Download Overcoming the Illusion of Will and Self-Fabrication

Overcoming the Illusion of Will and Self-Fabrication:
Going Beyond Naïve Subjective Personal Introspection to an
Unconscious/Conscious Theory of Behavior Explanation
Arch G. Woodside
Boston College
June 2005
June 12, 2005 7:15 PM FINAL
Correspondence regarding this article should be sent to Arch G. Woodside, Boston
College, Carroll School of Management, Department of Marketing, 450 Fulton Hall, 140
Commonwealth Avenue, Chestnut Hill, MA 02467; telephone / fax: +1-617-552-3069 /
6677 (email: [email protected]).
2
Overcoming the Illusion of Will and Self-Fabrication:
Going Beyond Naïve Subjective Personal Introspection to an
Unconscious/Conscious Theory of Behavior Explanation
ABSTRACT
Naïve subjective personal introspection includes the failure to recognize the confirmability
of one’s own attitudes and personal meanings learned explicitly from self-examining such
topics and explaining one’s own behavior. Unconscious/conscious theory of behavior
explanation follows from unifying the research on unintended thought-behavior with folk
explanations of behavior. This article describes advances in research confirming own
attitudes and personal meaning and suggests the need for applying multiple methods to
overcome the fundamental attribution error, inherent cultural prejudices, and the general
bias toward self-fabrication. The discussion is valuable for achieving a deep understanding
of how customers think, advancing from subjective to confirmatory personal introspection,
and understanding the need to apply research tools useful for enlightening knowledge and
overcoming the inherent bias within subjective personal introspection.
3
Overcoming the Illusion of Will and Self-Fabrication:
Going Beyond Naïve Subjective Personal Introspection to an
Unconscious/Conscious Theory of Behavior Explanation
In reviewing relevant literature Woodside (2004b) provides a series of telling
propositions regarding subjective and confirmatory personal introspection. The following
points reflect and extend these propositions. (1) The dominant logic in consumer research
includes asking questions that require some amount of subjective personal introspection by
a respondent—whether or not a separate individual is asking the question or the informant
both asks herself and answers the question and whether or not the informant answers the
question face-to-face with a researcher, reads the question in a survey, or ponders the
issue alone. (2) Because most thinking occurs unconsciously (Bargh 2002; Wegner, 2002;
Zaltman 2003) and the informant has limited access to her own unconscious thinking, the
informant is able to retrieve, interpret, and report (to herself and others) only a limited
amount of relevant knowledge and insight when answering questions and pondering a
specific topic. (3) The use of additional research tools (beyond self-interviewing and
meditation) aids in surfacing and confirming/refuting both events and personal meanings
that the informant otherwise concludes to be accurate answers reflecting her own prior,
and/or currently held, beliefs and attitudes.
The comment on Woodside (2004b) by Gould (2005) is useful for stimulating an
elaboration here on these three points—an elaboration leading to a call for crafting a
unified theory of how the mind explains behavior unconscious and consciously. Following
this introduction, section one of the present article describes the crux of Gould’s comment
4
and demonstrates its fallacy. Section two elaborates on advanced attribution errors, that
is, (1) personal denial of committing the fundamental attribution error, the illusion of will,
including unawareness that implicit (automatic, unconscious) meaning is relevant for
subjective personal introspection (SPI) and (2) disregard and denial of the usefulness of
member checks (independent assessments of an SPI for accuracy and completeness) and
other tools (e.g., implicit association tests (see Brunel, Tietje & Greenwald, 2004; Maison,
Greenwald, & Bruin 2004), the forced metaphor elicitation technique for uncovering
unconscious meaning, see Woodside, 2004b)—tools permitting access to unconscious
personal meaning that otherwise remain inaccessible to the researcher-informant. Section
three explicates the folk-conceptual theory of behavior explanation (see Hilton 1990;
McClure 2002; Malle, 1999; Malle, 2004a, 2004b; Malle & Knobe 1997) and emphasizes
this theory’s relevancy to SPI research and theory. Section four emphasizes the critical
importance of SPI and the use of mixed research designs in SPI research for (1) theory
building in consumer research, (2) deepening individual and group sense making, and (3)
aiding in preventing unfair, bad, and downright dangerous decision-making (e.g., Gaither,
2002; Kozak, 1996)—reasons that substantially extend Gould’s (1995) defense of SPI in
responding to Wallendorf and Brucks (1993) criticism of the method. The article closes
with conclusions and implications for theory and research.
DEEPENING SUBJECTIVE PERSONAL INTROSPECTION
Gould (2005)’s comment is incorrect in several aspects but the more important
flaw is its ignoring the main point in Woodside (2004b): Because subjective personal
introspection pervades consumer research and both the researcher and informant (whether
or not both are the same individual or different persons) are unable to examine relevant
5
unconscious and conscious data using only one method (e.g., Buddhist meditation),
applying mixed methods designs (Tashakkori & Teddlie, 1998) is essential for correcting
event memory failures and overcoming fundamental and advanced attribution errors.
