From a Brook to a Stream: The Case of Schema Research Ronald C. Goodstein Presentation to GMU...

Post on 16-Dec-2015

212 views 0 download

Transcript of From a Brook to a Stream: The Case of Schema Research Ronald C. Goodstein Presentation to GMU...

From a Brook to a Stream:From a Brook to a Stream:The Case of Schema ResearchThe Case of Schema Research

Ronald C. Goodstein

Presentation to GMU

December 2003

Road MapRoad Map

Schema Research

Application (s1)

Extension (s1 & s2)

Integration (s2)

Future Research

Basics of Schema TheoryBasics of Schema Theory

Stimulus Evoked Category

Match Mismatch

Schema Triggered AffectSchema Triggered Affect (Fiske 1982, Fiske and Neuberg 1990, Fiske and Pavelchak 1986)

“If relatively category-oriented processes are successful, then the perceiver goes no further toward more attribute-oriented processes.

Match MismatchCategory-based Piecemeal

Low motivation High motivation

Category Affect

Target AttributeEvaluation

Target

Might This Describe Ad Might This Describe Ad Processing?Processing?

Consumers exposed to 2000 ads dailyDevelop heuristic to ease the processing loadObservations are that default is to tune out, rather than to watch as we do in forced lab testsMotivation is needed to get consumers to processIncongruity is a motivating factor in processing.

Hypothesis 1 - ApplicationHypothesis 1 - Application

When an ad is discrepant from category expectation, relative to when it is consistent, it will motivate more extensive processing.

Might There Be Reasons to Might There Be Reasons to Watch a “Typical” Ad?Watch a “Typical” Ad?

A variety of factors might attenuate the relationship between incongruity and evaluations (Mandler 1982)

Strong Priors Meyers-Levy & Tybout 1989

Goals Keller 1991

Values

+

=

-

Hypothesis 2 - ExtensionHypothesis 2 - Extension

When an ad is consistent with category expectation, and as its category affect increases, it will motivate more extensive processing.

Hypothesis 3 - ExtensionHypothesis 3 - Extension

When an ad is consistent with category expectation, and processing goals are brand, versus ad-oriented, it will motivate more extensive processing.

MethodologyMethodologyPretest- Picking Ads

Phase 1- Ad Schema and Affect

Phase 2- 302 undergrads

6 ads – 3 typical/atypical

Instructions – ad vs. brand

Measures- Cognitive Responses; Catg. Responses; Affect Consistency; Evaluative Consistency; Time Watched; Recall

ResultsResultsHypothesis 1 – Atypical vs. Typical

More CRs; Fewer Catg. Resp; Less Affect Rltn.; Lower Evaluative Rltn.; Longer Viewing; Better Recall

Hypothesis 2 – Typical:+ vs. – Less Affect Rltn.; Lower Evaluative Rltn.;

Longer Viewing; Better Recall

Hypothesis 3 – Typical: Brand vs. AdLonger Viewing; Better Recall

Discussion & ImplicationsDiscussion & Implications

STA applies to advertising domain (CB)

STA ignores “affect” as a motivator (Psych)

Consumers’ reason for watching matters (CB)

In dichotomous world, categorization leads to positive evaluations…but multiple levels may exist (Mandler 1982; Meyers-Levy and Tybout 1989)

Don’t adopt models as “gospel” (next paper)

Negative Evaluations Inverted-U Relationship

What Happens to Moderately What Happens to Moderately Incongruent Stimuli?Incongruent Stimuli?

e.g., Carpenter and Nakamoto 1989; Ward and Loken 1987; Veryzer and Hutchinson 1998

e.g., Mandler 1982; Meyers-Levy and Tybout 1989;Peracchio and Tybout 1996

How Do We Reconcile the How Do We Reconcile the Differences? Differences?

Back to Mandler (1982)A variety of factors might attenuate the

relationship between incongruity and evaluations (Mandler 1982)

Strong PriorsGoalsValues

+

=

-

How Might This Work? How Might This Work? The process of resolving a moderate incongruity is seen as stimulating and enjoyable. Alter ability

Prior knowledge (Peracchio and Tybout 1996)

Alter enjoyment Risk (Campbell and Goodstein

1997) Alter motivation

Integrating Risk into the ModelIntegrating Risk into the Model

Risk is central to consumers’ evaluations (Dowling 1999)

High risk Brand names (Erdem 1998)

High risk Less variety seeking (Inman et al. JMR)

Risk Types – e.g., Financial; Social; Performance; Psychological

(Shimp and Bearden 1982)

Risk Moderates the Relationship Risk Moderates the Relationship

Moderately incongruent stimuli are evaluated negatively when social risk is high.

Low Risk High Risk

Summary of Study 1Summary of Study 1Study 1: 2 x 2 btw subjects

RiskLow = buy to have around the house

High = buy to take to a dinner at a potential employer’s home (p < .01)

CongruityCongruent = green, cylindrical

Moderate = green, triangular (p < .001)

MeasuresProduct attitudes ( = .95)

Purchase intentions

Manipulation checks (risk = .80; congru = .82)

Category experience

Age and gender

ResultsResults 5.0

4.0

3.0

2.0

4.71

3.63; F<1, n.s.

