I.“ z; ..
911

‘ s
.
v

[1

 

»t
Mn 3 .i: .2! ,
1-

!. CO
n».3fiu!fls..in!i .3
31.? f ‘5 it»

(v: 5.11»! ,
- I Al‘s: .- zlfiI‘v‘"
. 90:23.. I}:

p.

 

I: I. s:
L»..2v.sw.hmwmww.
.. l
. : s

I}! u n
{19.6. .
til :5“

. .r. 3‘ if r
l , : .1Z...,‘:..ur:§ia\‘

.u .‘T‘ .. .1

:93: .32... .2! .1 I... .

 

—__

l _!_."_?RARY fl
LMlchigan State ‘
University l

ii

This is to certify that the
dissertation entitled

BRANDING AGRl-FOOD PRODUCTS WITH CREDENCE
ATTRIBUTES

presented by
DOMENICO DENTONI

has been accepted towards fulfillment
of the requirements for the

PHD degree in AGRICULTURAL
ECONOMICS

 

 

 

 

 

0/
a r \‘Tr

Major Prof ssor’s Signature

 

1 2/1 7/2009

 

Date

MSU is an Afiirmative Action/Equal Opportunity Employer

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K:IProj/Acc&Pres/ClRC/DateDue.indd

 

 

BRANDING AGRI-FOOD PRODUCTS WITH CREDENCE ATTRIBUTES
By

Domenico Dentoni

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Agricultural Economics

2009

ABSTRACT
BRANDING AGRI-FOOD PRODUCTS WITH CREDENCE ATTRIBUTES
By

Domenico Dentoni

This study attempts to advance knowledge on how a firm can create value for customers
and pursue a benefit advantage strategy by adding credence attributes to its products.
Credence attributes are product features that consumers cannot verify before, during or
after consumption, but still can perceive and value. In global agri-food markets, “locally-

,9 ‘6

grown ,

’9 ‘6 9, 66 9, ‘6

place-of-origin , animal welfare , organic , eco-friendly”, “safe”, “natural”
are examples of credence attributes that consumer segments across the world increasingly
value when making their product and consumption choices. Given the unique nature of
credence attributes, analyzing how consumers build their perceptions and values for
products with credence attributes is of paramount importance for the firm pursuing such a
benefit advantage strategy. So far, a rich research strand in agricultural economics has
built upon consumer demand theory to analyze how adding credence attributes changes
consumers’ evaluations of generic agri-food products and the conditions under which
such an effect takes place.

In an attempt to integrate this research strand from agricultural economics, this

study builds upon the theory of attitude formation, developed in psychology and largely

Domenico Dentoni

applied to marketing, to understand how a firm can provide credence attribute
information to consumers differentiating its own individual brand from competitors. The
theory of attitude formation postulates that individuals not only develop evaluations for
attributes, but also form beliefs that these attributes are associated to objects and that
ultimately they create their attitudes and behavioral intentions based upon both attribute
evaluations and beliefs.

The three essays of this study provide the initial point towards building a theory
of branding products with credence attributes, which aims at understanding which
credence attribute information can differentiate a brand from its competitors and so allow
a firm to gain a benefit advantage. By analyzing the case of Michigan locally-grown
apples, the first essay starts exploring why consumers prefer products with credence
attributes and what is the direct and indirect effect of credence attributes. By tackling the
case of Liguria extra-virgin olive oil and of Southern Louisiana cream cheese, the second
essay starts exploring under which conditions generic credence attribute information and
brand information provide an advantage to an individual brand in terms of consumers’
attitudes and willingness-to-pay (WTP). By studying the case of fast food restaurants and
poultry welfare practices, the third essay starts exploring which positive brand
information can effectively mitigate the negative effect of an information shock relative
to a credence attribute. Data were collected from a series of on-line experiments on
college students and from US residents. Data analysis involved path modeling and
structural equation modeling, which provide the advantage of analyzing moderators and

mediators of the effect of credence attributes information on consumer WTP.

Copyright
DOMENICO DENTONI
2009

A Mamma, a Papa’, a Nonna, e a Claudia

ACKNOLEDGEMENTS

This dissertation is dedicated to my Mum and to my Dad, who made me always grow in a
happy, calm and enthusiastic environment and who give me full moral and practical
support in every choice I make. I hope you will enjoy this big book as my Christmas gift
this year. It is dedicated to my Grandma, who taught me how to serve the others with joy
through her example and experience. I look forward to translating this sentence to you
next week, possibly after eating a dish of your fantastic pasta. It is dedicated to Claudia,
who gave me her unconditioned love for such a long time and the motivation of getting
my studies done every semester to run back home, as soon as my exams were over, and
hug her. Today I feel very incomplete without you, but I hope that tomorrow we can
realize our dreams together.

If I said thank you to every person deserving it, I would probably need to write
two hundred more pages and make this volume extremely heavy. Yet there are people
that really deserve a Special mention. Thank you to my Major Professor, Dr. Chris
Peterson, for driving me consistently across this research work, for giving me trust and
freedom of exploring new disciplines whenever I felt it was impOrtant, and for providing
generous funding. Thank you to Dr. Glynn Tonsor for being an excellent guide and for
giving me an invaluable example that I will try to follow in my future career. Thank you

to Dr. Hamish Gow and to Dr. Tom Reardon for being an amazing source of inspiration

vi

and enthusiasm: I feel I have recorded in my brain every word you told me during these
years. Thank you to Dr. Roger Calantone for giving me the opportunity of gaining a
doctoral-level education in marketing and structural equation modeling within the MSU
Eli Broad Business School, along with generous funding and with a humor that probably
has Italian origins. I really hope that, outside my dissertation, I can keep working together
with all of you as we did so far.

I would like to thank also the participants and reviewers of the IAMA, SMS,
EAAE, IAAE and NEC-63/AAEA conferences. By interacting with you, I received great
feedback and developed many ideas. Thank you to Dr. Scott Loveridge and Dr. Steve
Hanson for their generous conference funding and thank you Debbie, Ann and Nancy for
processing all my paperwork with great efficiency and kindness. Thank you to all my
colleagues in the Department and special thanks to Byron, Nicole and Nicky. Without
your patient help, I would have left my program probably after two weeks of classes, and
certainly after two sessions of prelims. Thanks also to Adam, Alex and Ross: together we
have built an agribusiness student mafia that is both fun and effective.

Thank you to all the international friends I have lived and partied with during
these three years. Together we really enjoyed our lives. I would also like to acknowledge
the people at Chipotle, Espresso Royale and the Post for providing food, coffee and
music for thought and the daily energy to deal with all the challenges of a PhD life.

Finally, thank you to many people in Italy, which is my home. Thanks to my
Italian friends and relatives, who always encourage me and wait for me when I am

around the world and are also able to stand me once I am back home for more than one

vii

week. Thanks to the Italian olive oil producers and to all the people of goodwill that I met

during my researches. Together we will build a much better country.

viii

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................. xi

LIST OF FIGURES ......................................................................................................... xiii

KEY TO SYMBOLS AND ABBREVIATIONS ............................................................ xiv

INTRODUCTION .............................................................................................................. 1

CHAPTER 1

THE DIRECT AND INDIRECT EFFECT OF “LOCALLY-GROWN” ON

CONSUMERS’ ATTITUDES TOWARDS AGRI-FOOD PRODUCTS ........................ 11
Literature Review ................................................................................................ 14
Conceptual Framework and Hypotheses ............................................................. 20
Methodology ........................................................................................................ 24
Results .................................................................................................................. 29
Conclusions ......................................................................................................... 33

APPENDIX A

SURVEY INSTRUMENT — CHAPTER 1 ....................................................................... 36

REFERENCES - CHAPTER 1 ......................................................................................... 45

CHAPTER 2

BUILDING INDIVIDUAL BRANDS WITH PLACE-OF-ORIGIN INFORMATION:

IMPLICATIONS FOR THE INDUSTRY ........................................................................ 51
Literature Review ................................................................................................ 54
Theoretical Framework ....................................................................................... 60
Methods ............................................................................................................... 63
Results .................................................................................................................. 69
Conclusions ......................................................................................................... 80

APPENDIX B .

SURVEY INSTRUMENT - CHAPTER 2 ....................................................................... 85

APPENDIX C

OTHER TREATMENTS — CHAPTER 2 ....................................................................... 100

ix

APPENDIX D

METHODOLOGICAL NOTE - CHAPTER 2 .............................................................. 104

REFERENCES - CHAPTER 2 ....................................................................................... 110

CHAPTER 3

POSITIVE BRAND INFORMATION MITIGATING NEGATIVE SHOCKS ON

ANIMAL WELFARE ..................................................................................................... 1 17
Literature Review .............................................................................................. 120
Hypotheses Development ................................................................................... 127
Methods ............................................................................................................. 130
Results ................................................................................................................ I36
Conclusions ....................................................................................................... 1 4 7

APPENDIX E

SURVEY INSTRUMENT - CHAPTER 3 ..................................................................... 150

APPENDIX F

OTHER TREATMENT — CHAPTER 3 ......................................................................... 161

APPENDIX G

METHODOLOGICAL NOTE — CHAPTER 3 .............................................................. 163

REFERENCES - CHAPTER 3 ....................................................................................... 171

CONCLUSION ............................................................................................................... 177

LIST OF TABLES

Table 1 - Results of the Confirmatory Analysis ............................................................... 29
Table 2- Results of the Structural Eqaution Model .......................................................... 31
Table 3 - Impact of Attribute Information on Brand A and Brand B Olive Oil ............... 71
Table 4 - Impact of Attribute Information on Brand A and Brand B Cream Cheese ....... 73

Table 5 - Impact of Attribute Information from an Individual Firm and from a Collective
Organization on Brand Olive Oil ...................................................................................... 75

Table 6 - Impact of Attribute Information from an Individual Firm and from a Collective
Organization on Brand A Cream Cheese .......................................................................... 76

Table 7 - Impact of Attribute Information and Brand Information on Brand A Olive Oil78

Table 8 - Impact of Attribute Information on Brand A Cream Cheese ............................ 79
Table 9 - Simple path model testing hypothesis 1, Liguria olive oil .............................. 105
Table 10 - Simple path model testing hypothesis 2, Liguria olive oil ............................ 106
Table 11 - Simple path model testing hypothesis 3, Liguria Olive Oil .......................... 106
Table 12 - Computation of Total Causal Effects on INFO on WTPP ............................ 109

Table 13 - Multi-Group Associative LGM: Distracting versus Relevant Ex Ante Positive
Information ..................................................................................................................... 137

Table 14 - Multi-Group Associative LGM: Distracting versus Relevant Ex Post Positive
Information ..................................................................................................................... 141

Table 15 - Predictive LGM: Distracting versus Relevant Ex Ante Positive Information on
Consumer Attitudes ........................................................................................................ 144

Table 16 - Predictive LGM: Distracting versus Relevant Ex Post Positive Information on
Consumer Attitudes ........................................................................................................ 146

xi

Table 17 - Simple Piecewise LGM with Consumer Attitudes in Group 1 ..................... 165

Table 18 - Co-variance Matrix of the Associative LGM with Consumer Attitudes in
Group 1 ........................................................................................................................... 166

xii

LIST OF FIGURES

Figure 1 - The Direct Effect and the Indirect Effect of Credence Attributes on Consumer

Attitudes ............................................................................................................................ 22
Figure 2 - The Four Treatments Interacting Positive and Negative Information ........... 132
Figure 3 - The Generic Piecewise Latent Growth Model ............................................... 134

xiii

CFI

CSR

d.f.

EU

LGM

LM

MSU

NGO

PETA

POO

RMSEA

US

WTP

WTPP

KEY TO SYMBOLS AND ABBREVIATIONS

Comparative Fit Index

Corporate Social Responsibility
Degrees of freedom

European Union

Latent Growth Model

Lagrange Multiplier

Michigan State University
Non—Govemmental Organization

People for Ethical Treatment of Animals
Place-of-Origin

Root Mean—Square Error of Approximation
United States

Willingness-to-Pay

Willingness-to-Pay a Premium Price

xiv

INTRODUCTION

A fundamental question of agribusiness strategy and management is how a firm creates
value relative to competitors (Stewart, 1991) and ultimately gains a competitive
advantage (Porter, 1985). To this purpose, the firm can undertake a benefit advantage
strategy, which has the objective of creating more benefit to customers relative to its

competitors (Besanko et al., 1996).

This research work is designed to advance knowledge on how a firm can pursue a
benefit advantage strategy by adding credence attributes to its products. Credence
attributes are product features that consumers cannot verify either before, during or after
consumption, but still can perceive and value (Darby and Kami, 1973). In global agri-
food markets, “locally-grown”, “place-of—origin”, “animal welfare”, “organic”, “eco-
friendly”, “safe”, “natural” are examples of credence attributes that consumer segments
across the world increasingly values when making their product and consumption choices
(e.g., Nimon and Beghin, 1999; Loureiro and Umberger, 2007; Basu and Hicks, 2008;
Darby et al., 2008; Kanter et al., 2008 ; Frolich et al., 2009). Credence attributes also
create customer value for non-agricultural products, such as safety for surgery services,

coo-friendliness for cars, or place-of-origin for fashion articles (Darby and Kami, 1973).

1

Given the unique nature of credence attributes, analyzing how consumers build their
perceptions and values for products with credence attributes is of paramount importance

for any firm pursuing such a benefit advantage strategy.

To fully understand how consumers perceive and give value to products with
credence attributes, research needs to take into consideration both economics and
psychology theory. Consumer demand theory investigates what is the marginal value
increase consumers have when a certain attribute is added to a product (Lancaster, 1966).
Consumer psychology investigates how individuals process and use information about a
product, build their beliefs and attitudes towards it, and form their buying intentions and

actions (Fishbein, 1967; Fishbein and Ajzen, 1975; Lutz, 1991).

So far, a rich research strand in agricultural economics has built upon consumer
demand theory to analyze how adding credence attributes changes consulners’ evaluation
for generic agri-food products and the conditions under which such an effect takes place
(e.g., Thompson, 1998; Nimon and Beghin, 1999; Baker and Burnharn, 2001; Van der
Lans etal., 2001; Loureiro et al., 2002; Alfnes and Rickertsen, 2003; Lusk, Roosen and
Fox, 2003; De Pelsmacker et al, 2005; Loureiro and Umberger, 2007; Basu and Hicks,
2008; Darby et al., 2008; Ehmke et al., 2008; Kanter et al., 2008; Froelich et al., 2009).
However, this research does not provide a complete answer to how a firm should use
credence attributes to differentiate its own individual products, which are often
distinguished from other firms’ products by a brand (Aaker, 1991; Keller, 1993), and so

obtain a benefit advantage relative to competitors.

In an attempt to integrate this research strand from agricultural economics, this

research builds upon consumer psychology theory to understand how a firm can provide

2

credence attribute information to consumers that differentiates its own individual brand
from competitors. Specifically, the theory of attitude formation (Fishbein, 1967; F ishbein
and Ajzen, 1975; Lutz, 1991) provides the theoretical framework accompanying the three
essays that compose this research. The theory of attitude formation postulates that
individuals not only develop evaluations for attributes, but also form beliefs that these
attributes are associated to objects and that ultimately they create their attitudes and
behavioral intentions based upon both attribute evaluations and beliefs (Fishbein, 1967;

Fishbein and Ajzen, 1975).

The three essays of this research provide the initial point towards building a
theory of branding products with credence attributes, which aims at understanding which
credence attribute information can differentiate a brand from its competitors and so make
a firm gain a benefit advantage. In the first essay, the case of Michigan locally-grown
apples is used to start exploring why consumers prefer products with credence attributes
and what is the impact of credence attribute claims on consumers’ beliefs in the presence
of other product attributes. In the second essay, the case of Liguria extra-virgin olive oil
and of Southern Louisiana cream cheese is used to explore under which conditions
generic credence attribute information and brand information provide an advantage to an
individual brand in terms of consumers’ attitudes and willingness-to-pay (WTP). In the
third essay, the case of fast food restaurants and poultry welfare practices is used to
explore which positive brand information can effectively mitigate the negative effect of

an information shock relative to a credence attribute.

A multi-variate analysis approach is adopted to tackle the research questions of

the three essays. Specifically, the three studies are designed with structural equation

modeling, path modeling and latent growth modeling (Duncan et al., 1999; Hair et al.,
2006). These models share two key features. First, they provide a means to assess a set of
relationships simultaneously rather than in separate analyses (Hair et al., 2006). Second,
as a set of relationships can be assessed simultaneously, they also give the opportunity of
exploring the mediators and the moderators playing a role in explaining the impact of an

independent variable on a dependent variable (Kaplan, 2009).

Bringing these two features of multi-variate analysis to the ground of agri-food
marketing is crucial and timely. Most of the analysis conducted so far in this field
explored the impact of an independent variable, such as a credence claim, on adependent
variable, such as consumer WTP (e.g, Nimon and Beghin, 1999; Alfnes and Rickertsen,
2003; Lusk et al., 2003; Darby et al., 2008; Ehmke et al., 2008; Kanter et al., 2008;

F roelich et al., 2009 ). By using multi-variate techniques, this study integrates the extant
literature by exploring why and under which conditions a credence claim has an impact
on consumer WTP. Specifically, to explore why a credence claim has an impact on
consumer WTP, multi-variate techniques give the opportunity of analyzing the role of
consumers’ beliefs and attitudes as mediators of this relationship. To explore under which
conditions a credence claim has an impact on consumer WTP, multi-variate techniques
give the opportunity of analyzing the role of consumers characteristics and information
characteristics as moderators of the relationships between credence claims and beliefs,
between beliefs and attitudes, and between attitudes and WTP. Overall, applying multi-
variate techniques to the context of agri-food marketing provides the opportunity of
expanding knowledge on how consumers change their food perceptions and values, and

ultimately how they make their food buying and consumption decisions. Therefore, in a

market where firms are increasingly pushed to be consumer-responsive to develop a
benefit advantage, tackling research questions with multi-variate techniques provides key

response to the needs of marketing managers.

Data from students at Michigan State University and from US residents provided
the empirical evidence for tackling the research questions of the first two essays and of
the third essay, respectively. If the product is relevant to the sample, collecting data from
a population of students provides an ideal setting to test theory, because a quite
homogeneous group of respondents provides information in a controlled environment
without the “real-world” noise that the researcher cannot control (Calder et al., 1981;
Lynch, 1999; Winer, 1999). On the other hand, collecting data from a representative
population of US residents gives the opportunity of understanding how a research
question, tackled with a multi-variate analysis approach, can be of practical use for

marketing managers.

Data were collected with a series of on-line experiments between November 2008
and November 2009. Given their comparatively low costs and fast completion times, on-
line surveys are increasingly used by researchers (Hu et al., 2006; Louviere, 2008; Gao
and Schroeder, 2009). Researchers found that similar results are found from applying on-
line surveys, in-person interviews and conventional mail surveys (Marta-Pedroso et al.,
2007; Fleming and Bowden, 2009). Moreover, Hudson et a1. (2004) provided evidence

that on-line surveys do not have non-response bias.

The following chapters of this dissertation are organized as follows. The first
essay on the direct and indirect effects of “locally-grown” on consumers’ attitudes

towards apples is presented in chapter one, followed by a reproduction of the survey

5

instrument in the appendix. The second essay on building individual brands with place-
of-origin information constitutes chapter two, and appendices with a copy of the survey
instrument, other treatments and a methodological note follow. Chapter three presents the
third essay on brand information mitigating the negative impact Of information shocks on
animal welfare and is followed by appendices with a copy of the survey instrument, other

treatments and a methodological note. Conclusions are finally drawn in the last chapter.

REFERENCES

Aaker, D. (1991). Managing Brand Equity: Capitalizing on the Value of a Brand Name,
The Free Press, New York.

Alfnes, F. and K. Rickertsen (2003). “European Consumers’ Willingness to Pay for US.
Beef in Experimental Auction Markets.” American Journal of Agricultural Economics 85
(2), 396-405.

Baker, GA. and TA. Bumham (2001 ). “Consumer Response to Genetically Modified
Foods: Market Segment Analysis and Implications for Producers and Policy Makers.”
Journal of Agricultural and Resource Economics 26 (2), 387-403.

Basu, AK. and KL. Hicks (2008). “Label Performance and Willingness to Pay for Fair
Trade Coffee: A Cross-National Perspective.” Discussion Papers on Development Policy
125, Center for Development Research, Bonn, Germany.

Besanko, D., Dranove, D., and Shanley. M. (1996). The Economics of Strategy. New
York, Wiley & Sons.

Calder, B.J., Phillips, L.W. and Tybout, AM. (1981). “Designing Research for
Application.” Journal of Consumer Research 8, 197-207.

Darby, K., M. Batte, S. Ernst and B. Roe (2008). “Decomposing Local: A Conjoint
Analysis of Locally Produced Foods.” American Journal of Agricultural Economics 90
(2), 476-486.

Darby, M. and E. Karni (1973). “Free competition and the optimal amount of fraud.”
Journal of Law and Economics 16 (1), 67-88.

De Pelsmacker, P., L. Driesen, and G. Rayp (2005). “Do Consumers care about Ethics?
Willingness to Pay for Fair-Trade Coffee.” The Journal of Consumers Affairs 39 (2),
363-385.

Ehmke, M.T., J .L. Lusk and W. Tyner (2008). “The Relative Importance of Preferences
for Country-of-origin in China, France, Niger and the United States.” Paper presented at
the International Association of Agricultural Economists Conference, Gold Coast,
Australia, August 12-18, 2006.

Fishbein, M. (1967). “A Behaviour Theory Approach to the Relations Between Beliefs
about an Object and the Attitude Toward the Object.” Readings in Attitude Theory and
Measurement, John Wiley, New York, 389-400.

F ishbein, M. and I. Ajzen (1975). Belief Attitude, Intention and Behavior: An
Introduction to Theory and Research, Reading, Massachusetts: Addison-Wesley.

Fleming, CM. and Bowden, M. (2009). “Web-Based Surveys as an Alternative to
Traditional Mail Methods.” Journal of Environmental Management 90, 284-292.

Froelich, E.J., J .G. Carlberg and CE. Ward (2009). “Willingness-to-Pay for Fresh Brand
Name Beef.” Canadian Journal of Agricultural Economics 57, 119-137.

Gao, Z. and Schroeder, T.C. (2009). “Effect of Label Information on Consumer
Willingness-to-Pay for Food Attributes.” American Journal of Agricultural Economics.
In press, 2009.

Hair, J.F., Black, W.C., Babin, B.J., Anderson, RE. and Tatham, KL. (2006).
Multivariate Data Analysis (6th Edition), Upper Saddle River, NJ: Pearson Prentice Hall.

Hu, W., Adamowicz, W.L. and Veeman, M.M. (2006). “Labeling Context and Reference
Point Effects in Models of Food Attriubte Demand.” American Journal of Agricultural
Economics 88 (4), 1034-1049.

Hudson, D., Seah, L., Hite, D., Haab, T. (2004). “Telephone Presurveys, Self-Selection,
and Non-Response Bias to Mail and Internet Surveys in Economic Research.” Applied
Economics Letters 11, 23 7-240.

Kanter, C., K.D. Messer, and HM. Kaiser (2008). “Do rBST-Free and Organic Milk
Stigrnatize Conventionally Produced Milk?” Selected Paper prepared for presentation at

the American Agricultural Economics Association Annual Meeting, Orlando, FL, July 27-
29, 2008.

Kaplan, D. (2009). Structural Equation Modeling: Foundations and Extensions,
Thousand Oaks, CA: Sage Publications, Inc.

Keller, KL. (1998). Strategic Brand Management: Building, Measuring and Managing
Brand Equity, Upper Saddle River, NJ: Prentice Hall.

8

Lancaster, K]. (1966). “A New Approach to Consumer Theory.” The Journal of Political
Economy 74 (2), 132-157.

Loureiro, M.L., J .J . McCluskey and RC. Mittelhammer (2002). “Will Consumers Pay a
Premium for Eco-Labeled Apples?” The Journal of Consumer Affairs 36 (2), 203-219.

Loureiro, ML. and W.J. Umberger (2007). “Estimating Consumer Willingness to Pay for
Country-of-Ori gin Labeling.” Journal of Agricultural and Resource Economics 28 (2),
287-301.

Louviere, J .J ., Islam, T., Wasi, N., Street, D., Burgess, L. (2008). “Designing Discrete
Choice Experiments: Do Optimal Designs Come at a Price?” Journal of Consumer
Research 35, 360-375.

Lusk, J .L., J. Roosen and J .A. F ox (2003). “Demand for Beef fiom Cattle Administered
Growth Hormones _or Fed Genetically Modified Corn: A Comparison of Consumers in
France, Germany, the United Kingdom, and the United States.” American Journal of
Agricultural Economics 85, 16-29.

Lutz, R. (1991). “The Role of Attitude Theory in Marketing”, in: Kassarjian, Harold H.
and J .J . Robertson Editors, 1991. Perspectives in Consumer Behavior (4th ed.), Prentice-
Hall.

Lynch, J .G. (1999). “Theory and External Validity.” Journal of Academy of Marketing
Science 27 (3), 367-376.

Marta-Pedroso, C., Freitas, H. and Domingos, T. (2007). “Testing for the Survey Mode
Effect on Contingent Valuation Data Quality: A Case Study of Web Based versus In-
Person Interviews.” Ecological Economics 62, 388-398.

Nimon, W. and J. Beghin (1999). “Are Eco-Labels Valuable? Evidence from the Apparel
Industry.” American Journal of Agricultural Economics 81 (4), 801-811.

Porter, M. (1985). Competitive Advantage: Creating and Sustaining Superior
Performance. New York: Free Press.

Stewart, G. (1991). The Quest for Value. Harper-Collins, New York, NY.

Thompson, G. (1998). “Consumer Demand for Organic Foods: What We Know and
What We Need to Know.” American Journal of Agricultural Economics 80 (5), 11 13-
1118.

Van der Lans, 1., K. Van Ittersum, A. De Cicco, and M. Loseby (2001). “The role of
region of origin and EU certificates of origin in consumer evaluation of food products.”
European Review of Agricultural Economics 28 (4), 451-477.

Winer, RS. (1999). “Experimentation in the 21St Century: The Importance of External
Validity.” Journal of Academy of Marketing Science 27 (3), 349-358.

10

Chapter 1

THE DIRECT AND INDIRECT EFFECT OF “LOCALLY-GROWN” ON

CONSUMERS’ ATTITUDES TOWARDS AGRI-FOOD PRODUCTS

Growing segments of world consumers seek improved quality, healthiness and variety in
their food (e. g., Verbeke, 2005; IDDBA, 2008). Accordingly, demand of agri-food
products with credence attributes (e.g., place-of-origin, organic, locally-grown,
environment-friendly and fair trade) is increasing rapidly (e.g., Nimon and Beghin, 1999;
Loureiro and Umberger, 2007; Basu and Hicks, 2008; Darby et al., 2008; Kanter et al.,
2008; Frolich et al., 2009). This growing consumer demand has resulted in a large
literature studying a range of issues with credence attributes. Many studies suggest
credence attributes have an impact on some consumer groups’ buying intentions, and
specifically on the amount they are willing to pay for possessing products. However,
examining why consumers are willing to pay a premium price for credence attributes is
notably less prevalent in the literature. For example, Lusk et al. (2006) recognized this in
the context of country-of-origin labeling. In this study, we aim to begin filling this gap by
analyzing consumers’ motivations of buying agri-food products that are “locally-grown”.

We clarify if consumers are willing to pay a premium for “locally-grown” products

11

because they value the “locally-grown” attribute itself, or if they mainly value “locally-
grown” as a signal of other desirable product attributes, such as freshness or its

environmental-friendliness.