Here is the essence of Gould’s (2005) comment:
Thus, when people tell stories about their lives, confirmation could
mean that the researcher would seek a member check on the
accuracy of accounting for certain events. For example, did the
individual engage in a certain behavior or not? But there is no
member checking the meaning a person assigns in the story. This is
her own introspective story. People can compare stories with one
another but one does not generally ‘confirm’ one’s own story in
terms of personal meaning. When Holbrook [(2005)] is asking his
mother about certain remembrances, he is engaging in retrospection
in some sort of confirmation through his mother. But again, he is
not confirming the personal meanings of his experiences or of his
story. Thus, Woodside makes what is a major flaw [that] underlies
much of his argument in establishing the need for confirmation; he
conflates event-memory checking where confirmation may be useful
with meaning and experiential member checking where it is not.
A seemingly minor point: Because Holbrook (2005, p. 48) does not see the need,
his report does not include asking his mother anything, without him asking she volunteers.
“I often find myself musing over what rampant lack of self-confidence would encourage a
mechanical reliance on such self-imprisoning safeguards and such vision-restricting
6
formulas [e.g., member checks]. (The closest I have come to a member check has been
inviting my 91-year-old mother to attend a conference where I presented some of this
material and dutifully making revisions in my comments as she called out occasional
corrections from the audience.)”
Holbrook’s view is an example of an advanced overconfidence bias applied to
interpreting both factual accuracy and meaningfulness in his SPI. His SPI research would
benefit from replacing his musing about “rampant lack of self-confidence” with a more
mindful, complex model of explicit-implicit thinking (cf., Weick & Sutcliffe, 2001). His
use of a triangulation of data collection methods (i.e., his SPI, examining 2,300
photographs taken by his grandfather, and his mother’s member check) counters his
disparaging remarks about using additional data collection methods beyond SPI.
Overconfidence and self-fabrication rather than a rampant lack of self-confidence
are the dominant human tendencies in explaining our own behavior to ourselves and others
(cf., Langer, 1975; Wegner, 2002; Wilson 2002). Overconfidence bias in cognitive
science (e.g., Girgerenzer, 2000; Gilovich, 1991; Gilovich, Griffin, & Kahneman, 2002;
Lichtenstein, Fischhoff, & Phillips, 1982) refers to the human tendency to overestimate the
accuracy of one’s own answers. For instance, the study by Lichtenstein et al. (1982) gave
participants questions such as “Absinthe is (a) a precious stone or (b) a liqueur”; they
chose what they believed was the correct answer and then were asked for a confidence
rating in their answers, for example, 90% certain. When people said they were 100%
certain about individual answers, they had in the long run only about 80% correct answers;
when they were 90% certain, then had in the long run only 75% correct answers, and so
on. Lichtenstein et al. (1982) identify such discrepancy as the overconfidence bias and
7
explain its occurrence by general heuristics in memory search, such as confirmation biases,
or general motivational tendencies, such as the illusion of validity (see Girgerenzer, 2000).
Advanced overconfidence bias (AOB) goes beyond such cognitive science reports
of overconfidence bias. AOB includes the implicit—and often inaccurate—assumption
that the person holds unconsciously the meaning she reports in a SPI. The main point here
relates to McClelland, Koestner, and Weinberger’s (1992) evidence and conclusion
favoring dual (explicit and implicit) motives and goals—not that SPI reports are invalid
but that SPI and alternative introspection data collection methods tap different levels of
meaning and explanations. Not recognizing this possibility and not using mixed method
designs to acquire both explicit and implicit personal introspection data is committing
AOB.
Implicit motives are needs that people acquire in childhood that have become
automatic and nonconscious. Self-attributed motives are people’s conscious theories
about their needs that may often differ from their nonconscious needs. The picture
McClelland and his colleagues paints is of two independent explanatory systems that
operate in parallel and influence different types of behaviors. “In our terms, the adaptive
unconscious and the conscious explanatory system each has its own set of needs and
motives that influence different types of behaviors” (Wilson, 2002). The SPI reports in
the consumer research literature (e.g., Gould, 1991; Holbrook 2005) are akin to explicitly
interviewing yourself without recognizing the usefulness of applying implicit interviewing
methods to capture unconscious meanings and motives.
FUNDAMENTAL AND ADVANCED ATTRIBUTION ERRORS
8
The fundamental attribution error (Ross & Nisbett, 1991; Wilson, 2002) refers to
people overlooking situational influences on their actions and inferring that they acted on
the basis of their own internal states—inferring internal states via explicit interpretation
without the use of tools for learning implicit thinking. Choi, Nisbett and Norenzayan
(1999) demonstrate the people in Western cultures are especially prone to the fundamental
attribution error, and that people in East Asian cultures are less prone. The advanced
attribution error (AAE) includes denial that the fundamental attribution error is relevant
personally and that applying additional tools (a mixed methods research strategy) will help
overcome illusion of will and cultural bias that occur automatically during SPI.
