (1994) defined it as the
number of shots in each 30-second scene. In Thorson, Reeves and Schleuder
(1985) global complexity was coded as messages that contained many edits, scene
changes, zooms, pans, person or object movements. The ads chosen as exemplars
based on this scheme were then pretested by participants who rated each ad's
complexity on a 100-point scale after viewing "anchor" messages to get an idea
as to what was complex or not (see also Reeves & Thorson, 1986). The most
extreme ads were then used in their experiment. Lang (1990) and Schleuder
(1990) both relied on a global measure where the number of cuts, edits, zooms,
pans, and movement were used to rate an ad's visual complexity. Finally,
Hitchon, Duckler and Thorson (1994) measured complexity in a pretest using ads
that the authors created to meet high and low complexity criteria. Pretest
participants assessed complexity by four semantic differential scales: simple/
complex, many scenes/few scenes, slow moving/fast moving, lots of things
happening/few things happening.
Here, complexity is operationalized with a global measure consistent with
Thorson, Reeves and Schleuder (1985), Lang (1990), and Schleuder (1990), that
is, the number of edits/cuts and pans/zooms were counted. In addition, the
presence of multiple camera angles (yes/no) was coded. Further, the presence of
music has been found to lead to a significant physiological impact (Ries, 1969;
Zimny & Weidenfeller, 1963; Potter, Lang, & Bolls, 1998) that can tax a person's
cognitive capacity (Wheatley & Brooker, 1994) as well as lead to belief-based
change (Middlestadt, Fishbein, & Chan, 1994) regardless of whether a person
likes the music (Wheatley & Brooker, 1994). Thus, the presence of foreground
music was coded (yes/no). Finally, the number of facial closeups were counted.
Close-ups have been identified with paying more attention to television
commercials (Johnson-Cartee & Copeland, 1997) and have also been shown to be
remembered better and more strongly responded to (Reeves, Lombard & Melwani,
1992). The total score was used to create an index of complexity.
Table 1
Production Feature Complexity in Six Anti-Smoking Ads
Low Complexity
Medium Complexity
High Complexity
Boardroom
Aging
Soundbites
Thelon
Own Words
Dance
Edits/Cuts
4
0
9
11
14
19
Pans & Zooms
1
1
1
0
2
3
Multiple Camera Angles
No (0)
No (0)
Yes (1)
Yes (1)
Yes (1)
Yes (1)
Foreground Music
No (0)
No (0)
Yes (1)
No (0)
No (0)
Yes (1)
Facial Closeups
1
1
1
3
2
2
Complexity Points
6
2
13
15
20
24
Table 1 illustrates how the production feature complexity of the six
anti-smoking ads used were classified according to this coding scheme. In a
previous study (Zhou, et al., 1997) that differentiated message pacing, 0-7
edits was considered slow, 8-15 medium, 16 to 23 fast and 24 or more is very
fast. A similar approach is used here to differentiate levels of complexity.
However, since more than just edits in the ad are being counted, the bar has
been slightly raised. Thus, an ad with under 10 "complexity points" has been
categorized as having low complexity, an ad with 10 to 20 points has medium
complexity, while an ad with more than 20 points high complexity.
Smoking. Participants were asked indicate how often in the past month they had
smoked a cigarette. Participants who did not smoke were classified as
non-smokers, those who smoked 1 day a month, 1 to 3 days a month, 1 to 2 days a
week, 3 to 4 days a week, 5 to 6 days a week, or daily were smokers (based on
Grube, McGee & Morgan, 1986). Also, to get a better grasp on the type and
extent of the smoking behavior, several questions were asked for descriptive
purposes: "How many cigarettes did you smoke, on average, each day during the
past month," "In the eleven (11) months previous to last month, on average, how
often did you smoke cigarettes," "When you did smoke in the eleven (11) months
previous to last month, how many cigarettes did you smoke, on average, each day"
(Grube, McGee & Morgan, 1986). Participants were also the length of time that
they had smoked cigarettes.
Dependent variables. During each ad. Physiological attention to each ad was
operationalized by measuring the participants' heart rate during the duration of
each 30 second ad. The allocation of attentional resources to external stimuli,
such as a television commercial, results in the activation of the
parasympathetic nervous system (Graham, 1979). This activation can be indexed
by cardiac deceleration as measured by heart rate data (Lang, 1994).