To disentangle consumers’ motivations for buying “locally-grown” products, we
propose and test a model that separates the direct effect from the indirect effect of
“locally-grown” on consumers’ attitudes towards a product. Similarly to the distinction
suggested by Van der Lans et a1. (2001), we define direct eflect as the impact of “locally-
grown” on consumers’ attitudes towards a product, without any mediation. We instead
define indirect effect as the impact of “locally-grown” on consumers’ attitudes towards a
product mediated by their belief that other desirable product attributes (e.g., freshness or
environmental-friendliness) are present in the product. These product attributes that are
inferred from “locally-grown” may be either experience attributes, which are features that
can be verified by the consumer after disposal, or other credence attributes. For example,
some consumers may value “locally-grown” as a cue of product fieshness, which is an
experience attribute, or as a cue of environmental-friendliness, which is another credence

attribute.

We suggest that, along with “locally-grown”, any other credence attribute may
have a direct and indirect effect on consumers’ attitudes towards a product. For example,
some consumers may value the attribute “animal welfare” as a positive cue of desirable
“food safety” (which is, according to our definition, an example of an indirect effect),
while others may value the attribute “animal welfare” itself, because they really care
about the welfare of animals (which is an example of a direct eflect). Similarly, some

consumers’ may be willing to pay a premium for food “from France” either because they

12

believe that “from France” is a cue of “good flavor” or because they have a positive
reaction associated to the idea of France. Therefore, we suggest that the model applied to

“locally-grown” in this paper can be possibly tested also on other credence attributes.

Exploring whether “locally-grown” and other credence attributes have a direct or
indirect effect on consumers’ attitudes towards a product has important implications for
marketers, public agencies and non-govemmental organizations. Marketers
understanding why potential consumers are willing to pay a premium for credence
attributes can make their consumer-targeting strategies more effective. Public agencies
and non-govemmental organizations aiming at shifting consumer demand and enhancing
consumption of products with credence attributes for social welfare reasons could use the
model proposed in this paper to assess the effectiveness Of their promotion and awareness

programs.

The conceptual framework we propose is based upon the theory of attitude
formation, developed in psychology (Fishbein, 1967; F ishbein and Ajzen, 1975), adapted
to marketing theory (Lutz, 1991) and applied in a wide range of marketing contexts (e.g.,
Hoffman and Novak, 1996; Huang, 1996; Lee, 2000). Differently from existing economic
theories on signaling quality as a unique concept (Akerlof, 1970; Rosenman and Wilson,
1991), the theory of attitude formation enables us to study the problem of signaling
individual quality attributes by analyzing the relationships among consumers’ beliefs in
the presence of product attributes and their attitudes towards a product (Fishbein, 1967).
To test our conceptual framework we collected data from 60 students in an experiment

regarding “locally-grown” apples. We chose structural equation modeling as the

13

appropriate methodology to separate the direct from the indirect effect of “locally-grown”

on consumers’ attitudes towards apples.

The rest of this paper is organized as follows. In the next section, we review the
existing literature and propose our conceptual framework. Then, we develop and state our
hypotheses. After this, we describe our methods and present our results. In the last

section, we draw our conclusions from the results illustrated.

Literature Review

Credence attributes are quality features of a product or service that cannot be verified by
consumers neither before purchase nor after trial (Darby and Karni, 1973). Credence
attributes have different properties from search and experience attributes, as these are
features that consumers can verify before purchase and after purchase respectively, when
the product is used (Nelson, 1970). On the other hand, consumers cannot know with
certainty if a credence attribute is present within a product or service, as they do not
possess the technical expertise to make an assessment. In the context of food products,
credence attributes can be either features of the production process (i.e., country of origin
or organic practices) or of the chemical structure of a product material (i.e., calorie

content or the presence of chemical residues).

Both the agricultural economics and marketing literature have largely examined
the impact of several credence attributes on consumers’ intentions of buying products and
services. Since the 19805, a vast strand of the marketing literature has focused on the
impact of country-of-ori gin attributes on consumers’ evaluation of products (e. g.,

Peterson and Jolibert, 1995; Verlegh and Steenkamp, 1999; Pharr, 2005). These studies

14

found that the impact of country-Of-origin on consumer evaluations is significant in many
circumstances. More recently, the agricultural economics literature has analyzed the
impact of several credence attributes, including genetically-modified (e. g., Baker and
Bumham, 2001; Lusk et al., 2003), organic (e.g., Thompson, 1998; Kanter et al., 2008),
local or locally-grown (e.g., Darby et al., 2008; Froelich et al., 2009), environment-
friendly (e. g., Nimon and Beghin, 1999; Loureiro et al., 2002), place-of-origin (e.g., Van
der Lans et al., 2001; Alfnes and Rickertsen, 2003; Loureiro and Umberger, 2005 and
2007; Ehmke et al., 2008), fair trade (e.g., De Pelsmacker et al, 2005; Basu and Hicks,
2008) and hormone-free (e.g., Alfnes and Rickertsen, 2003; Kanter et al., 2008) on

consumers’ willingness-to-pay for agri-food products.

From these studies, researchers have found that the impact of many credence
attributes, such as the presence of procedures guaranteeing safety (Schroeder et al., 2007),
on consumers’ buying intentions has a positive direction. However, they have also found
that the impact of other credence attributes, such as “genetically-modified” (Lusk et al.,
2001), is sometimes negative. Researchers have often estimated the magnitude of the
impact of credence attributes on consumers’ willingness-to-pay (e.g., Alfnes and
Rickertsen, 2003; Lusk et al., 2003). Furthermore, some researchers has found that
credence attributes have a positive impact on consumer’s attitude towards a product (e. g.,
Ericksson et al., 1984), which in turn has a positive effect on consumers’ buying
intentions (F ishbein and Aj zen, 1975). Finally, researchers have analyzed how the impact
of credence attributes on consumers’ attitudes and buying intentions vary according to

consumers’ characteristics, such as their nationality (Tonsor et al., 2005; Basu and Hicks,

15

2008; Ehmke et al., 2008), level of income (Thompson, 1998; Pharr, 2005) and level Of

knowledge of the attribute (Baker and Bumham, 2001).

While much research has focused on measuring the magnitude and the direction
Of the impact of credence attributes on consumers’ buying intentions, a question that has
not been tackled systematically is why do credence attributes have such an impact? One
way to frame this broad question is analyzing whether consumers value credence
attributes because they are cues of other desirable attributes or because they are desirable
on their own. In order to analyze this specific question, we propose a conceptual

framework that builds upon the learning theory of attitude formation (Fishbein, 1967).
Consumers ’ Beliefs and Consumers ’ Attitudes towards a Product

There is a broad strand of the literature in consumer psychology analyzing the
relationship among consumers’ beliefs in the presence of product attributes to their
attitudes towards a product and their willingness to pay for it (F ishbein, 1967; Fishbein
and Ajzen, 1975; Ajzen, 1991 ;Eagly and Chaiken, 1993; Ajzen, 2005). Specifically, the
learning theory of attitude formation elaborated by Fishbein (1967) establishes the
relationship between a person’s beliefs in the presence of individual attributes of an
object and his overall attitude towards that object. An attitude towards an object is
defined as a “psychological tendency that is expressed by evaluating a particular entity
with some degree of favor or disfavor” (Eagly and Chaiken, 1993). There is large
evidence that a person’s attitude towards an Object is positively associated with actions
involving that Obj ect, even if there are social and personal factors that might weaken or

eliminate this relationship (Fishbein and Ajzen, 1980; McFadden, 1986).

16

In particular, the learning theory states that a person’s attitude towards an object is
the sum of his evaluative judgments for each attribute of the object times a consumer’s
belief strength that each attribute is actually in place. In formula, the attitude towards an

object is given by:
Attitude (Object) = Zni=1eibi (1)

where i is an attribute Of an Obj ect, n is the total number of attributes of the object, e, is a

person’s evaluative judgment for the attribute i and b, is a person’s belief strength that

attribute i is actually in place in the object. Both ei and bi can be thought as a scale of
values, rather than a yes/no value. A person’s evaluative judgment for an attribute
represents how much he cares about the presence of that attribute, while a person’s belief
strength represents how much he believes that that attribute is actually present within that

object.

The learning theory of attitude formation has borrowed itself to marketing theory
(Lutz, 1991) and has found application in a wide range marketing contexts (e. g., Hoffman
and Novak, 1996; Huang, 1996; Lee, 2000). Consumers build their attitudes towards a
product upon their own beliefs and evaluative judgments for each attribute of that product.
Then, consumers make their buying decisions by comparing their attitudes towards
competing products and by taking into account other personal and social factors that
might influence their decision (F ishbein and Ajzen, 1980). From this perspective,
marketing communication strategies have the goal of making consumers’ attitudes for
their product higher than their attitudes towards competing products (Lutz, 1991). To do

that, marketers have to decide whether they aim at changing consumers’ evaluative

l7

judgment for specific attributes or at changing their beliefs that a Specific attribute is in
place. To make this fundamental choice, it is crucial that marketers have an
understanding of how their own product attributes are perceived by consumers differently

from their competition.
Direct and Indirect Effect of Credence Attributes on Consumers ’ Attitudes

The impact of product attributes as signals, or cues, of consumers’ perceptions of quality
has been an important field of research in consumer psychology. Consumers use

attributes as cues when information is incomplete or difficult to obtain (Olson, 1978; ‘
Ericksson et al., 1984; Han and Terpstra, 1988; Rao and Monroe, 1989; Kinnani and Rao,
2000). In this study, we hypothesize that credence attributes have an impact on
consumers’ attitudes also because they are used as a one of desirable experience attributes

and other credence attributes.

Some research on the use of the credence attribute country-of-origin as a one of
other attributes has been already conducted in marketing literature, while comparatively
little work in this area has been done in the agri-food marketing field (Lusk et al., 2006),
with few exceptions (Umberger et al. 2003; Loureiro and Umberger, 2005). There is
evidence that the country-of-origin associated to a product has an important function in
increasing consumer’s beliefs in the presence of other experience attributes (e.g.,
Ericksson et al., 1984; Han and Terpstra, 1988; Hong and Wyer, 1989). For example, US
consumers considered TVS made in Japan more technologically advanced than domestic
TVS (e.g., Han and Terpstra, 1988). The effect of place of origin as a one of other
attributes has been defined by Van der Lans et al. (2001) as indirect effect, as the impact

of credence attributes on consumers’ willingness to pay for a product is mediated by

18

consumers’ perceived quality. Similarly, in this study, we propose that the effect of
credence attributes on consumers’ attitudes can be defined as “indirect” when it is
mediated by consumers’ beliefs in the presence of individual product attributes. Some
researchers have found that the impact of place of origin of a product on consumers’
attitudes is given only by the indirect effect as a mediation of consumers’ beliefs in the

presence of experience attributes (e.g., Ericksson et al., 1984).

Other researchers found that the idea of a place of origin on its own, when
attached to a product, can generate consumers’ positive affective feelings for the product
(Johansson and Nebenzahl, 1986; Van Ittersum et al., 1991; Verlegh and Steenkamp,
1999). These affective feelings are sometimes based on retrieval of personal past
experience with the place of origin (e.g., Oberrniller and Spangenberg, 1989; Li and
Wyer ,1994), while sometimes the place of origin can contribute to the creation of a
consumer’s self-image (Keller, 1998). In these circumstances, Van der Lans et a1. (2001)
claim that the place of origin has a direct effect on consumers’ attitude towards a product,
which means that the place of origin has an impact on consumers’ attitudes towards a
product without any mediation. Similarly, in this study, we propose that the effect of
credence attributes on consumers’ attitudes can be defined as “direct” when there is no
mediation in this relationship. Van der Lans et a1. (2001) found that direct and indirect
effect of region-of-origin attributes can coexist. However, other studies have found that
place of origin has sometimes no direct effect at all (Ericksson et al., 1984).

The Moderation Effect of Consumers ’ Familiarity with the Product
Consumers’ familiarity with a product is “the number of product-related experiences that

have been accumulated by the consumer” and is a major component of product

19

knowledge (Alba and Hutchinson, 1987). Familiarity with a product influences how a
person searches, uses and recalls information about that product (e.g., Park and Lessi g,

1981; Punj and Staelin, 1983; Johnson and Russo, 1984).

Researchers found that consumers with different levels of product familiarity use
different cues to form their beliefs about the quality of a product (Rao and Monroe, 1988).
Specifically, consumers with a lower familiarity with the product use cues that are
extrinsic to the product (Olson, 1977). For example, a consumer that has low familiarity
with wine is more inclined to evaluate quality from cues such as price, country-of-origin
or the name of the wine. In other words, consumers that are not familiar with a product
tend to use country-of-origin as a stereotype to evaluate a product, as they do not know
how to obtain more accurate information (Bodenhausen and Lichtenstein, 1987). On the
other hand, consumers with a higher familiarity with the product make a larger use of
cues that are intrinsic to the product, such as a wine’s color or flavor (Rao and Monroe,
1988). The theory on the consumers’ familiarity with a product (Rao and Monroe, 1988)
leads us to hypothesize that the indirect effect of credence attributes, which are extrinsic

cues, may vary according to the level of consumers’ familiarity with a product.

Conceptual Framework and Hypotheses

To explore why credence attributes have an impact on consumers’ attitudes, our
conceptual framework builds upon the theory of attitude formation (Fishbein and Ajzen,
1975), the theory of direct and indirect effect of place-of-origin attributes (Van der Lans

et al., 2001) and the theory of consumers’ familiarity with a product (Rao and Monroe,

20

1988) (Figure 1). In this study, we test our conceptual framework in the specific context

of “locally-grown” attributes.

Given the nature of credence attributes of being verifiable by consumer neither
before nor afier disposal, we firstly distinguish between the seller’s credence claim and
the buyers’ beliefs that the credence attribute is actually in place. In the case of “locally-
grown” attributes, as there is no current unambiguous definition of what is “local” or not
(Darby et al, 2008), consumers may perceive some products to be “more locally-grown”

or “less locally-grown”. The concept of consumers’ beliefs in the presence of the

j 4.—

“locally-grown” attribute as a scale of values is consistent with the learning theory of
attitude formation (Fishbein, 1967). Therefore, we propose that sellers’ “locally-grown”
claims and buyers’ beliefs are separate variables, and that sellers’ claims have an impact
on buyers’ beliefs. Along with sellers’ “locally-grown” claims, other attributes, such as
the color or the flavor of a product may have an impact on consumers’ beliefs in the
presence of the “locally-grown” attribute. In this study, we do not test this proposition but

do recognize this as an area for valuable future research.

Second, given the existing evidence from the place-of-origin literature (Van der
Lans et al., 2001), we hypothesize that “locally-grown” has both a direct effect and an
indirect effect on consumers’ attitude towards a product. Previous studies on consumers’
preferences for local products provide elements suggesting that “locally-grown” may
have this dual effect (Darby et al., 2008). Specifically, in Darby et a1. (2008), respondents
revealed that they value “locally-grown” strawberries mainly because they are fresher,
but also because they Simply like the idea of eating strawberries from their own land of

origin. This suggests that the credence attribute “locally-grown” is a cue of an experience

21

...8:£E< an .«o 8:8on 2: E max—om .flofiamcooz 38E ASBEEV flysom {25.200 ”vcowo:

 

 

:33:
l may—I30»
2.3.5.
m..—OE flan—60

 

 

 

 

 

 

 

 

 

835.
sec

 

 

 

 

 

 

woe—55¢. coEE—SU no flair—«3‘ 3:320 no «outm— «oohcau 2: can Beta «99:9 2:. - _ 25w:—

 

 

 

 

 

 

 

 

a23E<
8:320 350V
e23 8255
9.55350 5:5
§E=§a
9.585200
€330
t -2385
mug—om
9.52350
@2332 a .
occur—9&5 Us“ a
auzom .
mLoEsmcoU bEEEMm
9.58350

 

 

EEO
5.5.6

$.83

 

 

 

 

22

attribute, such as the freshness, as well as a direct driver of a consumer’s attitude towards
strawberries. Similarly, we hypothesize that the indirect effect of “locally-grown” is
mediated by a consumer’s beliefs in the presence of other credence attributes, such as

environmental-friendliness. In other words, we hypothesize that:

H1. Consumers’ beliefs that a product is “locally-grown” are positively associated

with their attitude towards the product.

H2. Consumers’ beliefs in the presence of experience attributes are partial
mediators of the effect of consumers’ beliefs that a product is “locally-grown”

attribute on their attitude towards the product.

H3. Consumers’ beliefs in the presence of other credence attributes are partial
mediators of the effect of consumers’ beliefs that a product is “locally-grown”

attribute on their attitude towards the product.

Finally, on the basis of evidence from the theory of consumers’ familiarity with a
product (Rao and Monroe, 1988), we hypothesize that consumers’ familiarity with the
product mitigates the indirect effect of “locally-grown” attributes on consumers’ attitudes
towards a product. As shown by Rao and Monroe (1989), a “locally-grown” attribute,
such as other cues that are extrinsic to the product, is more used by low-familiarity
consumers as a stereotype to infer product quality. Similarly, we hypothesize that a
“locally-grown” attribute is more used by low-familiarity consumers as a stereotype tO
evaluate the presence of other attributes of a product, such as its flavor or its safety. In

other words, we hypothesize that:

23

 

H4. Consumer’s familiarity with a product mitigates the indirect effect of “locally-
grown” attributes on consumers’ attitude towards the product mediated by
consumers’ beliefs in the presence of other credence attributes and experience

attributes.

Methodology

To test our hypotheses, data were collected through an on-line experiment administered
to a convenience sample of 60 undergraduate and graduate students enrolled at Michigan
State University, East Lansing, Michigan. The experiment was conducted during October
and November 2008. Students were recruited in two convenient campus locations. We
did not exclude any sub—group from the sample population of students. Out of the
students that undertook the questionnaire, 76% were graduate students and 24% were
undergraduates. Males were 59% of the sample, while US citizens were only 37% of the

sample.

We chose “locally-grown” apples as the product of interest of our study for
several reasons. First, apples represented a convenient product, as it is cheap and easy to
handle in an experimental setting. Second, there is a wide literature of experiments based
on apples that we could use as reference for our research design (e. g., Manalo, 1990;
DeEll and Prange, 1992; Mehinnagic et al., 2003). Third, as the location of this study is a
large producing and consuming state of apples, we assumed that our sample population,
on average, was familiar with the expression “locally-grown” apples, although not
univocally defined. Although 63% of the sample was from outside the US, 70% of them

were in the US for more than one year. Hence, we assume they were likely to have

24

acquired some familiarity with “locally-grown” products. For the same reason, we
assumed that respondents generally had enough involvement with the product to

undertake a fairly complex questionnaire.
Experimental Procedure

Out of these 60 respondents, 20 students undertook a pre-test questionnaire and 40

students completed the final questionnaire.

We performed a pre-test questionnaire to assess which attributes the respondents
most likely infer from “locally-grown” claims. Respondents were first asked to pick up to
three experience attributes that they infer when evaluating a “locally-grown” apple from a
list of eight suggested attributes. Second, they were asked to pick up to three credence
attributes that they infer when evaluating a “locally-grown” product from a list of twelve
suggested attributes. The lists of suggested experience and credence attributes were
created from previous research on consumers’ perceptions of attributes related to apples
(e.g., Manalo, 1990; DeEll and Prange, 1992; Mehinnagic et al., 2003). We found that,
when they observe a “locally-grown” apple, respondents most commonly infer credence
attributes such as pest and disease-free, pesticide and chemical-free, and healthy. Also,
they most commonly infer experience attributes such as firm, sweet and having good
flavor. Therefore, in our final experiment we used these three credence attributes and
three experience attributes as possible mediators of the relationship between “locally-

grown” attributes and 'consumers’ attitudes towards apples.

The final experiment involved two treatments with two levels each, giving four
stimuli in total. The first treatment is the credence claim that an apple is “locally-grown”,

where the two levels are presence or absence of the “locally-grown” claim. This

25

treatinent has the purpose of creating variation in the respondents’ beliefs that the apple is
locally-grown. The second treatment is the picture of an apple, where the two levels are

presence or absence of a picture of an apple. The purpose of this treatment is to introduce
a control variable in the model that may reduce the effect of the “locally-grown” attribute

on consumers’ beliefs and attitudes towards a product.

Students that accepted to participate in the final experiment were contacted by e-
mail and directed to an on—line experiment, which took on average 15 minutes. First of all,
respondents were asked demographic questions (e. g., gender, nationality, student year)
and eight questions measuring their familiarity with apples, such as “how frequently do
you consume apples, including both home and away from home?” and “Do you presently
have some apples with you at home?”. From these eight questions, we computed a
familiarity score for each respondent. Therefore, respondents were divided in two groups
and each respondent was administered two stimuli, which corresponds to one level for
each of the two treatments. As each of the 40 subjects was administered two stimuli, we
had a total of 80 observations from the final questionnaire. When the “locally-grown”
claim Was present, respondents were asked to “think about an apple that is claimed to be
locally-grown”. When this treatment was absent, respondents were simply asked to
“think about any apple that they would find in their shopping location”. When the apple
picture was present, respondents were asked to “look at the apple in the picture”. When
this treatment was absent, there was simply no mention of apple pictures in the

questionnaire.

After each stimulus, we measured beliefs in the presence of the “locally-grown”

attribute. We also measured beliefs in the presence of the other credence and experience

26

attributes that were previously selected in the pre-test. Beliefs were measured with a
seven-point Likert-scale question, where respondents were asked: “To what extent do you
believe that this apple is locally-grown?”. Finally, we measured respondents’ attitudes
towards apples, without any difference across groups. As commonly in use in the
literature, to assess consumers’ attitudes (Eagly and Chaiken, 1993), we asked: “How
would you describe your attitude towards this apple?” and then asked tO anSwer on four
seven-point Likert-scales, namely from bad (1) to good (7), from dislike to like, from
negative to positive and from unfavorable to favorable. At the end of the experiment,

each respondent received ten dollars compensation.
The Model

Data were analyzed with a structural equation model, based on a system of regressions
combining a factor model and a path model. In the factor model, the latent construct
“consumers’ attitude towards an apple” (F I ) is hypothesized to be a predictor of the four
measurable indicators of attitude: bad/good attitude (V1), dislike/like attitude (V2),
negative/positive attitude (V3) and unfavorable/favorable attitude (V4). Therefore, we

write:

v1= F1+el; (2)
V2: F1+e2; (3)
V3= F1+e3; (4)
V4: F1+e4. (5)

In these regressions, e. to e4 are the errors associated to each measured variable

V1 to V4.

27

In the structural model, consumers’ beliefs in the presence of the “locally-grown”
attribute [B(LG)], of other credence attributes [B(CredAttr)] and of experience attributes
[B(Eprttr)] are hypothesized to predict the construct “consumers’ attitude towards an
apple” (F1). Moreover, consumers’ beliefs in the presence of the “locally-grown”
attribute are predicted by the seller’s credence claim (LG), the picture ’of the apple (PIC)
and by the consumers’ familiarity with apples (FAM), as well as by their respective
interactions (LGPIC; FAMLG; FAMPIC; FAMLGPIC). Finally, consumers’ beliefs in
the presence of experience and other credence attributes are predicted by their beliefs in
the presence of the “locally-grown” attribute, by their familiarity with the product and by

picture of the apple, as well as by their interactions. Then, we write:

F1 = a5B(LG) + b5B(Eprttr) + c5B(OCredAttr) + d5PIC + cs; (6)
B(LG) = a6LG + b6PIC + c6LGPIC + d6FAM + f6FAMLG + g6FAMPIC +

+ h6FAMPICLG + e6; (7)
B(Eprttr)’ = a7BLG + b7FAM + c7PIC + d7FAMPIC + f7FAMBLG + e-;; (8)
B(OCredAttr)’ = agBLG + bgFAM + chIC + dsFAMPIC + stAMBLG + cs. (9)

In these regressions, B(Eprttr) and B(OCredAttr) represent 1x3 vectors, as

three experience attributes and three other credence attributes are considered in this study.

Therefore, b5 and c5 are also 1x3 vectors, while the predictors of B(Eprttr)’ and
B(OCredAttr)’ are 3x1 vectors. Finally, es and e6 represent the errors associated to

dependent variables F1 and B(LG), while e-; and eg represent the 3x1 vectors of errors

associated with the dependent variables B(Eprttr)’ and B(OCredAttr)’.

28

Results

Results from the confirmatory factor analysis are presented in Table 1. The latent
construct “consumers’ attitude towards the apple” loads to each of the four indicators of
attitudes that we have proposed, V1 to V4, with a statistical Significance at 5%. Therefore,
the four indicators of consumers’ attitudes towards a product are significant reflective
measures of the factor “attitudes towards the apple”. Moreover, as chi-square = 1.77 with
d.f. = 1 such that its p-value = 0.18, there is a good fit of the factor model with the data.
Therefore, we conclude that this factor model has convergent validity and we use this

“attitude towards apples” construct as dependent variable in the structural equation model.

Table 1 - Results of the Confirmatory Analysis

 

 

 

 

 

32531;:nt Independent Variables Errors R-Squared
V1 .950"I F1 .312: .903 _

V2 .950" F1 .3 14* .902

V3 .941" F1 .340: .885

V4 .927" F1 375* .859

 

 

 

 

 

Chi-Square = 1.767 based on 1 d.f.; P-Value = 0.18374.
RMSEA = 0.102. 90%, Confidence Interval = (0.000, 0.343)

 

Legend: V1 : Bad/Good Attitude Indicator; V2 : Dislike/Like Attitude Indicator; V3 :
Unfavorable/Favorable Attitude Indicator; V4 : Unfavorable/Favorable Attitude Indicator; F 1 :
“Consumer’s Attitude towards the Apple” Latent Construct.

Note: *Statistics significant at 5% level.

Results from the structural equation model are presented in Table 2. After
performing the Wald (W) test and the Lagrange Multiplier (LM) tests for respectively
dropping and including new free parameters, we decided to fix three parameters to zero.
Specifically, we dropped the variables “consumers’ familiarity with the product” (FAM)

and “apple picture” (PIC) from the regression on consumers’ beliefs in the presence of

29

 

the attribute “locally-grown” (BLG), as the W-test indicated that these two variables had
no impact on the dependent variable. For the same reason, we dropped the variable
“apple picture” also from the regression on consumers’ attitude towards the apple (F 1).
This result from the W-test suggests that introducing the variable “picture of an apple” as
a control in the model does not reduce the impact of a “locally-grown” attribute on

consumers’ attitudes towards a product.

The overall fit of the structural equation model with the data is low, as chi-square
= 1467 with d.f. = 124, such that its p-value < 0.01, while the root mean-square error of
approximation (RMSEA) is equal to 0.38. This problem might be caused by the small
sample size, which does not guarantee a sufficient power for testing the hypothesis of
exact fit of the model with the population. Looking at the specific regressions of the
model, sellers’ credence claim (LG), the apple picture (PIC) and respondents’ familiarity
with apples (F AM) do not explain much of the variation of consumers’ beliefs in the
presence of the attribute “locally-grown”, as R2(BLG) = 0.11 only. On the other hand,
goodness-to-fit measures of the other regressions of the model indicate that the
hypothesized predictors explain a large part of the variance of the respondents’ beliefs in
the presence of the experience and other attributes, as well as of their attitudes towards

the apples.

