Download as pdf or txt
Download as pdf or txt
You are on page 1of 14

Sympathy and Empathy: Emotional Responses

to Advertising Dramas
JENNIFER EDSON ESCALAS
BARBARA B. STERN*

This research examines differences in consumers' sympathy and empathy re-


sponses to teievised drama commercials. The research framework is multidisci-
plinary, for construct definition from humanities disciplines (aesthetics and philos-
ophy) grounds the empirical testing of sympathy and empathy responses to
advertising. Valid and reliable measurement instruments are developed to test
relationships between sympathy and empathy as responses to classical and vi-
gnette advertising dramas. Results of two experiments indicate that sympathy
responses mediate the effect of a drama advertisement's form on empathy re-
sponses, with both sympathy and empathy directly enhancing positive attitudes to
an advertisement.

T he purpose of this study is to distinguish between sym-


pathy and empathy as separate but related constructs
and to test the influence of each as emotional responses to
responses to the empirical measurement of response-side
effects. To do so, two new scales are developed—one for
sympathy and one for empathy—that enable the measure-
advertising dramas. The rationale for examining the linked ment of response differentiation, relatedness, and strength.
constructs is that prior empirical research has borrowed but The underlying assumption is that the different responses
one half of the sympathy/empathy system, treating empathy of sympathy and empathy are interconnected parts of a sys-
as a monolithic response measured on a bipolar continuum tem rather than mutually exclusive responses and that the
ranging from empathy to nonempathy (Boiler t990; Boiler, reconnection of sympathy to the empathy research stream
Babakus, and Olson 1989; Boiler and Olson 1991; Deighton allows for a more informed approach to emotional re-
and Hoch 1993; Deighton, Romer, and McQueen 1989; sponses. The empiricization of response differences that in-
Wells 1989). More recent empathy research follows suit by fluence attitudes to an ad is the basis for our construction
not including the role of sympathy as an emotional response of a response model rooted in a distinction between sym-
to persuasive appeals (see Aaker and Williams [1998] for pathy and empathy.
a review). The current study aims to expand the research The concept of emotional response differentiation rep-
domain to include both sympathy and empathy as a more resents a retum to the tradition in multidisciplinary research,
comprehensive system of emotional responses, to begin to where disciplines such as aesthetic criticism, philosophy,
explore ways in which sympathy and empathy differ and and psychology converge on the existence of two constructs
relate to each other, and to test sympathy and empathy in
(Morrison 1988). Differentiation centers on the contrast be-
an experimental setting. In so doing, we use Stem's (1994)
tween a self-conscious, emotionally cognizant, sympathetic
labels for different drama types (classical/vignette) and adapt
observer who remains "outside" a textual or real-life stim-
them to reflect Deighton and Hoch's (1993) distinction be-
ulus and an unself-conscious, emotionally absorbed, em-
tween advertising dramas that show emotions versus those
pathic participant who loses him/herself in it. Even though
that teach viewers how to have them.
both sympathy and empathy are considered feelings re-
Our study shifts the focus from the stimulus-side cate-
sponses to created works, they are categorized as different
gorization of dramas and interpretive analysis of consumer
response types. The logic is that people who experience
sympathy remain emotionally conscious of their personal
*Jennifer Edson Escalas is an assistant professor in the marketing de- lives and understand but do not directly reexperience an-
partment at the University of Arizona, Tucson, 85721-0108; jescalas®
bpa,arizona.edu, Barbara B, Stem is professor II and chair of the marketing
other's feelings (Langfeld [1920] 1967), whereas people
department at Rutgers, The State University of New Jersey, Faculty of who experience empathy completely forget their own per-
Management, Newark, NJ 07102; hbstem@aol,com. The authors thank sonal existence by sharing the feelings of the characters. In
Dawn Iacobucci, Jim Bettman, Shankar Ganesan, Hans Baumgartner, Scott consumer research, even though sympathy is not mentioned
McKenzie, the editor, associate editor, and three reviewers for their help
with this research.
by name, Deighton and Hoch make a similar distinction
between emotional responses communicated by displays of
566

© 2003 by JOURNAL OF CONSUMER RESEARCH, Inc, • Vol, 29 • March 2003


All rights reserved, 0093-5301/2003/2904-0009$10,00
SYMPATHY AND EMPATHY DRAMA AD RESPONSES 567

feelings whose intention and meaning can be interpreted by is not identical to the other's emotion, but consists of feel-
consumers—that is, sympathy—and emotional responses ings of sorrow or concem for another's welfare" (Eisenberg
elicited by feelings claims, following the empathic path to and Miller 1987, p. 92). Sympathy reflects "the difference
persuasion. The former response can properly be called sym- between the one sympathizing and the object of sympathy
pathy, for consumers can "conceivably be quite dispassion- . . . always somewhat present in consciousness" (Langfeld
ate" (Deighton and Hoch 1993, p. 267) rather than em- 1967, p. 138). That is, sympathy is a heightened awareness
pathically moved to manifest some of the physiological (Wispe 1986) of another person's state of mind and his or
correlates of emotion. In order to design scales that measure her circumstances (Mercer 1972) stemming from recogni-
the differences more precisely, we must first clarify the con- tion of his or her feelings.
fused terminological history of both constructs.
Empathy: In-Feeling
TWISTED fflSTORY "Empathy" is the more recent term in English, entering
The need for clarification stems from the constructs' the vocabulary only in the last century. The credit for trans-
twisted linguistic heritage, so convoluted that neither term lation into English goes to Edward Bradford Titchener
has necessarily meant the same thing to any two researchers. (1909a, 1909b), who coined the word empathy to express
For centuries, the terms have been so loosely used that mean- the difference between einfuhiung (in-feeling) and mitge-
ing is shifting and variable. To cut through the confusion, fuhiung (with-feeling) (Vischer [1873] 1994), which was
we re-ask the basic definitional questions—what is sym- already in the English lexicon as sympathy (Wispe 1987).
pathy? and what is empathy?—in an advertising context. Consumer research has drawn from the traditional defini-
We emphasize the importance of context-specific definitions, tion of the empathy response as an involuntary and unself-
for advertising stimuli exist in the domain of created media conscious merging with another's feelings. The term's his-
representations rather than that of real-life events, and re- torical and etymological roots lie in the pagan mystic phrase
sponses relate to media phenomena rather than to naturally that expresses this absorption—"I am you and you are I"
occurring events. The question of what to call the responses (see Buber [1937] 1970, pp. 133-134). From about 1800
has hampered knowledge transfer from multidisciplinary onward, the phrase appeared in German philosophy (Mall-
sources. Terminological inconsistency and mutability is part grave and Ikonomou 1994; Shipley 1931), and by the tum
of the baggage that modem psychology inherited (Strayer of the century, the original German word had entered re-
1987; Wispe, 1986, 1987), exacerbated by flip-flops in search on aesthetic psychology (Lipps 1897, 1903-1906a,
meaning during this century. Up to the 1950s, "sympathy" t903-1906b). Empathy refers to a person's capacity to feel
was the word of choice in research, even though it often within or in another person's feelings (Langfeld 1967, p.
referred to what is now called "empathy." However, from t38), and most researchers now consider it "an emotional
the late 1950s to the present, the situation was reversed, response that stems from another's emotional state or con-
with "empathy" being the word of choice, even though it dition and that is congment with the other's emotional state
often refers to what is now called "sympathy." The current or situation" (Eisenberg and Strayer 1987, p. 5). Eisenberg
confusion is a result of both the gobbling up of each term and Strayer note the current cross-disciplinary agreement
by the other and the flipped meanings at the century's half- that empathy is the act of feeling into another's affective
way mark. Keeping in mind that our study's focus is on experience (1987). Briefly, an empathy response is a per-
responses to created works, specifically ads, we tum to the son's absorption in the feelings of another.
most theoretically grounded source of construct differenti- To sum up, whereas sympathy stems from the perspective
ation—the late nineteenth century research of psychological of an observer who is conscious of another's feelings, em-
aestheticians (Vivas and Krieger 1953), who drew from pathy stems from that of a participant who vicariously
moral philosophy (ethics) and Aristotelian aesthetics. merges with another's feelings. In sympathy, rather than
one's feelings being merged in the object, they mn parallel
with the object (Langfeld 1967). The difference is analogous
SYMPATHY AND EMPATHY DEFINED to potential responses to watching an individual suck on a
lemon (Eisenberg and Miller 1987): a sympathetic response
Sympathy: With-Feeling consists of thoughts about what the lemon sucker might
In consumer research, whether or not sympathy is named, perceive or want, whereas an empathic response consists of
the emotional response is said to be a person's awareness an involuntary puckering and watering of the observer's
of the feelings of another, but not absorption in the feelings mouth. Table 1 presents a summary of the definitional issues
themselves (Deighton and Hoch 1993; Stem 1994). This in the sympathy/empathy contrast.
definition accords with the traditional definition of the term,
which entered English nearly three centuries ago via Hume's HYPOTHESES
A Treatise of Human Nature ([1739] 1968), Hume's work
set forth the logic of sympathy that entered the social sci- Hypothesis 1: Construct Differentiation
ences (Mercer 1972), referring to "an emotional response Let us begin by emphasizing that in the advertising con-
stemming from another's emotional state or condition that text, even though dramas contain the same attributes as film
568 JOURNAL OF CONSUMER RESEARCH