Bargh, Chen, and Burrows (1996) demonstrate how situation-message treatments
(i.e., marketing manipulations) can influence behavior directly without affecting
participants’ introspection. These researchers had college participants fill out a scrambledsentence that included words such as “wrinkled, gray, retired, wise,” and “old.” These
participants were thus primed with the stereotype of an old person, whereas other
participants in the study did not receive this version of the test. As each participant left
the experiment room, the person’s gait was measured surreptitiously. The individuals who
had been led to think about senior citizens walked more slowly than did those not primed
with this thought. The idea of the action arose from the stereotype and so influenced the
behavior directly, apparently without conscious will. Extensive postexperimental
interviews suggested that the participants were not particularly conscious of the aged
stereotype after the experiment. And even if they were, they were certainly unaware that
this might suggest they should walk at a different speed, or for that matter that their
9
walking speed was being assessed. Yet merely thinking of the kind of person who walks
slowly seemed to be sufficient to induce shuffling (Wegner, 2002, p. 128).
Gladwell (2005) offers several examples of how conscious and unconscious
thinking and deciding in the same person often diverge and conflict. The winning
trombone-playing performance of Abbie Conant, a professional concert musician, in a
behind-a-screen competitive audition is a case in point. Ms. Conant played sixteenth in the
thirty-three candidate audition. Her playing was so outstanding that the Philharmonic
music director, Sergiu Celibidache, cried out, “That’s who we want!” The competition
was stopped and the remaining seventeen players sent home.
There were two more rounds of auditions. Conant passed both
with flying colors. But once Celibidache and the rest of the committee saw
her in the flesh, all those long-held prejudices began to compete with the
winning first impression they had of her performance. She joined the
orchestra, and Celibidache stewed. A year passed. In May of 1981, Conant
was called to a meeting. She was demoted to second trombone, she was
told. No reason was given. Conant went on probation for a year, to prove
herself again. It made no difference. “You know the problem,”
Celibidache told her. “We need a man for the solo trombone.” (Gladwell,
2005, p. 247).
Conant took the case to court and over the next 13 years she won several rounds of
battles (in courtrooms and additional auditions) including reinstatement as first trombone
and pay equal to male colleagues (Gladwell, 2005).
10
Might reports of culture-trained automatic thinking differ in ways unrecognized in
existing SPI reports (e.g., Gould, 1991; Holbrook, 1986, 1995, 2003, 2005)? Would the
use of additional tools in a mixed methods design (e.g., screened and unscreened
auditions) sometimes confirm and sometimes disconfirm meanings expressed in explicit
thinking reports? In two studies Brunel, Tiethe, and Greenwald (2004) provide evidence
of such confirmation and disconfirmation within consumer research contexts. Their
studies compare participants’ explicit attitude reports with implicit association test (IAT)
findings. Based on computer-mediated response latency measurement protocols, the IAT
measure is computed by comparing the relative response times associated with several
categorization tasks (see Brunel et al. for details). Study 1 reports high efficacy of the
IAT as a measure of brand attitudes and brand relationship strength toward Macintosh and
Microsoft Windows-based PC machines: under conditions in which participants were not
expected to hide their beliefs, explicit brand attitudes were strongly correlated to implicit
attitudes and implicit brand relationship, thereby validating the IAT for brand evaluation.
Study 2 demonstrates no significant differences between explicit attitudes toward ads with
White spokespersons compared to ads with Black spokespersons suggesting that at the
explicit level, participants do not exhibit racial preferences; however, implicit measures of
attitude toward ads reveal strong preference for ads containing White versus Black
spokespersons. Brunel et al. (2004) conclude, “Consumers’ associative brand networks
may include concepts and associations that a consumer either cannot or will not report [to
others or themselves], but which may surface through the IAT.”
A number of additional studies examine the effect of introspection, or thinking
about reasons, on attitudes, judgments, and choices (e.g., Levine, Halberstadt, &
11
Goldstone, 1996; Simonson & Nowlis, 2000; Wilson, Hodges, & LaFleur, 1995). When
people are asked to explain their attitudes (or choices), they tent to focus on a subset of
the reasons that would otherwise (without the need to explain) influence their attitudes,
particularly reasons that are verbalizable, accessible, plausible, and/or self-enhancing.
Having to provide reasons or introspect can affect choices, because provided reasons are
typically only a subset of the factors that would otherwise influence preferences.
FOLK-CONCEPTUAL THEORY OF MIND AND BEHAVIOR EXPLANATION
Heretofore, the consumer research literature does not include the unique and
valuable advances in the literature of how people explain their own behavior (i.e., SPI).
Malle (1999, 2004) summarizes this body of work and develops the “folk-conceptual
theory” of mind and behavior explanation (folk model, for short). “It is not, however,
ordinary people’s own theory of explanation (they probably don’t have one), but rather a
genuine scientific theory” (Malle, 2004, p. 236). Malle points out that prior attribution
theory focuses introspection on people allegedly classifying causes of effect outcomes into
two major categories: person and situation causes—greatly simplifying the possible
conceptual framework in which explanations are embedded. Similarly, Gould’s (1991)
typology of energy states focuses on a small subset of human explanation—“bodily felt
experience of everyday consumption” (p. 205) rather than representing a sophisticated
folk model of mind and behavior.
The folk model categorizes behavior explanations into two major modes of
explanation—reason and cause—as well as two minor modes—causal history reasons
(CHR) and enabling factors (EF). Reason explanations are people’s explanations of an
intentional behavior that cite the agent’s reasons for acting that way; cause explanations
12
are people’s explanations of an unintentional behavior that cite the causes that brought
about the behavior.