Afier evaluating the overall fit of the model, we assess the significance of the
individual parameters. From the regression on respondents’ beliefs that an apple is
“locally-grown”, we found no variable having a Significant impact at the 5% statistical

Significance. From both the regressions on respondents’ beliefs in the presence of the

30

._o>o_ Sn 8 Enact—Mi 85:89.. .maotm :D can Emmi ”FEE a came. 2: Escozom {2:350
” Elm—mm ”825 2 can? on. 3.: mozom 958350 N HmmBmm agar.— vooO a ma: page 2: “2:0:an POE—.200 ” ><xEOm E8525 we no..."— a
can? 05 awning—om FEE—5:00 “ Emmom ”3335 EB 38.“— mo 02m n 23¢. 05 Eamozom Mcofizmcoo u hmmmm 33:82 a oEa< 2: HEDGE—om
mzofiameoo ” A<mzm Mqu can 2<m nozuEBE ” Dam—Zed MDE new $2... .5385:— ” DESI; Kaine. fits gfizfimm {on—350 u 2<m MOE
can 04 55825 n DEC: ”2%? 05.8 cosmtogo 32> “ OE ”GEE—~85 aBEDé—mooq n can? 2: 35 EEO N 235001— m§0c0.>__oo3
n 03%.. 2: 85 “2:5 {25580 ” 04m 58.350 E89: 03%.. 2: €333 3532 $583.50 ” _ ”— u88€£ ovBE< osfiogfiofiflgfleb

” v> £8865 agate. oEEcEEoEEoéED “ m> £886.: 0953.. 8:68:35 ” N> USNBvE 063:3 uooOBmm ” _> accumul—

<mm2¢ A<>MmH7= goo .mwmd H<mm§

:—

  

uc 3:53— -N 055,—.

31

experience attributes and of other credence attributes, we found that respondents’ beliefs
that an apple is “locally-grown” (BLG) and respondents’ familiarity with apples (FAM)
have a positive impact which is statistically Si gnificant at a 5% level. However, we found
that, inthe same regressions, the interaction between these two variables (BLG and
FAM), which is called FAMBLG, has a negative impact on ’respondents’ beliefs in the
presence of the experience attributes and of other credence attributes. This overall means
that respondents use the “locally-grown” attribute of an apple to infer sweetness, firmness,
flavor and healthiness of an apple, as well as the absence of pests/diseases and absence of
chemicals/pesticides in it. However, respondents’ familiarity with apples mitigates the
use of “locally-grown” as a cue of these attributes, as hypothesized (H4). Finally, we
found that the visual observation of the apple (PIC), while having a negative effect on

respondents’ beliefs, does not reduce the impact of “locally-grown”.

From the regression on respondents’ attitude towards the apple, we found that
consumers’ beliefs in the presence of the attributes “locally-grown” (BLG), absence of
pests and diseases (BPEST), good flavor (BGFLAV) and firmness (BFIRM) have a
positive impact which is statistically significant at a 5% level. Therefore, respondents’
beliefs in the presence of both the “locally-grown” attribute, experience and other
credence attributes (i.e. apples are free of pests and diseases (BPEST), have good flavor
(BGFLAV) and are firm (BF IRM)) have a positive impact on attitudes towards the apple.
On the other hand, we found the impact of consumers’ beliefs in the presence of the
attributes healthiness (BHEAL), absence of chemical residues (BCHEM), and sweetness

(BSWEET) on respondents’ attitude towards apples is not significant at a 5% level.

32

From these results, we therefore find evidence that respondents’ beliefs in the
presence of the “locally-grown” attribute have a direct effect on respondents’ attitude
towards the apples, and SO we find support to our hypothesis H1. On the other hand, we
find that respOndents’ beliefs in the presence of the “locally-grown” attribute have an
impact on their beliefs in the presence of other credence (BPEST) and experience
attributes (BGFLAV and BF IRM) that in turn have an impact on respondents’ attitudes
towards the apples. In other words, respondents’ beliefs in the presence of the experience
attributes and of other credence attributes are both partial mediators of the impact of
respondents’ beliefs in the presence of the “locally-grown” attribute on their attitude

towards the apple, as hypothesized (H2 and H3).

Conclusions

By analyzing the direct and indirect effect of credence attributes on consumers’ attitudes
towards a product, this study aims at bringing a conceptual and methodological
contribution to the existing agricultural economics literature. From a conceptual
standpoint, we introduced three novel constructs. First, by using the learning theory of
attitude formation (Fishbein, 1967), we introduced the distinction between a seller’s
credence claim and a consumer’s beliefs in the presence of credence attributes. Second,
by expanding the theory of the use of country-of-origin attributes as cues of perceived
quality (Pharr et al., 2005), we analyzed on the consumer’s use of credence attributes as
cues of other product attributes. Third, by building upon the theory of Rao and Monroe
(1988), we defined the role of a consumer’s familiarity with the product as a moderator of

the use of credence attributes as cues of other product attributes.

33

From a methodological standpoint, we suggested a quantitative methOd to
separate the direct and the indirect effect of credenceattributes on consumers’ attitudes
towards a product. By doing this, we introduced a more specific definition of indirect
effect of credence attributes cOmpared to the one proposed by Van der Lans et a1. (2001).
We proposed that the effect of credence attributes on consumers’ attitudes towards a
product is indirect when mediated by consumers’ beliefs in the presence of other product

attributes, either credence or experience.

Empirical evidence from this study Should be considered preliminary because of
the limited sample Size. Since a relatively complex model was estimated with 80
observations, we obtained a low power for testing the overall fit of the structural model to
the data. In future research, a larger sample should be used to test the overall fit of the

model proposed.

Although the power of the test is low, results provide empirical support to the four
hypotheses of this study. First of all, consumers’ beliefs in the presence of the credence
attribute “locally-grown” have both a direct and indirect effect on their attitudes towards
apples. As regards to the indirect effect, consumers’ beliefs in the presence of both
experience attributes and other credence attributes act as mediators of this relationship.
This confirms but also provides complementary detail to the evidence found by Van der
Lans et a1. (2001). Furthermore, consumers’ familiarity with apples acts as a negative
moderator of the impact of their beliefs in the presence of credence attributes as cues of
other attributes. This seems consistent with the conclusions by Rao and Monroe (1988),
who found that highly familiar consumers use extrinsic cues less than low familiar

consumers.

34

Mb»

.4“

L1

Results from this study have both managerial and policy implications. On one
hand, understanding why consumer value credence attributes is crucial for a firm’s
marketing communication strategies. By knOwing whether consumers value a credence
attribute on its own or rather use it as a cue of other valued attributes, a marketer can
make his communication more effective. On the other hand, public agencies and non-
profit organizations having the purpose of changing people’s buying, consumption and
disposal habits can learn from the distinction between direct and indirect effect of
credence attributes, as well as from the moderation role of people’s familiarity with a

product.

Future research in this area Should address the following limitations of this study.
First of all, this study has not analyzed the drivers of consumers’ beliefs in the presence
of a credence attribute, while focusing only on their effects. Future research should
address the impact of other product attributes as a major driver of consumers’ beliefs in
the presence of credence attributes. Second, this study is limited to the effects of the
credence attribute “locally-grown”, while other credence attributes may behave
differently from “locally-grown” attributes. By expanding the experiment to a broader set
of credence attributes, it would be possible to find and explain differences direct and
indirect effect of credence attributes on consumers’ attitudes towards a product. Third,
consumers’ personal values, which are largely studied in consumer psychology (e. g.,
Sheth et al., 1991; Schwartz, 1992), may explain a large part of the variation of the
magnitude of the direct and indirect effect of credence attributes on consumers’ attitudes
towards a product. Future research may analyze the role of consumers’ personal values as

a key moderator of this relationship.

Appendix A

SURVEY INSTRUMENT — CHAPTER 1

Demographics

Are you male or female?

Male

Female

What is the category of student that best represents you right now?

Freshman
Sophomore
Junior

Senior

WWDDW

Graduate

What is your country of citizenship?

!" United States
1" Other country (please specify country name)1

 

F' Other country (please specify country name, in case of double

 

citizenship)l

36

4. If you are citizen of a country different from United States, how long have you
been living within this country?

F Less than 3 months

Between 3 and 6 months
Between 6 and 12 months
Between 1 year and 2 years
Between 2 and 5 years

Between 5 and 10 years

3‘3’1713":

More than 10 years

Consumer Familiarity with Products

5. How frequently do you consume apples, including both home and away from
home?
Less
than Once 2:3 Once 2:3
Times Times .
Never Once a a a a Daily
a Month Week
Month Month Week
F F F f‘ F (" t“
6. How knowledgeable are you about the quality of apples when you make your

buying choice?

Very
Very Knowledgeable
Unknowledgeable
l 2 3 4 5 6 7
t“ C t" (" t" (" F
7. Do you presently have some apples with you at home?
(3*
Yes
F
No

37

8. Please list up to five varieties of apples that you know

Variety 1
Variety 2
Variety 3
Variety 4

Will

Variety 5

9. Please list up to three attributes that you consider DURING your consumption of
apples to evaluate its quality (for example, its juiciness).

W

10. Please list up to three visual attributes of the apple that you consider BEFORE
consuming an apple to evaluate its quality (for example, the presence of buises and
blemishes on its skin).

3

W

11. Please list up to three pieces information that you consider when you evaluate the
quality of apples (for example, whether it is organic or not)

1

W

3

12. Do apples from Washington have higher quality than apples from New York, or
vice versa?

Apples from New York have higher quality

38

Apples from Washington have higher quality

It depends
13. In the Lansing area, name up to five stores that you can think of that carry
"specialty apples".

lllll

1
2
3
4
5

14. In which specific country of the world are these varieties of apples traditionally
produced? (please do not guess nor look at the answer on-line)

Fuji

Jonathan I
l._______

Consumer Values
Please rate to what extent you agree or disagree with the following statements.

15. Usually, an important thing that I consider when I choose my food is its
nutritional value, its healthiness and its safety.

Strongly

Disagre 3:33?”
'1’ 2 3 4 5 6 7

t‘ F F t“ t“ (" t“

16. Usually, an important thing that I consider when I choose my food is its taste and
its aspect.

Strongly

. Strongly
Ersagre Agree
2 3 4 5 6 7

I

39

40

17. Usually, an important thing that I consider when I choose my food is if its
production process respects the environment, guarantees good standards of living to small
farmers, and/or contributes to the development of disadvantaged communities.

Sti°ngly Strongly
Disagre A
gree
‘1’ 2 3 4 5 6 7
F f‘ t“ (" (" 1’" P
18. Where would you place yourself on a scale from 1 to 7, if eating food that

respects farmers and/or the environment is a 1 and enjoying the food experience is a 7?

. Enjoying
Eating Food the Food
Respecting Ex rience
Environmen pe

t and/or

People 2 3 6

p r c o t‘ t‘ f“

19. Where would you place yourself on a scale from 1 to 7, if eating healthy food is a
l and eating food that respects farmers and/or the environment is a 7?

Eating Eating Food

Health Respecting

y Food the
Environmen

l 2 3 4 5 6 t7

F t“ F F t“ (‘ t“

20. Where would you place yourself on a scale from 1 to 7, if enjoying the food
experience is a 1 and eating healthy food is a 7?

Enjoying Eating
Food Health
Experience y Food
1 2 3 4 5 6 7

t“ t“ t“ (" (" t“ F

41

21. Please rate to what extent you agree or disagree with the following statements.

"If you eat well, you live well".

Strongly Strongly
Disagree 1 2 3 4 5 6 Agree 7

(" t“ F F r" t“ f‘

22. "If you are able to enjoy your food, you are able to enjoy your life".

Strongly Strongly
Disagree 2 3 4 5 6 Agree

t“ (" F t“ F F f“

23. "As a consumer, you have the power of awarding the companies that respects

small farmers and the environment, as well as punishing those that do not respect them".

Strongly Strongly
Disagree Agree

1 2 3 4 5 6 7

F t" (0‘ t“ (" (" 6‘

Treatment 1

Now imagine that you see an apple claimed to be "locally-grown" in your shopping
location.

24. Which claim did you just receive about this apple?

It is from Washington
It is "locally-grown"

It is organic

42

25. How would you describe your attitude towards an apple claimed to be "locally-

grown"?
Unfavorable Favorable
I 2 3 5 7
t" t“ F F t“
Dislike Like
1 2 3 5 7
C (" F (" t"
Negative Positive
1 2 3 5 7
F t“ F F (7‘
Bad Good
1 2 3 5 7
(T f“ t“ (‘ t“
26. To what extent do you believe that an apple claimed to be "locally-grown" is
really grown locally?
Very Strong
Very Weak Belief
Belief
l 5 7
t‘ t“ F t" (7
To what extent do you expect that an apple claimed to be "locally-grown"...
27. ...has a good flavor?
Very Low Very High
Expectation Expectation
1 4 5 7
t“ t" t“ (" t“ ("

43

28. ...is sweet?

Very Low Very High
Expectation 1 2 3 4 5 6 Expectation 7
t‘ t“ (" (‘ t" (‘ t"
29. ...is firm?
Very Low Very High
Expectation 2 3 4 5 6 Expectation
F t“ (7 F (‘ t" f‘
To what extent do you believe that an apple claimed to be "locally—grown" is...
30. ...healthy?
Very Weak Very Strong
Belief Belief
I 2 3 4 5 6 7
F F t" t" F (‘ F
31. ...free from chemical residues?
Very Weak Very Strong
Belief Belief
l 2 3 4 5 6 7
(s r r r c r r‘
32. ...free from pests and diseases?
Very Weak Very Strong
Belief Belief
l 2 3 4 s 6 7
f‘ r r r: r c r:

44

REFERENCES

Aaker, D. (1991). Managing Brand Equity: Capitalizing on the Value of a Brand Name,
The Free Press, New York.

Ajzen, I. (1991). “The Theory of Planned Behavior” Organi "ational Behavior and
Human Decision Processes 50, 179-211.

Ajzen, I. (2005). Attitudes, personality and behavior. Maidenhead [u.a.]: Open Univ.
Press, 2005.

Akerlof, G. (1970). “The Market for ‘Lemons’: Qualitative Uncertainty and Market
Mechanism.” Quarterly Journal of Economics 89, 488-500.

Alba, J .W. and J .W. Hutchinson (1987). “Dimensions of Consumer Expertise.” Journal
of Consumer Research 13 (3), 411-454.

Alfnes, F. and K. Rickertsen (2003). “European Consumers’ Willingness to Pay for US.

Beef in Experimental Auction Markets.” American Journal of Agricultural Economics 85
(2), 396-405.

Bodenhausen, G. and M. Lichtenstein (1987). “Social Stereotypes and Information-
Processing Strategies: The Impact of Task Complexity. ” Journal of Personality and
Social Psychology 52(5), 871 -880.

Baker, GA. and TA. Bumham (2001). “Consumer Response to Genetically Modified
Foods: Market Segment Analysis and Implications for Producers and Policy Makers.”
Journal of Agricultural and Resource Economics 26 (2), 387-403.

45

Basu, AK. and KL. Hicks (2008). “Label Performance and Willingness to Pay for Fair
Trade Coffee: A Cross-National Perspective.” Discussion Papers on Development Policy
125, Center for Development Research, Bonn, Germany.

Darby, K., M. Batte, S. Ernst and B. Roe (2008). “Decomposing Local: A Conjoint
Analysis of Locally Produced Foods.” American Journal of Agricultural Economics 90
(2), 476-486.

Darby, M. and E. Karni (1973). “Free competition and the optimal amount of fraud.”
Journal of Law and Economics 16 (1), 67-88.

DeEll, J .R. and R.K. Prange (1992). “Post-harvest quality and sensory attributes of
organically and conventionally grown apples.” HortScience 27, 1096-1099.

De Pelsmacker, P., L. Driesen, and G. Rayp (2005). “Do Consumers care about Ethics?
Willingness to Pay for Fair-Trade Coffee.” The Journal of Consumers Affairs 39 (2),
363-3 85.

Eagly, AH. and S. Chaiken (1993). The psychology of attitudes, Fort Worth, Texas,
Harcourt Brace.

Ehmke, M.T., J .L. Lusk and W. Tyner (2008). “The Relative Importance of Preferences
for Country-of-origin in China, France, Niger and the United States.” Paper presented at

the International Association of Agricultural Economists Conference, Gold Coast,
Australia, August 12-18, 2006.

Ericksson, G., J. Johansson and P. Chao (1984). “Image Variables in Multi-Attribute

Product Evaluations: Country-of-Origin Effects.” Journal of Consumer Research 11 (2),
694-699.

F ishbein M. (1967). “A Behaviour Theory Approach to the Relations Between Beliefs
about an Object and the Attitude Toward the Object.” Readings in Attitude Theory and
Measurement, John Wiley, New York, 389-400.

F ishbein, M. and I. Ajzen (1975). Belief, Attitude, Intention and Behavior: An
Introduction to Theory and Research, Reading, Massachusetts: Addison-Wesley.

F ishbein, M. and I. Ajzen (1980). Understanding Attitudes and Predicting Social
Behavior, New Jersey: Prentice Hall.

Froelich, E.J., J .G. Carlberg and CE. Ward (2009). “Willingness-to-Pay for Fresh Brand
Name Beef.” Canadian Journal of Agricultural Economics 57, 119-137.

46

The Food Institute (2008). Consumer Understanding of Buying Local, The Food Institute,
Washington DC. Website: www.foodinstitute.com/buyinglocal.cfm

Han, M. and V. Terpstra (1988). “Country-Of-Origin Effects for Uni-National and Bi-
National Products.” Journal of International Business Studies 19 (2), 235-255.

Hoffman, DH. and T.P. Novak (1996). “Marketing in Hypermedia Computer-Mediated
Environments: Conceptual Foundations.” Journal of Marketing 60 (3), 50-68.

Hong, S. and RS. Wyer (1989). “Effects of country-of-origin and product-attribute
information on product evaluation: an information processing perspective.” Journal of
Consumer Research 16, 175—187.

Huang, CL. (1996). “Consumer preferences and attitudes towards organically grown
produce.” European Review of A gricultural Economics 23 (3) 331-350.

IDDBA (2008). What ’s In Store Report 2008, International Dairy Deli Bakery
Association, Madison, Wisconsin. Website: www.iddba.org.

Johansson, J. and I. Nebenzahl (1986). “Multinational production: effect on brand value.”
Journal of International Business Studies 17 (3), 101—126.

Johnson, El and J .E. Russo (1984). "Product Familiarity and Learning New
Information." Journal of Consumer Research 11 (6), 542-550.

Kanter, C., K.D. Messer, and HM. Kaiser (2008). “Do rBST-Free and Organic Milk
Stigmatize Conventionally Produced Milk?” Selected Paper prepared for presentation at

the American Agricultural Economics Association Annual Meeting, Orlando, 'FL, July 27-
29, 2008.

Keller, K. L. (1998). Strategic Brand Management. Building, Measuring and Managing
Brand Equity, Upper Saddle River, NJ: Prentice Hall.

Kirmani, A. and AR. Rao (2000). “No pain, No Gain: A Critical Review of the
Literature on Signaling Unobservable Product Quality.” Journal of Marketing 64, 66-79.

Lee, D.Y. (2000). “Retail bargaining behavior of American and Chinese customers.”
European Journal of Marketing 34 (1), 190-206.

Li, W. and R. Wyer (1994). “The Role of Country of Origin in Product Evaluations:
Informational and Standard-of- Comparison Effects.” Journal of Consumer Psychology,
3 (2), 187-212. v

47

Loureiro, M.L., J .J . McCluskey and RC. Mittelhammer (2002). “Will Consumers Pay a
Premium for Eco-Labeled Apples?” The Journal of Consumer Affairs 36 (2), 203 —219.

Loureiro, ML. and S. Hine (2002). “Discovering Niche Markets: A Comparison of
Consumer Willingness to Pay for Local (Colorado Grown), Organic, and GMO—Free
Products.” Journal of Agricultural and Applied Economics 34 (3), 477-487.

Loureiro, ML. and W.L. Umberger (2005). “A Choice Experiment Model for Beef: What
US Consumer Responses Tell Us about Relative Preferences for Food Safety, Country-

of-Orgin Labeling and Traceability.” Journal of A gricultural and Applied Economics 37
(1), 49-63.

Loureiro, ML. and W.J. Umberger (2007). “Estimating Consumer Willingness to Pay for
Country-of-Origin Labeling.” Journal of A gricultural and Resource Economics 28 (2),
287-301.

Lusk, J .L., M.S. Daniel, D.R. Mark and CL. Lusk (2001). “Alternative calibration and
auction institutions for predicting consumer willingness-to-pay for non-genetically
modified corn chips.” Journal of Agriculture and Resource Economics 26, 40-57.

Lusk, J .L., J. Roosen and J .A. Fox (2003). “Demand for Beef from Cattle Administered
Growth Hormones or Fed Genetically Modified Corn: A Comparison of Consumers in
France, Germany, the United Kingdom, and the United States.” American Journal of
Agricultural Economics 85, 16-29.

Lusk, J .L., J. Brown, T. Mark, I. Proseku,R. Thompson and J. Welsh (2006). “Consumer
Behavior, Public Policy, and Country-of-Origin Labeling.” Review of Agricultural
Economics 28 (2), 284-292.

Lutz, R. (1991). “The Role of Attitude Theory in Marketing”, in: Kassanian, Harold H.
and J .J . Robertson Editors, 1991. Perspectives in Consumer Behavior (4th ed.), Prentice-
Hall.

Manalo, AB. (1990). “Assessing the importance of apple attributes: an agricultural
application of conjoint analysis.” Agricultural and Resource Economics Review 19 (2),
118-124.

Mehinnagic, E., G.E. Royera, D. Bertrand, R. Symoneaux, F. Laurens and F. Jourjon
(2003). “Relationship between sensory analysis, penetrometry and visible NIR
spectroscopy of apples belonging to different cultivars.” Food Quality and Preference 14,
473-484.

48

McFadden, D. (1986). “The choice theory approach to marketing research.” Marketing
Science 5, 275—297.

Nelson, P. (1970). “Information and Consumer Behavior.” The Journal of Political
Economy 78 (2), 311-329.

Nimon, W. and J. Beghin (1999). “Are Eco-Labels Valuable? Evidence from the Apparel
Industry.” American Journal of Agricultural Economics 81 (4), 801-811.

Obermiller, C. and E. Spangenberg (1989). “Exploring the effects of origin labels: an
information processing framework.” Advances in Consumer Research 16, 454-459.

Olson, J. (1978). "Inferential Belief Formation in the Cue Utilization Process." Advances
in Consumer Research 5, 706-713.

Park, CW. and VP. Lessig (1981). "Familiarity and Its Impact on Consumer Decision
Biases and Heuristics." Journal of Consumer Research 8 (9), 223-230.

Peterson, RA. and A.J.P. Jolibert (1995). “A Meta-Analysis of Country of Origin
Effects.” Journal of International Business Studies 49 (4), 883-900.

Pharr, J. (2005). “Synthesizing country-of-origin research from the last decade: Is the
concept still salient in an era of global brands?” Journal of Marketing Theory and
Practice 13 (4), 34-45.

Punj, ON. and R. Staelin (1983). "A Model of Consumer Information Search Behavior
for New Automobiles." Journal of Consumer Research 9 (1), 366-3 80.

Rao, AR. and KB. Monroe (1988). “The Moderating Effect of Prior Knowledge on Cue
Utilization in Product Evaluations.” Journal of Consumer Research 15 (2), 253-264.

Rao, AR. and KB. Monroe (1989). “The Effect of Price, Brand Name, and Store Name
on Buyers’ Perceptions of Product Quality: An Integrative Review.” Journal of
Marketing Research 26 (3), 351-357.

Rosenman, RE. and W. Wilson (1991). “Quality Differential and Prices: Are Cherries
Lemons?” The Journal of Industrial Economics 39 (6), 649-658.

Schroeder, T., G. Tonsor, J. Pennings and J. Mintert (2007). “Consumer Food Safety
Risk Perceptions and Attitudes: Impacts on Beef Consumption across Countries.” The
BB. Journal of Economic Analysis and Policy 7 (1), 19-32.

49

Schwartz, S. H. (1992). “Universals in the content and structure of values: Theory and
empirical tests in 20 countries”, in M. Zanna (Ed), Advances in experimental social
psychology 25, 1—65. New York: Academic Press.

Sheth, J .N., B.I. Newman and BL. Gross (1991). Consumption Values and Market
Choices: Theory and Applications, Cincinnati, Ohio: Southwestern Publishing Company.

Thompson, G. (1998). “Consumer Demand for Organic Foods: What We Know and
What We Need to Know.” American Journal of A gricultural Economics 80 (5), 1113-
1118.

Tonsor, G., T. Schroeder, J. Fox and A. Biere (2005). “European Preferences for Beef
Steak Attributes.” Journal of Agricultural and Resource Economics 30 (2), 367-380.

Umberger, W.J., D.M. Feuz, C.R. Calkins, and BM. Stiz (2003). “Country-of-Origin
Labeling of Beef Products: US. Consumers Perceptions.” Journal of Food Distribution
Research 34, 103-116.

Van der Lans, 1., K. Van Ittersum, A. De Cicco, and M. Loseby (2001). “The role of
region of origin and EU certificates of origin in consumer evaluation of food products.”
European Review of Agricultural Economics 28 (4), 451-477.

Van Ittersum, K., M.J.J.M. Candel, and M.T.G. Meulenberg (2003). “The influence of
the image of a product's region of origin on product evaluation.” Journal of Business
Research 56 (3), 215-226.

Verbeke, W. (2005). “Consumer acceptance of functional foods: socio-demographic,
cognitive and attitudinal determinants.” Food Quality and Preference 16, 45-57.

Verlegh, P. and J .B. Steenkamp (1999). “A review and meta-analysis of country-of-origin
research.” Journal of Economic Psychology 20 (5), 521-546.

50

Chapter 2

BUILDING INDIVIDUAL BRANDS WITH PLACE-OF-ORIGIN INFORMATION:

IMPLICATIONS FOR THE INDUSTRY

Growing segments of world consumers seek better quality, healthiness and larger variety
in their food consumption (Verbeke, 2005; IDDBA, 2008). As part of this process,
consumers’ attention for place-of-origin (POO) attributes as part of the demand of agri-
food products is increasing (Grunert, 2005). The major dimension of P00 attributes that
have been studied in agricultural economics and marketing literature since the 19603 is
country-of-origin (Dichter, 1962; Schooler, 1965; Peterson and Jolibert, 1995; Verlegh
and Steenkamp, 1999; Loureiro and Umberger, 2005; Ehmke etal., 2008) while in the
last decade region-of-origin attributes have been studied separately, in relation with
consumers’ values for tradition and authenticity of agri-food products (Kuznesof et al.,

1997; Tregear et al., 1998; Van der Lans et al., 2001; Scarpa et al., 2005).

These strands of the literature have primarily analyzed the impact of POO
attribute information on consumers’ product evaluation (Peterson and J olibert, 1995;
Verlegh and Steenkamp, 1999; Loureiro and Umberger, 2005) and the factors explaining
how this impact varies, including consumers’ characteristics (Knight and Calantone, 2000;

Scarpa et al., 2005; Ehmke et al., 2008; Gao and Schroder, 2009), consumers’
51

motivations (Verlegh and Steenkamp, 1999; Van der Lans et al., 2001; Loureiro and
Umberger, 2005; Lusk etal., 2006) and product characteristics (Ward et al., 2003). While
these studies found that POO attributes have a significant impact, either positive or
negative, on consumers’ evaluation of a generic product, only a few studies analyzed how
POO attribute information can differentiate an individual brand from the other brands
within a product category. The problem of brand differentiation within a product category
with a POO attribute - for example, differentiation of the brand Sun Maid California
raisins from other California raisins (Crespi and Marette, 2002) - has been already tackled
in the literature (Brester and Schroeder, 1995; Kaiser and Liu, 1998; Chung and Kaiser,
2000; Crespi and Marette, 2002). However, these studies are based upon estimations of
the aggregate demand elasticity for various types of P00 information, but they have not

analyzed individual consumers’ perceptions and buying intentions.

This study has the purpose of starting to fill this gap by analyzing which POO
information, if any, has an exclusive impact on an individual brand and differentiate it
from other brands within a product category with a POO attribute. Analyzing which POO
information differentiates a brand from other brands with the same POO attribute has
important implications for firms within an industry. As suggested by Crespi and Marette
(2002), through POO information a firm owning a brand can create a competitive
advantage over other firms selling products with the same POO attribute. This can be
crucial as firms may not always cooperate with its neighbors to jointly increase
consumers’ evaluation for its POO, but they may also be forced to compete with them to

get access to a limited number of buyers (Steenkamp and Van Trijp, 1996).