TABLE 1

DEFINITIONAL ISSUES

Empathy Sympathy

Primary component Affective Cognitive


Feeling Thinking
Control Involuntary Voluntary
Self-other differentiation Absent Present
Self-other orientation Self Other
Loses self in other Distance from other
Relation to other Merging with other Understanding of other
Opposite Nonempathy Antipathy
Indifference Hostility
Attitudinal effects Direct Indirect

and theatrical performances, they are not equivalent in model of drama forms, it does not follow the parallel response
length, depth, or complexity (Mick 1987). The 60-second dichotomy. Rather, we treat sympathy and empathy as dif-
television commercial lacks what Aristotle called the "mag- ferent but related responses to classical drama, viewing
nitude" necessary to achieve the full panoply of emotional Stem's association of sympathy with vignettes as overly sche-
responses (Fergusson 1961, p. 66). Notwithstanding adver- matic. The rationale is that only the classical form, unlike
tising's brevity, the structural similarity between drama vignettes (Stem 1994), possesses a unified plot (characters-
commercials and its lengthier antecedents suggests that anal- in-action), linear development, causality, and interaction. Sup-
ogous sympathy and empathy responses, albeit less pro- port for the influence of traditionally stmctured forms on
foundly felt, are likely to be elicited. The following dis- consumer responses is corroborated by analyses of story
cussion is based on the structural similarities, but ratchets grammars found in text linguistics, discourse analysis, and
down drama theory to fit advertising constraints. semiotics (Mick 1987). Here, story grammars represent in-
In consumer drama research. Wells (1989) was the first dividuals' tacit knowledge about the characteristic structures
to introduce the construct of empathy and its connection to of traditional stories, comparable to cohesively structured
advertising drama, pointing out that drama's performative classical dramas. Individuals invoke story grammars to enable
aspect draws viewers into an ad. His work and that of later "rapid, schema-driven comprehension under time compressed
researchers such as Deighton and Hoch (1993) and Stem conditions" (Mick 1987, p. 265) inherent in television ads. It
(1994) draw from the Aristotelian definition of drama (Fer- follows that individuals faced with the nontraditional vignette
gusson 1961) as a form of representation designed to be structure, in which rapid jumps from unrelated characters and
performed for viewing audiences (shown) rather than nar- events occur, are less likely to invoke schemas for compre-
rated to them (told). Stem focuses on structure (as distinct hending emotion because the viewer is unable to invoke sche-
from content), positing a dichotomous typology in which mas for comprehending emotion, let alone lose him/herself
classical drama—the traditional form—is characterized by in what s/he sees (Shweder 1994).
unified linear plots leading toward a resolution, causality, We find further support for the association between clas-
and characters who interact and change. sical drama and both response types in the consumer and
The other dramatic form in Stem's dichotomy is labeled advertising research literature. For example, findings from
"vignette," and her tables 2 and 3 (1994) present the dif- research on ads that tell a story indicate that well-developed
ferences in detail. Briefiy, the vignette's stmctural attributes stories are better able to hook ad viewers into the commercial
differ from those of the classical form as follows: vignettes (a concept that shares the absorption characteristic of em-
are characterized by multiple unconnected episodes (rather pathy), as well as elicit higher levels of positive emotions,
than a single unified plot), repetitive organization (rather compared to poorly developed stories (Escalas, Moore, and
than linear), and characters contained within each episode Edell, forthcoming). Drama/empathy consumer research
(rather than interacting with those in other episodes). Here, also supports the power of classical drama to evoke sym-
the repetitive episodic form and the multitude of different pathy and empathy insofar as its structural attributes have
characters lack the "linear order, sustained human interac- the capacity to draw consumers into comprehending or vi-
tion, and causal change" (Stem 1994, p. 613) necessary to cariously participating in a commercial (Deighton and Hoch
absorb viewers. This form is said to evoke the response of 1993; Wells 1989). Hence, our first hypothesis predicts that
sympathy rather than empathy, because viewers are detached classical dramas will evoke responses that are both more
observers of the lives of others and remain self-aware rather sympathetic and more empathetic than responses to
than absorbed in the performance. That is, a response di- vignettes.
chotomy of sympathy or empathy is posited as parallel to
the stimulus dichotomy. HIA: Classical dramas are more likely to evoke sym-
However, whereas our study accords with the dichotomous pathy than vignette dramas.
SYMPATHY AND EMPATHY DRAMA AD RESPONSES 569