Accordingly, such data provide a reliable and valid measure of physiological
attention. Physiological arousal during each ad was operationalized by the
participants' skin conductance responses (SCR) which is a reliable and valid
measure of arousal in response to media messages (Hopkins & Fletcher, 1994;
Martin & Venables, 1983). There are three ways SCR can be measured: as a
conductance level response (e.g., a change from a resting baseline level),
through its amplitude, and through its frequency. Frequency of SCR is used
here. It should be noted that SCR measures only degree of arousal, not its
valence (e.g., negative or positive).
After each ad. Attention to the ad was measured by three likert items
(attention to, interest in, and involvement with the ad). Arousal caused by
the ad was measured by one semantic differential scale: excited/not excited. Ad
evaluation was measured by seven semantic differential scales (persuasive/not
persuasive, convincing/not convincing, likeable/not likeable, favorable/not
favorable, powerful/not powerful, and valuable/not valuable). Finally, ad
importance was measured by two semantic differential scales (relevant/not
relevant, of concern/of no concern). All semantic differential scales were
7-point. The alphas on the multi-item scales (i.e., attention to the ad and ad
evaluations) on each separate ad were satisfactory, ranging from 0.73 to 0.92.
Ad importance received high correlations for each ad, in all cases with p<.0001.
Sociodemographic questions. For descriptive purposes, participants answered
questions about their age, gender, race, parental income, grade point average,
university year, and fraternity/sorority membership.
RESULTS
Of the participants, 37% were male and 63% were female; 88.7% were Caucasians.
They had a mean age of 20 (sd. 1.81). Most were in their second year of
university (M = 2.19, sd. 1.12) and had a mean grade point average of 2.99 (sd.
0.45); only 23% were members of fraternities or sororities. The participants
came from households with a mean income of between $80,000 and $100,000 (M =
5.05, sd. 2.08). Of the 62 participants, 62.9% (n = 39) were non-smokers while
37.1% (n = 23) were smokers. Of those who smoked in the past month, they smoked
between seven and nine cigarettes per day (M = 2.70, sd. 2.58). During the 11
previous months, the participants reported that they smoked slightly more than
one to two days per week (M = 3.35, sd. 2.25) on which occasions they consumed
between four and six cigarettes (M = 2.35, sd. 2.42). The mean length of
smoking experience was almost 2.5 years (M = 2.43, sd. 2.04).
Tests of the hypotheses and answers to research questions. Hypothesis 1
suggested that participants would exhibit higher levels of autonomic attention,
as measured by heart rate, to anti-smoking ads that are more complex. This
hypothesis was not supported. There was no significant effect of complexity on
heart responses, nor were there any significant interactions.
Hypothesis 2 suggested that participants would exhibit higher levels of
autonomic arousal, as measured by the frequency of skin conductance responses,
to anti-smoking ads that are more complex. This hypothesis was supported.
Complexity had a main effect on SCR (F(2,65) = 2.82, p<.066). Low complexity
and medium complexity ads had means of 1.14 and 1.13 respectively while high
complexity ads had a mean of 1.55. It appears then that high levels of
complexity, as opposed to low or medium levels, leads to significantly more
autonomic arousal. This main effect was qualified by an interaction between
complexity and smoking status (F(2,65) = 3.58, p<.033; Figure 1). Here, smokers
displayed a pattern inconsistent with non-smokers. Thus, smoker means were
0.94, 0.69, and 1.00 for the low, medium and high conditions while the means for
non-smokers were 1.20, 1.27, and 1.79 for the three conditions. Consistent with
Antecol (1998), this result is particularly interesting because it suggests that
different processes may be occurring at the autonomic level in smokers and
non-smokers.
Research Question 1a asked how complexity would affect the participants'
self-reported levels of attention and arousal while Research Question 1b asked
if participants' self-reported levels of attention and arousal would be
consistent with their autonomic responses. Consistent with the heart rate date,
complexity did not lead to any significant effects on self-reported attention.
Inconsistent with the SCR data, complexity did not lead to lead to any
significant effects on self-reported arousal. This result suggests that effects
at the autonomic level may be importantly different from those at the
self-report level.
Hypothesis 3 suggested that participants would rate complex ads as more
likeable and more important than simple ads. This hypothesis was supported.