52

To analyze the impact of P00 information on consumers’ evaluation for an
individual brand, we found appropriate to develop a theoretical framework that builds
upon the theory of attitude formation, developed in psychology and largely applied to
marketing (Fishbein, 1967; F ishbein and Ajzen, 1975; Lutz, 1991). According to the
theory of attitude formation, consumers form their attitude towards a brand and
ultimately their intentions to buy it not only by evaluating how much they like each brand
attribute, but also by assessing how strongly they believe that the brand is associated to
each attribute (Fishbein, 1967; Lutz, 1991). Therefore, to improve the understanding of
consumers’ choice of individual brands with POO attributes, we analyze which POO
information has an impact on consumers’ belief strength of the association between the
individual brand and the P00 attribute, on their attitudes towards a brand and ultimately

on their WTP a premium price for a brand relatively to its competitors in the market.

We test our proposed theoretical framework with data collected from 241
graduate students at Michigan State University through an intemet-based experiment. In
particular, we assess differences in respondents’ beliefs, attitudes and WTP a premium
price for a brand across groups receiving different sets of P00 information treatments

with a series of path models.

This article is organized as follows. In section 2, we review the relevant literature
on POO attributes and formation of attitudes. Building upon this literature, we present
our theoretical framework and hypotheses in section 3. In section 4, we describe the
method we use to test these hypotheses. Results are presented in section 5, while our

conclusions are drawn in section 6.

53

Literature Review

The. concept of “place-of-origin” attributes has often been used in the literature (Moore,
1980; Hong and Wyer, 1990; Bertozzi, 1995; Van der Lans et al., 2001; Skuras and
Vakrou, 2002) to generalize the analysis across country-of-origin attributes and region-
of-origin attributes. Country-of-origin attribute are typically communicated through the
phrase "made in” a specific country and they are an extn'nsic product cue similarly to
price, brand name or warranty, as none of these directly bear on product performance
(Peterson and Jolibert, 1995). Region-of-origin attributes are similar extrinsic product
cues that, differently from country-of-origin attributes, are usually strongly associated to
culture, history and people of a geographical area (Kuznesof et al., 1997; Van der Lans et
al., 2001). Wines, cheeses, onions, grapefruits, wooden furniture, perfumes, and cigars
are all examples of products being marketed as national products (Centner et al., 1989;
Kotler et al., 1993; Papadopoulos, 1993), but the same influence is expected in other
places of origin such as regions and provinces (Hauser, 1993; Balling, 1995) but also

trade zones and continents (Papadopoulos, 1993; Smith, 1993).

From the 19603, many studies in marketing and agricultural economics literature
have agreed that POO attributes may have a positive impact on consumers’ evaluation of
products (Dichter, 1962; Schooler, 1965; Peterson and Jolibert, 1995; Verlegh and
Steenkamp, 1999; Balabanis and Diamantopoulos, 2004; Loureiro and Umberger, 2005;
Ehmke et al., 2008), but also a negative impact when country image is not favorable to ’
consumers or when the image of the place does not fit with the product (Van Ittersum etb

al., 2000). However, the impact of POO may vary according to various consumer

54

characteristics (Bonnet and Simioni, 2001; Scarpa et al., 2005) and product

characteristics (Agrawal and Kamakura, 1999;.Ward et al., 2003).

These studies have analyzed the impact of adding a POO attribute on consumers’
attitudes and buying intentions for a generic product, but not for an individual brand
(V erlegh and Steenkamp, 1999; Piron, 2000; Pecotich and Rosenthal, 2001). However, in
the marketplace brands can be more or less associated to the place of origin by consumers,
according to the other brand attributes and to brand information, which act as signals of
the P00 attribute (Keller, 1993). Therefore, the impact of adding POO attributes on
consumers’ attitudes and buying intentions may vary significantly according to factors
related to the individual brands. In this study, we attempt to integrate the extant literature
by exploring which information influences the impact of P00 attributes on consumers’
beliefs, attitudes and intentions to buy a brand and so differentiates an individual brand

from its’competitors with the same POO attributes.
Attribute and Brand Place-afiOrigin Information

To explore which information has an impact on consumers’ belief strength of the
association between a brand and the POO, we first distinguish between the concepts of
attribute P00 information and brand P00 information. This distinction is similar to the
distinction between generic advertising and branded advertising introduced by Crespi and

Marette (2002).

Attribute and brand POO information have the same ultimate goal, that is
increasing consumers’ evaluation and WTP for brands with the POO attribute, but differ
as they provide fundamentally differentmessages to consumers. In particular, attribute

P00 information aims at increasing consumers’ evaluations for brands with the POO

55

attribute by emphasizing the benefits given by the POO attribute and creating
associations between the POO attribute and other favorable attributes. For example,
attribute information about California raisins aims at increasing consumers’ evaluation
for the attribute “California” when associated to the product “raisins”. Also, this
information aims at increasing consumers’ beliefs that the attribute “California” is
associated with other quality attributes such as “good flavor” and “eco-friendly”. On the
other hand, brand P00 information aims at increasing consumers’ evaluations for brands
with the POO attribute by strengthening consumers’ beliefs that a particular brand is
associated with the POO attribute and with other favorable attributes. For example, brand
information about “Sun Maid Raisins” aims at increasing consumers’ association
between Sun Maid and California, as well as other favorable associations such as

between Sun Maid and “good flavor” or between Sun Maid and “sunny land”.

Generic advertising, “place branding” and in large part also POO certifications
and labels release attribute information, as they aim at increasing consumers’ evaluation
for POO attribute and at strengthening the association between the P00 and other
favorable attributes. When promoting a product with a POO attribute, such as California
Raisins, generic advertising aiming at shifting consumer demand for the entire product
category with that attribute. (Brester and Schroeder, 1995; Piggott et al., 1996; Kinnucan
et al., 1997; Kaiser and Liu, 1998; Chung and Kaiser, 2000; Crespi and Marette, 2002)
and at creating a favorable product-country image (Laroche et al., 2005, Lusk etal.,
2006), whereas an image can be defined as a set of strong and consistent associations that

reinforce each other (Keller, 1993). Generic advertising can be implemented both by a

56

private firm, a private group of firms or by a public entity representing the firms within a

territory.

Similarly, “place branding” (Kotler et al., 1993; Kavaratzis, 2005; Iversen and
Hem, 2008) are promotion activities aiming at building an image and reputation across all
the products and services offered within the POO, assuming that there might be spillover
effects across different products and services from the same POO. Place branding
activities are usually implemented by multiple private actors within a POO as well as

funded or sometimes even managed by public entity.

Promotion and advertising activities implemented by firms, inter-firm
organizations or public entities can be classified as means to release either attribute

information or brand POO information, or potentially to release both.

Information on POO certifications and labels, such as Protected Denominations of
Origin in Europe (Van Ittersum et al., 2000; Bonnet and Simioni, 2001; Verbeke and
Ward, 2006) or voluntary Country-of-Origin Labeling (COOL) (Loureiro and Umberger,
2005; Lusk et al., 2006; Verbeke and Ward, 2006), usually has a more ambiguous role in
changing consumers’ perceptions. On one hand, POO certifications and labels are seals
guaranteeing consumers that a specific branded product is indeed from the P00 and is
produced according to certain quality standards, and therefore would classify as brand
information. On the other hand, most of the information on POO certifications and labels
usually aims at convincing consumers that the entire product category with the POO
attribute is controlled and selected according to high quality standards, and therefore

classifies as an attribute information, as it aims at increasing consumers’ evaluation for

the POO attribute.

57

In marketing literature, the most cited example of means releasing brand POO
information is a company’s advertising of its brand(s) (Leclerc et al., 1994; Shocker et al.,
1994; D’Astous and Ahmed, 1999; Haubl and Elrod, 1999; Clarke etal., 2000; Andrews
and Kim, 2007; Pappu et al., 2007). This is what Crespi and Marette (2002) call branded
advertising, providing the example of “California Sun Maid” brand advertising. Pace
Picante sauce from Texas, Coors beer from Colorado (Takor and Lavack, 2003) and
Zespri kiwi from New Zealand (Beverland, 2001) are examples of other brands with POO
attributes that have been heavily publicized with mass media advertising. In these cases,
advertising has created a strong consumers’ association of the brand name, the company
ownership or the source of components with the POO attribute. In the agri-food markets,
another commonly cited means to release brand information are private third-party
certifications (Farina and Reardon, 2000; Giovannucci and Reardon, 2000; Reardon et al.,
2001; Konefal et al., 2005). Private third-party certifications are used by food
manufacturers and retailers to provide quality, origin and safety assurance to their
consumers and therefore to complement their brands (Hatanaka et al., 2005). Although
they often proved to be effective, mass media advertising and private certification
systems require a large financial investment which is unbearable by the large majority of

agri-food firms aiming at differentiating their product from competitors.

Third-party endorsements and appraisals from actors with high status in the
marketplace — such as chefs, cultural associations linked to particular territory and
journalists in the food sector — are alternative means to provide positive information
about an individual brand (Dean, 1999, Huffman et al., 2004; Andrews and Kim, 2007)

and so can be classified as providing brand information. As theorized by Podolny (1993),

58

a seller’s social ties with actors with high status in the market can act as signals of
product quality or of specific product attributes. Third-party endorsements and appraisals
from actors with high market status can isolate a brand from the others as these actors
cannot endorse all the brands in the marketplace. However, if they would endorse too
many brands, the third-party endorsers would risk losing their own reputation and.
credibility (Dean, 1999). Similarly to third-party endorsements, the information that a
retailer or a buyer with high reputation and an image related to a place of origin, as well
as the participation to events, food competitions and fairs related to a place of origin can

classify as brand information (Aaker, 1991; Keller, 1993).

In this study, we analyze if attribute information differentiates a brand from its
competitors with the same POO attribute or if brand POO information has a higher
impact on consumers’ beliefs, attitudes and intentions to pay a premium price for a brand

relatively to its competing brands.
Consumers ’ Beliefs, Attitudes and Buying Intentions

According to the learning theory (F ishbein, 1967), consumers’ attitudes towards a brand
and their decision to buy it are driven by the evaluation of the individual brand attributes.
In particular, based upon their prior beliefs and by processing new information,
consumers form their evaluations for the single brand attributes and their beliefs that the
brand is associated with the attribute (F ishbein, 1967). In the case of P00 attributes, for
example, consumers form their attitudes towards “Zespri” kiwi brand by assessing how
much they like kiwi from New Zealand as well as how much they believe that the brand
“Zespri” is really associated with New Zealand kiwis. As POO attributes are credence

attributes, which means that they are verifiable by consumers neither before nor after

59

consumption (Darby and Karni, 1973), information play a key role in determining
consumers’ beliefs and, in turn, consumers’ beliefs are crucial to establish their attitudes

towards products (Coswell and Mojduszka, 1996).

However, researchers found that consumers’ attitudes towards a brand do not
predict buying behavior accurately (Fishbein and Ajzen, 1975). Instead, consumers’
attitudes towards the action of buying the brand, moderated by their subjective norms,
predict buying intentions much more accurately, as tested in the theory of attitude
formation (Fishbein and Ajzen, 1975; Sheppard et al., 1988). Buying intentions predict
behavior “unless intent changes prior to performance” or “unless the intention measure
does not correspond to the behavioral criterion in terms of action, target, context, time-
frame and/or specificity” (Sheppard et al., 1988). The intention of buying a brand has
various dimensions. The most generally accepted is the willingness to do an effort to
perform to the buying action (F ishbein and Ajzen, 1975; Eagly and Chaiken, 1993). The
nature of such an effort may vary according to the context: it may be the WTP to obtain a
product from that brand, the likelihood to pay a premium for that brand, or the likelihood
to buy the product even if it is not sold in the most favorite purchasing location. A second
key dimension of buying intentions is the choice of the brand among alternatives (Ajzen
and Fishbein, 1980), which is the process of comparing and selecting among the ‘

intentions associated with each alternatives in the choice set.

Theoretical Framework
To analyze which POO information gives higher consumers’ attitudes towards a brand

and WTP a premium price and so differentiates a brand from all its competing brands -

60

including those with the same POO attribute - we propose a theoretical framework that
builds upon the learning theory of attitude formation (Fishbein, 1967) and the theory of

reasoned action (Fishbein and Ajzen, 1975).

First of all, we assume that both individual firms and collective organizations ,
representing a place of origin can give attribute information, while we assume that only
individual firms can give brand information. As a matter of facts, while collective
organizations have the incentive of promoting the region they represent as a whole, they
should have no incentive to promote individual brands within their region but not the
others. Also, we assume that firms generally have lower costs when they provide POO
information through a collective organization than by themselves, because of economies

of scale in advertising and marketing.

Based upon these assumptions, we first develop hypotheses about attribute POO
information. In particular, through the first two hypotheses (HI-H2), we tackle the initial
questions: do firms giving attribute POO information create a benefit advantage for their
own brand over other brands? Therefore, should an individual firm give attribute POO
information or should giving attribute information be an exclusive task of collective inter-

firm organizations?

Similarly to previous literature on generic advertising and brand differentiation
(Kaiser and Liu, 1998; Crespi and Marette, 2002), we hypothesize that attribute
information, either given by an individual firm or an inter-firm organization representing
some or all the firms within the P00, have no different impact on consumers’ beliefs and

attitudes towards competing brands within the same POO, and so does not create any

61

differentiation among them. In particular, assuming that Firm A owns Brand A and Firm

B owns Brand B, we hypothesize:

H1. Given Firm A and Firm B being two firms from the same POO, attribute POO
information given by Firm A has no different impact on consumer willingness

to pay a premium price for “Brand A” relative to “Brand B”.

H2. Given an individual firm owning “Brand A” and a collective inter-firm
organization being from the same POO, attribute POO information given by
the individual has no different impact than the same information provided by

the collective inter-firm organization on the willingness to pay a premium price

for “Brand A”.

If we found that empirical evidence supporting these two hypotheses, we would
then imply that Firm A should not provide attribute information but should let a

collective inter-firm organization give this information, whenever the latter exists.

Furthermore, we develop our third and last hypothesis (H3) to compare the impact
of brand POO information and attribute POO information on consumers’ beliefs and
attitudes towards a brand. Similarly to Crespi and Marette (2002), who provide evidence
that generic advertising giving attribute information reduce brand differentiation within
firms of the same POO, we hypothesize that brand POO information has significantly
larger positive impact on consumers’ attitudes towards a brand than attribute POO
information. However, we hypothesize that this would hold only when consumers have
been previously exposed to some attribute POO information, no matter what is the source:

consistently with Fishbein (1967), if consumers have strong brand origin associations

62

relative to a place that they completely ignore, brand POO information may not have a

high impact on consumers’ attitudes. Therefore, we hypothesize that:

H3. Once consumers have already received attribute POO information, the impact
of brand POO information on the willingness to pay for a brand is higher than

the impact of any further attribute POO information.

If we found empirical evidence supporting this third hypothesis, we would imply
that a firm should give brand POO information to consumers rather than attribute POO
information, whenever other firms or a collective inter-firm organization provides

attribute POO information.

Methods

To test our hypotheses, we collected data through an intemet-based experiment
administered to 241 graduate students from Michigan State University. When sending an
e-mail advertisement to recruit students to undertake the test, we voluntarily asked for
people that “are interested in food from different places and cultures”, as we wanted to
attract respondents that are highly motivated in learning and potentially buying new

products from scarcely known places of origin.

To perform such an experiment, we chose our products of interest using three key
criteria. First, we looked for products that, when associated to a particular place of origin,
are completely unknown to our sample. In this way, we attempt to give information
treatments to respondents that have very weak prior beliefs regarding to the products

associated to the POO attribute, as respondents’ prior beliefs may largely vary according

63

to their individual experiences (John et al., 1986). As we expect that prior beliefs are
weak, we could assume that respondents’ initial-beliefs and attitudes towards an unknown
product with POO attribute are very similar to each other. Second, we looked for
products that are quite regularly used by the majority of consumers, although their
familiarity towards the product may largely vary. Third, we looked for ”products that are
commonly promoted with POO information, both by individual firms and national or

regional collective associations of producers.

We finally choose Creole cream cheese from Southern Louisiana (USA) and
extra-virgin olive oil from Riviera Ligure (Italy) as our products of interest. We found
that 85% of our respondents consume olive oil and 75% of them consume cream cheese
at least once a month. However, only four of them (i.e., 0.02% of our sample) have heard
before about olive oil from Riviera Ligure and only seven of them (i.e., 0.03% of our
sample) have heard before Creole cream cheese from Southern Louisiana. Therefore, we
assumed that our sample has some basic knowledge of and involvement with olive oil
and cream cheese as generic products, while they have weak prior beliefs on these
products when associated to places of origin such as Riviera Ligure and Southern

Louisiana.

Experimental Procedure

The experiment was conducted in June 2009. Respondents were recruited through an
email advertisement by the researchers, with the support of the MSU Office of the

Registrar. Each respondent undertook a questionnaire divided in an initial demographics

64

section plus two sections with information treatments and measurements. Each section of

the experiment is used to collect data to test different hypotheses.

In the initial demographics section, respondents were asked preliminary questions
about their gender and nationality, as well as their initial attitude towards and their use of
olive oil and cream cheese. In the first section, we collect data on the impact of product-
level information (H1 and H2). Respondents were divided in four groups: the first group
of respondents received firm A’s attribute POO information treatment and their POO
beliefs, attitudes and buying intentions for Brand A were measured. The second group of
respondents received the same treatment as the first group, but their POO beliefs,
attitudes and buying intentions were measured for Brand B, which has the same POO
attribute. The third group of respondents received a collective inter-firm organization’s
attribute POO information treatment and their POO beliefs, attitudes and buying
intentions for Brand A were measured. Finally, the fourth group received only a brief
description of the product with the attribute, and their POO beliefs, attitudes and buying
intentions for both Brand A and Brand B were measured. Both firm A’s and collective
inter-firm organization’s attribute POO information were manipulated with a 5-row very
positive description of the product together with an information about a POO certification
of the product that, as assumed, provided attribute information. Consumers’ belief
strength of the association between the P00 and the brand was measured with a single
seven-point Likert-scale item (strongly disagree/strongly agree). Consumers’ attitudes
were measured with a similar seven-point Likert-scale item, from very negativeto very

positive. Consumers buying intentions were measured with an individual question such as

65

“Would you pay a premium to have brand A rather than another brandfrom the same
place-of-origin?”, where the possible answers were yes, no or “I don’t know”.

In the second section of the questionnaire, we collected data on the impact of
brand POO information and attribute POO information on consumers’ beliefs, attitudes
and WTP a premium price. First of all, the group of respondents that did not receive any
attribute POO information in section 1 of the experiment received it, such that every
respondent received some attribute POO information before the start of section 2.
Therefore, respondents were divided in two groups: the first group of respondents was
given another firm (say, Firm C)’s product-level POO information, and then their POO
beliefs, attitudes and buying intentions for Brand C were measured as in section 1. The
second group of respondents was instead given Firm C’s brand POO information, and
then their beliefs, attitudes and buying intentions for Brand C are measured as in section
1. Brand information consists of a set of endorsements to Brand C from a set of
information sources (including three sources among private 3rd party certifiers,

distributors, chefs, journalists, tasters or non-govemmental organizations).
The Model

To establish the relationships among the administered information treatments and the
measured variables, we analyzed the data collected with a series of three path models.
Each path model is used to evaluate each hypothesis.

A path model represents a set of regressions where independent variables are both
exogenous and endogenous, which means that some of them in turn maybe being caused
by other variables (Hair et a1. 2006). In other words, within a path model independent

variables in one equation may be dependent variables in another equation. The model is

66

usually depicted in a circle-and-arrow figure in which single-headed arrows indicate
causation. Path analysis requires the usual assumptions of regression. It is particularly
sensitive to model specification because failure to include relevant causal variables or
inclusion of extraneous variables often substantially affects the path coefficients, which
are used to assess the relative importance of various direct and indirect causal paths to the
dependent variable. Differently from structural equation models, path models estimate the
relationship among measurable variables rather than among latent factors (Hair et a1.
2006)

The information treatments administered in our experiment are obviously
exogenous variables in these path models. These are modeled as binary variables, named
INF 0, whereas the “0” and “1” values have a different meaning in each path model built.
Consistently with the three hypotheses tested, in the first path model the value “0” of the
binary variable represents the impact of POO attribute information on “Brand A”, while

66199

the value represents the impact of P00 attribute information on “Brand B”. In the
second path model, the value “0” of the binary variable represents POO attribute
information from an individual firm and the value “1” represents POO attribute
information from a collective organization. Finally, in the third path model the value “0”
represents POO attribute information from an individual firm, while the value “1”
represents POO brand information.

Variables describing respondents’ gender, nationality, initial attitudes and habits
towards the products are also exogenous variables in the path models, while respondents’

beliefs, attitudes and willingness to pay a price premium (WTPP) are endogenous

variables. Therefore, the path models can be generically described as:

67

BELSEL= B HINFO+BnDEM+B131NIT+81 (1)
BELFLAV = lelNFO + [322 BELSEL + B 23DEM + B24 INIT + 82 (2)
ATT = B3IINFO + [332 BELSEL + [333 BELFLAV + B 34DEM + B351NIT + 83 (3)

WTPP = [34.INFO + B42BELSEL + [343 BELSEL + BMDEM + B451NIT + 84, (4)

where BELSEL stands for respondents’ beliefs that the branded product is from the most
carefully selected product from the place-of-origin, BELFLAV means respondents’
beliefs that the branded product has a good flavor, ATT stands for respondents’ attitudes
towards the brand and WTPP stands for the willingness to pay a premium price for the
brand relatively to other brands from the same place of origin]. Moreover, DEM and
[NIT are vectors of exogenous variables including demographics and initial attitudes and
habits of respondents. Finally, the binary variable INFO has a different meaning
according to the hypothesis tasted, as described above.

The relationships among these variables are estimated through the path I
coefficients ([33). These are standardized regression coefficients showing the direct effect
of an independent variable on a dependent variable in the path mOdel. Thus when the
model has two or more causal variables, path coefficients are partial regression
coefficients which measure the extent of effect of one variable on another in the path
model controlling for other prior variables. As an intercept is not usually included in path
models, [33 describe deviations from the mean (Bentler 2004). The residual error terms or
disturbance terms, described by 3., 82, 33 and 84 reflect unexplained variance plus
measurement error.

The path analysis has been conducted with the computer package EQS 6.1

(Bentler, 2004). Satorra-Bentler robust maximum likelihood estimation (MLE) has been

68

used to obtain path coefficients, as various binary categorical variables are present in the
model, including INFO, gender and nationality (Bentler, 2004). The dependent variable
WTPP is binary but has been treated as continuous in these path models. Treating a
dependent binary variable as continuous gives the opportunity of including it in the path
model and so of analyzing the impact of endogenous variables on it, which would have
not been possible otherwise (Bentler ,2004). While treating dependent binaryvariables as
continuous is generally inappropriate and can lead to meaningless interpretations
(MacCallum et al., 2002), such as estimated probabilities that “1” happens below zero or
above 1, this is generally considered as acceptable when the sample size is not too large
and the quantity of “1” within the sample is quite low (Bentler, 2004), as it is our samplez.
Seven-point Likert-scale measured variables have been treated as continuous variables as
well, consistently with use largely made in the literature (Likert ,1932; Cronbach, 1950;

Peabody, 1962).

Results

With our first path model, we analyze if attribute POO information about a place of
origin given by a firm does provide an advantage to its own brand relative to competing
brands within the same place of origin. With the second model, we analyze if attribute
POO information from a collective organization has a different impact on respondents’
WTP a premium from the same type of information from an individual firm. Finally, with
our third path model, we analyze if brand POO information has a different impact on

respondents’ WTP a premium relatively to attribute POO information.

69

Impact of attribute information on competing brands

Results show that the attribute POO information about Riviera Li gure olive oil does have
a significantly higher impact on respondents’ willingness to pay a premium for the brand
of the firm releasing the information (“Brand A”) relatively to other brands from the
same place of origin (“Brand B”). The direct effect of a firmgiving attribute information
on respondents’ willingness to pay a premium for “Brand B” relatively to “Brand A” is -
0.10 percentage points, while its indirect effect is equal to -0.05 percentage points. As
common in path analysis (Bentler 1994), we computed the indirect effect as the sum of
the products of all the path coefficients linking the information treatment to the
willingness to pay a premium for the brands. The sum between the indirect and the direct
effect results in a total causal effect of -0. 1 5 percentage points (see table 3). This means
that when attribute information is provided by the firm owning “Brand A”, the expected
percentage of respondents that are willing to pay a premium for “Brand B” is 15% lower
than the mean, that is 25%-15%=10%. The overall fit of this model is very good, with
Comparative Fit Index (CFI)=1 and RMSEA=0, and it is comparatively better than the fit
of a path model where any direct or indirect path between the information treatment and
the final dependent variable is fixed to zero. Based on this evaluation of the goodness-to-
fit, we consider this differential impact of attribute POO information on respondents’
willingness to pay a premium significant.

In the case of cream cheese, the advantage provided by attribute POO information about
Southern Louisiana cream cheese to “Brand A” relatively to “Brand B” is much smaller
than in the case of Riviera Ligure olive oil. The total causal effect is equal to -0.02, which

is given by the sum of the direct effect (0.06) and the indirect effect (-0.08) (see table 4).

70

Table 3 - Impact of Attribute Information on Brand A and Brand B Olive Oil

 

 

 

 

 

 

 

 

 

 

 

 

 

Dependent Independent Variables Standardized Path Estimates Std.

Variable Errors

BELSEL Mean 5.05
INFO (0=on brand A; l=on
brand b) -0.21* 0.18
MALE -O. 13 0.20
ATTEVOO 0.23" 0.11
E] 0.94

BELFLAV Mean 4. 80
INFO (0=on brand A; 1=on
brand b) 0.15* 0.14
BELSEL 0.66“ 0.07
MALE 0.08 0.14
EUROPE 0.13 0.89
E2 0.75

ATT Mean 5.4 I
INFO (0=on brand A; 1=on
brand b) -0.15* 0.14
BELSEL 0.32”“ 0.11
BELFLAV 040* 0.09
MALE 0.13 0.13
ATTEVOO 0.21 0.07
E3 0.66

WTPP Mean 0.25
BELSEL 0.10 0.05
BELFLAV 0.22 0.05
ATT 0.16 0.04
INFO (0=on brand A; 1=on
brand b) -0.10 0.07
MALE 0.18"' 0.08
E4 0. 89

ATTEVOO Mean 6. [8
MALE -0.29* 0.20
E5 0.96

Indirect Effect of INFO on WTPP -0.05

Direct Effect of INFO on WTPP -0.10

Total Causal Effect of INFO on WTPP -0.15

Satorra-Bentler scaled Chi-Square= 6.68 with 10 d.f.; P-value = 0.76

CF 1 = 1.000; RMSEA = 0.000; RMSEA 90% Confidence Interval = (0.000; 0.071)

 

 

 

Note: The asterisk (*) indicates significance at 95% level; the bold font highlights the major variable of
interest INFO.

71

This means that when the attribute information is provided by the firm owning “Brand A”,
the expected percentage of respondents that are willing to pay a premium for “Brand B”
is 2% lower than the mean, that is 20%-2%=18%. Although the differential impact of
attribute POO information on respondents’ willingness to pay a premium statistically
significant also in the case of cream cheese (as the overall fit of the model is good with
CFI = 0.98 and RMSEA = 0.04 and it is comparatively better than with a similar model
with path coefficients between INFO and WTPP fixed to zero), we realize that in the case
of cream cheese the differential impact of attribute POO information on respondents’
willingness to pay a premium for two brands from the same place of origin is not
practically significant.