HIB: Classical dramas are more likely to evoke em- FIGURE 1


pathy than vignette dramas. FOUR-STAGE PATH TO AD ATTITUDE

Hypotheses 2-4: Construct Relationships Advertising Sympathy Empathy Ad Attitude


Drama Type
Sequentiality: Sympathy and Empathy Responses to
Advertising Dramas. Hypothesis 2 treats sympathy and
empathy as sequentially related emotional responses to clas- that in an American individualistic cultural context, the rec-
sical dramas, proposing that sympathy is a precursor of em- ognition of other-focused emotions has a significant positive
pathy. The relationship shown in the first three stages in effect on A^^ (their term "empathy" refers to what we call
figure 1 is one of progression, in which classical dramas sympathy). Research by Stout and others further supports
evoke sympathy as a process toward empathy. The proces- the idea that emotion recognition enhances ad attitudes
sual concept is found in many research literatures, with em- (Stout, Homer, and Liu 1990; Stout and Leckenby 1986).
pathy always posited as an end state preceded by sympathy. However, prior research has not treated the traditional
This relationship is posited on the basis of hierarchically hierarchical ordering of sympathy and empathy, and our
ranked, sequential responses, with empathy consistendy study proposes the existence of a sequential relationship.
conceived of as a more highly ranked end state that works Recall that of the two emotions, empathy is consistently
through a less highly ranked sympathy. viewed as stronger than sympathy in the elicitation of pleas-
In aesthetics, sympathy and empathy are described as ure. Empathy is conceived of as an end state of absorption
stages that the same person can go through in experiencing or immersion likely to engender more positive attitudes than
drama, with empathy the ultimate stage of drama reception sympathy, a more self-aware preliminary state. A multidis-
(Langfeld 1967). In psychology, the process is conceived ciplinary research tradition converges on this premise, with
of as one that moves from sympathy to empathy, with rec- empathy regarded as a more strongly positive response from
ognition of emotions conveyed by contextual and verbal early aesthetic criticism to current consumer research on
content (Strayer 1987) and culminating in vicarious emotion advertising effects.
sharing. In advertising and consumer research, a parallel A necessarily brief summary of multidisciplinary confir-
progression is described in the high-involvement hierarchy- mation begins with the aesthetic drama critic Langfeld
of-effects model (Lavidge and Steiner 1961; McGuire 1978; 1967), who described sympathy ("I understand your pain")
Ray 1973). Note that this hierarchy refers only to high- as a weaker and more detached response than empathy ("I
involvement advertising (Krugman 1966-1967), which our share your pain"). Empathy enables viewers "to live in the
study's experimental environment produces by requiring re- character" (Langfeld 1967, p. 137) and derive pleasure from
spondents to focus closely on the ads. In the high-involve- the performance. In later aesthetic film criticism (Aumont
ment situation, the hierarchical model describes thinking et al. [1983] 1992; Metz 1974), the empathic response is
responses, which we view as more characteristic of sym- described as the ultimate experience in which the spectator
pathy, as a precursor of feeling responses, which we view shares the emotions of characters and thus becomes the cen-
as characteristic of empathy. Mick's subjective comprehen- tral figure in the performance. Similarly, in rhetorical crit-
sion model (1992) also supports the processual ordering, icism, Barthes's pleasure or enjoyment of the text is said to
because it describes a more cognitive surface-level objective result in a more positive attitude toward the ad (McQuarrie
comprehension of a message as necessary to the more af- and Mick 1996).
fective deeper-level subjective responses. Communication research echoes the concept of an in-
volved and empathetic reader who experiences the mecha-
H2: Classical dramas will generate sympathy responses nism of transportation, defined as "immersion into a text"
as a precursor to generating empathy responses. (Green and Brock 2000, p. 702; see also Gerrig 1994).
Whereas the more cognitive response of elaboration "leads
Hierarchical Ranking: Effects of Sympathy and Em- to attitude change via logical consideration and evaluation
pathy Responses on Ad Attitudes. Turning from hy- of arguments," the transportation process leads to persuasion
pothesis 2 to hypothesis 3, we now explore the subsequent through reduced negative cognitive responding, realism of
effect of sympathy and empathy on attitude toward the ad experience, and strong affective responses (Green and Brock
(A^d). It is clear from consumer advertising research that 2000, p. 702).
empathy will have a positive effect on A^d, because empathic Findings in the general emotion and advertising literature
sharing is by its nature a felt affective response. The ad- lend further support to the hierarchical concept. Stout and
vertising literature is consistent in reporting that affective colleagues (Stout et al. 1990; Stout and Leckenby 1986)
ad responses lead to positive A^d (e.g.. Brown, Homer, and have found that even though recognized emotions (sym-
Inman 1998; Edell and Burke 1987; Holbrook and Batra pathy) have a significant effect on positive attitudes to an
1987; Stayman and Aaker 1988). Moreover, findings also ad, personally felt emotions (empathy) have an even greater
yield evidence that sympathy responses have a positive re- infiuence. Findings in the more specific drama/empathy re-
lationship with ad attitudes. Aaker and Williams (1998) find search literature also corroborate the sympathy/empathy hi-
570 JOURNAL OF CONSUMER RESEARCH

erarchy, with empathy presumed to be the stronger response and measure the effects of classical and vignette television
to advertising dramas (Deighton et al. 1989). The reasoning commercials on viewers' empathy and sympathy responses.
is that higher empathy is likely to drive more positive ad In each experiment, participants watched a different set of
attitudes, because when viewers are absorbed or immersed classical and vignette drama advertisements and recorded
in a dramatic ad, they are more inclined to develop positive their sympathy, empathy, and attitude responses after each
attitudes toward the ad that encouraged this response. In ad.
fact. Boiler and Olson (1991) have called empathic projec-
tion onto advertising characters the heart of the dramatic ad
persuasion process. EXPERIMENT 1
In light of this multidisciplinary convergence, we propose Table 3 presents the eight television ads used in experi-
that responses in high-involvement situations will move ment 1. The advertisements were selected by the researchers
from sympathy to empathy in a process that elicits positive from locally recorded television ads in 1998 and 1999 and
A^j. Thus, we move forward from hypothesis 2, in which from Clio-award winning ads in 1997 and 1998. The authors
the assumption of sequentiality was treated, to hypothesis used Stem's differentiating criteria (1994) to categorize the
3, in which the infiuence on A^d is treated. Hypothesis 3 ads as classical or vignette dramas, including attributes such
draws from the same multidisciplinary theoretical base to as single versus multiple plots, chronological versus cyclical
propose that the response process in which sympathy re- progression, and central versus numerous unrelated char-
sponses fiow through empathy responses leads to positive acters. To ensure that the stimuli had face validity and to
effects on A^^. ascertain overall ad quality, a pretest was conducted in which
22 undergraduate participants were asked to evaluate the
stimulus ads.
H3: Sympathy ad responses will infiuence ad attitudes
Participants' evaluations confirm the categorization
via empathy ad responses.
scheme, with the four classical drama ads rated as having
significantly more classical attributes (measured with three
Effect on A^^. Hypothesis 4 combines hypotheses 2 and seven-point scales: the ad told a story; the ad had a begin-
3 to propose a four-stage path to positive A^d in which the ning, middle, and end; and the ad showed the personal evo-
classical drama ad form will have a positive effect on sym- lution of one or more characters, a = .79; M^^^^^-^^^^ =
pathy responses, which in tum will have a positive effect 5.77, M,i^„„„ = 4.02, F(l,151) = 113.87, p < .001) and
on empathy responses, which in tum will have a positive the four vignette ads rated as having significanUy more vi-
effect on ad attitudes, as shown in figure 1. gnette attributes (measured with three seven-point scales:
many unrelated scenes, many different characters, and no
SCALE DEVELOPMENT chronological order, a ^ .64; M,ig„„„ = 4.82, M,,,,,;,,, =
2.50, F(l, 151) = 223.56, p < .001). Both ad types are con-
Pretests sidered to be high quality, with the vignette dramas some-
what higher than the classical dramas (three seven-point
The introduction of sympathy and empathy into consumer scales: professional, high quality, and realistic, a = .75;
research requires measurement instruments for both con- M.i,„„„ = 5.58, M,,,,,,,,, = 5.45, F(l,151) = 9.15, p<
structs, necessitating the development of two new scales to .01). Thus, the set of eight stimulus ads has face validity as
test the hypotheses. Since prior research measured empathy clear exemplars of classical and vignette dramas.
alone on a single scale (Deighton et al. 1988), we needed
two unique scales to measure sympathy and empathy as
separate responses to advertisements. The scales are called Method
Ad Response Sympathy (ARS) and Ad Response Enipathy A total of 115 undergraduate students at a large, public
(ARE; see table 2). southwestern university participated in a one-hour experi-
A review of sympathy and empathy scales used in psy- mental task as part of a requirement for the introductory
chological research (Davis 1983; Mehrabian and Epstein marketing course. Before viewing any ads, the participants
1972) and consumer behavior research (Boiler et al. 1989) completed a questionnaire about prior attitudes toward prod-
yielded 20 items likely to refiect the degree to which con- ucts and brands (Prior A,,) shown in the ad stimulus set (plus
sumers experience empathy and sympathy while watching filler brands). The participants then viewed the eight ads
a commercial. To construct the most parsimonious item set, (table 3) in two randomly assigned orders. After viewing
we conducted two pretests in which 147 participants were each ad, participants rated A^j and the degree of sympathy
exposed to six different television ads and responses were and empathy evoked by each advertisement. Half of the
measured after each exposure. Based on a series of factor participants filled out sympathy/empathy responses first and
analyses, LISREL analyses, and ANOVA analyses, 10 ad ad attitudes second, and the other half did the reverse.
response items (five sympathy, five empathy) were selected
for the ARS and ARE scales. Variables. Ad attitudes were measured by three seven-
Next, we present two experiments that test the scales for point scale items anchored by very favorable/very unfavor-
unidimensionality, reliability, and empirical distinctiveness. able, very negative/very positive, and very bad/very good
SYMPATHY AND EMPATHY DRAMA AD RESPONSES 571

TABLE 2

AD RESPONSE SYMPATHY (ARS) AND AD RESPONSE EMPATHY (ARE) SCALE ITEMS AND
AVERAGE STANDARDIZED FACTOR LOADINGS

Experiment 1 Experiment 2

ARS items:
1. Based on what was iiappening in
tiie commercial, I understood what
the characters were feeiing. .99 .91
2. Based on what was happening in
the commercial, I understood what
was bothering the characters. .95 .99
3. While watching the ad, I tried to
understand the events as they
occurred. .61 .97
4. While watching the ad, I tried to
understand the characters'
motivation. .61 .96
5. i was abie to recognize the prob-
lems that the characters in the ad
had. .92 .94
ARE items:
1. Whiie watching the ad, I experi-
enced feeling as if the events were
realiy happening to me. .93 .63
2. Whiie watching the ad, I feit as
though I were one of the
characters. .95 .88
3. Whiie watching the ad, I felt as
though the events in the ad were
happening to me. .96 .89
4. Whiie watching the commercial, I
experienced many of the same
feeiings that the characters
portrayed. .81 .77
5. While watching the commercial, I
feit as if the characters' feelings
were my own. .88 .93
NOTE.—The following instructions were used for these statements: "For the television commercial you just saw,
please rate how descriptive each of the following statements is of how you personally reacted to this ad," measured
on a seven-point scale anchored by Not at all descriptive/Very descriptive.