4.58

2.56; F = 21.36, p < .001

Attd.

Congruent Moderately Incongruent

F = 6.01, p < .02

Low Risk

High Risk

Summary of Study 2Summary of Study 2Study 2: 2 x 2 btw subjects

Risk Low = buy to have at home High = buy to take to a picnic with friends of significant other... (p < .01)

Congruity …9 point scale Congruent = 12 oz. can (2.90) Moderate = 12 oz. sports bottle (5.01; F = 13.80,

p < .001)Measures

Product attitudes ( = .94)Purchase intentions

Manipulation checks (risk = .85; congru = .70)Covariates

5.0

4.0

3.0

2.0

4.66 4.67; F<1, n.s.

5.10

4.00; F = 8.82, p < .004

Attd.

Congruent Moderately Incongruent

F = 4.23, p < .04Low Risk

High Risk

ResultsResults

DiscussionDiscussion

Congruent and moderately incongruent packages evaluated similarly under low risk.

Congruent packages are preferred under high risk.

No “moderate incongruity effect!” Tybout (1997)…Tybout (1997)…

““Did they resolve?”Did they resolve?”

Alternate RationalesAlternate RationalesCongruity is unresolved

Perhaps risk makes it too difficult to resolve the incongruity … ability (Tybout 1997)

Consumer is overly stimulatedPerhaps risk + novelty = too much, so cut off

processing with risk … motivation (Steenkamp and Baumgartner 1992)

Conservatism prevails Risk yields a preference for the norm …enjoyment (Erdem 1998)

Reviewer Police

Study 3: 3 x 2 btw subjectsRisk

No = simple evaluation Low = buy to have at home High = buy to take to a picnic with friends

of significant other... (p < .01)

Congruity …9 point scale Congruent = 12 oz. can (2.90) Moderate = 12 oz. sports bottle (p < .001)

MeasuresSame… plus COGNTIVE RESPONSESCOGNTIVE RESPONSES

Summary of Study 3Summary of Study 3

5.0

4.0

3.0

2.0

4.504.07

5.10; F = 3.50, p < .06

4.69

4.57; F < 1, ns

3.54; F = 3.11, p < .08Attd.

Congruent Moderately Incongruent

F = 2.68, p < .10

F = 11.25, p < .001

No Risk

Low Risk

High Risk

ResultsResults

Congruent c Moderatec

No Risk Unresolved .10 .19

Low Risk

Unresolved .07 .12

High Risk

Unresolved .00 .07

a = p < .01b = p < .05c = p < .10

Cognitive Response Analysis - Cognitive Response Analysis - ResolutionResolution

Congruent Moderate

No Risk Totalc 1.56 1.47

Curious .30 .44a

Low Risk

Totala, c 1.79 1.71

Curiousb .39b .20a, b

High Risk

Totala 1.34 1.35

Curiousb .19 .07

a = p < .01b = p < .05c = p < .10

Cognitive Response Analysis Cognitive Response Analysis – Optimal Stimulation– Optimal Stimulation

a = p < .01b = p < .05c = p < .10

Congruent Moderate

No RiskConservative .03 .00

Low Riska

Conservative .04a .05a

High Riska

Conservative .52a, b .66a, b

Cognitive Response Analysis - Cognitive Response Analysis - ConservatismConservatism

Study 4: 2 x 2 btw subjectsRisk

No = simple evaluation High = buy to take to a picnic with friends

of significant other... (p < .01)

Congruity …9 point scale Congruent = 12 oz. can (2.90) Moderate = 12 oz. sports bottle (p < .001)

MeasuresSame… plus preference for norm scales

… Change Seeking Index

Summary of Study 4Summary of Study 4

5.0

4.0

3.0

2.0

4.20

3.64

4.72; F = 3.90, p < .05

2.97; F = 6.43, p < .01

Attd.

Congruent Moderately Incongruent

No Risk

High Risk

ResultsResults

Process AnalysisProcess Analysis

CSI … No

Preference for norm… Yes, matched evaluation analyses.

Views of Incongruity… No

DiscussionDiscussion

Moderate incongruity effect found in packaging domain.

Moderate incongruity effect occurs for judgment, but not choice.

Reason… Conservatism = Preference for

the Norm

ConclusionsConclusions“Mandler effect” may be accepted too liberally in consumer domain.Moderate incongruity effects are strongly attenuated by any social risk.Conservatism has the power to explain many CB effects e.g., COO, brand preference, variety seekingNeed to include purchase occasions in choice processes as this is an important positioning strategy.

What’s on the Horizon?What’s on the Horizon?Look for areas from psychology or marketing that integrate with the schema (fit) idea.Ethnicity in advertising (with Del

Vecchio)Thematic matching (with Kalra)Cue consistency (with Miyazaki

and Grewal)Looking for new ideas!!