Therefore, we conclude that data evidence does not support the first hypothesis
(H1) in the case of olive oil but it supports it in the case of cream cheese. That is, an olive
oil firm releasing attribute POO information may expect to obtain an advantage in terms
of price premium for its own brand, although the information provided to respondents is
about the generic place-of-ori gin product rather than about the individual brand. On the

other hand, a cream cheese firm should not expect to obtain such a market advantage.

The impact of attribute information from a collective organization

Results show that attribute POO information about Riviera Ligure olive oil provided by

an individual firm has a significantly higher impact on respondents’ willingness to pay a

premium for its own brand than the same information provided by a collective

organization representing all the firms from the place of origin. The total causal effect of
attribute POO information from an individual firm on respondents’ willingness to pay a

premium for its brand is +0.31 percentage points, which is given by the sum of the direct

72

Table 4 - Impact of Attribute Information on Brand A andBrand B Cream Cheese

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Dependent Independent Variable Standardized Path Std. Errors

Variable Estimates

BELSEL Mean 4.34
FAMCC 0.37* 0.06
INFO (0=on brand A; 1=on
brand b) -0.30* 0.14
E l 0. 88

BELFLAV Mean 4.59
BELSEL 0.47“ 0.10
INFO (0=on brand A; 1=on
brand b) 0.10* 0.14
ENJFOOD 0.20 0.05
E2 0.87

ATT Mean 4. 78
BELSEL 0.16 0.08
BELFLAV 053* 0.08
IN F0 (0=on brand A; 1=on
brand b) -0 .05 0.12
ATTCC 020* 0.05
E3 0.74

WTPP Mean 0.14
BELSEL 0.14 0.05
BELFLAV 0.13 0.05
ATT 0.26“ 0.05
INFO (0=on brand A; 1=on
brand b) 0.06 0.06
E4 0.90

FAMCC Mean 2.27
ATTCC 0.48" 0.06
ES 0.88

Indirect Effect of INFO on WTPP -0.08

Direct Effect of INFO on WTPP +0.06

Total Causal Effect of INFO on WTPP -0.02

Satorra-Bentler scaled Chi-Square= 16.86 with 14 d.f.; P-value = 0.263

CFI = 0.982; RMSEA = 0.043; RMSEA 90% Confidence Interval = (0.000; 0.105)

 

Note: The asterisk (*) indicates significance at 95% level; the bold font highlights the major variable of

interest INFO.

73

 

effect (+0.21) and the indirect effect mediated by respondents’ beliefs and attitudes
(+0.10) (see table 5). This means that when attribute information is provided by an
individual firm, the expected percentage of respondents that are willing to pay a premium
for its brand is 31% higher than the expected percentage across groups, that is
17%+31%=48%. As the overall fit of this model is good (CFI=O.98 and RMSEA=0.05)
and comparatively better than the fit of a path model where any direct or indirect path
between the information treatment and the final dependent variable is fixed to zero, we
consider this differential impact of attribute POO information from an individual firm on
respondents’ willingness to pay a premium significant.

Differently from the case of Riviera Ligure olive oil, in the case of Southern
Louisiana cream cheese the impact of attribute information practically does not vary
depending on its source. The total causal effect of attribute POO information from an
individual firm on respondents’ willingness to pay a premium for its brand is less than
+0.01 percentage points. This means that, no matter if the attribute POO information
about Southern LOuisiana cream cheese is provided by an individual firm or by a
collective organization, the expected percentage of respondents that are willing to pay a
premium for its brand is still 14% (see table 6). Although the difference among
information treatments is not practically significant, the overall path model including the
information treatment fits the data well (CF I=O.92 and RMSEA=0.075) and
comparatively better than a similar model where the individual paths between the
information treatment and the final dependent variable are fixed to zero.

Therefore, we conclude that data evidence does not support our second hypothesis

(H2) in the case of olive oil but it supports it in the case of cream cheese. That

74

Table 5 - Impact of Attribute Information from an Individual Firm and from

a Collective Organization on Brand Olive Oil

 

 

 

 

 

 

 

 

 

 

 

 

Dependent Independent Variable Standardized Path Std.

Variable Estimates Errors

BELSEL Mean 5. 00
INFO (0=from Coll.Org.; l=from
Firm) 0.25* 0.18
MALE -0.13 0.19

- E I 0. 96

BELFLAV Mean 4.65
BELSEL 0.68* 0.07
INFO (0=from Coll.Org.; l=from
Firm) -0.33 0. 14
MALE 0.06 0.14
EUROPE 0. 16* 0.69
E2 0.73

ATT Mean 5.24
BELSEL 050* 0.09
BELFLAV 0.27* 0.07
IN F0 (0=from Coll.Org.; l=from
Firm) 0.18* 0.13
MALE 0.10 0.12
ATTEVOO 0.19" ’ 0.05
E3 0.62

WTPP Mean 0.20
BELSEL 0.42 0.05
BELFLAV 0.21 0.05
ATT 0.14 0.05
INFO (0=from Coll.Org.; l=from
Firm) 0.21* 0.06
MALE 0.04 0.07
E4 0.89

Indirect Effect of INFO on WTPP +0.10

Direct Effect of INFO on WTPP +0.21

Total Causal Effect of INFO on WTPP +0.31

Satorra-Bentler scaled Chi-Square = 16.08 with 12 d.f.; P-value = 0.19

CFI = 0.978; RMSEA = 0.053; RMSEA 90% Confidence Interval = (0.000; 0.112)

 

 

 

Note: The asterisk (*) indicates significance at 95% level; the bold font highlights the major variable of
interest INFO.

75

Table 6 - Impact of Attribute Information from an Individual Firm and from
a Collective Organization on Brand A Cream Cheese

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Dependent Independent Variable Standardized Path Estimates Std.

Variable ' Errors

BELSEL Mean 4.43
INF O (0=from Coll.Org.; l=from
Firm) -0.26* 0.14
SOUTHAM 0.18* 0.36
ATTCC 0.29"“ 0.05
E I 0.90

BELFLAV Mean 4.59
BELSEL 0.47“ 0.10
ENJFOOD 0.20““ 0.05
INFO (0=from Coll.Org.; l=from
Firm) 0.10 0.13
E2 0.87

ATT Mean 4. 78
BELSEL 0.19" 0.08
BELFLAV 0.49“ 0.09
ENJFOOD 0.13 0.04
ATTCC 0.19* 0.05
E3 0. 72

WTPP Mean 0.14
BELSEL 0.14 0.05
BELFLAV 0.13 0.05
ATT 026* 0.05
INFO (0=from Coll.Org.; l=from
Firm) 0.06 0.06
E4 0.90

FAMCC Mean 2.26
ATTCC 0.48" 0.06
E5 0. 88

ENJFOOD Mean 5.98
SOUTH -0.20* 1.41
E6 0. 98

Indirect Effect of INFO on WTPP 005

Direct Effect of INFO on WTPP +0.06

Total Causal Effect of INFO on WTPP +0.01

Satorra-Bentler scaled Chi-Square = 32.69 with 20 d.f.; P-value = 0.036

CFI = 0.920; RMSEA = 0.075; RMSEA 90% Confidence Interval = (0.019; 0.120)

 

Note: The asterisk (*) indicates significance at 95% level; the bold font highlights the major variable of

interest INFO.

76

 

is, an oliVe oil firm may expect to obtain a larger advantage in terms of price premium for
its brand when it releases attribute POO information rather than when the same
information is released by a collective organization representing a group of firms from
the same place of origin. On the other hand, a cream cheese firm should not expect to
obtain a larger market advantage when providing attribute information on its own rather
than letting a collective organization representing them providing the same to

respondents.

The impact of brand information

Results show that brand POO information about a Riviera Ligure olive oil brand has a
significantly higher impact on respondents’ willingness to pay a premium for its own
brand than attribute POO information about generic Riviera Li gure olive oil. The total
causal effect of brand POO information is +0.27 percentage points, which is given by the
sum of the direct effect (+0.17) and the indirect effect mediated by respondents’ beliefs
and attitudes (+0.10) (see table 7). This means that an individual firm provides brand
information to a group of respondents, the expected percentage of respondents that are
willing to pay a premium for its brand is 27% higher than the expected percentage across
groups, that is 17%+27%=44%. The good overall fit of the model with the data both in
absolute terms (CFI=O.93 and RMSEA=0.075) and relatively to a path model where any
direct or indirect path between the information treatment and the final dependent variable
is fixed to zero indicate that the differential impact of brand and attribute POO

information respondents’ WTP a premium is statistically significant.

77

Table 7 - Impact of Attribute Information and Brand Information on Brand A Olive

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Oil

Dependent Independent Variable Standardized Path Std.

Variable Estimates Errors

BELSEL Mean 4. 71
INFO (0=attribute info; l=brand
info) 0.19* 0.13
MALE -0.12* 0.14
ASIA 0.12 0.22
ATTEVOO 0.19* 0.07
E] 0.95

BELFLAV Mean 4. 79
BELSEL 0.56”“ 0.05
INFO (0=attribute info; l=brand
info) 0.15* 0.10
FAMEVOO 0. 16* 0.04
MALE -0. 10* 0.11
EUROPE 0.10"“ 0.35
E2 0.74

ATT Mean 5.03
BELSEL 0.29* 0.06
BELFLAV 0.47 * 0.06
ATTEVOO 0.09* 0.04
E3 0.70

WTPP Mean 0. l 7
BELSEL 0.18* 0.03
BELFLAV 022* 0.04
ATT 0.03 0.03
INFO (0=attribute info; l=brand
info) 0.17* 0.04
MALE 0.12* 0.05
E4 0. 89

FAMEVOO Mean 2.81
ATTEVOO 0.46* 0.06
EUROPE 0.12* 0.30
ASIA -0.1 1 0.26
E5 0.86

ATTEVOO Mean 6. 05
MALE -0.13* 0.14
ASIA -0. 14* 0.25
E6 0. 98

Indirect Effect of INFO on WTPP +0.10

Direct Effect of INFO on WTPP +0.17

Total Causal Effect of INFO on WTPP +0.27

Satorra-Bentler scaled Chi-Square = 53.65 with 23 d.f.; P-value = 0.000

CFI = 0.932; RMSEA = 0.075; RMSEA 90% Confidence Interval = (0.049; 0.100)

 

Note: The asterisk (*) indicates significance at 95% level; the bold font highlights the major variable of

interest INFO.

78

 

Table 8 - Impact of Attribute Information on Brand A Cream Cheese

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Dependent Independent Variable Standardized Path Estimates Std.

Variable Errors

BELSEL Mean 4.52
INF O (0=attribute info; l=brand
info) 0.08 0.10
MALE -0. 17* 0.11
EUROPE 0.15* 0.11
AFRICA 0.13* 0.05
E l 0. 96

BELFLAV Mean 4. 73
BELSEL 0.51* 0.06
INFO (0=attribute info; l=brand
info) 0.01 0.10
EUROPE -0.10 0.09
EVPOO -0.10 0.03
ENJFOOD 0.15* 0.03
E2 0.84

ATT Mean 4. 79
BELSEL 0.38* 0.07
BELFLAV 0.29* 0.07
INFO (0=attribute info; l=brand
info) 0.07 0.09
SOUTHAM 0.08 0.17
ATTCC 0.17* 0.04
E3 0.78

WTPP Mean 0.18
BELSEL -0.02 0.04
BELFLAV 0.08 0.03
ATT 0.24* 0.03
F AMCC 0.21* 0.02
INFO (0=attribute info; l=brand
info) 0.13* 0.05
AFRICA 0.14* 0.06
ATTCC -0. 14* 0.02
E4 0.92

FAMCC Mean 2.47
ATTCC 0.47* 0.04
E5 0. 88

Indirect Effect of INFO on WTPP +0.03

Direct Effect of INFO on WTPP +0.13

Total Causal Effect of INFO on WTPP +0.16

Satorra-Bentler scaled Chi-Square = 81.57 with 56 d.f.; P-value = 0.014

CFI = 0.912; RMSEA = 0.044; RMSEA 90% Confidence Interval = (0.020; 0.063)

 

Note: The asterisk (*) indicates significance at 95% level; the bold font highlights the major variable of

interest INFO.

79

 

Similarly to the case of Riviera Ligure olive oil, the impact of brand POO
information about a Southern Louisiana; cream cheese brand is also significantly higher
than the impact of attribute POO information. The total causal effect of brand POO
information on respondents’ WTP a premium for its brand is +0.16 percentage points.
This means that when a Southern Louisiana cream cheese firm provides brand
information to a group of respondents, the expected percentage of respondents that have a
WTP a premium for that brand relatively to other cream cheese brands from Southern
Louisiana is 34%, which is a percentage 16% points higher than the mean across groups
of respondents receiving either brand or attribute information (see table 8). The overall
path model including the information treatment fits the data well (CFI=O.92 and
RMSEA=0.075) and comparatively better than a similar model where the individual
paths between the information treatment and the final dependent variable are fixed to
zero.

We conclude that data evidence does support our third hypothesis (H3) both. in the
case of Riviera Ligure olive oil and Southern Louisiana cream cheese. That is, a firm
producing either olive oil or cream cheese should expect to obtain a larger advantage in
terms of price premium for its brand when it releases brand POO information rather than
attribute information, when the information recipient has already received some

information about the generic place-of-origin product.

Conclusions
Agricultural economics literature largely explored the effect of generic advertising and

collective certification schemes on consumers’ evaluations for an entire place-of-origin

80

product category, but rarely studied how a firm can create a competitive advantage for its
individual brand relatively to the other brands within the same place of origin. Borrowing
from the theory of attitude formation developed in the field of psychology (Fishbein,
1967; Fishbein and Ajzen, 1975), in this study we attempted to start filling this gap by
analyzing which place-of-origin information increases consumers’ buying intentions for
an individual brand while differentiating it from the other brands from the same place-of-
origin. We introduced a distinction between attribute and brand place-of-origin
information, which differ in the objective of their messages. While the former aims at
increasing consumers’ evaluation towards the place-of-origin attribute, the latter

strengthens the association between a specific brand and the place-of-origin.

By building upon the distinction between attribute and brand information and on
the basis of the data evidence from the experiment conducted, we obtained three main
findings from this study. First, when developing an olive oil brand with a place-of-origin
attribute, a firm can obtain a market advantage in terms of consumers’ WTP a premium
price by providing attribute information on its own. On the other hand, when developing
a cream cheese brand with a place-of-origin attribute, a firm does not obtain a market
advantage over the other brands from the same place of origin when providing attribute
information. In this case, assuming that performing communication and promotion
activities is cheaper collectively than individually, a firm should choose to provide
attribute information as part of a collective organization. Finally, when providing brand
place-of-origin information to potential consumers that have already received some

information about a regional product, a firm obtains a significantly higher percentage of

81

information recipients that are willing to pay a premium for its brand rather than when

providing attribute information.

We believe that these findings could be generalized further from the specific cases
of olive oil and cream cheese. One way to interpret the different impact of attribute
information on potential consumers’ buying intentions across olive oil and cream cheese
is that olive oil can be considered a “credence good” (Darby and Karni, 1973), i.e. a
product whose perceived quality is strongly related to intangible attributes such as the
place of origin, while cream cheese can be considered an “experience good”, i.e. a
product whose perceived quality more strongly depends on tangible quality attributes
(such as flavor, color, texture). Given the nature of credence goods, information about the
product and the source of information has a higher impact on consumers’ evaluation than
in the case of experience goods, where instead the product trial is crucial to determine
consumers’ evaluations (Darby and Karni, 1973). Further research may verify if this
interpretation generalizing further the cases of olive oil and cream cheese holds or not by
testing the same hypotheses in the case of other “credence goods” and “eXperience

goods”.

There are two major limitations in this research that could be overcome in future
research. First, we established only if potential consumers have higher attitudes towards
brands or the WTP a premium price for them, but we did not try to estimate how much
they are willing to pay more when they receive POO information. By estimating the
impact of various types of POO information on consumers’ WTP, fiiture research may
provide a quantitative insight on the benefits of various branding strategies for an agri-

food firm. Second, we crafted the brand POO information treatments on Riviera Ligure

82

olive oil and Southern Louisiana cream cheese by composing a set of information
reported from a heterogeneous set of Sources, including retailers, opinion leaders’
endorsements and food competition awards won. We estimated the impact of the
treatment as a whole, without trying to estimate the marginal impact of each piece of
information depending on its source. By estimating the marginal impact of different
sources of information about a place of origin, future research may also provide insights
on the factors determining variation of the marginal impact of information sources on

consumers’ attitudes and buying intentions for agri-food products with POO attributes.

83

Footnotes

‘ We decided to model W T PP as a binary variable, where the value “1” stands for
respondents that are willing to pay a premium price, where the value “0” stands for
respondents that have answered either “no” or “I don’t know”. We decide to count the “I
don’t know” responses as “0” values because, although the percentage of participants
responding “I don’t know” is around 30%, from a t-test we found no significant variation

in the change of this percentage across groups.
2 To evaluate the appropriateness of treating the dependent binary variable WTPP as

continuous, we also checked and found that results are stable when we excluded WTPP

from our path model and had only the 7-point scale ATT as dependent variable.

84

Appendix B

SURVEY INSTRUMENT — CHAPTER 2

Food Values

1. When I choose the food I eat, an important thing I consider is the country or
region where it is produced.

Strongly Strongly
Disagree Agree

1 2 3 4 5 6 7

f" F f“ 1’" C F F

2. I enjoy discovering and tasting food from other regions of the world.

Strongly Strongly
Disagree Agree

1 2 3 4 5 6 7

f" f“ f” f" (" (" f“

3. How would you describe your attitude towards extra-virgin olive oil in general on
a scale from very negative (1) to very positive (7)?

Very Bldg-hive
Negative

1 2 3 4 5 7

f" (‘ t" 1" t" f‘ f"

4. Have you ever heard about or tasted an extra-virgin olive oil from Riviera Ligure?

Yes No I don't remember
(" f“ t"

85

5. How would you describe your attitude towards cream cheese in general from a
very negative (1) to very positive (7)?

Very Very
Negative ' Positive
1 2 3 4 5 6 7

(7 ('~ (" C (" ("‘ C

6. Have you ever heard about or tasted Creole Cream Cheese from Southern

Louisiana?
Yes No I don't know
f‘ (7 f”

Familiarity with Cream Cheese

Please answer the following questions about cream cheese. Feel free to leave the
spaces blank whenever you do not know what to answer in the open questions.

7. How frequently do you consume cream cheese, including both home and away

from home?
Less than Once a 2-3 times a Once a 2-3 times a .
Never Dally
once a month month week week
month
F f” f“ 1’" t“ f‘ t‘
8. How knowledgeable are you about the quality of cream cheese?
Very
Very
Knowledgeable
Unknowledgeable2 3 4 5 6 7
(" t" 1’“ (7 f” (" 1'"

9. Do you presently have some cream cheese with you at home?
No .
Yes

f“
f‘

86

10. Please list up to five brands of cream cheese that you have heard (for example,
"Philadelphia" brand).

You can leave the fields blank if you do not know the answer.

 

Brand II
Brand 2'
Brand 31
Brand 4[

Brand 5'

11. Please list up to three attributes that you consider DURING your consumption of
cream cheese to evaluate its quality (for example, its flavor).

 

 

 

You can leave the fields blank if you do not know.
1I
2|
3|

12. Please list up to three attributes that you consider BEFORE your consumption of
cream cheese to evaluate its quality (for example, its color).

 

You can leave the fields blank if you do not know.
1

2

3|"TT—_

13. Please list up to three pieces of information on the label of a cream cheese that
you consider when you evaluate its quality (for example, if it is organic or not).

 

 

 

You can leave the fields blank if you do not know.
1

2

{——

 

 

 

87

14. Does cream cheese from England have higher quality than cream cheese from
Italy, or vice versa?

Cream cheese from England has higher quality
Cream cheese from Italy has higher quality

It depends

I don't know

3333

15. In the Lansing area, name up to five stores that you can think of that carry a wide
selection of "specialty cream cheeses".

You can leave the fields blank if you do not know the answer.
Store 1|

Store 2|
Store 3
Smm4
Store 5|

 

 

 

 

 

 

16. In which specific country of the world are these cream cheese varieties
traditionally produced?

You can leave the fields blank if you do not know the answer.
Neufchatel r

Mascarpone I

 

 

Familiarity with Olive Oil
Please answer the following questions about extra-virgin olive oil (for simplicity, we will
just call it "olive oil"). Feel free to leave the spaces blank whenever you do not know

what to answer in the open questions.

17. How frequently do you consume olive oil, including both home and away from

home?
Less than Once a 2-3 times a Once a 2-3 times a .
Never once a Dally
month month week week
month
6‘ t“ (7 (3 t" 1"“ f‘

88

18. How knowledgeable are you about the quality of olive oil?

very Very
Knowledgeable

Iljnknowledgeable2 3 4 5 6 7

r“ r r‘ F ‘7 I F F

19. Do you presently have some olive oil with you at home?
Yes
No

C
(3

20. Please list up to five brands of olive oil that you may know (for example,
"Bertolli" olive oil brand).

You can leave these fields blank if you do not know the answer.

 

Brand 1I
Brand 2]

Brand 3
Brand 4
Brand 5|

 

 

 

 

 

21. Please list up to three attributes that you consider DURING your consumption of
olive oil to evaluate its quality (for example, its flavor).

You can leave some fields blank if you do not know.
1|
2l
3l

 

 

89

22. Please list up to six pieces of information on the label of an olive oil bottle that
you consider when you evaluate the quality of an olive oil (for example, if
weather it is obtained from hand-picked olives or not).

You can leave some fields blank if you do not know.

 

 

 

 

 

23. Does olive oil from California have higher quality than olive oil from France, or
vice versa?

Olive oil from California has higher quality
Olive oil from France has higher quality

It depends

I don't know

3333

24. In the Lansing area, name up to five stores that you can think of that carry a wide
selection of "specialty olive oil".

You can leave some fields blank if you do not know.
Store ll

Store 2|
Store 3
Store 4
Store 5|

 

 

 

 

 

 

25. In which specific country of the world are these olive varieties traditionally
produced?

You can leave the fields blank if you do not know the answer.
Frantoio I

Kalamatal

 

 

90

Treatment 1: Firm A's Attribute-Level Information

Please read the following piece of information on extra-virgin olive oil.
"Pietrantica" is an extra-virgin olive oil from Riviera Li gure.

Riviera Ligure is an astonishing Italian region where mountains touch the sea, little
fisherman villages keep their ancient beauty and olive trees grow in stone-made terraces.

"Pietrantica" extra-virgin olive oil carries a Protected Denomination of Origin (P.D.O.).
Contrary to other oils on the market, Protected Denomination of Origin (P.D.O.) products
are always subject to checks in the field, documentary verification, sensorial and
chemical-physical analyses and, only upon completion of this sequence of controls may
an oil bear that precious European Community label.

Therefore, consumers can be sure that "Pietrantica" extra-virgin olive oil is obtained from

the healthiest and carefully selected olives of the Riviera Li gure.

A picture with Olive on and Olive Field Landscape is placed here'.

Measurement 1: Beliefs, Attitudes and Buying Intentions

Please respond to the following questions about the "Pietrantica" P.D.O. extra virgin
olive oil from Riviera Ligure.

26. How would you describe your attitude towards the "Pietrantica" P.D.O. extra
virgin olive oil from Riviera Ligure?

Very Very

Negative Positive
2 4

1 3 5 6 7

(" 1" f" f‘ f" F F

Please rate to what extent you agree or disagree with the following statement.

91

27. I believe that "Pietrantica" extra virgin olive oil is obtained from the most
carefully selected olives of Riviera Ligure.

Strongly Strongly
Disagree Agree

1 2 3 4 5 6 7

f“ t“ t" (7 f‘ (3 (7

28. I believe that "Pietrantica" extra virgin olive oil has a great flavor.

Strongly Strongly
Disagree Agree

1 2 3 4 5 6 7

F 6‘ f‘ (" (‘ F f‘

29. Would you pay a higher price for "Pietrantica" compared to another extra virgin
olive oil from Riviera Ligure?

Yes No I don't know

(3 f“ F

30. Would you pay a higher price for "Pietrantica" compared to another Protected
Denomination of Origin (P.D.O.) extra virgin olive oil from Riviera Ligure?

Yes No I don't know

(" 1" f"

92

31. Would you pay a higher price for a P.D.O. extra virgin olive oil rather than for a
non-P.D.O. extra virgin olive oil from Riviera Li gure?

Pay more for P.D.O.Pay the same I don't know

(" C C

Treatment 2: Firm C's Brand-Level Information

Now please read this other piece of information on extra-virgin olive oil.

The Cipriani Hotel in Venice, the Orient Express luxury train, the Splendido Hotel in
Portofino, the Quirisana Hotel in Capri... "Costa dei Rosmarini" is a delicate extra-virgin
olive oil from Riviera Ligure chosen by the best Chefs and the most elegant Hotels in
Italy. "Costa dei Rosmarini" has been also received a special mention from the famous
Michelin restaurant guide. It has been already chosen by top gourmet shops such as
Whole Foods, Harrods in London and La Grande Epicerie in Paris.

This is the result of four decades of work and passion for olives of the Petrelli family, that
obtains their "Costa dei Rosmarini" extra virgin olive oil from the most healthy and
carefully selected olives of Riviera Ligure.

A picture with a chef cooking with olive oil is placed here.

A Whole Foods Market logo is placed here.

A Venice Simplon Orient-Express logo is placed here.

A picture with olive oil farmer and olive oil tree is placed here.

Measurement 2: Beliefs, Attitudes and Buying Intentions

Please respond to the following questions about the "Costa dei Rosmarini" extra virgin
olive oil from Riviera Ligure.

32. How would you describe your attitude towards the "Costa dei Rosmarini" extra
virgin olive oil from Riviera Ligure?

93

Very Very

Negative Positive
4 6

1 2 3 5 7

C C ' f” C (7 t" (7

Please rate to what extent you agree or disagree with the following statement.

33. I believe that "Costa dei Rosmarini" extra virgin olive oil is really obtained from
the most carefully selected olives of Riviera Ligure.

Strongly Strongly
Disagree Agree

1 2 3 4 5 6 7

f“ F f“ t" l" (‘ f”

34. I believe that "Costa dei Rosmarini" extra virgin olive oil has a great flavor.

Strongly Strongly
Disagree Agree

1 2 3 4 5 6 7

f“ (3 (7 (~ 1" f‘ C

35. I believe that "Costa dei Rosmarini" extra virgin olive oil has a great flavor.

Strongly Strongly
Disagree Agree

1 - 2 3 4 5 6 7

(‘ i” F t“ t" 1'" t"

36. Would you pay a higher price for "Costa dei Rosmarini" compared to another
extra virgin olive oil from Riviera Ligure?
Yes No I don't know

C (7 (3

94

37. Would you pay a higher price for "Costa dei Rosmarini" compared to a Protected
Denomination of Origin (P.D.O.) extra virgin olive oil from Riviera Ligure?
Yes No I don't know

(7 F (‘

Treatment 1 Cheese: Collective Organization's Attribute-Level Information

Now please read the following piece of information on cream cheese.

The Louisiana Dairy Products Association is proud to promote the Creole Cream Cheese,
a legendary local delicacy from Southern Louisiana.

Whether eaten as a part of a meal savory or sweet, Creole Cream Cheese is a celebrated
part of the New Orleans culinary tradition that dates back 150 years to the region’s first
French settlers. This cheese is similar to Neufchatel and other fresh farmhouse style
cheeses with a taste somewhere between ricotta and creme fraiche, and with an
underlying hint of buttermilk. Creole Cream Cheese is customarily served with a sprinkle
of sugar, drizzle of syrup, or mixed with fresh fruit, as well as eaten spread on bread or
crackers.

Today, the authentic Creole Cream Cheese from Southern Louisiana carries
the "Real Louisiana Cheese seal", adopted as part of a strategic promotion effort for
the Louisiana cheese industry.