(a = .95). Participants' sympathy and empathy responses to CEI = .97, and RMSEA = .08), indicating that the factor
the ads were measured by the straight average of the five- structure of the scales is robust across the eight ads. (The
item, seven-point scales developed above (OARS = -88 and chi-square statistic is significant [x^(240) = 484.68, p <
a^RE = -95; table 2). .001], which is not unusual given this test statistic's sen-
sitivity to sample size.) The average standardized factor
Covariates. Participants' prior attitudes toward the ad- loadings are given in table 2 (all f's > 3.73).
vertised products and brands were measured using a single
five-point scale anchored by very favorable/very unfavor- To provide evidence of discriminant validity between the
able. Ad order and questionnaire order are also included as two constructs, we ran a LISREL model constraining the
covariates in the ANOVA models. two scales to be perfectly correlated. The fit of the model
constraining the sympathy and empathy constructs' corre-
lation to be 1.0 is significantly worse than the unconstrained
Results model (x^(8) difference = 204.50, p < .001). Comparing
Scale Verification. Analysis of the 10 ad response the average variance extracted estimates for sympathy and
items shows that ARS and ARE are unidimensional, re- empathy to the square of the latent construct intercorrelation
liable, and empirically distinct. Confirmatory factor anal- results in every pair of constructs passing this test, indicating
ysis provides evidence of construct validity. The uncon- that sympathy and empathy are empirically distinct. The two
strained LISREL model, with eight groups representing combined measures used in subsequent analyses are signif-
each of the eight stimulus ads (and parameter estimates icantly but moderately correlated (r = .50, p< .001). Ei-
free to vary except the factor structure), fits according to nally, Cronbach alphas for the two scales indicate they have
three of the four most common fit indices (GEI = .91, solid internal consistency (a,y^p,,hy = .88, Oempathy = -95).
572 JOURNAL OF CONSUMER RESEARCH

Hypotheses lA and IB. The prediction that classical FIGURE 2


dramas will evoke more sympathy (hypothesis 1 A) and more
MODEL SUPPORTED BY EXPERIMENTS 1 AND 2
empathy (hypothesis IB) than vignette dramas is supported.
Results are assessed by a within-subject ANOVA model,
including the between-subjects covariates (ad and question- EXPERIMENT 1

naire order), a subject class variable. Prior A^ covariate, a


dummy variable for the advertised brand (nested in ad type),
and the dichotomous ad type variable (classical drama vs. Ad Type Sympathy Empathy Ad AtUtude
.45 .72 .40
vignette drama). We find that classical dramas evoke more
ad response sympathy than vignettes (M,,,,,,^^, = 5.26, OFI = .99, CFI = ,94. RMSEA = .07, x\ = 8.60 (p < .01)
^vigneue = 3.96, f ( l , 798) = 301.78, p < .001, w' = 0.57),
and they also evoke more ad response empathy
(A^ciassicai = 3.13, M,;^^,,^ = 2.65, F(l,798) = 36.88, p<
.001, w^ = 0.14). In addition, for both ad types, the sym- EXPERIMENT 2
pathy response is significantly higher than the empathy re-
sponse (?„,,„„„= 18.42, p < . 0 0 1 , f,,a,,ica, = 27.36, p< .23
.001). Thus, classical drama ads evoke more sympathy and
empathy responses, whereas vignettes evoke less of both, Ad Type Sympathy Empathy Ad AtUtude
.92 .72 .40
despite the pretest evaluation of the vignette ads as higher
quality.
GFI = 1.0, CFI = 1.0, RMSEA = .04, x\ = 4.22 (p > .10)
Hypothesis 2. Hypothesis 2 predicts that sympathy re-
sponses will precede empathy responses to advertising dra-
mas (i.e., an indirect path from ad type to sympathy to We again use LISREL mediation tests, in this case com-
empathy). We test this assertion with mediation tests con- paring the model in figure 1 to a model with an additional
ducted in LISREL, comparing the sequential model in figure direct path from ad response sympathy to ad attitudes (see
1 to a model with an additional direct path from ad type to fig. 2). We find that allowing the additional path significantly
empathy.' The model shown in figure 1 is supported by two improves the model fit (x(l) difference = 44.39, p<
of four common goodness of fit measures (GFI = .97, .001, GFI = .99, CFI = .94, RMSEA = .07). The direct
CFI = .94), but not by two others (RMSEA = .15, effect of ARS on A^j in a model that does not include
X^(3) = 52.99, p < .001). However, all three relationships empathy ((3 = .53, t = 17.37, p < .001) is also significant
in the model are positive and significant (marginal in the in the model that includes empathy (j3 = .24, t = 6.76,
case of ad type on ARS): ad type on ARS, x = -45, t = p< .001). However, the indirect effect of ARS on A^^, is
5.24 ip < .06); ARS on ARE, 0 = .72, t = 22.46 (p < also significant in the latter model (Z = 4.56, p < .001),
.05); ARE on A^,, 0 = .53, t = 21.36 (p < .05). providing evidence that the effect of sympathy on A^d is
The model that includes a direct path from dramatic partially mediated by empathy.
ad type to empathy fits better than the model with only
an indirect path from ad type to sympathy to empathy, Hypothesis 4. Hypothesis 4 combines hypotheses 2 and
based on the chi-square difference test and CFI score, al- 3 into one direct effects model incorporating all four stages
though it does not have a better GFI or RMSEA score shown in figure 1. We have tested this hypothesis in hy-
(x'(l) difference = 8.55, p < .01, CFI = .95, GFI = .87, pothesis 2 and hypothesis 3 above. As shown in figure 2,
RMSEA = .16). The direct effect of ad type on empathy the best-fitting model is the one that emerges from hypoth-
found in hypothesis 1 (0 = .10) remains significant in esis 3, which includes a significant direct path from sym-
the model that includes sympathy (ad type on ARE: j3 = pathy to A^d. Thus, hypothesis 4 is partially supported in
- . 2 3 , t — 2.93, p< .01), however the indirect path from that the indirect effect of ad type on sympathy on empathy
ad type to sympathy to empathy is significant (Z = 4.16, on A^d is significant (;8 = .24, Z = 5.01,p< .001; see table
p < .001; Kenny, Kashy, and Bolger 1998), providing some 4). Furthermore, comparing the model in figure 2 to a model
support for hypothesis 2.^ where sympathy and empathy are not related reinforces the
superiority of the figure 2 model (x^(2) = 394.07, p<
Hypothesis 3. Hypothesis 3 asserts that sympathy ad .001, GFI = .82, CFI = .47, RMSEA = .47; AIC =
responses will infiuence A^^ by way of empathy responses 323.22 vs. AIC model in fig. 2 = 24.55). Table 4 sum-
(i.e., an indirect path from sympathy to empathy to A^^). marizes the direct and indirect effects of the model in figure
2.
'The four steps shown in fig. 1 are included in all the models that test
mediation to avoid model saturation in the models that include both direct
and indirect effects. Discussion of Results
^The LISREL mediation results presented for hypothesis 2 and hypoth-
esis 3 in both experiments have all been replicated using the four-step This study empirically demonstrates that sympathy and
regression analysis approach recommended by Baron and Kenny (1986). empathy make separate contributions to emotional responses
SYMPATHY AND EMPATHY DRAMA AD RESPONSES 573