Causal history of reason explanations provide an explanatory link between reasons
and their own causal history, citing factors that preceded and thus brought about the
reasons for an action. These explanations literally describe the causal history of reasons,
which could lie in childhood, in cultural training, in personality traits, or in a situational
cue that triggered a particular desire (Malle, 1999). Without direct reference to the folk
model literature, Allen (2002) demonstrates the dominance of causal history of reason and
enabling factor explanations in his “fits-like-a-glove (FLAG) framework choice of
postsecondary education. Similarly, Holbrook’s (2005) SPI application of an eight-cell
value typology, extrinsic-intrinsic, self-other oriented, and active-reactive provides a
causal history of reason explanation—rich in value interpretation but very narrow in
coverage of behavior explanation.
Enabling factors explanations refer to the agent’s skill, efforts, opportunities, or to
removed obstacles (see McClure & Hilton, 1997; Turnbull, 1986); these explanations take
it for granted that the agent had an intention (and reasons) to perform the behavior; what
they try to clarify is how it was possible that this behavior was in fact performed” (Malle
1999, p. 31). Enabling factor explanations only explain the action’s occurrence—they
cannot be sued to explain why the agent formed the intention in the first place (which is
what reason explanations do). For example, “How come John aced the exam?”—He’s a
stats whiz.” Such enabling factor explanations refer to the agent’s skills, efforts,
opportunities, or to removed obstacles Malle, 1999, p. 31; also, McClure & Hilton, 1997;
13
Turbull, 1986). Gould’s (1991) eight cell typology of energy states: tense-calm,
energized-tired, absorbed-not absorbed is an enabling-factor explanations framework.
Malle (2004) contends that two broad motivations for explaining behavior exist:
finding meaning (i.e., sense making (cf. Weick 1995)) and managing social interaction.
These two broad motivations correspond to the two forms in which explanations exist in
the world: as cognitive representatives and as communicative acts.
Strong self-serving biases tend to occur in behavior explanations that favors selffabrication over self-revelation (Wilson, 2002). On one level of analysis (type of behavior
explained), explanations of actions are most likely to enhance or diminish one’s self-image
because they are observable, and hence accessible for evaluation by others, and intentional
(thus fully subject to either praise or blame). Conversely, experiences are unobservable
(thus difficult to evaluate by others) and unintentional (hence easier to excuse from
responsibility), so explanations of experiences should be at least susceptible to self-serving
biases. On another level of analysis (mode of explanation), self-serving actors should
explain their positive behaviors with reasons (implying intentionality) and their negative
behaviors with causes (implying lack of intentionality) because intentionality intensifies
praise and blame. On a third level of analysis, citing situation causes for negative
behaviors reduces blame and citing person causes increases blame, both when using cause
explanations and when using enabling factor explanations (see Malle, 1999; Malle &
Bennett, 1988; Weiner, 1995).
Examining the preconditions (e.g., meta thinking by the actor of her perceived
intentionality of a behavior—“did I really intend to do that?) affects which behavior
explanation category most likely applies to explain behavior. If the self-informant
14
perceives that the behavior was unintentional, situation/person cause explanations most
often occur. If perception is that the behavior was intentional, belief/desire reason
explanations most frequently occur and CHR and EF explanations also occur but with less
frequency.
Alternative framing of the behavior explanation issue affects the category applied
in SPIs. “How was this possible?” is a frame that increases EF explanations that otherwise
are rare when the framing is a motivationally biased question—“Why?” or “What for?”
Malle et al. (1998) supports this hypothesis, finding that enabling factor explanations
occur 4 to 12 times more frequently in response to a “How possible?” question than in
response to any other explanatory question. Moreover, enabling factor explanations
should be frequent when the behavior is difficult to perform but rare when the behavior is
easy to accomplish (McClure & Hilton, 1997; Malle, 1999).
Exhibit 1 illustrates how different ways of framing issues influence behavior
explanations. For example, several researchers (Becker, 1998; Thompson, Locander, &
Pollio 1989; Woodside 2004a) caution against using “Why?” framing questions for several
reasons, for example, the question implies intention when the behavior occurred
unintentionally; informants, even self-informants, become defensive because “Why?”
requires an answer that makes sense, one that does not reveal logical flaws and
inconsistencies; “How?” questions give people more leeway, are less constraining, and
encourage telling a story that includes a chain of events, thinking, and evaluating.
However, Becker (1987, p. 60) emphasizes an important exception to his condemnation of
“why” questions: sometimes the self-informant or other researcher want to know, exactly,
what kinds of reasons the informant gives for what she has done or think she might do as
15
part of a description that guides thinking. But, in his study investigating choice of
postsecondary education, Allen (2002) provides examples of informants finding great
difficulty in answering “why” questions but their difficulty vanishes quickly when
providing in situ descriptions regarding “who, what, where, when” issues.
__________
Exhibit 1 here.