Complexity had a significant effect on ad evaluation (F(2,65) = 18.59, p<.0001)
such that low complexity ads were evaluated the lowest (M = 3.87), followed by
medium complexity ads (M = 4.27), and high complexity ads (M = 4.89). All means
were significantly different. No other effects were significant. Complexity
also had a significant effect on ad importance (F(2,65) = 3.30, p<.0383). Here,
however, the pattern was not as clear. Low complexity ads were rated lowest (M
= 4.02) followed by high complexity (M = 4.37) and medium complexity ads (M =
4.60). Only low complexity ads were significantly different from medium
complexity ads; high complexity ads were not significantly different from the
other two. No other effects were significant.
DISCUSSION
These findings are in line with previous research that complexity is an
important influence on the effectiveness of a television message (e.g., Geiger &
Reeves, 1993; Hockberg & Brooks, 1978; Reeves, Thorson & Schleuder, 1986;
Thorson, Reeves & Schleuder, 1985). That influence appears to be concentrated
on autonomically-measured arousal (SCR) as opposed to autonomic attention (heart
rate); a finding that reflects previous findings (e.g., Lang, Bolls, Potter &
Kawahara (in press); Zhou, et al., 1997). Non-smokers show a greater arousal
with greater complexity while non-smokers do not. This result would seem to
suggest that different processes are occurring at the autonomic level for the
smokers and non-smokers (Antecol, 1998).
The study also showed that measuring attention and arousal autonomically
produces different results than by measuring them through self reports.
Although self-reported and autonomic attention produced similar results in that
in both cases complexity did not lead to a significant main effect,
self-reported and autonomic arousal produced inconsistent findings. While there
was a clear autonomic arousal pattern, the self-reported responses showed no
differences. This finding suggests that viewers are not aware of how particular
anti-smoking ads are arousing them. However, this finding may be an artifact of
the one-item measure used here. As a result, a more precise measure will be
used in the next experiment.
Finally, as predicted and consistent with Jackowitz, Schooler and Flora (1997),
increasing complexity in anti-smoking ads led to more positive ad evaluations as
well increased perceived importance. This finding suggests that increasing the
complexity of an ad, in addition to increasing its arousingness, can also lead
to thoughts that are consistent with a higher level of motivation to process the
ad more fully. Such processing has been linked to attitude change (Chaiken,
Wood & Eagly, 1996; Petty & Priester, 1994).
Having shown that complexity does make a difference with respect to viewing
anti-smoking ads, we turn now to the second aspect of this study, namely whether
these findings could be replicated in an experiment that includes ad type (i.e.,
individual blame vs. industry blame ads) as an additional factor.
EXPERIMENT 2
General design. To again test the main effect of complexity on self-report
measures (H3, RQ1a), to evaluate the main effect of ad type (RQ2), and to
ascertain if there is an interaction between complexity and ad type (RQ3) this
experiment was designed as a mixed 2 (smoker: smoker vs. non-smoker) x 2
(complexity: high vs. low) x 2 (ad type: individual blame vs. industry blame)
analysis of variance (ANOVA). Smoking was a between-subjects variable.
Complexity and ad type were within-subjects variables. Participants (N = 362)
representing the 18-24 year-old demographic were recruited from journalism
classes at two large Midwestern universities. They participated in groups of
approximately 20.
Procedure: Each participant saw the two individual blame and the two industry
blame ads. One ad in each condition was a high complexity ad while the other
was a low complexity ad. Before and after viewing each anti-smoking ad, the
participant viewed one filler commercial for general household products. These
filler ads were placed around the anti-smoking ads so that participants saw
commercials other than the ones under study. To ensure there were no order
effects, the order of the ads were counter-balanced. After viewing each ad, the
participants answered ad-specific pencil-and-paper questions regarding the ad.
After viewing all the ads, the participants completed the remaining self-report
measures.
Independent variables. Ad type and complexity. Four of the six ads used in
the previous experiment were used here. The two anti-smoking ads chosen to
represent the individual blame condition were Dance and Aging. Aging was a low
complexity ad while Dance was a high complexity ad (see Table 1). The two ads
chosen to represent the industry blame condition were Own Words and Boardroom.
Boardroom was a low complexity ad while Own Words was a high complexity ad (see
Table 1). The four ads were classified as either individual blame or industry
blame based on a pre-test of 63 different journalism students (44 females, 19
males) who viewed 21 various anti-smoking commercials. Details of this pretest
are available upon request.