TABLE 3 (n = 239), or they do not respond at all on these dimensions


(n = 83). However, it appears that in very few instances
STIMULUS ADVERTISEMENTS FOR EXPERIMENT 1
will individuals experience high levels of empathy without
also experiencing high levels of sympathy (n — 5, with the
Brand Description of the ad Ad type
lowest sympathy score quite high at 3.4 and an average of
Kodak film Daddy's little girl gets Classical drama
3.7).
married, snapshots We find partial support for hypothesis 3 and hypothesis
of their past life 4 in that we find both an indirect effect of sympathy re-
together sponses on A^d, working sequentially through empathy re-
One-Step e.p.t. preg- Couple unsuccessful in Classical drama sponses, and also a direct effect, which may be a function
nancy tests attempt to get
pregnant of a number of factors, including cognitions arising from
Prudential's Living Heart transplant recipi- Classical drama the experience of sympathy and differences in the degree
Needs Benefit ent's life story to which viewers experience empathy responses to a par-
Weyerhaueser (en- After a first little girl re- Classical drama ticular advertising drama. We discuss the direct effect of
dorsing recycling) jects him, boy recy-
cles love note to
sympathy responses on A^d in greater detail in the dis-
second little girl cussion.
Apple computers Scenes of geniuses Vignette drama One limitation in this experiment is the dominance of
and famous individu- poignant and warm ads in the stimulus set to the exclusion
als who were consid- of other types such as funny ones. However, since many
ered independent
thinkers advertisements are designed to be funny as well as to evoke
Nike shoes Scenes of many TV Vignette drama positive feelings, we decided to replicate the experiment
viewers watching Mi- using another stimulus set that included drama ads with
chael Jordan dunk- humorous appeals. Thus, experiment 2 examines sympathy
ing in slow motion
Scenes of volunteers Vignette drama
and empathy responses to a wider array of advertising ex-
United Way
with recipients of aid, ecutions.
message of thank
you
Volvo automobiles Scenes of individuals Vignette drama EXPERIMENT 2
who believe Volvo
saved their lives This study uses a new set of eight advertisements that
include humor appeals and upbeat music as well as a new
group of participants to examine the effects of classical and
to advertising. Support for hypothesis 1 indicates that clas- vignette television commercials on sympathy and empathy
sical drama ads are better at eliciting both sympathy and responses.
empathy responses, compared to vignette ads, a finding con-
trary to Stem's prediction (1994), and suggests that vignettes Stimuli
may elicit responses on dimensions not measured in the
experiment, such as role projection, a state that requires the Table 5 describes the eight new television ads selected from
cognitive skills of role taking and perspective taking (Strayer the same ad database used in experiment 1. These ads were
1987). Additionally, hypothesis 1 results indicate that sym- classified by the authors as being either classical or vignette
pathy appears to be the "easier" emotional response to evoke dramas. We followed the same pretest procedure, using the
in an advertising context, which seems reasonable insofar same measures with a similar subject pool to ensure that the
as complete absorption in a 60-second advertisement (re- stimuli had face validity and to ascertain overall ad quality.
quired by empathy) may be more difficult to evoke than Thirty-three undergraduate participants rated the four classi-
recognition of the emotions (sympathy). cal ads as having significantly more classical attributes than
Findings for hypothesis 2 indicate that there is an indirect the four vignette ads (a = .87; A/,,^,,i,^, - 5.34, M,,^^^,, =
effect of dramatic ad type on empathy responses through 4.23, F(l,228) = 33.41, p < .001), and the four vignette
sympathy responses. The classical drama's chronologically ads as having more vignette attributes (a = .68; M^^^^^,,^ =
organized plot enables viewers to recognize the feelings of 3.41, M,,,,,i,,, = 2.84, F(l,228) = 17.33, p < .001). We
central characters (sympathy) as well as to share them (em- found a marginal difference in the quality of the ad types,
pathy), which lends support to the idea that classical drama with vignettes again slightly higher (a = .78; M^ign^K =
ads evoke empathy by way of sympathy. The pattem of 5.24, Af,,,,,i,^, = 5.04, F(l, 228) = 2.83,p < .10). Thus, the
results that emerges from performing midpoint splits on the ads were rated as high quality and were shown to have face
seven-point ARS and ARE scales confirms the sequential validity as clear exemplars of classical and vignette dramas.
process proposed by hypothesis 2. We find that, of 460 In addition, averaging across all participants and all ads,
observations (115 participants across four classical drama the new set of advertisements evoked significantly more
ads), ad viewers respond in some instances with high levels upbeat feelings than the first set (A/^pteat filings, experiment i =
of sympathy and empathy (n = 133), while in other in- 2.38, M^,,,^, ,,,,„,,, ^p,Hn,em 2 = 2.70, t = 4.82, p<.OOl).
stances, they respond sympathetically but not empathetically The new set also evoked significantly lower warm
574 JOURNAL OF CONSUMER RESEARCH

TABLE 4

DIRECT AND INDIRECT EFFECTS OF MODEL SHOWN IN FIGURE 2

Experiment 1 Experiment 2
Standardized Standardized
Hypothesized relationship estimate Z-value estimate Z-value

Ad type -• Sympathy .45 5.24" .92 11.20"*


Sympathy -• Empathy .72 22.46" .72 22.06***
Sympathy -i- A ^ .24 6.76" .23 6.29"*
Empathy -• A^^ .40 12.77" .40 12.99*"
Ad Type -• Sympathy -• Empathy .33 5.10" .67 9.99***
Sympathy -> Empathy -• A^^ .29 11.10" .29 11.19*"
Ad Type -• Sympathy -• Empathy .24 5.01" .48 9.32***
***p < .001

feelings than the first set (M^^.^ f^^^^gj experiment l = 2.94, p < .001] due to sample size, and the RMSEA = .13.)' The
^warm feelings, experiment 2 = 2.78, f = " 2 . 4 0 , ^ < .05), nOt SUr- average standardized factor loadings are given in table 2
prising in that humor and upbeat appeals are less likely to (all f's > 7.77).
evoke warmth than poignant appeals (using five-point scales We ran a LISREL model constraining the two scales to
based on Goodstein, Edell, and Moore [1990]). be perfectly correlated to provide evidence of discriminant
validity between ARS and ARE. The fit of the model where
the correlation between sympathy and empathy is con-
Method strained to be 1.0 is significantly worse than the correspond-
ing unconstrained model (x^(8)difference = 344.02, p <
Following the same method used in experiment 1, 92 .001). Comparing the average variance extracted estimates
undergraduate students from the same university partici- for sympathy and empathy to the square of the latent con-
pated in a one-hour experimental task as part of a require- struct intercorrelation results in every pair of constructs
ment for the introductory marketing course. Before viewing passing this test except one, indicating that sympathy and
any ads, the participants completed a questionnaire about empathy are empirically distinct. The two combined ARS
prior attitudes toward products and brands shown in the ad and ARE measures are significantly but moderately corre-
stimulus set (plus filler brands). The participants then viewed lated (r = .59,p< .001). Finally, Cronbach alphas for the
the eight ads (table 5) and, after viewing each ad, rated their two scales indicate they have solid internal consistency
feelings responses, ad attitudes, and the degree of sympathy ("sympathy = -90, Oempathy = -96).
and empathy evoked by each advertisement.
Hypotheses lA and IB. These hypotheses are as-
sessed by a within-subject ANOVA model, including the
Variables. As in experiment 1, participants' ad re- subject class variable. Prior Ag covariate, a dummy vari-
sponses were measured by the same scales: ARS (a = .89), able for the advertised brand (nested in ad type), and the
ARE (a = .96; table 2), and A^d (three items, a = .96). ad type variable (classical drama vs. vignette drama).
The results support the prediction that classical dramas
Covariates. Prior attitudes toward the advertised prod- will evoke more sympathy (hypothesis lA) and more em-
ucts and brands were measured using a single five-point pathy (hypothesis IB) than vignette dramas (ARS:
brand attitude scale anchored by very favorable/very unfa- M,,,,3ica, = 4.83, M„,^„,„, = 3.91, F(l,635) = 126.44, p <
vorable. .001, co^ = 0.41; ARE: M,,,,,,,,, = 3.82, M.,,„,,, = 3.28,
F(l,635) = 26.89, p < . 0 0 1 , CJ' = 0.12), replicating our
findings from experiment 1. In addition, for both ad types,
Results the sympathy response is significantly higher than the em-
pathy response (/,,^„„„ = 6.32, p<.00\, ?„„,,,,„ = 7.58,
Scale Verification. As in experiment 1, analysis of the p < .001). Experiment 2 confirms that classical drama ads
10 ad response items shows that ARS and ARE are uni- evoke more sympathy and empathy responses than vignette
dimensional, reliable, and empirically distinct. Confirmatory drama ads.
factor analysis provides evidence of construct validity: the
Hypothesis 2. Hypothesis 2 predicts that classical dra-
unconstrained LISREL model, with eight groups represent-
mas will generate sympathy as a precursor to empathy. We
ing each of the eight stimulus ads (and parameters free to
vary except the factor structure), fits according to two of
'The poor RMSEA score is likely due to the fact that ARS and ARE
the four most common fit indices (GFI = .92, CFI = .94). are related constructs, therefore, the nontarget loadings are not strictly zero
(The chi-square statistic is significant [x^(240) = 765.17, as specified in the model.
SYMPATHY AND EMPATHY DRAMA AD RESPONSES 575