__________
Malle (2004, p. 122) emphasizes that the influence of explanatory choices can be
found at every level of analysis, “People increase their use of causal history explanations
when accounting for negative actions (Nelson & Malle,2003); they increase their use of
belief reasons when trying to appear rational (Malle et al., 2000), and they explicitly add a
mental marker to their belief reasons when they want to distance themselves from the
agent [or situation] (e.g., ‘Why is he looking at apartments?’—‘He thinks I am moving in
with him’; Malle et al., 2000).” The use of mental state markers on reasons occurs in
introspection reports, for example, when a person distances current belief or want from
reports of belief and want in the prior situation that the researcher-informant is describing
(e.g., “I searched on-line because I thought I could get a better deal”) versus not marking
the reason (e.g., “I searched on-line to get a better deal”).
Note Exhibit 1 includes a feedback relationship between belief and desire reasons
to describe the finding that valuing occurs automatically from a belief and that humans are
culturally trained to perceive and categorize information, and form beliefs, based on
evaluations—judgments of familiar and acceptable behaviors, desires, and dislikes.
Valuing is finding benefit from achieving the goal implied by a belief.
16
What should be done to overcome framing biases? A useful initial answer: because
explicating descriptive details may help reduce the influence of the fundamental attribution
error and aids in uncovering the if-then contingencies in behavior explanations (see
Woodside & Wilson, 2003), self-informant behavior explaining benefits from first asking
“what, how, where, and when” questions before asking “who and why” questions. A
useful follow-up answer: let us recognize the need to apply multiple framing questions for
the same behavior explanation issue because all questions include biases (e.g., see Clark &
Schober, 1992). A third answer: because behavior explanation includes both unconscious
and conscious intentional and unintentional thinking that the self-informant is unable to
uncover adequately from any one inquiry method, SPI benefits from employing mixed
method designs for confirming events and personal meanings and identifying the
applicability of explicit versus implicit meanings and attitudes for specific behaviors.
MIXED METHODOLOGY IN SUBJECTIVE PERSONAL INTROSPECTION
Gladwell (2005, pp. 81-84) dramatically illustrates the use of a mixed method
design for (dis)confirming meanings expressed in his explicit SPI. He reports, “I’ve taken
the Race IAT [available at www.implicit.harvard.edu] on many occasions, and the result
always leaves me feeling a bit creepy. At the beginning of the test, you are asked what
your attitudes towards blacks and whites are. I answered, as I am sure most of you
would, that I think of the races as equal. Then comes the [implicit association] test.
You’re encouraged to complete it quickly… I took the test a second time, and then a third
time, and then a fourth time, hoping that the awful feeling of bias would go away. It made
no difference. It turns out that more than 80 percent of all those who have ever taken the
test end up having pro-white associations, meaning that it takes them measurably longer to
17
complete answers when they are required to put good words into the “Black category than
when they are required to link bad things with black people. I didn’t do quite so badly. On
the Race IAT, I was rated as having a ‘moderate automatic preference for whites.’ But
then again, I’m half black. (My mother is Jamaican.)”
As Bargh, et al. (1996), McClelland et al. (1992), and Wilson (2002) stress and
empirically support, the most important point being made here is that attitude, motive, and
meaning self-report measures do not correspond to implicit (e.g., IAT, TAT, and
FMET—forced metaphor elicitation technique, Woodside 2004b) measures. Gould (2005)
mistakes Woodside (2004b) as seeing “one of the problems with introspection as being its
subjectivity and the need to reign that in with various confirmatory approaches”.
Woodside (2004b) focuses on broadening SPI—not trying to reign it in—by making the
case for combining explicit and implicit methods in SPI reports.
Applying the useful proposition that research to learn explicit or implicit meanings
held by an informant applies to SPI, Exhibit 2 is a typology of research methods that
provides insights for designing mixed method designs and for building SPI theory that
covers both explicit and implicit meaning. The typology includes the eight possible
combinations of verbal vs. nonverbal, explicit vs. implicit, and positivistic vs. interpretative
methods; examples of each of the eight, and combinations of using two of each, are
available in the consumer research literature.
___________
Exhibit 2 here.
___________
18
Cell 1, the fixed-point survey method, is the dominant research method for most
studies attempting to measure informant-held meaning. Major advantages of this method
include ease of data collection of explicit meanings held individually for a large number of
informants and the applicability of statistical hypothesis testing. Limitations of this
method include the failure to recognize that most informant-thinking is done
unconsciously and that implicit thinking may differ substantially from explicit thinking,
along with social-desirability biases in answering questions explicitly (see Fisher 1993).
Cell 2, existential-phenomenological reports (see Thompson, et al., 1989) includes
the researcher starting with a very broad question and allowing the informant to reframe
and pose additional questions as well as answers. Such interviews provide “thick
descriptions” of situations, thinking processes, and some amount of both explicit and
implicit meanings held by the informant. Self-editing of responses before the informant
shares them with the researcher likely limits the value of the method along with the
inability of the informant to retrieve a substantial share of implicit meaning held in
memory. However, especially when the practice includes multiple interviews over several
weeks with the same informant, the method offers the advantage of self-revelation of
implicit meanings the informant rarely becomes aware when completing fixed-point
surveys (e.g., see Cox 1967).