Smoking. Participants were again asked to indicate how often in the past month
they had smoked a cigarette. Participants who did not smoke were classified as
non-smokers, those who smoked 1 day a month, 1 to 3 days a month, 1 to 2 days a
week, 3 to 4 days a week, 5 to 6 days a week, or daily were smokers (based on
Grube, McGee & Morgan, 1986). Again, to get a better grasp on the type and
extent of the smoking behavior, several questions were asked for descriptive
purposes: "How many cigarettes did you smoke, on average, each day during the
past month," "In the eleven (11) months previous to last month, on average, how
often did you smoke cigarettes," "When you did smoke in the eleven (11) months
previous to last month, how many cigarettes did you smoke, on average, each day"
(Grube, McGee & Morgan, 1986). Participants were also asked the length of time
that they had smoked cigarettes.
Dependent variables. Attention to the ad was measured by three items
(attention to, interest in, and involvement with the ad). Arousal caused by
the ad was measured, unlike the simple one-item scale used in Experiment 1, by
four semantic differential scale: distressed/not distressed, jittery/not
jittery, nervous/not nervous, and fearful/not fearful. Ad evaluation was
measured by seven semantic differential scales (persuasive/not persuasive,
convincing/not convincing, likeable/not likeable, favorable/not favorable,
powerful/ not powerful, and valuable/not valuable). Finally, ad importance was
measured by two semantic differential scales (relevant/not relevant, of
concern/of no concern). All scales were 7-point. The alphas for these measures
on each separate ad were again satisfactory, ranging from 0.72 to 0.89.
Similarly, ad importance received high correlations for each ad, in all cases
with p<.0001.
Sociodemographic questions. For descriptive purposes, participants answered
questions about their age, gender, race, parental income, grade point average,
university year and fraternity/sorority membership.
RESULTS
Of the participants, 34.5% were male and 65.5% were female; 86.2% were
Caucasians. They had a mean age of 20.61 (sd. 2.78). Most were in their second
year of university (M = 2.53, sd. 0.95) and had a mean grade point average of
3.20 (sd. 0.60); 31.2% were members of fraternities or sororities. The
participants came from households with a mean income of close to $80,000 and
$100,000 (M = 4.67, sd. 1.87). Of the participants, 60.5% (n = 219) were
non-smokers while 39.5% (n = 102) were smokers. Of those who smoked in the past
month, they smoked between 10 and 12 cigarettes per day (M = 3.98, sd. 1.95).
During the 11 previous months, the participants reported that they smoked
slightly more than one to two days per week (M = 3.51, sd. 2.02) on which
occasions they consumed between four and six cigarettes (M = 2.14, sd. 1.97).
The mean length of smoking experience was almost 3 years (M = 34.73 months, sd.
22.11).
Tests of the hypotheses and answers to research questions. Research Question
1a asked how an ad's complexity affects the participants' self-reported levels
of attention and arousal. As in Experiment 1, ad complexity had no significant
main effect on self-reported attention; nor did complexity interact with smoking
status. Similarly, complexity had no significant main effect for self-reported
arousal. However, it did interact with smoking status (F(1,360) = 3.37,
p<.0673; Figure 2). As with the skin conductance responses in Experiment 1,
smokers showed almost no differences as a function of complexity (low M = 3.64,
high M = 3.60). In contrast, non-smokers reported more arousal for low
complexity ads (M = 3.39) than for high complexity ads (M = 3.51). It should be
noted that, although the smokers were again flat in their responses, they were
higher than for non-smokers.
Hypothesis 3 suggested that participants would evaluate complex ads more
positively than simple ads. Inconsistent with Experiment 1, complexity had no
significant effects on ad evaluation. In addition, complexity had no
significant effects on ad importance. Hypothesis 3 was not supported.
Research Question 2 asked how ad type would affect participants' ad-specific
responses. Ad type had a significant effect on attention (F(1,360) = 5.93,
p<.0154) such that individual blame ads received more attention (M = 5.60) than
industry blame ads (M = 5.44). Ad type did not interact with smoking status.