TABLE 5 path, in this case from sympathy to ad attitudes (see fig. 2).
We find that the inclusion of the direct path significantly
STIMULUS ADVERTISEMENTS FOR EXPERIMENT 2
improves the model fit (x^(l) difference = 38.55, p <
.001, GFI = 1.0, CFI = 1.0, RMSEA = .04). The direct
Brand Description of the ad Ad type
effect of sympathy on A^a in a model that does not include
Classical drama
empathy (j3 = .52, r = 16.81, p < .001) is also significant
Polaroid Man changes mind
about coming home in the model that includes empathy (J3 = .23, t - 6.29,
for lunch when he p < .001), however, the indirect effect is also significant
finds photo of wife in (Z = 4.56, p < .001). Thus, consistent with our experiment
his briefcase 1 findings, the effect of sympathy on A^d is partially me-
e-toys Man buys firefly cage Classical drama
on-line after playing
diated by empathy.
in the yard with his
son
Hypothesis 4. Hypothesis 4 combines hypotheses 2 and
Levi's jeans Man and woman make Classical drama 3 into one direct effects model incorporating all four stages
eye contact in the el- shown in figure 1. We tested this hypothesis in hypothesis
evator, imagine their 2 and hypothesis 3 above, where we found that the best-
married life together
Lab researcher reports Classical drama
fitting model includes a significant direct path from sym-
Diet Pepsi
chimpanzee study a pathy to A^d, as shown in figure 2. Hypothesis 4 is partially
failure, while chimps supported by the significant indirect effect of ad type to
buy her a Diet Pepsi sympathy to empathy to AAd (|8 = .48, Z = 9.32, p <
Cisco Systems International children Vignette drama .001; see table 4). Furthermore, comparing this model to a
ask "Are you ready?"
for the internet model where sympathy and empathy are not related rein-
Nationsbank Scenes of Atlanta resi- Vignette drama forces the superiority of the fig. 2 model (x^(2) = 349.80,
dents on the day the p < .001, GFI = .84, CFI = .60, RMSEA = .43; AIC =
1996 Olympic site 295.35 vs. AIC model in fig. 2 = 20.21).
was announced
Dr. Pepper Scenes of a "salsa" Vignette drama
party featuring danc- Discussion of Experiment 2 Results
ing and Dr. Pepper
Midas Scenes of individuals Vignette drama Experiment 2 empirically replicates the findings of ex-
enjoying their (well-
periment 1. It reaffirms support for hypothesis 1 and extends
maintained) cars
the applicability of the ad response scales (ARS and ARE)
to a wider range of ad types. It also provides support for
hypothesis 2: sympathy responses mediate the effect of dra-
use LISREL mediation tests fo compare the sequential model
matic ad type on empathy responses. Midpoint splits on the
in figure 1 to a model with an additional direct path from
seven-point ARS and ARE scales across 372 observations
ad type to empathy. The model shown in figure 1 receives
(93 participants each viewed four classical drama ads) reveal
support according to two of four common goodness of fit
that in 148 instances, participants reported high sympathy
measures (GFI = .97, CFI = .95; however, RMSEA =
and high empathy; in 97 instances, low levels of both sym-
.13, •)^^(3) = 42.77, p < .001). All three relationships in this
pathy and empathy; and in 107 instances, high sympathy
model are significant and positive (ad type on ARS: j3 =
with low empathy. In only 20 cases did participants report
.92, t = 11.20, p<.05; ARS on ARE: 0 = .72, t =
high empathy with low sympathy (and of these, the lowest
22.06, p < .05; ARE on A^^: 0 = .53, t = 21.26, p < .05).
sympathy score was 2.4, with a mean of 3.5), supporting
The fit of the model that includes a direct path from ad
our sequential model that predicts that empathy should not
type to empathy is marginally superior to the sequential
occur without sympathy.
model in terms of the chi-square difference test and CFI
measure (x'(l) difference = 2.94, p<.lO, CFI = .96), Support is found for hypotheses 3 and 4, as we find ad
however, it is not better on the two other standard goodness response empathy partially mediates the effect of ad re-
of fit measures (GFI = .97, RMSEA = .16). Further, the sponse sympathy on ad attitudes. However, as in the case
significant direct effect of ad type on empathy found in of experiment 1, there exists a significant direct effect of ad
hypothesis 1 (|8 = .54) is only marginally significant in the response sympathy on A^j, in addition to the hypothesized
model that includes sympathy (fi = - . 1 5 , t = 1.72, p< indirect effect through ad response empathy. In sum, our
.10), while the indirect path is significant (Z = 8.73, p< overall pattem of results supports the relationships predicted
.001). Therefore, hypothesis 2 is supported: the effect of ad by the four-stage, sequential model, with the addition of the
type on empathy is mediated by sympathy. direct effect of ad response sympathy on ad attitudes.

Hypothesis 3. Hypothesis 3 proposes an indirect effect DISCUSSION


of sympathy responses on ad attitudes via empathy re-
sponses. We again use LISREL mediation tests to compare This study contributes to an expansion of the research
the model in figure 1 to a model with an additional direct domain of advertising responses by including both sympathy
576 JOURNAL OF CONSUMER RESEARCH