Cell 3, automatic thought-retrieval research, includes asking what benefits and
beliefs evoke which brands (e.g., see Thelen & Woodside, 1997). The method is useful
for measuring implicit meanings associated as well as not associated with a given brand
versus competing brands. The method has the advantages of limiting the occurrence of
self-desirability bias, the collection of implicit data for a large number of informants, and
19
applicability of statistical hypothesis testing. Limitations being unable to uncover the
causal history reasons that support the automatic brand retrievals.
Cell 4 include TAT (see McClelland, et al., 1992), and the FMET (see Woodside,
2004b. McClelland provides compelling evidence supporting the external validity of TAT
data and offers mixed method reports comparing explicit and implicit informant-held
meanings. Cell 4 methods often require extensive training of researchers and often
substantial effort in analyzing data.
Cell 5, direct observation-based frequency recordings include researcher
assumptions about the implicit meanings held by consumers. For example, Wells and
Losciuto (1966) infer some consumers examined packages to learn price information while
others did not by direct observation. The method has all the advantages associating from
“being there” and not relying on informant’s memory. Playing back video or tape
recordings of consumers in naturally-occurring situations to learn their explicit
interpretations of their own behavior represent a mixed method design (e.g., see Taylor &
Woodside). Like all interpretive methods cell 5 requires very substantial effort for both
data collection and analysis.
Cell 6, direct observation meaning reports, includes the informant creating a
tangible expression of meanings and interpretations that she wishes or is able to share with
the researcher. The collages from magazine and newspaper images that Zaltman’s
informants create as part of the ZMET are examples of nonverbal data. Doyle and Sims’
(2002) advances in the “cognitive sculpting” technique provide another example. Data
collection usually extends to include both etic (researcher) and emic (informant) explicit
20
interpretations of the nonverbal expressions; thus, a mixed method design results actual
applications of cell 6 methods.
Cell 7 includes IAT reports (e.g., Brunel, et al., 2004; Masion et al., 2004). Brunel
and his colleagues (2004) demonstrate the value of collecting both explicit and implicit
meanings held by informants—applying the mixed method strategy of combining cells 1
and 7. Gladwell’s (2005) commentary of his own prior explicit-survey and IAT-implicit
associations represents a three-method design: cells 1, 2, and 7.
Cell 8, behavioral drama enactments, includes video-tapping and interpreting
informants’ enactments of themselves as inanimate objects, for example, “using facial and
body gestures and motions with any available props that you might care to use, describe
yourself as a typewriter (or dog, cat, sports utility vehicle, ice cream), for examples, see
Dichter (1964). To expand on etic interpretations of informant enactments, asking
informants to observe video-tapes and interpret their own performances offers the
opportunity of expanding from a single to a mixed-method design, for example, combining
cell 8 with cell 2.
CONCLUSIONS AND IMPLICATIONS FOR THEORY AND RESEARCH
SPI is pervasive in consumer research—answering questions asked by ourselves or
others requires SPI. SPI researchers (e.g., Gould and Holbrook) need encouragement to
look beyond single-method applications to mixed-method designs as Woodside (2004b)
illustrates in SPI reporting. This point is extremely important for building and testing
scientific theory, building in safeguards to protect ourselves from our own biases and selffabrications when making decisions affecting ourselves and others, as well as increasing
21
mindfulness (Gaither, 2002; Kozak, 2006; Weed, 1991; Weick & Sutcliffe, 2001) while
reducing arrogance.
Along with ignoring the multiple methods Woodside (2004b) describes and
applies, the main point Gould (2005) suggests is inaccurate: “But there is no member
checking the meaning a person assigns in the story. This is her own introspective story.”
Own explicit self-reports often do not express the same implicit meaning held by the same
informant—our own implicit introspection stories often differ dramatically from own
explicit storytelling.
Single-method SPI reporting and criticism of calls for using multiple-method,
explicit-implicit tools in SPI studies illustrate more than the fundamental attribution error
and overconfidence bias. Such criticism is illustrative of not doing the homework
confirming that humans are incapable of explicitly fully uncovering their own-held meaning
without resorting to mixed-method explicit-implicit research designs. Fortunately, the
work of Bargh, et al. (1996), Brunel, et al. (2004), McClelland et al. (1992), Wegner
(2002), Wilson (2002), Woodside (2004b), and Zaltman (2003) advance SPI research
methodology beyond single-method explications of meaning. The work of Allen (2002),
Hilton (1990), Malle (1999, 2004), McClure and Hilton (1997), Wegner (2002), and
Wilson (2002) advances behavior explanation research well beyond person-situation
attribution theory (Kelley, 1967, 1973) and person-trait only typologies (e.g., Gould,
1991; Holbrook, 2005). Thus, the future is bright and the possibilities are profound for
advancing SPI in consumer research
22
REFERENCES
Allen, D. E. (2002). Toward a theory of consumer choice as sociohistorically shaped
practical experience: The fits-like-a-glove (FLAG) framework. Journal of Consumer
Research, 28, 515-532.
Bargh, J. A. (2002). Losing consciousness: Automatic influences on consumer judgment,
behavior, and motivation. Journal of Consumer Research, 29, 280-285.
Bargh, J. A., Chen, M., & Burrow, L. (1996). Automaticity of social behavior: Direct
effects of trait construct and stereotype activation on action. Journal of Personality &
Social Psychology, 71, 230-244.