The "Real Louisiana Cheese" is a certification mark that assures consumers they are
purchasing natural cheese, made in Southern Louisiana exclusively with Southern
Louisiana milk. With this seal, the Louisiana Dairy Products Association has been able to
consistently and effectively promote many styles and varieties of cheese

from Louisiana cheesemakers who qualify for and use it on their packaging.

A picture of cream cheese with strawberries is placed here.

95

Measurement 1 Cheese: Brand A

"Mauthe's Dairy" is a Creole Cream Cheese from Southern Louisiana that carries the
"Real Louisiana Cheese seal".

Please respond to the following questions about the "Mauthe's Dairy" Creole Cream
Cheese from Southern Louisiana.

38. How would you describe your attitude towards the "Mauthe's Dairy" certified
Creole Cream Cheese from Southern Louisiana?

Very Very
Negative Positive
1 2 3 4 5 6 7

f‘ (7 t" f” f" f" f“

Please rate to what extent you agree or disagree with the following statement.

39. I believe that "Mauthe's Dairy" certified Creole Cream Cheese is obtained from
the most carefully selected milk of Southern Louisiana.

Strongly Strongly
Disagree Agree

1 2 3 4 5 6 7

f‘ (3 (" F (3 (" t“

40. Would you pay a higher price for "Mauthe's Dairy" Creole Cream Cheese
compared to any other Creole Cream Cheese from Southern Louisiana?

Yes No I don't know

f” 6" f”

96

41. I believe that "Mauthe's Dairy" certified Creole Cream Cheese has a great flavor.

Strongly Strongly
Disagree 1 2 3 4 5 6 Agree 7
1" F (3 (3 (7 C ("

42. Would you pay a higher price for "Mauthe's Dairy" Creole Cream Cheese
compared to another "Real Louisiana Cheese"-certified Creole Cream Cheese?

Yes No I don't know

(7 C 1"

43. Would you pay a higher price for a "Real Louisiana Cheese"-certified Creole
Cream Cheese rather than for a non-certified Creole Cream Cheese?

Pay more for "Real Pay the same I don't know
Louisiana Cheese"-

certified

(" f‘ (‘

Treatment 2 Cheese: Firm C's Attribute-Level Information

Now please read this last piece of information on cream cheese.
"Bittersweet" is another Creole Cream Cheese from Southern Louisiana.

This farmer's-style cheese indigenous to Southern Louisiana is made with a blend of skim
milk and half & half cream. This style of cheese was brought to Southern Louisiana,
many think, from the Brittany and Burgundy regions of France and became extremely
popular in New Orleans from the 18003 to the 19803 when production ceased. We
resurrect this Creole icon for your enjoyment. It is best eaten with a sprinkle of sugar or
fresh fruit.

A picture with cream cheese and strawberries is placed here.

97

Measurement 2 Cheese: Brand C

Please respond to the following questions about the "Bittersweet" Creole Cream
Cheese from Southern Louisiana.

44. How would you describe your attitude towards the "Bittersweet" Creole Cream
Cheese from Southern Louisiana?

Very Very
Negative 1 2 3 4 5 6 Positive 7
C F (3 f“ 6' t‘ (7

Please rate to what extent you agree or disagree with the following statement.

45. I believe that "Bittersweet" Creole Cream Cheese is obtained from the most
carefully selected milk of Southern Louisiana.

2 3 4 5 6

Strongly Strongly
Disagree 1 Agree 7
(" 1" (3 (" f" f‘ t"

46. Would you pay a higher price for "Bittersweet" Creole Cream Cheese compared
to any other Creole Cream Cheese from Southern Louisiana?

Yes No I don't know

(7 (‘ f”

47. Would you pay a higher price for "Bittersweet" Creole Cream Cheese compared
to another "Real Louisiana Cheese"-certified Creole Cream Cheese?

Yes No I don't know

(3 I“ f“

98

Footnotes

1To have access to the pictures in the questionnaire or to the original questionnaire with

the pictures, please contact the author at dentonid@msu.edu.

99

Appendix C

OTHER TREATMENTS — CHAPTER 2

In this appendix, I provide the information treatments that did not appear in the survey
instrument for Group 1 but that were administered to either Group 2, 3 or 4. Each

treatment was presented to respondents within the same page.

Attribute Information from Multi-Firm Organization (Olive Oil)
Please read the following piece of information on extra-virgin olive oil.

Founded in 2001, the Riviera Ligure olive oil Consortium, with the strength deriving
from the participation of the olive growers, presses and bottlers, is enthusiast to promote
and protect the reputation of the extra-virgin olive oil from Riviera Li gure to offer
consumers an oil of a quality sought by connoisseurs.

Riviera Ligure is an astonishing Italian region where mountains touch the sea, little
fisherman villages keep their ancient beauty and olive trees grow in stone-made terraces.

Today, the authentic extra-virgin olive oil from Riviera Ligure is certified the "Protected
Denomination of Origin" (P.D.O.), which controlled and promoted by the Riviera Ligure
olive oil Consortium.

Contrary to other oils on the market, Protected Denomination of Origin (P.D.O.)
products are always subject to checks in the field, documentary verification, sensorial and
chemical-physical analyses and, only upon completion of this sequence of controls may
an oil bear that precious European Community label.

Therefore, consumers can be sure that P.D.O. extra-virgin olive oil is obtained from the
healthiest and carefully selected olives of the Riviera Ligure.

100

A picture with olive oil and landscape is placed here.

Attribute Information Treatment from Individual Firm (Olive Oil)
Now please read this other piece of information on extra-virgin olive oil.

"Costa dei Rosmarini" is another extra virgin olive oil from Riviera Ligure. Here is a
description that the Petrini family, producers of "Costa dei Rosmarini", give about their
extra-virgin olive oil.

From Capo Cervo to the border with France, Riviera Ligure is a continuum of bays,
traditional ports and rocks on the sea that suddenly become valleys and mountain peaks.
Riviera Li gure’s history is rooted in a tradition in which the production of olive oil has
played a central role for thousands of years. The “taggiasca” tree produces a unique olive
and imparts Riviera Li gure olive oil with a full, delicate and well-rounded flavor, and a
deep color.

Here in Riviera Li gure, everything is different. Time does not matter. Also people are
different: they built miles of stone walls to sustain stripes of cultivations all along the
steepest mountains. They found the space for little squares of land among mountains, and
took care of the hard soil for centuries to plant olives. And this produced a wonder: the
extra-virgin olive oil.

In this magnificient land, Riviera Ligure, the Petrini family produces the "Costa dei
Rosmarini" extra-virgin olive oil.

A picture with olive oil farmer and olive oil tree is placed here.

Attribute Information Treatment from Individual Firm (Cream Cheese)
Now please read this piece of information on cream cheese.

"Mauthe's Dairy" is a Creole Cream Cheese, a legendary local delicacy from Southern
Louisiana.

Whether eaten as a part of a meal savory or sweet, Creole Cream Cheese is a celebrated
part of the New Orleans culinary tradition that dates back 150 years to the region’s first
French settlers. This cheese is similar to Neufchatel and other fresh farmhouse style
cheeses with a taste somewhere between ricotta and creme fraiche, and with an
underlying hint of buttermilk. Creole Cream Cheese is customarily served with a sprinkle

101

of sugar, drizzle of syrup, or mixed with fresh fruit, as well as eaten spread on bread or
crackers.

Today, the "Mauthe's Dairy" Creole Cream Cheese from Southern Louisiana also carries
the "Real Louisiana Cheese" seal.

The "Real Louisiana Cheese" is a certification mark that assures consumers they are
purchasing natural cheese, made in Southern Louisiana exclusively with Southern
Louisiana milk. With this seal, the Louisiana Dairy Products Association has been able to
consistently and effectively promote many styles and varieties of cheese

from Louisiana cheesemakers who qualify for and use it on their packaging.

A picture with cream cheese and strawberries is placed here.

Treatment 2 Cheese: Firm C's Brand-Level Information (Cream Cheese)
Now please read this last piece of information on cream cheese.

As the primary producer of Artisan cheeses showcasing Louisiana's rich culinary heritage,
“Bittersweet” Dairy produces a Creole Cream Cheese that synthesizes the finest flavors
of the Creole culinary heritage.

“Bittersweet” Creole Cream Cheese is delighting America. This year, “Bittersweet” won
the prestigious American Artisanal Treasure Award and the “Butters Up” American
Cheese Society Award. Chef John F olse, who crafted the best Creole Cream Cheese
recipe for "Bittersweet", received the Southern Foodways Alliance Lifetime Achievement
Award for his creation!

“Bittersweet” Creole Cream Cheese has been selected by top specialty stores such Whole
Foods Market, the famous Artisan Cheese Gallery in San Francisco, the Cheese Store of
Beverly Hills, and obviously the finest delis of New Orleans.

"Bittersweet" is the result of 150 years of passion in producing cream cheese and Creole
culinary heritage.

A picture with cream cheese producer winning an award is placed here.

A Whole Foods Market is placed here.

102

An Artisan Cheese Gallery is placed here.

A picture with cream cheese and strawberries is placed here.

103

Appendix D

METHODOLOGICAL NOTE —- CHAPTER 2

This methodological note provides a detailed report of the analysis conducted in Chapter
2. Results obtained from the path analysis described within the chapter are derived after
undertaking the following intermediate steps:

0 Path analysis without demographic and attitudinal predictors;

0 Path analysis with demographic and attitudinal predictors;

0 Test of robustness of results with and without the WTPP variable;

0 Computation of Total Causal Effects of INFO on WTPP.

This methodological note presents the steps undertaken only with respondents’
data on Liguria extra-virgin olive oil. The same steps have been undertaken with
respondents’ data on Southern Louisiana cream cheese. The entire analysis has been
performed with the structural equation program EQS, copyright by RM. Bentler,

Multivariate Software, Inc., Version 6.1, l985-2006 (B91).

104

Path analysis without demographic and attitudinal predictors

Building path models without demographic and attitudinal predictors is a preliminary
step to analyze if the results on the impact of information treatments are supporting
hypotheses without control variables and how much variance of dependent variables
remains unexplained before inserting the control variables in the model (Hair et al., 2006).

Table 9, 10 and 11 illustrate the results that test respectively the three hypotheses of this

 

 

 

 

 

 

 

 

 

study.
Table 9 - Simple path model testing hypothesis 1, Liguria olive oil
Dependent Variable Independent Variables Standardized Path Estimates
BELSEL Mean 5.04
INFO (0=on brand A; 1=on brand b) -0.40*
E I 0. 90*
BELFLAV Mean 4. 79
INFO (0=on brand A; 1=on brand b) 0.64*
BELSEL 030*
E2 0. 99*
ATT Mean 5.41
INFO (0=on brand A; 1=on brand b) -0.30*
BELSEL 0.34*
BELFLAV 042*
E3 070*
WTPP Mean 0.25
BELSEL 0.02
BELFLAV 0.09
ATT 0.08
INFO (0=on brand A; 1=on brand b) -0.10
E4 028*
Chi-square .000 based on -4 d.f. CFI = 0.942.

 

 

Note: Chi-square is null and degrees of freedom are negative because the model is under-identified.
The asterisk (*) indicates significance at 95% level; the bold font highlights the major variable of
interest INFO.

105

Table 10 - Simple path model testing hypothesis 2, Liguria olive oil

 

 

 

 

 

Defindent Variable Independent Variables Coefficients
BELSEL Mean 5.00
INFO (0=from Coll.Org.; l=from Firm) 052*
E l 0. 85*
BELFLAV Mean 4. 65
INFO (0=from Coll.Org.; l=from Firm) 0.08
BELSEL 066*
E2 1.0
ATT Mean 5.24
INFO (0=from Coll.Org.; l=from Firm) 0.48*
BELSEL 0.48*
BELFLAV 029*
E3 0. 59*
WTPP Mean 0.20
BELSEL 0.01
BELFLAV 0.08
ATT 0.05
INFO (0=from Coll.Org.; l=from Firm) 0.l7*
E4 0. 94*

 

 

 

 

Chi-square .000 based on -4 d.f. CF I = 0.937.

 

Note: Chi-square is null and degrees of freedom are negative because the model is under-identified.
The asterisk (*) indicates significance at 95% level; the bold font highlights the major variable of
interest INFO.

Table 11 - Simple path model testing hypothesis 3, Liguria Olive Oil

 

 

 

 

 

Dependent Variable - Independent Variables Coefficients
BELSEL Mean 4. 71
INFO (0=attribute info; l=brand info) 0.44*
E I 1.00*
BELFLAV Mean 4. 79
INFO (0=attribute info; l=brand Info) 0.38
BELSEL 0.59
E2 0.78
ATT Mean 5.03
INFO (0=attribute info; l=brand info) 0.03
BELSEL 0.28
BELFLAV 0.49
E3 [.14
WTPP Mean 0. I 7
BELSEL 0.06
BELFLAV ' 0.07
ATT 0.01
INFO (0=attribute info; l=brand info) 0.14
E4 2.85

 

 

 

 

Chi-square .000 based on -4 d.f. CFI = 0.960.

 

Table 11 - Simple path model testing hypothesis 3, Liguria olive oil

Note: Chi-square is null and degrees of freedom are negative because the model is under-identified;
The asterisk (*) indicates significance at 95% level; the bold font highlights the major variable of
interest INFO.

106

 

 

These three models are under-identified as there are too many parameters to be
estimated compared to the known variables. As predictors are added to the model, the
number of known variables relative to free parameters to be estimated increases and the

model becomes identified.

Path analysis with demographic and attitudinal predictors
First, an initial model is tested controlling for all the predictors available, including
respondents’ gender, origin, food values and familiarity with the product.
After evaluating this initial model by observing the goodness-to-fit and after identified
and dropping the parameters that do not contribute to make the chi-square decrease with
the Wald test, the following second model is tested:
BELSEL=BHINFO+81; (1)
BELFLAV = BziBELSEL + 132sz0 + stEUROPE + [324MALE + 32; (2)
ATT = BngELSEL + B3ZBELFLAV+ [333INFO+ 034MALE+ BgsATTEVOO+ 83; (3)
WTPP = B4lBELSEL + B42BELFLAV + B43ATT + B44INFO + BasMALE + 34. (4)
This model does not have a perfect fit with data, as chi-square is 29.81 with d.f.
and its p-value < 0.01, but the multivariate LM test suggests that freeing the path
coefficients of the variables MALE on AITEVOO, MALE on BELSEL and ATTEVOO
on BELSEL would improve the goodness-to-fit by dropping the chi-square of 10.34, 4.98
and 9.44 points, respectively. 1
Therefore, a third unrestricted model is run, where these three path coefficients

are freed, which is presented in Table 3 and interpreted in the chapter text. The same

107

procedure of goodness-to-fit evaluation and application of suggestions from Wald test

and LM test leads to the models presented in Table 5 and Table 7 within the chapter text.

Test of robustness of results with and without the WTPP variable

Recognizing in the chapter text that modeling a binary dependent variable WTPP as
continuous is not ideal although acceptable in this case, the significance and the direction
of the model path coefficients is tested, with the independent binary variable WTPP
dropped. This contributes to evaluate if modeling WTPP in such a way provides stable
results or if instead its introduction or removal changes the interpretation of the other
coefficients drastically.

Therefore, a model which is equal to the one presented in Table 1 but without the
independent variable WTPP is run. Results show that the variable INFO has a negative
effect (with a path coefficient equal to -0.45) and ATTEVOO has a positive effect on
BELSEL (+0.25); the variables BELSEL (+0.62) and INFO (+0.31)'both have a positive
effect on BELFLAV; and the variables BELSEL (+0.31), BELFLAV (+0.41), MALE
(+0.29) and ATTEVOO (+0.22).have a positive effect on ATT, while INFO has a
negative effect (-0.31) on ATT. As direction and significance of the path coefficients
towards BELSEL, BELFLAV and ATT is the same when adding the binary variable
WTPP or not, it is concluded that keeping the WTPP in the final model in Table 1
provides very stable results. The same procedure is followed to analyze stability of results

for hypotheses 2 and 3 with similar conclusions.

108

Computation of Total Causal Effects of INFO on WTPP

To interpret the magnitude of the effect of the variables INFO on WTPP in the path
model testing hypothesis 1, the total causal effect is computed by summing the direct
effect of INFO on WTPP and the total indirect effects linking INFO and WTPP, as
commonly done in path analysis (Hair et al., 2006). The total indirect effects in each path
are computed by multiplying each individual path coefficient from the variable INFO to
the variable WTPP (see Table 12). The same procedure is used to compute the total

causal effects in the path models testing hypotheses 2 and 3.

Table 12 - Computation of Total Causal Effects on INFO on WTPP

 

 

 

 

 

 

 

 

 

 

 

 

Path Path Coefficient
INFO -> BELSEL -> BELFLAV -> ATT -> WTPP -0.01
INFO -> BELSEL -> ATT -> WTPP -0.01
INFO -> ATT -> WTPP -0.02
INFO -> BELFLAV -> ATT -> WTPP 0.01
INFO -> BELSEL -> WTPP -0.02
INFO -> BELSEL -> BELFLAV -> WTPP -0.03
INFO -> BELFLAV -> WTPP 0.03
Total Indirect Effect of INFO -> WTPP -0.05
Direct Effect of INFO -> WTPP -0.10
Total Causal Effect of INFO -> WTPP -0.15

 

 

109

 

REFERENCES

Aaker, D. (1991). Managing Brand Equity." Capitalizing on the Value of a Brand Name,
The Free Press, New York.

Agrawal, J. and Karnakura, WA. (1999). “Country of Origin: A competitive advantage?”
International Journal of Research in Marketing 16, 255-267.

Ajzen, 1., and Fishbein, M. (1980). Understanding attitudes and predicting social
behavior. Englewood Cliffs, NJ: Prentice.Hall.

Andrews, M. and Kim, D. (2007). “Revitalising suffering multinational brands: an
empirical study.” International Marketing Review 24 ( 3), 350-3 73.

Balling R. (1995). “Der Herkunftsaspekt als Erfolgsfaktor fur das
Lebensmittelmarketing.” Ber Landwirtsch 73, 83-108.

Bentler RM. (2004). EQS Structural Equations Program Manual. Encino, CA:
Multivariate Software, Inc.

Bertozzi, L. (1995). “Designation of origin: Quality and specification.” Food Quality and
Preference 6 (3), 143-147.

Beverland, M. (2001). “Creating value through brands: the ZESPRI kiwi fruit case.”
British Food Journal 103 (6), 383 - 399.

Bonnet, C., and Simioni, M. (2001). “Assessing consumer response to protected
designation of origin labeling: A mixed multinomial logit approach.” European Review of
Agricultural Economics 28 (4), 43 3—449.

110

Brester, G.W. and Schroeder, T.C. (1995). “The Impacts of brand and Generic
Advertising on Meat Demand.” American Journal of A gricultural Economics 77 (4), 969-
979.

Centner T.J., Turner S.C., Bryan J.T. and Mizelle W.O. (1989). “Product differentiation
protection: Vidalia onions as a case study of regional specialty crops.” Research report,
University of Georgia, College of Agriculture, Agricultural Experiment Stations, 576-595.

Chung, C. and Kaiser, HM. (2000). “Distribution of Generic Advertising Benefits across
Participating Firms.” American Journal of Agricultural Economics 82 (3), 659-664.

Clarke, 1., Owens, M. and Ford, J .B. (2000). “Integrating Country of Origin into global

Marketing Strategy: A Review of US Marking Statuses.” International Marketing Review
17 (2), 114-126.

Caswell, J .A. and Mojduszka, EM. (1996). “Using Informational Labeling to Influence
the Market for Quality in Food Products.” American Journal of Agricultural Economics
78 (5), 1248-1253.

Channel4 (2009). Information from Channel4 Web Site at: www.channel4.com.

Crespi, J .M. and Marette, S. (2002). “Generic Advertising and Product Differentiation.”
American Journal of Agricultural Economics 84 (3), 691-701.

Cronbach, L.J. (1950). “Further evidence on response sets and test design.” Educational
and Psychological Measurement 10 (1), 3-31.

Darby, MR. and Karni, E. (1973). “Free Competition and the Optimal Amount of
Fraud.” The Journal of Law and Economics 16, 67-88.

D’Astous, A. and Ahmed, SA. (1999). “The Importance of Country Images in the
Formation of Consumer Product Perceptions.” International Marketing Review 16 (2),
108-125.

Dean, DH. (1999), “Brand endorsement, popularity, and event sponsorship as advertising
cues affecting consumer pre-purchase attitudes.” Journal of Advertising 39 (3), 1-12.

Dichter, E. (1962). “The world customer.” Harvard Business Review 40 (4), 113-122.

Eagly, AH. and S. Chaiken (1993). The psychology of attitudes, Fort Worth, Texas,
Harcourt Brace.

111

Ehmke, M.D., Lusk, J .L. and Tyner, W. (2008). “Measuring the relative importance of
preferences for country of origin in China, France, Niger, and the United States.”
Agricultural Economics 38 (3), 277 — 285.

Farina, E. and Reardon, T. (2000). “Agrifood grades and standards in the extended

Mercosur: their role in the changing agrifood system.” American Journal of Agricultural
Economics 82 (5), 1170—1176.

Fishbein, M. (1967). “A Behaviour Theory Approach to the Relations Between Beliefs
about an Object and the Attitude Toward the Object.” Readings in Attitude Theory and
Measurement, John Wiley, New York, 389-400.

Fishbein, M. and I. Ajzen (1975). Belief Attitude, Intention and Behavior: An
Introduction to Theory and Research, Reading, Massachusetts: Addison-Wesley.

Gao, Z. and Schroeder, T.C. (2009). “Consumer responses to new food quality
information: are some consumers more sensitive than others?” Agricultural Economics
40(3), 339-346.

Giovannucci, D. and Reardon, T. (2000). “Understanding grades and standards and how
to apply them”. In: Giovannucci, D. (Ed.), A Guide to Developing Agricultural Markets
and Agro-enterprises. The World Bank, Washington, DC.

Grunert, KG. (2005). “Food quality and safety: consumer perception and demand.”
European Review of Agricultural Economics 32 (3), 369—391.

Hair, J .F ., Black, W.C., Babin, B.J., Anderson, RE. and Tatham, KL. (2006).
Multivariate Data Analysis (6th Edition), Upper Saddle River, NJ: Pearson Prentice Hall.

Haubl, G. And Elrod, T. (1999). “The Impact of Congruity between Brand Name and
Country of Production on Consumers‘Product Quality Judgments.” International Journal
of Research in Marketing 16, 199-215.

Hatanaka, M, Bain, C. and Busch, L. (2005). “Third-party certification 1n the global
agrifood system.” Food Policy 30 (1), 354-369

Hauser A. (1993). “Verbraucherpraferenzen fur Nahrungsmittel aus der Naheren
Umgebung. Analyse einer Reprasentativenbefragung bei Nordrhein-Westfalischen
Verbrauchemt.” In: Agrarwirtschaft Zeitschrift fur Betriebswirtschafi, Marktforschung
und Agrarpolitik, 141-159.

112

Hong, ST. and Wyer RS. (1990). “Determinants of Product Evaluation: Effects of the
Time Interval between Knowledge of a Product's Country of Origin and Information
about Its Specific Attributes.” Journal of Consumer Research 17 (3), 277-288.

Huffman, W.E., Rousu, M., Shogren, J .F . and Tegene, A. (2004). “Who do consumers
trust for information? The case of genetically modified foods.” American Journal of
Agricultural Economics 86 (5), 1222-1229.

John, D.R., Scott, C .A. and Bettman, JR. (1986). “Sampling Data for Covariation
Assessment: The Effect of Prior Beliefs on Search Patterns.” Journal of Consumer
Research 13, 36-47.

Kaiser, HM. and Liu, DJ. (1998). “The Effectiveness of Generic versus Brand
Advertising: The Case of US. Dairy Promotion.” Review of Agricultural Economics 20
(1), 69-79.

Kavaratzis, M. (2005). “Place Branding: A Review of Trends and Conceptual Models.”
The Marketing Review 5 (4), 329-342.

Keller, KL. (1993). “Conceptualizing, Measuring and managing Customer-Based Brand
Equity.” Journal of Marketing 57 (1), 1-22.

Kinnucan, H., Xiao, H., Hsia, CH. and Jackson, J. (1997). “Effects of health Information
and generic advertising on US meat demand”. American Journal of A gricultural
Economics 79, 13-23.

Knight, G. and Calantone, R. (2000). “A Flexible Model of Consumer Country-of-Origin
Perceptions.” International Marketing Review 17 (2), 127-145.

Konefal, J., Mascarenhas, M. and Hatanaka, M. (2005). “Governance in the global agro-
food system: backlighting the role of transnational supermarket chains.” Agriculture and
Human Values 22 (3), 291-302.

Kotler P., Haider D.H., Rein I. (1993). Marketing places: attracting investment, industry,
and tourism to cities, states and nations. New York: Free Press.

Kuznesof, S., Tregear, A. and Moxey, A. (1997). “Regional foods: a consumer
perspective.” British Food Journal 99 (6), 199 — 206.

IDDBA (2008). What ’3 In Store Report 2008, International Dairy Deli Bakery
Association, Madison, Wisconsin. Website: www.iddba.org.

113

Iversen, NM. and Hem, LE. (2008). “Provenance associations as core values of place
umbrella brands: A framework of characteristics.” European Journal of Marketing 42 (5),
603-626.

Laroche, M., N. Papadopoulos, L.A. Heslop, and M. Mourali. 2005. “The Influence of
Country Image Structure on Consumer Evaluations of Foreign Products.” International
Marketing Review 22(1), 96-115.

Leclerc, F. Schmitt, B.H. and Dube, L. (1994). “Foreign Branding and Its Effects on
Product Perceptions and Attitudes. ” Journal of Marketing Research 3(2), 263-270.

Likert, R. (1932). “A technique for the measurement of attitudes.” Archives of
Psychology 140.

Loureiro, ML. and Umberger, W.J. (2005). “Assessing Consumer Preferences for
Country-of-Origin Labeling.” Journal of Agricultural and Applied Economics 37 (I), 49-
63.

Lusk, J.L., Brown, J., Tyler, M., Proseku, 1. Thompson, R. and Welsh, J. (2006).
“Consumer Behavior, Public Policy, and Country-of-Origin Labeling.” Review of
Agricultural Economics 28 (2), 284—292.

Lutz, R. (1991). “The Role of Attitude Theory in Marketing”, in: Kassarjian, Harold H.
and J .J . Robertson Editors, 1991. Perspectives in Consumer Behavior (4th ed.), Prentice-
Hall.

MacCallum, R.C., Zhang, S., Preacher, K.J., Rucker, D.D.'(2002). “On the practice of
dichotomization of quantitative variables.” Psychological Methods 7 (1), 19—40.

Moore, W.L. (1980). “Levels of Aggregation in Conjoint Analysis: An Empirical
Comparison.” Journal of Marketing Research 17 (4), 516-523.

Papadopoulos N. (1993). “What product and country images are and are not.”, 3— 38, in:
Papadopoulos N., Heslop L.A., Product-country images: impact and role in international
marketing. New York: Haworth Press, 1993.

Pappu, R., Quester, P.G. Cooksey, R.W. (2007). “Country image and consumer-based
brand equity: relationships and implications for international marketing.” Journal of
International Business Studies 38, 726-745.

Peabody, D. (1962). “Two components in bipolar scales: Direction and extremeness.”
Psychological Review 69, 65-73.

114

Pecotich, A. and Rosenthal, M.J. (2001). “Country of Origin, Quality, Brand and
Consumer Ethnocentrism.” Journal of Global Marketing 15 (2), 31-60.

Peterson, RA. and A.J.P. Jolibert (1995). “A Meta-Analysis of Country of Origin
Effects.” Journal of International Business Studies 49 (4), 883-900.

Piggott, N., Chalfant. J.. Alston. J. and Griffith. G. (1996). “Demand response to
advertising in the Australian meat industry.” American Journal of Agricultural
Economics 78, 268-279.