and empathy in a comprehensive system of emotional re- nition of the emotions of the characters in the ad may in-
sponses. Results indicate that the restored linkage can fluence viewers' cognitive responses to the ad, something
broaden consumer research about drama effects by allowing that we did not measure in our studies. For example, a
for the extension and modification of extant drama adver- sympathetic ad viewer may begin to think about whether a
tising theory. The development and testing of two new scales character's goals and motives are the same as his or her
(ARS and ARE) reveal a cohesive pattem of results across own, and this self-awareness may generate positive A^d on
two experiments and 16 drama advertisements, which sub- its own by eliciting role taking or projection (Strayer 1987).
stantiates the differentiation between sympathy and empathy In the context of studying feelings responses to advertising,
and the influence of each as emotional responses to classical Edell and Burke find that advertisements affect cognitive ad
advertising dramas. We consistently find that classical dra- judgments directly (1987) and indirectly through feelings
mas are better than vignettes at generating both sympathy responses (Burke and Edell 1989). An ad viewer experi-
and empathy responses, which confirms the assumption that encing sympathy may engage in more positive adjudgments
it takes well-developed characters and a linear plot to evoke and fewer negative ones, such as source derogations and
sympathy and empathy. We also find an indirect effect in counterarguments (Wright 1973), resulting in net positive
that classical dramas work through sympathy responses as ad attitudes.
a precursor to empathy responses. These findings reveal a Second, there may exist individual response differences
consistent pattem of results, with sympathy the first and in the sympathy/empathy sequence. That is, if the emotions
more easily achieved emotional response and empathy the and attitudes are unfolding in a sequential or logical causal
later response of some ad viewers for some ad dramas. chain progression as figure 1 shows, some viewers may not
In addition, interesting findings not predicted in our hy- reach the empathy stage for a particular classical drama ad.
potheses are the ones that show a direct as well as an indirect When an empathy response is not achieved, sympathy alone
effect of sympathy in the empathic process. Recall that in may have a positive effect on A^d. Although sympathy may
the indirect route, sympathy responses lead to empathy re- be less emotionally involving than empathy, it is, nonethe-
sponses that in turn lead to positive ad attitudes. This process less, an emotional response to a drama advertisement. This
supports the idea that both sympathy and empathy are stages explanation is speculative insofar as the covariance matrices
that an individual goes through in experiencing drama (De- used in the LISREL analysis of the empathic path to per-
lacroix [1927] 1953), progressing from early recognition of suasion collapse the data across ads and individuals: they
portrayed emotions to later absorption in sharing those emo- cannot show whether or not those instances where the ad
tions (Strayer 1987). However, our results indicate that there viewer does not reach the empathy stage of the sequential
is also a direct route in which sympathy responses alone process are the same instances where sympathy has a direct
enhance ad attitudes. effect on ad attitudes.
The direct route seems to exemplify the idea that some-
thing is better than nothing; that is, even though an empathy Limitations
response is preferred, a sympathy response alone also en-
hances A^d. In empathy research, even though sympathy is Possible limitations to our research conclusions might in-
not mentioned by name, Deighton and Hoch (1993) point clude the following. First, as is true of any cross-sectional
out that even quite dispassionate consumers can understand design, the experiments reported do not provide as strong
the intention and meaning of emotional responses com- evidence of the direction of causality between sympathy,
municated by displays of feelings (our sympathy). That is, empathy, and A^^ as we might have liked. We are certainly
whereas empathy with ads is an emotional response on the able to say that our proposed sequence is better supported
part of some consumers (Deighton and Hoch 1993, p. 278), by the data than alternative models in which empathy pre-
sympathy is also an emotional response on the part of others cedes sympathy, but were the experiments to unfold over
who are not empathically moved by a particular ad. Wells's time, for example, it would bolster our claim toward cau-
research (1989) reiterates this conjecture, pointing out that sality. Second, the experimental setting might be character-
structured classical dramas have tiie capacity to draw con- ized as unrealistic in that participants are forced to pay at-
sumers into comprehending and/or vicariously participating tention to ads that they might otherwise zap or screen out.
in a commercial. Advertising researchers (Aaker and Wil- However, in our defense, even though no laboratory setting
liams 1998; Stout et al. 1990) also find that recognized represents a real-life viewing situation, consumer behavior
emotions (our sympathy) have a significant effect on positive researchers have long used experiments to study theoretical
attitudes to an ad. In the context of the poignant (experiment concepts without contamination by noise inevitably found
1) and upbeat (experiment 2) feelings displayed by the char- in an externally valid real-world setting. A related limitation
acters in our stimulus ads, the recognition of the actors' is that the stimulus sets include only award-winning ads,
positive feelings should generate a positive response that assumed to be "good," which excludes possible altemative
directly enhances A^j responses to "bad" or "not-so-good" ads. Thus, the findings
cannot be generalized to ads that may evoke neither empathy
We propose that the positive effect of sympathy on ad nor sympathy, or to alternative consumer reactions such as
attitudes is a consequence of several mechanisms. First, the nonresponsiveness. Finally, our within-subjects design may
sympathetic experience may directly affect A^d, for recog- have created a common method variance problem attrib-
SYMPATHY AND EMPATHY DRAMA AD RESPONSES 577

utable to the experimental condition, in which all of the data Baron, Reuben M. and David A. Kenny (1986), "The Moderator-
was obtained from the same subjects at the same time, using Mediator Variable Distinction in Social Psychological Re-
the same questionnaire items repeatedly. Although we can- search: Conceptual, Strategic, and Statistical Considerations,"
Journal of Personality and Social Psychology, 55 (6),
not rule out this problem entirely, the LISREL results re-
1173-1182.
ported in this article have been replicated using more tra-
Boiler, Gregory W. (1990), "The Vicissitudes of Product Experi-
ditional repeated measures ANOVA analyses. ence: 'Songs of Our Consuming Selves' in Drama Ads," in
Advances in Consumer Research, Vol. 17, ed. Marvin E.
Directions for Future Research Goldberg, Gerald Gorn, and Richard W. Pollay, Provo, UT:
Association for Consumer Research, 621-626.
The provocative finding of a direct effect of sympathy in Boiler, Gregory W., Emin Babakus, and Jerry Olson (1989),
the empathic sequence requires additional investigation. One "Viewer Empathy in Response to Drama Ads: Development
possible explanation is that this direct effect may stem from of the VEDA Scale," Working Paper No. 402-89, Working
Series Collection, Fogleman College of Business and Eco-
the fact that some individuals do not reach the empathy
nomics, Memphis State University, Memphis, TN 38152.
stage, only reaching the sympathy stage, which in itself
Boiler, Gregory W. and Jerry Olson (1991), "Experiencing Ad
directly affects A^j. This explanation encourages further Meanings: Crucial Aspects of Narrative/Drama Processing,"
study of individual differences in empathic responsiveness, in Advances in Consumer Research, Vol. 18, ed. Rebecca H.
with some individuals tending to respond more empatheti- Holman and Michael R. Solomon, Provo, UT: Association
cally to media representations in general and, thus, to ad- for Consumer Research, 172-175.
vertisements in particular. Whereas some people easily lose Brown, Steven P., Pamela M. Homer, and J. Jeffrey Inman (1998),
themselves in movies, others remain detached no matter "A Meta-Analysis of Relationships between Ad-Evoked Feel-
what is on the screen. It seems likely that only those indi- ings and Advertising Responses," Journal of Marketing Re-
viduals who are predisposed to be empathic responders will search. 35 (February), 114-126.
proceed along the sympathy/empathy path in our sequential Buber, Martin ([1937] 1970), / and Thou, trans. Walter Kaufmann,
model. However, so little is known about the relationship New York: Scribner's.
Burke, Marian Chapman and Julie A. Edell (1989), "The Impact
between individual differences in media and advertising re-
of Feelings on Ad-Based Affect and Cognition," Journal of
sponses that further research is essential.
Marketing Research, 26 (February), 69-83.
Another untouched area for future research is the ex- Davis, Mark H. (1983), "The Effects of Dispositional Empathy on
amination of sympathy/empathy unresponsiveness: that is, Emotional Reactions and Helping: A Multidimensional Ap-
consumers' failure to respond with either sympathy or em- proach," Journal of Personality, 51 (June), 167-184.
pathy to dramatic advertising appeals. Based on the premise Deighton, John and Stephen J. Hoch (1993), "Teaching Emotion
that both sympathy and empathy are desired by advertisers, with Drama Advertising," in Advertising Exposure, Memory,
it seems reasonable to assume that negative responses are and Choice, ed. Andrew A. Mitchell, Hiilsdale, NJ: Lawrence
undesirable. Further study is required to examine the neg- Erlbaum Associates, 261-282.
ative states of nonempathy and antipathy and to extend the Deighton, John, Daniel Romer, and Josh McQueen (1989), "Using
postulation of linked constructs to unresponsiveness to ads. Drama to Persuade," Journal of Consumer Research, 16 (De-
In this regard, just as sympathy and empathy differ from cember), 335-343.
Delacroix, Henri ([1927] 1953), "Varieties of Aesthetic Experi-
each other, so too, we predict, would their oppositional
ence," in The Problems of Aesthetics, ed. Eliseo Vivas and
states. Nonempathy, the opposite of empathy, is defined as Murray Krieger, New York: Holt, Rinehart & Winston,
a lack of engagement with a stimulus, and antipathy, the 279-284.
opposite of sympathy, is defined as a hostile response to a Edell, Julie A. and Marian Chapman Burke (1987), "The Power
specific ad and/or to advertising in general. Future research of Feelings in Understanding Advertising Effects," Journal
that measures the nature, degree, and power of avoidance of Consumer Research, 14 (December), 421^33.
versus antipathy responses can increase understanding of Eisenberg, Nancy and Paul A. Miller (1987), "Empathy, Sympathy,
emotional responses or the lack thereof to advertisements and Altruism: Empirical and Conceptual Links," in Empathy
and attitudes toward them. and Its Development, ed. Nancy Eisenberg and Janet Strayer,
Cambridge: Cambridge University Press, 292-316.
Eisenberg, Nancy and Janet Strayer (1987), "Critical Issues in the
[Received December 1999. Revised August 2002. David Study of Empathy," in Empathy and Its Development, ed.
Gien Mick served as editor and John Deighton served as Nancy Eisenberg and Janet Strayer, Cambridge: Cambridge
associate editor for this article.] University Press, 3-16.
Escalas, Jennifer Edson, Marian C. Moore, and Julie A. Edell
(forthcoming), "Fishing for Feelings: A Hook Helps !"yoMrna/
REFERENCES of Consumer Psychology.
Aaker, Jennifer L. and Patti Williams (1998), "Empathy versus Fergusson, Francis (1961), Aristotle's Poetics, trans. S. H. Butcher,
Pride: The Influence of Emotional Appeals across Cultures," New York: Hill & Wang.
Journal of Consumer Research, 25 (December), 241-261. Gerrig, Richard J. (1994), "Narrative Thought?" Personality and
Aumont, Jacques, Alain Bergala, Michel Marie, and Marc Vernet Social Psychology Bulletin, 20 (6), 712-715.
([1983] 1992), Aesthetics of Film, trans, and revised Richard Goodstein, Ronald C , Julie A. Edell, and Marian Chapman Moore
Neupert, Austin: University of Texas Press. (1990), "When Are Feelings Generated? Assessing the Pres-
578 JOURNAL OF CONSUMER RESEARCH