Becker, H. S. (1998). Tricks of the trade. Chicago: University of Chicago Press.
Brunel, F. F., Tietje, B. C., & Greenwald, A.G. (2004). Is the implicit association test a
valid and valuable measure of implicit social cognition? Journal of Consumer
Psychology, 14, 385-404.
Choi, I., Nisbett, R. E., & Norenzayan, A. (1999). Causal attribution across cultures:
Variation and universality. Psychological Bulletin, 125, 47-63.
Clark, H. H., & Schober, M. F. (1992). Asking questions and influencing answers. In
Tanur, J. M. (Ed.), Questions about questions (pp. 15-48). New York: Sage.
Cox, D. F. (1967. Risk handling in consumer behavior—an intensive study of two cases.
In D. F. Cox (Ed.), Risk taking and information handling in consumer behavior (pp.
34-81). Boston: Harvard University, Graduate School of Business Administration.
Dichter, E. (1964). Handbook of consumer motivations: The psychology of the world of
objects. New York: McGraw-Hill.
23
Fisher, R. (1993). Social desirability bias and the validity of indirect questioning. Journal
of Consumer Research, 20, 303-315.
Gaither, C. (2002). How much do doctors really know? The Boston Globe Magazine,
July 14, 12-14, ff.
Gigerenzer, G. (2000). Adaptive thinking: Rationality in the real world. New York:
Oxford University Press.
Gilovich, T. (1991). How we know what so isn’t: The fallibility of human reason in
everyday life. New York: Free Press.
Gilovich, T., Griffin, D., & Kahneman, D. (Eds.) (2002). Heuristics and biases.
Cambridge, UK: Cambridge University Press.
Gladwell, M. (2005). Blink: The power of thinking without thinking. New York: Little,
Brown.
Gould, S. (1991). The self-manipulation of my pervasive, perceived vital energy through
product use: An introspective-praxis perspective. Journal of Consumer Research, 18,
194-207.
Gould, S. (1995). Researcher introspection as a method in consumer research:
Applications, issues, and implications. Journal of Consumer Research, 21, 719-722.
Gould, G. (2005). Comparing, not confirming personal introspection: A comment on
Woodside. Psychology & Marketing, 22, forthcoming.
Hilton, D. J. (1990). Conversational processes and causal explanation. Psychological
Bulletin. 107, 86-81.
24
Holbrook, M. B. (1986). I’m hip: An autobiographical account of some consumption
experiences. In Lutz, R.J., (Ed.). Advances in consumer research, 13, (pp. 614-618).
Provo, UT: Association for Consumer Research.
Holbrook, M. B. (1995). Consumer research: Introspective essays on the study of
consumption. Thousand Oaks, CA: Sage.
Holbrook, M. B. (2003) Time travels in retrospace: Unpacking my grandfather’s trunk—
Some introspective recollections of Brule River. In S. Brown & J. Sherry (Eds.),
Retroscapes (pp. 218-241). Thousand Oaks, CA: Sage.
Holbrook, M. B. (2005). Customer value and autoethnography: Subjective personal
introspection and the meanings of a photograph collection. Journal of Business
Research, 58, 45-61.
Kelley, H. H. (1967). Attribution theory in social psychology. In Levine, D. (Ed.),
Nebraska Symposium on Motivation, 15 (pp. 192-240). Lincoln: University of
Nebraska Press.
Kelley, H. H. (1973). The process of causal explanations in interaction: Some dialogues
on dialogue. In Harvey, J. H., Ickes, W.J., & Kidd, R. F. (Eds.), New directions in
attribution research, 3 (pp. 307-328). Hillsdale, NJ: Lawrence Erlbaum.
Kozak, A. (1996). Local clinicians need knowledge tools. American Psychologist, 51,
1335-1336.
Langer, E. J. (1975). The illusion of control. Journal of Personality & Social Psychology,
32, 311-328.
25
Levine, G. M., Halberstadt, J. B., & Goldstone, R. L. (1996). Reasoning and the
weighting of attributes in attitude judgments. Journal of Personality & Social
Psychology, 70, 230-240.
Lichenstein, S. Fischoff, B., & Phillips, L. (1982). Calibration and probabilities: The State
of the art in 1980. In D. Kahneman, P. Slovic, & A. Teversky (Eds.), Judgment under
uncertainty: Heuristics and biases (pp. 306-334). Cambridge, UK: Cambridge
University Press.
Malle, B. F. (1999). How people explain behavior: A new theoretical framework.
Personality and Social Psychology Review, 3, 23-48.
Malle, B. F. (2004). How the mind explains behavior. Cambridge, MA: Bradford Book,
MIT Press.
Malle, B. F., & Bennett, R. E. (1998). People’s praise and blame for intentions and
actions: Implications of the folk concept of intentionality. Unpublished manuscript.
Malle, B. F., & Knobe, J. (1997). The folk concept of intentionality. Journal of
Experimental Social Psychology, 33, 101-121.
Masion, D., Greenwald, A.G., & Bruin, R.H. (2004). Predictive validity of the implicit
association test in studies of brands, consumer attitudes, and behavior. Journal of
Consumer Psychology, 14, 405-415.