Ad type also led to a significant main effect on arousal (F(1,360) = 105.99,
p<.0001). Industry blame ads (M = 3.81) were more arousing than individual
blame ads (M = 3.22). There was no interaction with smoking status. Likewise,
there was a main effect on ad importance (F(1,360) = 27.51, p<.0001). Again,
industry blame ads were thought to be more important (M = 4.70) than individual
blame ads (M = 4.39). This effect was modified by an interaction with smoking
status (Figure 3). Here, smokers thought industry blame ads were much more
important (M = 5.26) than individual blame ads (M = 4.80). The same
relationship appeared for non-smokers although it was less pronounced: industry
blame ads were thought to be more important (M = 4.33) than individual blame ads
(M = 4.11). Ad type had no significant main effect on ad evaluation. Overall,
then, it appears that when ad complexity is included with ad type, it is
industry blame rather than individual blame ads that rated higher.
Research Question 3 asked if blame focus would interact with complexity level.
Blame focus did interact with complexity on attention (F(1,360) = 65.77,
p<.0001; Figure 4). Individual blame ads received more attention when they were
low complexity (M = 5.44) as opposed to high complexity (M = 5.76). Industry
blame ads showed the opposite pattern. Here, low complexity ads received more
attention (M = 5.67) than high complexity ads (M = 5.20). Blame focus also
interacted with complexity on arousal (F(1,360) = 31.43, p<.0001; Figure 5).
The same pattern was again apparent. Individual blame ads were rated as more
arousing when they were high in complexity (M = 3.36) then when they were low in
complexity (M = 3.08). The reverse pattern was seen for industry blame ads,
although it was less pronounced. Here, low complexity ads were rated as more
arousing (M = 3.90) than high complexity (M = 3.72). This interaction also
occurred on ad evaluation (F(1,360) = 8.07, p<.0047; Figure 6) with the same
cross-over interaction. Individual blame ads led to more positive evaluations
when they were high in complexity (M = 4.87) then when they were low in
complexity (M = 4.74). Industry blame ads showed the reverse pattern. Here,
low complexity ads were evaluated more favorably (M = 4.96) than high complexity
ads (M = 4.76). Finally, there was a marginally significant interaction on ad
importance (F(1,360) = 2.79, p<.0955; Figure 7) also with the crossover effect.
Individual blame ads were rated as more important when they were high in
complexity (M = 4.42) then when they were low in complexity (M = 4.36). For
industry blame ads, low complexity ads were rated as more important (M = 4.78)
than high complexity ads (M = 4.63). However, on ad importance there was also a
three-way interaction between smoking, ad type and complexity (F(1,360) = 4.54,
p<.0338; Figure 8 and 9). For non-smokers, the same cross-over pattern was
evident. For individual blame ads, those high in complexity were rated as more
important (M = 4.16) than those low in complexity (M = 4.07). For industry
blame ads, those low complexity were rated as more important (M = 4.48) than
those high in complexity (M = 4.48). However, for smokers complexity did not
matter: industry blame ads were rated as more important (Low M = 5.23, High M =
5.28) than individual blame ads (Low M = 4.80, High M = 4.81).
DISCUSSION
The results of this experiment demonstrated that structural complexity
interacts with blame focus in anti-smoking ads. Industry blame ads have more
impact on attention, arousal, ad evaluation and ad importance if their
structural complexity is low. Individual ads have more impact if they are high
in complexity. Taking complexity account seems especially valid for smokers who
found the industry blame ads to much more important than the individual blame
ads (although it should be pointed out that they found both types of ads more
important than non-smokers regardless of their complexity level). Thus, one way
to draw smokes into an anti-smoking message would be to create industry blame
ads that are low in complexity.
There are at least two possible explanations for this difference in
effectiveness. First, it may be that the industry blame ads which, because they
were novel for this group of participants (indeed, because they are a relatively
new phenomenon, this type of ad is likely novel for most viewers), did not need
increased complexity to draw in the participants. In contrast, although the
actual individual blame ads used here were unfamiliar to the participants, these
ads represented the status quo approach (Dorfman & Wallack, 1993) in televised
anti-smoking campaigns. As such, the participants, while they may not have been
familiar with the actual ads, were familiar with this type of ad and thus needed
higher complexity levels to draw them in. Second, because only four
experimental ads were used, there was only one ad in each cell of the
interaction. Thus, this result may be specific to the ads used here. Although
it is belief that this is not the case, especially because the ads used here
were judged to be exemplars in our pretest, it nevertheless would be necessary
to replicate these results using a wider universe of anti-smoking ads.