ence and Reliability of Feelings Based on Storyboards and titudes, and Memory," Joumal of Consumer Research, 18
Animatics," in Emotion in Advertising: Theoretical and Prac- (4), 411^24.
tical Explorations, ed. Stuart J. Agres, Julie A. Edell, and Morrison, Karl F. (1988), "I Am You": The Hermeneutics of Em-
Tony M. Dubitsky, Westport, CT: Quorum Books, 175-193. pathy in Western Literature, Theology, and Art, Princeton,
Green, Melanie C. and Timothy C. Brock (2000), "The Role of NJ: Princeton University Press.
Transportation in the Persuasiveness of Public Narratives," Ray, Michael L. (1973), Marketing Communications and the Hi-
Joumal of Personality and Social Psychology, 79 (5), erarchy of Effects, Cambridge, MA: Marketing Science
701-721. Institute.
Holbrook, Morris B. and Rajeev Batra (1987), "Assessing the Role Shipley, Joseph Twaddell (1931), The Quest for Literature: A Sur-
of Emotions as Meidators of Consumer Responses to Ad- vey of Literary Criticism and the Theories of the Literary
vertising," Joumal of Consumer Research, 14 (December), Forms, New York: R. R. Smith.
404^20. Shweder, Richard A. (1994), "'You're Not Sick: You're Just in
Hume, David ([1739] 1968), A Treatise of Human Nature, ed. L. Love': Emotion as an Interpretive System," in The Nature of
Shelby-Bigge, London: Oxford University Press. Emotion: Fundamental Questions, ed. Paul Ekman and Rich-
Kenny, David A., Deborah A. Kashy, and Niall Bolger (1998), ard J. Davidson, New York: Oxford University Press, 32-44.
"Data Analysis in Social Psychology," in Handbook of Social Stayman, Douglas M. and David A. Aaker (1988), "Are All the
Psychology, ed. Daniel T Gilbert, Susan T Fiske, and Gardner Effects of Ad-Induced Feelings Mediated by AA^?" Joumal
Lindzey, 233-265. of Consumer Research, 15 (3), 368-374.
Krugman, Herbert E. (1966-1967), "The Measurement of Adver- Stem, Barbara B. (1994), "Classical and Vignette Television Ad-
tising Involvement," Public Opinion Quarterly, 30 (Winter), vertising Dramas: Structural Models, Formal Analysis, and
583-596. Consumer Effects," Joumal of Consumer Research, 20
Langfeld, Herbert Sidney ([1920] 1967), The Aesthetic Attitude, (March), 601-615.
Port Washington, NY: Kennikat. Stout, Patricia A., Pamela M. Homer, and Scott S. Liu (1990),
Lavidge, Robert and Gary Steiner (1961), "A Model for Predictive "Does What We See Influence How We Feel? Felt Emotions
Measurements of Advertising Effectiveness," Joumal of Mar- versus Depicted Emotions in Television Commercials," in
keting, 25 (October), 59-62. Emotion in Advertising: Theoretical and Practical Explora-
Lipps, Theodor (1897), Raumasthetik und Geometrische-optische tions, ed. Stuart J. Agres, Julie A. Edell, and Tony M. Dub-
Tauschungen (Aesthetics of space and geometric-illusions), itsky, New York: Quorum Books, 195-210.
Leipzig: J. A. Barth. Stout, Patricia A. and John D. Leckenby (1986), "Measuring Emo-
(1903-1906a), Aesthetik: Psychologie des Schonen und tional Response to Advertising," Joumal of Advertising, 15
der Kunst (Aesthetics: The psychology of beauty and art), (4), 35-42.
Hamburg: L. Voss. Strayer, Janet (1987), "Affective and Cognitive Perspectives on
- ([1903-1906b] 1926), Psychological Studies, 2d ed., trans. Empathy," in Empathy and Its Development, ed. Nancy Ei-
Herbert C. Sanborn, Baltimore: Williams & Wilkins. senberg and Janet Strayer, Cambridge: Cambridge University
Mallgrave, Harry Francis and Eleftherios Ikonomou (1994), "In- Press, 218-244.
troduction," in Empathy, Form, and Space: Problems in Titchener, Edward Bradford (1909a), Elementary Psychology of
German Aesthetics, 1873-1893, ed. and trans. Harry Francis the Thought Processes, New York: Macmillan.
Mallgrave and Eleftherios Ikonomou, Chicago: University of (1909b), A Text Book of Psychology, New York: Macmillan.
Chicago Press, 1-75. Vischer, Robert ([1873] 1994), "On the Optical Sense of Form: A
McGuire, William J. (1978), "An Information-Processing Model Contribution to Aesthetics," in Empathy, Form, and Space:
of Advertising Effectiveness," in Behavioral and Manage- Problems in German Aesthetics, 1873-1893, ed. and trans.
ment Science in Marketing, ed. Harry L Davis and Alvin J. Harry Francis Mallgrave and Eleftherios Ikonomou, Chicago:
Silk, New York: Ronald, 156-180. University of Chicago Press, 98-173.
Vivas, Eliseo and Murray Krieger, eds. (1953), "Section 5: The
McQuarrie, Edward F. and David Glen Mick (1996), "Figures of
Aesthetic Experience," in The Problems of Aesthetics, ed.
Rhetoric in Advertising Language," Joumal of Consumer Re-
Eliseo Vivas and Murray Krieger, New York: Holt, Rinehart
search, 22 (March), 424- 438.
& Winston, 277-279.
Mehrabian, Albert and N. Epstein (1972), "A Measure of Emo-
Wells, William D. (1989), "Lectures and Dramas," in Cognitive
tional Empathy," Journal of Personality, 40 (September),
and Affective Responses to Advertising, ed. Patricia Cafferata
522-533.
and Alice M. Tybout, Lexington, MA: Lexington Books,
Mercer, Philip (1972), Sympathy and Ethics: A Study of the Re-
13-20.
lationship between Sympathy and Morality with Special Ref-
Wispe, Lauren (1986), "The Distinction between Sympathy and
erence to Hume's Treatise, Oxford: Clarendon.
Empathy: To Call Forth a Concept, a Word Is Needed," Jour-
Metz, Christian (1974), Language and Cinema, trans. Donna Jean nal of Personality and Social Psychology, 50 (August),
Umiker-Sebeok, The Hague: Mouton. 314-321.
Mick, David Glen (1987), "Toward a Semiotic of Advertising Story (1987), "History of the Concept of Empathy," in Empathy
Grammars," in Marketing and Semiotics: New Directions in and Its Development, ed. Nancy Eisenberg and Janet Strayer,
the Study of Signs for Sale, ed. Jean Umiker-Sebeok, Berlin: Cambridge: Cambridge University Press, 17-37.
Walter de Gruyter, 249-278. Wright, Peter L. (1973), "The Cognitive Processes Mediating Ac-
(1992), "Levels of Subjective Comprehension in Adver- ceptance of Advertising," Joumal of Marketing Research, 10
tising Processing and Their Relations to Ad Perceptions, At- (Februrary), 53-62.

You might also like