Piron, F. (2000). “Consumers’ Perceptions of the Country-of-Origin Effect on Purchasing
Intentions of (In)conspicuous Products.” Journal of Consumer Marketing 17 (4), 308-321.

Podolny, J.M. (1993). “A Status-Based Model of Market Competition.” American
Journal of Sociology 98 (4), 829-872.

Reardon, T., Codron, J .-M., Busch, L., Bingen, J. and Harris, C. (2001). “Global change
in agrifood grades and standards: agribusiness strategic responses in developing
countries.” International Food and Agribusiness Management Review 2 (3), 421—435.

Scarpa, R., Philippidis, G. and Spalatro, F. (2005). “Product-Country Images and
Preference Heterogeneity for Mediterranean Food Products: A Discrete Choice
Framework.” Agribusiness 21 (3), 329—349

Schooler RD. (1965). “Product bias in the central American common market.” Journal
of Marketing Research 2, 394—397.

Sheppard, B.H., Hartwick, J. and Warshaw, PR. (1988). “The Theory of Reasoned
Action: A Meta-Analysis of Past Research with Recommendations for Modifications and
Future Research.” Journal of Consumer Research 15 (3), 325-341.

Shocker, A.D., Srivastava, R.K. and Ruekert, R.W. (1994). “Challenges and
Opportunities Facing Brand Management: An introduction to the Special Issue.” Journal
of Marketing Research 31, 149-158.

Skuras, D. and Vakrou, A. (2002). “Consumers’ Willingness to Pay for Origin Labelled
Wine.” British Food Journal 104 (11), 898-912.

Smith W.R. (1993). “Country-of-ori gin bias: a regional labeling solution”. International
Marketing Review 10 (6), 4— 12.

115

Steenkamp, J-B.E.M. and Van Trijp, H.C.M. (1996). “Quality Guidance: A Consumer-
Based Approach to Food Quality Improvement Using Partial Least Squares.” European
Review of Agricultural Economics 23, 195-215.

Takhor, M. and Lavack, AM. (2003). “Effect of Perceived Brand Origin Assications on
Consumer Perceptions of Quality.” Journal of Product and Brand Management 12 (6),
394-407.

Tregear, A., Kuznesof, S. and Moxey, A. (1998). “Policy initiatives for regional foods:
some insights from consumer research.” Food Policy 23 (5), 383—394.

Van der Lans, I., K. Van Ittersum, A. De Cicco, and M. Loseby (2001). “The role of
region of origin and EU certificates of origin in consumer evaluation of food products.”
European Review of Agricultural Economics 28 (4), 451-477.

Van Ittersum, K., Candel, M.J.J.M. and Thorelli, F. (2000). “The Market for PDO/PGI
protected regional products: Consumer Attitudes and Behavior”. In: B. Sylvander, D.
Barjolle and F. Arfini (eds), The Socio-Economics of Origin Labelled Products in Agri-
Food Supply Chians: Spatial, Institutional and Co-Ordination Aspects. Paris: INRA-ESR,
Actes et Communciations No. 17-1, 209-221.

Verbeke, W. (2005). “Consumer acceptance of functional foods: socio-demographic,
cognitive and attitudinal determinants.” Food Quality and Preference 16, 45-57.

Verbeke W. and Ward, R.W. (2006). “Consumer interest in information cues denoting
quality, traceability and origin: An application of ordered probit models to beef labels.”
Food Quality and Preference 17, 453—467.

Verlegh, P. and J.B. Steenkamp (1999). “A review and meta-analysis of country-of-origin
research.” Journal of Economic Psychology 20 (5), 521-546.

Ward, R.W., Briz, J. and de Felipe, I. (2003). “Competing Supplies of Olive Oil in the
German Market: An Application of Multinomial Logit Models.” Agribusiness 19 (3),
393—406.

116

Chapter 3

POSITIVE BRAND INFORMATION MITIGATING NEGATIVE SHOCKS ON

ANIMAL WELFARE

Animal welfare has recently become one of the most contentious issues in animal
agriculture (American Veterinary Medical Association, 2006, Farm Foundation, 2006).
While there appears to be no standardized definition of “animal welfare”, ongoing public
discussions and agricultural economics literature generically use this phrase to define the
subject of how production practices impact the treatment of farm animals In the US,
residents have recently expressed ethical concerns for animal welfare issues with
successful ballot initiatives banning the use of gestation crates in swine production in
three states (Videras, 2006). In the European Union (EU), the Commission signed a
protocol in 2006 recognizing that animals are sentient beings and obliging the European
Institutions to pay full regard to the welfare requirements of animals when formulating
and implementing Community legislation (EU Commission, 2009), partly in response to
pressure exerted by consumer activist groups. Public initiatives are accompanied also by
private. initiatives of major players within the meat supply chain. The European retailers’
association GLOBALGAP, which by de facto controls the access of the majority of food

imports in Europe (Reardon et al., 2010), has set animal welfare species-specific

117

 

standards at the production and processing level. Global fast food chains such as
McDonald’s and Burger King are sourcing an expanding share of their food from crate
free sources (Martin, 2007). As new legislation and private standards on animal welfare
bring additional costs for producers, such as the cost of switching to new practices and
increasing the level of controls, which in turn increase the price paid by final consumers
(Henson and Traill, 2000; Stott et al., 2005), understanding consurmrs perceptions and
buying intentions for meat products with animal welfare attributes becomes of great

importance for the industry.

A large recent strand of the literature has evidenced that a segment of consumers
are willing to pay a premium for pork, chicken and beef with animal welfare attributes
(Harper and Nilsson, 2006; Lagerkvist et al., 2006; Carlsson et al., 2007; Lijenstolpe,
2008; Tonsor et al., 2009a; Tonsor et al., 2009c). Results from this research strand are
consistent with qualitative studies on consumers’ attitudes and perceptions for “animal
welfare” products (Harper and Makatouni, 2002; Schroder and McEachern, 2004).
Consumers’ preferences for animal welfare do not seem to vary significantly depending
on demographic variables (N ilsson et al., 2006; Carlsson et al., 2007; Tonsor et al.,
2009c), although they may vary according to their altruism and tendency of free riding

(Lusk et al., 2007).

Although literature on animal welfare is rapidly expanding, a knowledge gap that
remains unexplored is how consumers change their perceptions and preferences for meat
products when receiving information on animal raising, handling and processing
practices. Only Tonsor et al. (2009b) appear to have so far explored the impact of media

coverage with animal welfare information on consumer preferences for meat products.

118

study were collected from 460 US residents through an experiment on fast food chicken

breast sandwiches and then analyzed with a set of Latent Growth Models (LGM).

The remainder of this paper is organized as follows. In the next section, the
literature on negative information shocks, the role of positive brand information and the
theory of attitude formation are reviewed. Hypotheses are developed in the following
section, while in section four the research methods and the model are presented. After

illustrating the results in section five, conclusions are provided in the last section.

Literature Review

Negative information shocks can be defined as strong evidence from a well defined
source that suddenly makes an attribute salient to consumers (Dawar and Pillutla, 2000;
Klein and Dawar, 2004; Roehm and Tybout, 2006). In the field of agricultural economics,
researchers have analyzed the impact of negative information shocks on consumer
demand for food and agricultural products (Brown, 1969; Dahlgran and F airchild, 1987;
Smith et al., 1988; Robenstein and Thurman, 1996; Piggott and Marsh, 2004;
Kalaitzandonakes et al., 2004). These studies have analyzed the impact of information
shocks on food safety and healthiness, but not on animal welfare issues. In marketing,
researchers have found negative information shocks can create negative brand
associations (Klein and Dawar, 2004), affect consumers’ attitudes toward the brand, and

ultimately harm brand equity (Dawar and Pillutla, 2000).

Negative shocks can stem from media information of bad outcomes of the

consumption of a brand’s product, in the case of product-harm crises (Klein and Dawar,

120

2004) such as food-borne disease outbreaks. Negative shocks can also be brought about
by negative publicity of non-govemmental organizations (N GOs) advocating against an
industry or company practices, such as unethical treatment of workers (Elliott and
Freeman, 2003). However, negative information can also come from word-of-mouth
(Scott and Tybout, 1981; Tybout et al., 1981; Smith and Vogt, 1995) and rumors, when
the source of information transmitted through the word-of-mouth is not well defined
(Kamins et al., 1997). There is evidence that word-of-mouth has a stronger negative

effect on consumers’ evaluation of an object than rumors (Smith and Vogt, 1995).

The magnitude of the effect of negative information shocks on consumers’ brand
evaluations depends on various factors. First of all, it depends on the content of the
information shock, which means whether the negative information is a product-harm
crisis (Klein and Dawar, 2004) or a scandal (Roehm and Tybout, 2006). In the case of
product-harm crises, such as the consumer outrage at contaminated Coca-Cola cans in
Belgium and France in 1999 (The Economist, 1999), consumers may perceive a threat for
themselves that they were unaware of (Klein and Dawar, 2004), experience fear and
develop responses to c0pe with it (Rogers, 1975; Floyd et al., 1990; Tanner et al., 1991).
In the case of scandals revealing that a firm harms other entities, such as other people
(Elliott and Freeman, 2003), animals, or the environment, consumers may perceive
compassion or solidarity (Batson, 1998), as well as egregiousness towards the harming
firm (Klein et al., 2003; Klein et al., 2004), which may lead to brand boycotting (Klein et
al., 2004). However, consumers may also create inferences between scandals and
product-harm crises. In the case of animal welfare, researchers have found consumers

associate scandals about firms mistreating animals with food safety concerns and

121

2004): such as food-borne disease outbreaks. Negative shocks can also be brought about
by negative publicity of non-govemmental organizations (N GOs) advocating against an
industry or company practices, such as unethical treatment of workers (Elliott and
Freeman, 2003). However, negative information can also come from word-of-mouth
(Scott and Tybout, 1981; Tybout et al., 1981; Smith and Vogt, 1995) and rumors, when
the source of information transmitted through the word-of-mouth is not well defined

(Kamins et al., 1997). There is evidence that word-of-mouth has a stronger negative

 

effect on consumers’ evaluation of an object than rumors (Smith and Vogt, 1995). L,

The magnitude of the effect of negative information shocks on consumers’ brand
evaluations depends on various factors. First of all, it depends on the content of the
information shock, which means whether the negative information is a product-harm
crisis (Klein and Dawar, 2004) or a scandal (Roehm and Tybout, 2006). In the case of
product-harm crises, such as the consumer outrage at contaminated Coca-Cola cans in
Belgium and France in 1999 (The Economist, 1999), consumers may perceive a threat for
themselves that they were unaware of (Klein and Dawar, 2004), experience fear and
develop responses to cope with it (Rogers, 1975; Floyd et al., 1990; Tanner et al., 1991).
In the case of scandals revealing that a firm harms other entities, such as other people
(Elliott and Freeman, 2003), animals, or the environment, consumers may perceive
compassion or solidarity (Batson, 1998), as well as egregiousness towards the harming
firm (Klein et al., 2003; Klein et al., 2004), which may lead to brand boycotting (Klein et
al., 2004). However, consumers may also create inferences between scandals and
product-harm crises. In the case of animal welfare, researchers have found consumers

associate scandals about firms mistreating animals with food safety concerns and

121

specifically to product-harm crises (Verbeke and Viaene, 2000; Harper and Makatouni,

2002)

A second key factor driving the magnitude of the effect of negative information
shocks on consumers’ brand attitudes is the initial equity of the targeted brand
(Ahluwalia et al., 2000; Dawar and Pillutla, 2000; Pullig et al., 2006). In particular, when
consumers have a strong positive attitude towards the targeted brand (Petty and Krosnick,
1995) or commitment for it (Ahluwalia etal., 2000), negative information shocks have a
weaker effect. Moreover, differentiation of a brand from competitors can limit the
negative spillover from information shocks targeting a competing brand (Roehm and
Tybout, 2006). For example, the presence of strong consumers’ beliefs that a brand
owner follows corporate social responsibility (CSR) principles is likely to mitigate the
effect of negative information shocks about that brand, when the negative information is

unrelated to the CSR principles.

A third important factor that explains variation in the effect of negative
information shocks on a brand is the target of the information shock. That is if the
information shock targets the brand directly, one of its competing brands within the same
industry, or instead the whole industry, without any specification about individual brands
(Roehm and Tybout, 2006). In some circumstances, the negative information shocks
targeting a competing brand (Brand B) may have a negative effect on Brand A. In this
case, an information shock on Brand B has a “negative spillover” on Brand A (Roehm
and Tybout, 2006), whereas “spillover” is commonly defined as any phenomenon in
which information influences beliefs that are not directly addressed in a communication

(Ahluwalia et al., 2000; Balachander and Ghose, 2003).

122

Relative to this literature on negative information shocks, this research provides
contributions in the following three areas. First, an analysis is presented on how the
impact of negative information shocks on consumers’ attitudes varies in the context of a
scandal on animal welfare practices. Second, an analysis is done on how the impact of
such a negative information shock varies when positive information is given beforehand.
Third, an analysis is provided on how the effect of the negative information shock on
consumers’ attitudes varies according to whether the ex ante positive information is

distracting or is directly relevant to animal welfare issues.
Positive Brand Information

Positive information about the brand can stem from the firm owning the brand, through
advertising (Weinberger et al., 1981), or from external sources that are tied to the firm,
such as sponsors or CSR partners (Klein and Dawar, 2004). Positive brand information
usually has the effect of creating or strengthening positive brand associations (Keller,
1993) but it has also the role of moderating the effect of negative information shocks
about the same brand (Weinberger et al., 1981; Okada and Reibstein, 1998). In the
agricultural economics literature, many studies on the interaction between negative and
positive information has been applied to the case of genetically-modified food products
(Fox et al., 2002; Rousu et al., 2002; Lusk et al., 2004; Wachenheim and VanWechel,
2004; Nayga et al., 2005). Positive information usually has an impact weaker than
negative information shocks (Smith and Vogt, 1995; Fox et al., 2002), as it is recognized
to attract less attention than negative information shocks (Scott and Tybout, 1981; Tybout

et al, 1981).

123

When it is used to moderate the effect of negative information shocks on
consumers’ brand attitudes, positive brand information has a different outcome according
to two major dimensions: the order in which the positive information is received (Smith,
1993; Smith and Vogt, 1995) and the distance in the content of positive and negative
information, that is, whether the two pieces of information strictly contradict each other
or are about different brand attributes (Tybout et al., 1981; Okada and Reibstein, 1998;
Klein and Dawar, 2004). When provided ex ante, positive information generally mitigates
the negative effect of word-of-mouth (Smith and Vogt, 1995) and negative product trial
(Smith, 1993), even if the positive and the subsequent negative information contradict
each other. When the positive information is provided ex post and denies a negative
information shock or a rumor, it might be ineffective in moderating the negative brand
association or even strengthening it (Tybout et al., 1981; Okada and Reibstein, 1998).
When creating positive associations that are distant from the negative associations, ex
post positive information moderates the effect of negative information shocks (Tybout er

al., 1981; Klein and Dawar, 2004).

A third factor explaining variability of the positive information in mitigating
negative shocks to competing brands is the initial brand differentiation (Roehm and
Tybout, 2006), which means having strength and uniqueness of brand associations
(Keller, 1993). When Brand A is not clearly differentiated from the brand targeted by the
negative shock (Brand B) and the positive information on Brand A is an ex post denial
message - such as “the bad thing happened to Brand B has not happened to our Brand A”
- then the positive information can reduce or eliminate the negative spillover effect

(Roehm and Tybout, 2006). However, in the same circumstance, when Brand A is clearly

124

 

 

 

differentiated from Brand B, positive information on Brand A that denies what happened
to Brand B can create a negative spillover that would not otherwise exist and ultimately

damage Brand A (Roehm and Tybout, 2006).

Relative to this literature on the role of positive brand information mitigating
negative information shocks, this research provides a contribution in the following two
areas. First, an analysis is provided on how the mitigating role of positive information
varies in the context of a scandal on animal welfare practices. Second, an analysis is
presented on how the effect of positive information on consumers’ beliefs, attitudes and
buying intentions varies according to whether its content is distracting from the Subject of

the scandal or directly relevant to it.

In the attempt to bring such a contribution to the animal welfare debate and to the
literature on negative and positive information, this study proposes and tests a theoretical
framework that builds upon the theory of attitude formation (Fishbein, 1967; Fishbein

and Ajzen, 1975).
Consumers ’ Brand Beliefs, Attitudes and Buying Intentions .

Consumers’ cognitive process to create their attitudes towards brands and ultimately to
establish their buying behavior usually starts from evaluating brand attributes (Fishbein,
1967). By processing information about the attributes of a brand, consumers establish
both evaluations and belief strengths for each attribute, such that the combination of the
two determines their attitudes towards the brand (Fishbein, 1967). Brand attributes are a
category of brand associations, which in turn are a key dimension of brand equity: when a
brand has strong, favorable and unique associations, then it is clearly differentiated from

other brands (Aaker, 1991; Keller, 1993). Brand attributes may be observed before
125

consumption (search attributes) or only after consumption (experience attributes, Nelson,
1970), but some of them may not be visible either before or after consumption (credence
attributes, Darby and Karni, 1973). In the case of credence attributes, consumers’ belief
strengths play a crucial role in establishing their attitudes towards products, and brand

information has a crucial importance in determining consumers’ beliefs.

However, consumers’ attitudes towards a brand do not always predict buying
behavior (Fishbein and Ajzen, 1975). On the other hand, consumers’ attitudes towards
buying the brand, moderated by their subjective norms, predict buying intentions much
more accurately (Fishbein and Ajzen, 1975; Sheppard et al., 1988). In turn, buying
intentions predict behavior “unless intent changes prior to performance” or “unless the
intention measure does not correspond to the behavioral criterion in terms of action,
target, context, time-frame and/or specificity”. The intention of buying a brand has
various measurable dimensions. The most general one is the willingness to do an effort to
perform to the buying action (Fishbein and Ajzen, 1975; Eagly and Chaiken, 1993),
whereas the nature of the effort may vary according to the context: it may be the
willingness to pay to obtain a product from that brand, the likelihood to pay a premium
for that brand, or the likelihood to buy the product even if it is not sold in a favorite
purchasing location. A second key dimension of buying intentions is the choice of the
brand among alternatives (Fishbein, 1980; Fishbein and Ajzen, 1980), which is the
process of comparing and selecting among the intentions associated with each alternative

in the choice set.

126

This study borrows from these theories predicting the formation of attitudes and
buying intentions to use the concepts of consumers’ beliefs in the presence of an attribute

associated to the brand (Fishbein, 1967) and attitudes towards a brand (Fishbein, 1967).

Hypotheses Development

The conceptual framework of this study is built upon the theory of attitude formation
(Fishbein, 1967; F ishbein and Ajzen, 1975) and the theories of the interaction between
positive and negative information shocks developed in consumer economics (Fox et al.,
2002; Rousu et al., 2002; Lusk et a1. 2004; Wachenheim and Van Wechel, 2004; Nayga

et al., 2005) and consumer psychology (Tybout et al., 1981; Smith, 1993; Smith and Vogt,

1995; Okada and Reibstein, 1998; Klein and Dawar, 2004; Roehm and Tybout, 2006).

When analyzing the interaction between the negative shocks and the positive
brand information, two assumptions are made based on the extent literature. First,
negative information has a stronger marginal impact than positive information, no matter
the information sequence nor the content of positive information, as already found by
Smith and Vogt (1995), Fox et a1. (2002) and Lusk et a1. (2004). Second, ex ante positive
information has a larger effect on mitigating the effect of the negative shock than ex post
positive brand information, as already tested in extant literature (Smith, 1993; Smith and
Vogt, 1995; Klein and Dawar, 2004). This assumption is also consistent with the theory
explaining the impact of prior beliefs and the order of information on consumers’

evaluations of objects (Russo et al., 1998; Carlson and Pearo, 2004; Carlson et al., 2006).

127

Building upon these assumptions, two major hypotheses are tested. First, ex ante
brand information which is directly relevant to the content of the following negative
shock is more effective in moderating the negative effect of the negative shock than
brand information which aims at distracting from that content. Providing positive
information on environment, social welfare and animal welfare attributes of a brand and
of the brand owner may be considered the strategy of companies that are trying to
minimize the future risk of being affected by future negative information shocks caused
by advocating Non-Govemmental Organizations (N GOs) or other civil society
organizations. From this perspective, major food companies that joined multi-stakeholder
dialogue initiatives such as the Sustainable Agriculture Initiative Platform (SAI, 2009),
may be interested in developing positive brand information on sustainability issues even
if their consumers value other attributes of their brands more. Therefore, it is

hypothesized:

H1. Consumers receiving ex ante positive information relevant to animal welfare
discount the following negative information shock on animal welfare more

than consumers receiving ex ante distracting positive information.

This hypothesis juxtaposes with findings from previous literature suggesting that
positive information is more effective when it “distracts” consumers from the negative
shock, as it creates negative associations or rational suspiciousness (Tybout et al., 1981;
Okada and Reibstein, 1998; Roehm and Tybout, 2006).. If data provide evidence
supporting this hypothesis, then providing ex ante positive information on issues that are
relevant to future information shocks may be considered as a form of insurance for

protecting the brand from scandals. Moreover, if the positive brand information has the

128

strength of differentiating the brand from competitors, then the brand may become
immune to any negative information shocks affecting its industry, consistent with the

finding of Roehm and Tybout (2006).

On the other hand, how should a company act when it has already been affected
by a negative information shock? Should it react by developing brand information related
to the content of the negative information, or should it choose to provide distracting
positive information? Consistent with existing literature on product crises (Tybout et al.,
1981; Okada and Reibstein, 1998), which highlights the risk that ex post information
relevant to the negative shock just strengthen consumers’ negative associations, it is
hypothesized here that distracting positive information has a more positive effect on

consumers’ attitudes than relevant positive information. In other words:

H2. Consumers receiving ex post positive information distracting from animal
welfare issues after a negative information shock have a stronger increase in
attitudes than consumers receiving ex post information relevant to animal

welfare.

After these two hypotheses are tested, further exploration will be made of which
consumers’ demographic and attitudinal characteristics significantly explain variation
across the effects of positive brand information relevant to or distracting from animal

welfare issues.

129

Methods

To test the hypotheses, data was collected from an on-line experiment focused on fast

food boneless chicken sandwiches and animal welfare issues administered to 460 US-

based residents in November 2009. Data was collected randomly from a representative
sample recruited according to state, age, ethnic group and education level criteria by a

professional survey company. Response rate to the experiment was around 20%, while
on-line questionnaire completion rate was around 75%. Average length of the

questionnaire was around 14 minutes.

A fast food brand was chosen as the object of our experiment because, similarly
to other private actors within the meat industry, they have been recently targeted by
negative information shock about their animal welfare practices by advocating NGOs
(Hudson and Lusk, 2004; Martin, 2007). Although other negative information hit both
fast foods and other actors competing in different industries, the case of animal welfare
and fast foods was chosen because it is a relatively new issue, where respondents are less
likely to have strong beliefs prior to the experiment. Therefore, we expect to find more
variation after each information treatment on animal welfare than for after treatments on,
say, environmental issues, labor issues or genetically-modified issues. On these latter
issues, US respondents received a much heavier information load in the past five to ten
years and so they are likely to have stronger prior beliefs (Fox et al., 2002; Rousu et al.,
2002; Lusk et al. 2004). Furthermore, fast food restaurants have been already the object
of previous studies on negative information regarding different attributes (Roehm and

Tybout, 2006). Finally, chicken boneless sandwiches were chosen as the product of

130

interest because various fast food brands offer a similar product and because many ethical

concerns were focused on the quality of life of chickens.
Research Design

After accepting the invitation to participate in this study, respondents were redirected to a
web link with the questionnaire page. The experiment was divided in three major parts.
First, participants answered questions on demographics, on their food value and their
consumption habits related to chicken consumption. In the initial demographics section,
along with a few preliminary questions about age, gender, ethnic group and nationality,
respondents were asked how much they value origin, naturalness, sustainability and taste
when purchasing and consuming food. Moreover, they were asked how often they
consume chicken products. Every question has been measured with a seven-point Likert-

scale item.

Second, respondents were divided into four groups, each receiving a different set
of treatments. The four treatments consisted of positive information distracting from or
relevant to animal welfare issues, as well as provided before a negative information
shock (i.e. ex ante) or after the same shock (i.e. ex post) (see Figure 3). The positive
brand information consisted of a set of reported declarations from differences sources: an
advocating NGO (Greenpeace), a certifying NGO (Animal Welfare Society), a university
expert on meat and animal welfare and a self-claim from McDonald’s. The negative
information treatment, published by the People for Ethical Treatment of Animals (PETA),
denounced that McDonald’s suppliers mistreat chicken and inflict them terrible pains

while stocking, transporting and slaughtering them.

131

Figure 2 - The Four Treatments Interacting Positive and Negative Information

 

Group 1

Ex Ante Positive Information
Distracting from Animal

Group 3

Ex Ante Positive Information
Relevant to Animal Welfare

 

Welfare Issues Issues
Group 2 Group 4
Ex Post Positive Information Ex Post Positive Information
Distracting from Animal Relevant to Animal Welfare
Welfare Issues Issues

 

 

 

 

Third, after each treatment,participant responses were elicited on animal welfare
beliefs, attitudes towards McDonald’s chicken sandwiches and willingness-to-pay a
premium price (WTPP). Respondents’ belief strength in the association between animal
welfare and the brands was measured with a seven-point Likert-scale, where the
respondents are asked to strongly disagree/strongly agree with the following statement: “I
believe that McDonald’s takes effective measures to provide proper animal welfare to
chickens and hens raised, transported, and processed for production of food products sold
in their restaurants.” Respondents’ attitudes towards the brands were measured with one
seven-point Likert-scale question asking “How would you describe your attitudes
towards McDonald’s?” where the scale was from very negative to very positive. WTPP

has been elicited with two consecutive'questions. First, respondents were simply asked

132

whether they were willing to pay a premium price or not for a McDonald’s chicken
sandwich, compared to a similar sandwich by a competing fast food brand. Participants
responding “yes” were then asked which interval of price premium, expressed in
percentage terms, were willing to pay. Therefore, we modeled WTPP as a continuous
variable where the participants responding “no” had a zero value, while the participants
responding “yes” had a value equal to the average value of the interval of price premium
chosen. As the distribution of the variable WTPP was strongly skewed to the right, we
added one point to each value and took the natural logarithm in order to make the WTPP

distribution more normally distributed.
The Model

In order to capture the dynamic nature of the data we have collected, an analysis was
through a set of latent growth models (LGMs) (Duncan et al, 1999). LGMs canbe
considered a specific category of structural equation models (SEMs) where the latent
factors are the intercept and the slope of the growth of a variable across a group of
individuals. Compared to longitudinal panel modes, LGMs have the advantage of both
describing single individual’s development trajectory of variables and capturing
individual differences in these trajectories over time (Duncan et al, 1999). In particular,
the latter characteristic allows the researcher to explore the factors moderating the
intercept and slope of the development trajectory. Similarly to SEMs, limitations of
LGMs include the assumption of multi-norrnally distributed variables and the necessity

of large samples (Duncan et al, 1999).

133

Figure 3 - The Generic Piecewise Latent Growth Model

Cov(F1,F3)

    
 

 
 

M3
D3

Ml
DI

\.

Cov(Fl,F2) Cov(F2 F3)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fl M2 F2 D2 F3
Intercept Slope, Slope,
Up Down
0
1 o o
1 0.5 0 -1
l -

V1 V2 V3
Attitudes Attitudes Attitudes
(Time 0) (Time 1) (Time 2)

E1 E2 E3

Legend: V1: Initial Consumers’ Attitudes; V2: Consumers’ Attitudes after receiving Positive Information;
V3: Consumers’ Attitudes after receiving Negative Information

134

 

As common in use in LGMs (Duncan et al., 1999), the loadings are fixed from
factors to the measured variables at arbitrary values. In this study, the parameters were
instead freed of the factors’ means and variances, as well as the co-variances among
factors. The factors’ mean indicates the expected difference between the measurable
variables at two different times, while the factors’ variance indicates the inter-individual
variability around the mean. Finally, the co-variance among factors indicates weather the
initial levels of beliefs and attitudes are significantly associated with future changes or

not.