Importantly, we found that the self-reported arousal measure used in this
experiment was successful. Unlike the first experiment, where there was no
self-report effect for arousal, in this experiment there was a significant
interaction. That finding suggests that if this scale had been used in the
previous experiment, those participants may have been able to accurately
self-report their arousal levels in a manner consistent with the autonomic
response data. Future research is required to test this point.
Overall then the results of Experiment 1 showed that complexity does have an
effect on the effectiveness of anti-smoking ads, both at an autonomic and
self-report level. In Experiment 2, complexity was found to interact with ad
type such that individual blame ads high in complexity and industry ads low in
complexity were rated the most effective. Together, the results point to a need
to be more sophisticated in looking at different possible dimensions that could
influence the effectiveness of an anti-smoking ads. For example, not including
the dimension of complexity may be the reason for the inconsistent results of
Antecol (1998), Goldman and Glantz (1998), and Pechmann (1998) rather than the
different measures and samples used by those experimenters. The experiments
also pointed to the need to include both physiology and self-report data, as
often self-report data can be unreliable in that participants cannot accurately
self-report their autonomic activity.
Implications for health practitioners. The results presented here place health
practitioners in somewhat of a quandary. Despite some findings that television
anti-smoking campaigns involving individual blame ads have been successful
(e.g., Boddewyn, 1994; Moskowitz, 1983; Pierce at al., 1994; Popham et al.,
1994; Weis & Burke, 1986), the weight of the evidence suggests that these
campaigns have been ineffectual in achieving direct or sustained behavioral
change (e.g., Bauman et al., 1989, 1991; Canadian Long Range Planning Branch,
1977; O'Keefe, 1971; Wallack, 1981; Wallack & Corbett, 1987; Warner, 1977). The
reason for this lack of success, based on our results, may be that the wrong
types of ads were being directed at smokers. Rather than the individual blame
of unknown complexity that were studied by the above authors, health
practitioners may want to consider campaigns that include industry blame ads
that are low in complexity or individual blame ads that are high in complexity.
It was these combinations that led to responses on our various scales that are
consistent with increased motivation to process the ads. Such in-depth
processing has been linked to attitude change (Chaiken, Wood, & Eagly, 1996;
Petty & Priester, 1994) a necessary precursor of behavior change and/or
modification
Limitations and Future Research. The two studies represent the first step in
attempting to quantify the differences between individual and industry blame
anti-smoking ads that are either high and low in complexity. However, as such a
beginning there are several limitations. First, it was not possible to explore
every possible avenue. As a result, health-relevant outcomes (e.g.,
anti-smoking attitudes, norms, personal smoking policies, public smoking
policies, and smoking intentions) and actual behaviors were not measured. Thus,
future research should these variables. Second, the possibility exists, based
on the differences in the self-reported arousal results, that the self-report
measures may not be sensitive enough. Future research should seek out
additional and/or more refined measures to ascertain if the same pattern of
results manifests itself. One potential source of proven measures, for example,
is the Velicer et al.'s (1994) smoking policy inventory. This inventory would
be especially useful in measuring health-relevant outcomes. Third, only
autonomic arousal was measured in our first experiment. Future research should
also consider the valence of that arousal. Fourth, although comparable results
were achieved in the two experiments, different procedures were used. It is
possible that the procedures used in one or both of the experiment may have been
problematic. In either case, this is not how individuals generally watch
television. Therefore, future research should explore these issues in a more
realistic setting. Fifth, these studies used only one demographic and only a
particular element of that demographic. Future research should replicate these
studies with different demographics to ascertain if the results are consistent
across populations. Finally, these studies were purposefully restricted to the
area of anti-smoking ads. Future research should examine if the same results
would ensue in different domains. For example, would the same results occur if
alcohol or drugs were the subject of the ads? There is some evidence to suggest
that differential blame foci lead to different perceptions of responsibility.
Thus, Iyengar (1990, 1991) found, on issues of poverty, television news viewers
of society-oriented stories tended to be more sympathetic to poor individuals
than viewers who saw individually-oriented stories. References
Antecol, M. (1998). Effects of individually-focused vs. structurally-focused
arguments in anti-smoking television commercials. Unpublished Ph.D.
dissertation, School of Journalism, University of Missouri - Columbia, Columbia,
MO.