McClelland, D. C., Koestner, R., & Weinberger, J. (1992). How do self-attributed and
implicit motives differ? In C. P. Smith (Ed.), Motivation and personality: Handbook of
thematic content analysis (pp. 49-72). Cambridge, UK: Cambridge University Press.
26
McClure, J., & Hilton, D. J. (1997). Because you can’t always get what you want: When
preconditions are better explanations than goals. British Journal of Social Psychology,
36, 223-240.
Ross, L., & Nisbett, R. E. (1991). The person and the situation. New York: McGrawHill.
Simonson, I., & Nowlis, S. M. (2000). The role of explanations and need for uniqueness
in consumer decision making: Unconventional choices based on reasons. Journal of
Consumer Research, 27, 49-68.
Tashakkori, A., & Teddlie, C. (1998). Mixed methodology: Combining qualitative and
quantitative approaches. Thousand Oaks, CA: Sage.
Taylor, J. L., & Woodside, A. G. (1981). Exchange behavior among salesmen and
customers in natural settings. In P. Reingen & A. G. Woodside (Eds.), Buyer
interactions: Empirical research and normative Issues (pp. 45-63). Chicago:
American Marketing Association.
Thelen, E. M., & Woodside, A. G. (1997). What evokes the brand or store? International
Journal of Research in Marketing, 14, 125-143.
Thompson, C. J., Locander, W.B., & Pollio, H. R. (1989). Putting consumer experience
back into consumer research: The philosophy and method of existentialphenomenology, 133-146.
Turnbull, W. (1986). Everyday explanation: The pragmatics of puzzle resolution. Journal
for the theory of social behavior, 16, 141-160.
Wallendorf, M. & Brucks, M. (1993). Introspection in consumer research:
Implementation and implications. Journal of Consumer Research, 20, 339-359.
27
Weed, L. L. (1991). Knowledge coupling: New premises and new tools for medical care
and education. New York: Springer-Verlag.
Wegner, D. M. (2002). The illusion of conscious will. Cambridge, MA: Bradford Books,
The MIT Press.
Weick, K. E. (1995). Sensemaking in organizations. Thousand Oaks, CA: Sage.
Weick, K. E., & Sutcliffe, K. M. (2001). Managing the unexpected: Assuring High
Performance in an age of complexity. San Francisco, CA: Jossey-Bass.
Weiner, B. (1995). Judgments of responsibility. New York: Guilford.
Wilson, T. D. (2002). Strangers to Ourselves: Discovering the adaptive unconscious.
Cambridge, MA: Belknap Press of Harvard University Press.
Wells, W., & Losciuto, L. (1966). Direct observation of purchasing behavior. Journal of
Marketing Research, 3, 227-229.
Wilson, T. D., Hodges, S. D., & LaFleur, S. J. (1995). Effects of introspecting abut
reasons: inferring attitudes from accessible thoughts. Journal of Personality & Social
Psychology, 69, 16-28.
Woodside, A.G. (2004a). Advancing means-end chains by incorporating Heider’s balance
theory and Fournier’s consumer-brand relationship typology. Psychology &
Marketing, 21, 279-294.
Woodside, A.G. (2004b). Advancing from subjective to confirmatory personal
introspection in consumer research. Psychology & Marketing, 21, 987-1010.
Woodside, A. G., & Wilson, E. J. (2003). Case study research for theory-building.
Journal of Business & Industrial Marketing, 18, 493-508.
Zaltman, G. (2003). How customers think. Boston: Harvard Business School Press.
trait CHR
(e.g.,
informant
commits act
because she is
friendly)
offer person
CHR
offer
situation
implies
causal
intentionality
Who?
actor’s desired outcome,
often called goal, aim,
end, or purpose (e.g.,
“to
want to ” “to feel like ”
marker
?
offer desire
reason
trait
EF
person
EF
situation
EF
implies
enabling
intentionality
implies
belief/
desired
valuin
g
How?
Fram
e
Why?
conscious / unconscious
knowledge, what actor
finds true/ false (e.g.,
“he thought,” “she
knew”)
marker
?
offer belief
reason
Key: CHR, causal history reason
• valuing--positive or negative affect
toward the action or its outcome
• marker--mental state verb used to distance
oneself from a prior held belief or desire
(e.g., “I feared,” “I thought,” “I wanted”)
trait cause (e.g.,
low/high selfesteem)
person
cause:
conscious /
situation
cause:
conscious /
offer cause: actor
mentions factors causing
the
implies
unintentionalit
y
What?
When?
Exhibit 1. Framing Questions Influence on Directing Subjective Personal Introspection in Explaining Behavior
Source: Inspired, in part, by Figure 5.1, Malle (2004, p. 119).
Nonverbal
Responses
Verbal
Responses
6. Direct
observation
meaning reports;
ZMET
8. Behavioral
drama
enactments
5. Direct
observation
frequency
recordings
7. Implicit
association
test
Explicit
Implicit
4. TAT;
FMET
3. Automatic
thought
retrievals
Implicit
Explicit
2. Existentialphenomenological
reports
1. Fixed-point
survey
responses
Interpretive
Positivistic
Exhibit 2. Measuring Informant-Held Explicit and Implicit Associations
29
30