Basil, M.D. (1994). Multiple resource theory II: Empirical examination of
modality specific attention to television scenes. Communication Research, 21,
208-231.
Bauman, K. E., LaPrelle, J., Brown, J. D., Koch, G. G., & Padgett, C. A. (1991).
The influence of three mass media campaigns on variables related to adolescent
cigarette smoking: Results of a field experiment. American Journal of Public
Health, 81(5), 597-604.
Bauman, K. E., Padgett, C. A., & Koch, G. G. (1989). A media-based campaign to
encourage personal communication among adolescents about not smoking cigarettes:
Participation, selection and consequences. Health Education Research, 4(1),
35-44.
Boddewyn, J. J. (1994). Cigarette advertising bans and smoking: The flawed
policy connection. International Journal of Advertising, 13, 311-332.
Canadian Long Range Planning Branch and Non-Medical Use of Drugs in Directorate
(1977). Smoking and health in Canada. Ottawa, ON: Department of National Health
and Welfare.
Chaiken, S., Wood, W., & Eagly, A. H. (1996). Principles of persuasion. In E. T.
Higgins & A. Kruglanski (Eds.), Social psychology: Handbook of basic principles
(pp.702-742). New York: Guilford Press.
Cook, E.W.I. (1985). VPMANLOG (Version 5.0). Birmingham, AL.
Graham, F. K. (1979). Distinguishing among orienting, defense, and startle
reflexes. In H. D. Kimmel, E. H. Van Olst, & J. F. Orlebeke (Eds.), The
orienting reflex in humans (pp. 137-167). Hillsdale, NJ: Lawrence Erlbaum
Associates, Inc., Publishers.
Hitchon, J., Duckler, P., & Thorson, E. (1994). Effects of ambiguity and
complexity on consumer response to music video commercials. Journal of
Broadcasting & Electronic Media, Summer, 289-306.
Iyengar, S. (1991). Is anyone responsible? How television frames political
issues. Chicago: University of Chicago Press.
Iyengar, S. (1990). Framing responsibility for political issues: The case of
poverty. Political Behavior, 12, 19- 40.
Johnson-Cartee, K. S., & Copeland, G. (1997). Manipulation of the American
voter: Political campaign commercials. Westport, CT: Praeger.
Lang, A. (1990). Involuntary attention and physiological arousal evoked by
structural features and emotional content in TV commercials. Communication
Research, 17, 275-299.
Lang, A., Bolls, P. D., Potter, R. F., & Kawahara, K. (in press). The effects of
production pace and arousing content on the information processing of television
messages. Journal of Broadcasting and Electronic Media.
Moskowitz, J. (1983). Preventing adolescent substance abuse through drug
education. In T. G. Glynn, C. G. Leukefeld & J. P. Ludford (Eds.), Preventing
adolescent drug abuse: Intervention strategies. Research monograph 47.
Rockville, MD: National Institute of Drug Abuse.
Middlestadt, S.E., Fishbein, M., & Chan, D K-S. (1994). The effect of music on
brand attitudes: Affect- or belief-based change? In E.M. Clark, T.C. Brock, &
D.W. Stewart (Eds.), Attention, Attitude, and Affect in Response to Advertising
(pp.149-167). Hillsdale, NJ: Lawrence Earlbaum.
O'Keefe, M. T. (1971). The anti-smoking commercials: A study of television's
impact on behavior. Public Opinion Quarterly, 35, 242-248.
Petty, R. E., & Priester, J. R. (1994). Mass media attitude change:
Implications of the elaboration likelihood model of persuasion. J. Bryant & D.
Zillman, (Eds.). Media effects: Advances in theory and research (pp.91-122).
Hillsdale, NJ: LEA.
Pierce, J. P., Evans, N., Farkas, A. J., Cavin, S. W., Berry, C., Kramer, M.,
Kealey, S., Rosbrook, B., Choi, W., & Kaplan, R. M. (1994). Tobacco use in
California: An evaluation of the Tobacco Control Program, 1989-1993. La Jolla,
CA: University of California, San Diego.
Popham, W. J., Muthen, L. K., Potter, L. D., Duerr, J. M., Hetrick, M. A., &
Johnson, M. D. (1994). Effectiveness of the California 1990-1991 tobacco
education media campaign. American Journal of Preventive Medicine, 10(6),
319-326.
Potter, R. F., Lang, A., & Bolls, P. D. (1998). Orienting to structural features
in auditory media messages. Psychophysiology, Supplement, (S66).
Reeves, B., & Thorson, E. (1985). Watching television: Experiments on the
viewing process. Communication Research, 13, 343-361.
Reeves, B., Lombard, M., & Melwani, G. (1992, May). Faces on the screen:
Pictures or natural experience? Paper presented to the Mass Communication
division of the International Communication Association, Miami.
Schleuder, J. (1990). Effects of commercial complexity, the candidate, and issue
vs. image strategies in political ads. In Advances in Consumer Research 17,
159-168.
Ries, H.A. (1969). GSR and breathing amplitude related to emotional reactions to
music. Psychonomic Science, 14, 62-64.
Thorson, E., Reeves, B., & Schleuder, J. (1986). Attention to local and global
complexity in television messages. In M.L. McLaughlin (Ed.), Communication
Yearbook 10. Beverly Hills, CA: Sage.
Thorson, E., & Lang, A. (1992). Effects of television video graphics and lecture
familiarity on adult cardiac orienting responses and memory. Communication
Research, 19(3), 346-369.
Thorson, E., Reeves, B., & Schleuder, J. (1985). Message complexity and
attention to television. Communication Research, 12, 427-454.
Velicer W. F., Laforge R. G., Levesque D. A., & Fava J. L. (1994). The
development and initial validation of the smoking policy inventory. Tobacco
Control, 3, 347-355.
Wallack, L. M. (1981). Mass media campaigns: The odds against finding behavior
change. Health Education Quarterly, 8(3), 209-260.
Wallack, L. M. , & Corbett, K. (1987). Alcohol, tobacco and marijuana use among
youth: An overview of epidemiological, program and policy trends. Health
Education Quarterly, 14(2), 223-249.
Warner, K. E. (1977). The effects of anti-smoking campaigns on cigarette
consumption. American Journal of Public Health, 67, 645-650.
Weis, W., & Burke, C. (1986). Media content and tobacco advertising: an
unhealthy addiction. Journal of Communication, 36, 59-69.
Wheatley, J.J., & Brooker, G. (1994). Music and spokesperson effects on recall
and cognitive response to a radio advertisement. In E.M. Clark, T.C. Brock, &
D.W. Stewart (Eds.), Attention, Attitude, and Affect in Response to Advertising
(pp.149-167). Hillsdale, NJ: Lawrence Earlbaum.
Yoon, K., Bolls, P., & Lang, A. (1998). The effects of arousal on liking and
believability of commercials. Journal of Marketing Communications, 4, 101-114.
Zimny, G.H., & Weidenfeller, E.W. (1963). Effects of music upon GSR and heart
rate. American Journal of Psychology, 76, 311-314.
Zhou, S., Schwartz, N., Bolls, P. D., Potter, R. F., Lang, A., Trout, G.,
Funabiki, R., Borse, J., & Dent, D. (1997, August). When an edit is an edit, can
an edit be too much?: The effects of edits on arousal, attention, and memory
for television messages . Paper presented to the Theory and Methodology Division
of the Association for Education in Journalism and Mass Communication, Chicago,
IL.
Key Words:
Ads/Commercials
Television/TV
Smoking/Anti-smoking
Individual Blame/Focus
Structural Blame/Focus
Complexity and Blame Focus in Anti-Smoking Television Commercials:
The Role of Complexity and Individual vs. Industry Blame
on Smokers and Non-Smokers
Michael Antecol1, Esther Thorson2 , Annie Lang3, Robert F. Potter4, June Flora1,
Lisa Henriksen1
1 Stanford Center for Research in Disease Prevention
School of Medicine, Stanford University
2 Center for Advanced Social Research
School of Journalism, University of Missouri
3 Institute for Communication Research
Department of Telecommunications, Indiana University
4 Institute for Communication Research
Telecommunication and Film Department,
University of Alabama
Correspondence to the first author at:
Stanford Center for Research in Disease Prevention
Stanford University School of Medicine
1000 Welch Road
Palo Alto, CA 94304-1825
phone: (650) 725-5011
fax: (650) 725-6906
[log in to unmask]
Submitted to the Advertising Division
of the Association for Education in Journalism and Mass Communication
to be Considered for Presentation at the 1999 Conference
April 1, 1999
|