In this study, to compare the impact of positive information distracting from and
relevant to animal welfare issues, the LGM was built in four sequential steps: (1) with a
simple piece-wise LGM, (2) with an associative LGM, (3) with a multi-group LGM and
(4) with a predictive LGM. Data from the simple piece-wise LGM provide preliminary
evidence that positive brand information has a positive impact on consumers’ beliefs,
attitudes and WTPP, no matter if it is distracting or relevant to animal welfare or given ex
ante or ex post. Moreover, the variance of the latent growth and decrease factors is
significantly large, which means that there is high inter-individual variation which

justifies the use of LGM in this setting.

The generic simple piece-wise LGM applied to the case of ex ante positive
information treatments has the following form, consistent with LGM literature (Duncan

et al, 1999) (see Figure 4):
V1:111131 +121F2+131F3+el; (1)

V2 = 112F1 + 122F2 + l32F3 + e2; (2)

135

V3 = 113131 +123F2 +133F3 +63; (3)

F1=31M1+b1D1; (4)
F2 = azMz + szz- (5)
F3 = a3M3 + b3D3. (6)

In these expressions, V1, V2 and V3 stand for the measured variables of interest

(beliefs, attitudes and WTPP) at time 0, time 1 and time 2. F1, F2 and F3 represent
respectively the intercept, the growth factor caused by the positive information and the

decrease factor caused by the negative information. Moreover, lij represent the loadings
from the factors to the measured variables and e, are the errors. Loadings and variable
errors are fixed in order to make the model perfectly identified. Moreover, M; are the
inter-individual means of the intercept and the slope, while D, are the inter-individual

variances of the intercept of the slope to be estimated. Finally, Cov(Di,DJ-) is estimated to

understand if intercept and slope are significantly associated.

Results
The Impact of Distracting versus Relevant Ex Ante Positive Information

Results from the associative LGM with data from two groups of respondents reveal that
consumers’ beliefs on animal welfare, attitudes and willingness to pay a price premium
increase significantly at 95% level both when consumers receive distracting and relevant

positive information before the negative information (see Table 13).

136

Table 13 - Multi-Group Associative LGM: Distracting versus Relevant Ex Ante
Positive Information

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Distracting Info Relevant Info Equality LM Test (Chi-
Square)

Mean Var. Mean Var. Mean Var.
AWBeliefO 3.41 * 1.894 * 3.76 * 2.515 * 3.98 ** 0.98
Attitude0 4.06 * 2.679 * 4.53 * 2.427 * 4.26 ** 0.55
WTPPO 2.8% * 0.007 * 2.0% * 0.003 * 0.56 28.10 **
AWBeliefl 3.92 * 4.868 * 4.79 * 7.892 * 14.59 ** 4.40 **
Attitudel 4.46 * 2.435 * 4.93 * 4.765 * 0.19 14.36 **
WTPP] 4.6% * 0.019 * 3.6% * 0.015 * 0.06 0.74
AWBelien 2.91 * 2.357 * 3.29 * 2.953 * 0.01 1.43
Attitude2 3.36 * 2.070 * 3.63 * 2.901 * 0.93 3.12 **
WTPP2 2.4% 0.005 * 1.8% 0.003 * 0.19 9.93 **
Overall Fit Indexes: -
Chi- 805.25 with 45 d.f. 745.97 with 45 d.f. 1551.23 with 90 d.f.
Square
CFI 0.920
RMSEA 0.148

 

 

 

 

 

 

Note: *95% probability that the parameter is significantly different from zero; **90% probability of
significant drop of chi-Square when the equality constraint is removed.

When consumers receive positive information distracting from animal welfare
issues at McDonald’s, their animal welfare beliefs increase on average from 3.41 points
to 3.92 and then decrease to 2.91 points when negative information on animal welfare is
provided. This may seem odd, as the provided information aimed at distracting
consumers from animal welfare issues, but it is likely that positive information about
healthiness of McDonald’s products has been used as a cue to increase beliefs on animal
welfare. Also, their attitude towards the McDonald’s product increase on average from
4.06 to 4.46 points and then decrease to 3.36 points, while their willingness to pay a
premium for it increases from 2.8% to 4.6% and then decreases to 2.4%. Similarly, when

137

consumers receive relevant positive information on animal welfare practices at
McDonald’s, their beliefs increase on average from 3.76 points to 4.79 and then decrease
to 3.29 points when negative information on animal welfare is provided. Also, their
attitude towards the product increase on average from 4.5 3 to 4.93 points and then
decrease to 3.63 points, while their willingness to pay a premium increases from 2.0% to
3.6% and then decreases to 1.8%. However, the analysis reveals that the decrease of
consumers’ willingness to pay a premium that received the negative information is not
significant at 95% level, either when they ex ante received distracting or relevant positive
information. This is probably driven by high censoring of WTPP at 0%, which takes

place around 85% of respondents.

The significance at 95% level of the variance of all the variables indicates that
there are significant inter-individual variation around the average increase and decrease
of consumers’ perceptions and buying intentions which could be explained by adding
predictors to the model. The two associative models with distracting and relevant positive
information have both a good overall fit with the data, as their chi-square is respectively

805.25 and 745.97 with 45 degrees of freedom (d.f.).

Although the overall trend of increase and decrease is similar, results from the
multi-group LGM provide evidence that there are significant differences between the

impacts of distracting versus relevant ex ante positive information (see Table 1).

First of all, the overall fit of the restricted multi-group model with the data is bad

(chi-square is 1551.23 with 90 d.f., CFI=0.920 and RMSEA=0.148), indicating that the
two models with distracting and relevant positive information cannot be constrained to be

equal. Specifically, when consumers receive relevant positive information, their animal

138

welfare beliefs are significantly higher than when they receive distracting positive
information, as the Lagrange Multiplier (LM) test indicates that the overall fit of the
model would increase significantly (with a drop equal to 14.59 chi-square points) if this
equality constraint is removed. Moreover, the initial attitudes and animal welfare beliefs
are significantly higher for the group receiving the relevant positive information. We
claim that this difference across group is casual rather than due to demographic
differences across the two groups, as the differences across average age, income,
education, gender and state of residency are not significant. However, from descriptive
statistics, we found that the group receiving the relevant positive information had both
higher initial attitudes for sustainability, naturalness and taste related to the other group,

but obviously this was not possible to be controlled with the sample selection.

Important differences in the impact of distracting and relevant positive
information are not only related to the means of the intercept, the increase and the
decrease factors, but also to their variances. The LM test provides evidence that when
consumers receive relevant positive information, the variance of the increase and
decrease factors in attitudes is significantly larger than when they receive distracting
information. Moreover, the variance of the increase in their animal welfare beliefs is
higher and the variance of the decrease in their willingness to pay a premium is smaller.
Overall, these differences in variance show that relevant positive information on animal
welfare causes a larger variation of responses compared to distracting positive
information. This further justifies the search for variables explaining the change in

perceptions caused by relevant positive information on animal welfare practices.

139

On the other hand, as the LM test does not show that overall fit would improve
significantly when the equality constraints of the increase and decrease factors’ means
were released, results from the multi-group LGM do not provide evidence that relevant
positive information has a stronger mitigating effect on the following negative
information shock than distracting positive information. Therefore, these results provide

no evidence supporting hypothesis H1.
The Impact of Distracting versus Relevant Ex Post Positive Information

Similarly to the results with data from the first two groups of respondents, results of the
associative LGM with data from the other two groups of respondents reveal that
consumers’ beliefs on animal welfare and attitudes increase significantly at 95% level
both when they receive distracting and relevant positive information, even when positive

information follows negative information (see Table 14).

When consumers receive positive information distracting from animal welfare
issues at McDonald’s after the negative information shocks, their animal welfare beliefs
increase from 3.18 to 3.56 points but are still lower than their initial beliefs before
receiving the negative information shock (3.91 points). Similarly, their attitudes towards
the McDonald’s product and their willingness to pay a price premium increase, but they
are still lower than their initial attitudes and WTPP before receiving the negative
information shock. However, the analysis reveals that the increase of consumers’ WTPP
receiving the positive information is not significant at 5% level, either when this is
distracting or relevant to animal welfare issues. When instead consumers receive ex post

relevant positive information on animal welfare practices at McDonald’s, their beliefs

140

.-—...J

 

Table 14 - Multi-Group Associative LGM: Distracting versus Relevant Ex Post

Positive Information

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Distracting Positive Info Relevant Positive Info Equality LM Test (Chi-
Square)
Mean Var. Mean Var. Mean Var.
AWBeliefO 3.91 * 2.484 * 3.53 * 2.216 * 2.95 0.27
Attitude0 4.44 * 2.144 * 4.64 * 2.267 * 0.84 0.88
WTPPO 2.5% * 0.006 * 2.8% * 0.008 * 0.06 1.77
AWBeliefl 3.18 * 2.199 * 2.94 * 2225* 0.44 0.00
Attitude] 3.47 * 2.640 * 3.85 * 2.412 * 2.57 1.49
WTPP] 2.1 % 0.002 * 1.5% * 0.005 * 4.60 ** 5.30 **
AWBeliefZ 3.56 * 6.399 * 3.66 * 9.765 * 6.18 ** 5.46 **
Attitude2 4.20 * 6.077 * 4.30 * 8.319 * 0.19 1.18
WTPP2 2.2% 0.008 * 2.2% 0.022 * 0.05 5.02 **
Overall F it Indexes:
Chi-Square 735.56 with 45 d.f. 661.90 with 45 d.f. 1715.96 with 90 d.f.
CF I 1.000
RMSEA 0.000

 

 

 

 

Note: *95% probability that the parameter is significantly different from zero. "90% probability of

significant drop of chi-Square when the equality constraint is removed.

increase from 2.94 to 3.66 points, which is higher than their initial beliefs before

receiving the negative information shock (3.53 points). On the other hand, consumers’

attitudes towards the McDonald’s product and their willingness to pay a price premium

for it increase, but they are still lower than their initial attitudes and WTPP before

receiving the negative information shock.

Again, the significance at 95% level of the variance of all the variables indicates

that there are significant inter-individual variation around the average increase and

decrease of consumers’ perceptions and buying intentions. The two models have both a

141

 

good overall fit with the data, as their chi-square is respectively 735.56 and 661.90 with

45 d.f..

The similarity in the increase and decrease trends across the two groups receiving
distracting and relevant positive information after a negative information shock is

confirmed by the results from the multi-group LGM (see Table 14).

First of all, the overall fit of the restricted multi-group model with the data is
excellent (as CFI=1.000 and RMSEA=0), indicating that the two models with distracting
and relevant positive information can be effectively constrained to be equal. However,
the LM test suggests removing a few equality constraints across the two groups. First, the
average increase in consumers’ animal welfare beliefs is significantly higher for
consumers receiving relevant information than for those receiving distracting information,
as removing the equality constraint would lead to a drop of 6. l 8 chi-square points. In
particular, after receiving both the negative and the relevant positive information on
animal welfare, consumers have higher beliefs that the McDonald’s product has the
animal welfare attribute than initially, while this does not happen in the case of
consumers receiving distracting positive information. Moreover, the decrease of
willingness to pay when negative information is provided is significantly higher in one of

the two groups, although no difference in treatments was given beforehand.

Along with differences in means, the two models with ex post distracting and
relevant positive information present a few differences also in variances. Specifically, the
variance of the WTPP decrease factor and the variance of the beliefs and WTPP increase
factors is significantly higher in the group receiving the ex post relevant positive

information. These differences in variance confirm that relevant positive information on

142

animal welfare causes a larger variation of responses compared to distracting positive
information, which further justifies the search for variables explaining the change in

perceptions caused by relevant positive information on animal welfare practices.

However, as LM test do not show that overall fit would improve significantly
when the equality constraints of the increase and decrease factors’ means were released,
results from the multi-group LGM do not provide evidence that distracting positive
information has a more positive effect on consumers’ attitudes than relevant positive
information when provided after the negative information. Therefore, our results provide

no evidence supporting hypothesis H2.

Predictors of the Impact of Distracting versus Relevant Positive Information

Results from the predictive LGM provide evidence that variables at individual level have
a different effect on the intercept, grth and decrease factors according to whether the

positive information provided is distracting from or relevant to animal welfare issues.

First, when positive information about McDonald’s is given ex ante and it is
distracting from animal welfare issues, age, gender and income play a practically
significant role, i.e. each of these variables help improving the overall fit of the model,
although they are not always statistically significant at 95% level. In particular,
consumers with higher income tend to be significantly more sensitive to positive
distracting information at 95% level and to discount negative information on animal
welfare, while males tend to discount positive distracting information, which is relative to
the healthiness of McDonald’s products. The overall fit of this predictive LGM with the

data is good, as CFI is 0.989 and RMSEA is 0.097 (see Table 15).

143

Table 15 - Predictive LGM: Distracting versus Relevant Ex Ante Positive

Information on Consumer Attitudes

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Distracting Indep. Coeff. Std. Relevant Indep. Var. Coeff. Std.
Info Var. Err. Irjo Err.
Intercept Mean 4.23 * 0.60 Intercept Mean 5.85 * 0.45
(Fl) Male 0.32 0.35 (F4) Education -0.26 * 0.09
Age 0.06 0.12 Age -0.01 0.09
Income -0.13 0.09 Ev.Sustainable -0.15 0.08
Ev. Taste 0.28 0.14
Growth (F2) Mean 0.20 0.49 Growth Mean 0.62 0.64
Male -0.74 * 0.29 (F5) Education 0.01 0.13
_Age 0.05 0.10 _Age 0.34 * 0.12
Income 0.20 * 0.07 Ev.Sustainable 0.06 0.11
Ev. Taste -0.07 0.20
Decrease (F3) Mean 1.44 * 0.49 Decrease Mean 2.73 * 0.57
Male -0.31 0.29 (F6) Education 0.21 * 0.09
Age 0.05 0.10 Age -0.22 * 0.09
Income -0.19 * 0.07 Ev.Sustainable 0.19 * 0.08
Ev. Taste 0.36 * 0.15
Covariance Matrix: Covariance Matrix:
F1 F2 F3 F4 F5 F6
F1 2.91 * F4 2.19 *
F2 -0.70 * 1.92 * F5 -0.71 * 4.45 *
F3 0.93 * -0.31 1.91 * F6 1.00 * -0.48 * 2.38"“
Overall Fit Indexes: Overall Fit Indexes:
Chi-Square 235.80 with 18 degrees of Chi-Square 184.96 with 24 degrees of
freedom freedom
CFI 0.989 CFI 1.000
RMSEA 0.097 RMSEA 0.000

 

 

 

 

Note: In the Predictive LGM, n=93 because there are 22 cases with missing income data that were excluded
from the analysis.

When positive information about McDonald’s is given ex ante and it is relevant to animal
welfare issues, education and age are associated with changes in consumers’ attitudes.
Moreover, consumers’ evaluation of sustainability and taste attributes in food had a
practical impact on consumers’ attitudes. Specifically, consumers with higher education

have lower initial attitudes towards McDonald’s chicken sandwich and are more sensitive

144

 

to negative information on animal welfare issues. On the other hand, elder consumers
tend to be significantly more sensitive to positive relevant information while they tend to
discount negative information. Finally, consumers who value sustainability and taste
attributes in food consumption tend to be more sensitive to negative information on
animal welfare. The overall fit of this predictive LGM with the data is good, as CFI is

1.000 and RMSEA is 0.000.

Moreover, when consumers receive information relevant to animal welfare is
given, the stronger their growth in attitudes with ex ante positive information, the
smoother their decrease in attitudes following the negative information shock (as the
covariance between F5 and F6 is -0.48 and significant at the 95% level), as the
correlation matrix in Table 3 shows. The same negative association is not significant in
the case of consumers receiving distracting positive information. Finally, consumers with
higher initial attitudes tend to discount both negative and positive information, no matter

whether it is distracting or relevant to animal welfare.

When instead positive information about McDonald’s is given ex post and it is
distracting from animal welfare issues, income and frequency of consumption
significantly explain positive and negative changes in consumers’ attitudes (see Table 16).
In particular, consumers with higher income tend to discount negative information on
animal welfare, while people consuming chicken more frequently tend to be more
sensitive to negative information on animal welfare. This direct association between
frequency of meat consumption and sensitiveness to negative information on animal
welfare seems to contradict the recent research that found that frequent consumers of

meat tend to discount information on animal welfare.

145

Table 16 - Predictive LGM: Distracting versus Relevant Ex Post Positive
Information on Consumer Attitudes

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Distracting Indep. Var. Coeff. Std. Relevant Indep. Var. Coeff. Std.
Info Err. Info Err.
Intercept Mean 4.34 * 1.07 Intercept Mean 5.22 0.43
(Fl) Income on 0.07 (F4) Education -0.22* 0.10
Freq. Cons. 0.19 0.17 Ev.Sustainab1e 0.02 0.08
Ev.Sustainable 0.01 0.08
Ev. Taste -0.08 0.09
Decrease Mean 0.08 1 . 12 Decrease Mean 0.94 0.45
(F2) Income -0.20* 0.07 (F5) Education -0. l 9 0.10
Freq. Cons. 0.36 * 0.17 Ev.Sustainab1e 0.11 0.08
Ev.Sustainable 0.13 0.08
Ev. Taste -0.09 0.09
Growth (F3) Mean -0.13 0.11 Growth Mean -2.03 0.84
Income 0.20 0.11 (F6) Education 0.42 * 0.19
Freq. Cons. -0.47 0.27 Ev.Sustainable 0.04 0.15
Ev.Sustainable 0.19 0.13
Ev. Taste 0.08 0.14
Covariance Matrix: Covariance Matrix:
F1 F2 F3 F4 F5 F6
F1 2.06 * F4 2.16 *
F2 0.60 * 2.28 * F5 0.78 * 2.30 *
F3 -0.63 -1.83 * 5.24 * F6 1.57 * -2.76* 7.97 *
Overall Fit Indexes: Overall Fit Indexes:
Chi-Square 163.92 with 24 degrees of Chi- 144.26 with 13 degrees of
freedom ngare freedom
CFI 0.916 CFI 1.000
RMSEA 0.145 RMSEA 0.000

 

 

 

 

Note: In the Predictive LGM, n=93 because there are 22 cases with missing income data that were

 

 

excluded from the analysis.

A possible explanation of this association may be that frequent meat consumers in
the US are strengthening their inferences across the animal welfare attributes and both
food safety and taste, which are obviously salient attributes for frequent meat consumers.

The overall fit of the model is acceptable, as RMSEA=0.145 and CFI=0.916.

When positive information about McDonald’s is given ex post and it is relevant to

animal welfare issues, consumers’ education and evaluation of sustainability attributes

146

play a practically significant role in explaining the positive and negative changes in
attitudes. Specifically, consumers with higher education have lower initial attitudes
towards McDonald’s products and they are more sensitive to positive information on
animal welfare. The overall fit of the model is excellent, as RMSEA=0.000 and

CFI=1 .000.

Finally, correlation matrices in Table 4 suggests that the stronger the decrease in
attitudes when negative information is given, the weaker the following effect of positive
information, no matter if distracting or relevant to animal welfare issues. Moreover, the
higher are the initial attitudes towards the McDonald’s chicken sandwich, the stronger is
the effect of ex ante negative information. At the same time, when positive information is
relevant to animal welfare issues, initial attitudes are positively associated with the
attitude growth, as the covariance between F4 and F6 shows. This effect is not present in
the case of distracting positive information. This illustrates that relevant positive
information on animal welfare can be more useful than distracting information to restore
the initial attitudes of consumers that really like a branded product, once a negative shock

occurred.

Conclusions
In the new era of global food systems, effective communication of food quality attributes
to final consumers through brands is becoming a managerial task that goes far beyond

meeting public and private standards imposed by governments and retailers.

147

 

This study provides implications for fast food company managers that are
responsible for communicating the quality attributes of their brands to final consumers
and that need to tailor brand information to specific consumer characteristics by
indicating which content of positive brand information is more effective to protect a
brand from information shocks on animal welfare and which consumers are more

sensitive to different information content.

Tackling such a research question provides a contribution to the rapidly
expanding animal welfare literature (Lagerkvist et al., 2006; Carlsson et al., 2007;
Lijenstolpe, 2008; Tonsor et al., 2009a; Tonsor et al., 2009c), where only a few studies
have so far analyzed how media coverage affects consumers’ preferences for meat
products (Tonsor et al., 2009b). Specifically, this appears to be the first study analyzing
the interaction of positive and negative information about animal welfare on consumers’
perceptions and intentions to buy a product. Outside the boundaries of the animal welfare
literature, this study also attempts to integrate current knowledge on the impact of
sequences of positive and negative information shocks on consumer behavior, developed
across the fields of economics (Fox et al., 2002; Rousu et al., 2002; Lusk et a1. 2004;
Wachenheim and Van Wechel, 2004; Nayga et al., 2005) and psychology (Russo et al.,
1998; Smith and Vogt, 1995; Roehm and Tybout, 2006), by analyzing inter-individual
and inter-group differential effects with a Latent Growth Modeling (LGM) approach

(Duncan et al., 1999).

Results show that the means of consumers’ attitude growth and decrease do not
differ significantly across different content of information, but the variance of consumers’

attitude grth and decrease is significantly higher when information relevant to animal

148

welfare is provided. This highlights that different consumers have very different reactions
when exposed to animal welfare information, consistent with the findings of Lusk et a1.
(2004), who found that consumers with stronger priors are less sensitive to genetically-
modified information. Specifically, we found that age and education have a significant
predictive power on the reaction to positive information relevant to animal welfare. On
the other hand, we found that income and gender can explain a significant part of the
growth variance when consumers receive positive information distracting from animal

welfare issues.

Although results have useful managerial implications, the analysis of this study is
limited to the context of fast food industry and to the case of animal welfare. Future
research should seek for a generalization of these results across industries and across
content of attribute information. For example, it would be interesting to test if the same
conclusion could be drawn in the same industry when consumers are exposed to
environmental friendly production or on labor conditions. Moreover, it would be
interesting to test if, when exposed to the same animal welfare attribute negative and
positive information, consumers’ perceptions change across meat products, across
individual brands or across different levels of the supply chain of the product. Finally, it
would be useful to analyze how different contents and different sources of positive
information act on mitigating the negative impact of information shocks. These
hypotheses could be tested in future research by applying the multi-group LGM analysis

introduced in this study and changing the set of information treatments.

149

Appendix E

SURVEY INSTRUMENT — CHAPTER 3

Thank you for participating to this research study. This study is conducted by the
Department of Agricultural, Food and Resource Economics and the Department of
Marketing at Michigan State University. Mr. Domenico Dentoni is the research
coordinator and Prof. Christopher H. Peterson is the responsible principal investigator.

From this study, we hope to learn insights on how consumers perceive various attributes
of meat products and process product information. You will be asked questions about
both beef steak and chicken breast. Your participation to this research project is
completely voluntary and we will preserve the confidentiality of your information. Your
participation in this study will take no more than 20 minutes.

Feel free to ask the researchers any questions you may have at the following contacts:

0 Mr. Domenico Dentoni, 409 Agricultural Hall, Michigan State University, 48825
East Lansing, Michigan. Email: dentonid@msu.edu. Phone: 517-488-9277.

0 Prof. Christopher H. Peterson, 83 Agriculture Hall, Michigan State University,
48825, East Lansing, Michigan. Email: petersl7@msu.edu. Phone: 517-355-1813.

Demographics
1. I am: _ Male __ Female
2. I am years old (fill-in the blank or drop down).

3. The best description of my educational background is:

a. Did not graduate from high school
b. Graduated from high school, Did not attend college

150

 

c. Attended College, No Degree earned

d. Attended College, Associates or Trade Degree earned

e. Attended College, Bachelor’s (BS. or BA.) Degree earned

f. Graduate or Advanced Degree (M.S., Ph.D., Law School)

g. Other (please explain):
4. There are __ adults and __ children living in my household (please fill-in the

two blanks) '

5. My ZIP code is:

6. What best describes your race?

 

White, Caucasian
Black, African American
Asian, Pacific Islander
Mexican, Latino
American Indian

Other (please describe):

reap-99‘s»

 

Food Attitudes and Values

7. How frequently do you consume the following meat products at any meal, either
at home or away from home consumption:

 

4 or more 2-3 times Once per 2-3 times Once per Never
times per per week week per month month or
week less

 

Chicken

 

 

Beef

 

 

 

 

 

 

 

 

8. How much time have you spent residing outside the US during your entire life?

None, I’ve always lived in the US
Between 1 month and 6 months
Between 6 months and 1 year
Between 1 year and 2 years

999‘?”

151

e. Between 2 years and 5 years
f. Between 5 years and 10 years
g. Between 10 and 20 years

 

Please rate to what extent you agree or disagree with the following statements:

9. When Ichoose the food I eat, an important thing I consider is the country or
region where it is produced. (Seven-point scale, from 1. Strongly Disagree to 7.
Strongly Agree)

10. When I choose the food I eat, an important thing I consider is if it is natural (that
is, if it is produced without modern technologies) (Seven-point scale, from 1.
Strongly Disagree to 7. Strongly Agree)

D'L. may] -

 

11. When I choose the food I eat, an important thing I consider is if it is "sustainable”
(that is, if it is produced by a company that respects the social and environment
conditions within the area of production). (Seven-point scale, from 1. Strongly
Disagree to 7. Strongly Agree)

12. When I choose the food I eat, an important thing I consider is its taste and
appearance (Seven-point scale, from 1. Strongly Disagree to 7. Strongly Agree)

Initial McDonald’s Brand Equity

Please answer the following questions about McDonald's.

A McDonald’s logo is placed here]. 1

13. How would you describe your attitude towards McDonald's? (Seven-point scale,
from 1. Very Negative to 7. Very Positive)

Please rate to what extent you agree or disagree with the following statement.

14. I believe that McDonald's takes effective measures to provide proper animal
welfare to chickens and hens raised, transported, and processed for production of
food products (e.g., chicken nuggets and eggs) sold in their restaurants. (Seven-
point scale, from 1. Strongly Disagree to 7. Strongly Agree)

15. Do you believe that McDonald’s takes MORE, EQUAL or LESS effective
measures to provide proper animal welfare to chickens and hens raised,

152

transported, and processed for production of food products (e. g., chicken nuggets
and eggs) sold in their restaurants relative to its competitors?

a. More
b. Equal
0. Less
(1. Idon’t know

Now please answer the following questions about Burger King.

A Burger King logo is placed here.

16. How would you describe your attitude towards Burger King? (Seven-point scale,
from 1. Very Negative to 7. Very Positive)

Please rate to what extent you agree or disagree with the following statement.

17. I believe that Burger King takes effective measures to provide proper animal
welfare to chickens and hens raised, transported, and processed for production of
food products (e.g., chicken nuggets and eggs) sold in their restaurants. (Seven-
point scale, from 1. Strongly Disagree to 7. Strongly Agree)

18. Do you believe that Burger King takes MORE, EQUAL or LESS effective
measures to provide proper animal welfare to chickens and hens raised,
transported, and processed for production of food products (e.g., chicken nuggets
and eggs) sold in their restaurants relative to its competitors?

a. More
b. Equal
c. Less
d. Idon’t know

19. If the price of a Boneless Chicken Sandwich were the same across the following
brands, which brand would you choose?

a. McDonald’s
b. Burger King
0. Kentucky Fried Chicken

153

d. Wendy’s
e. Others
f. None

20. Would you be willing to pay a premium if it costs more to purchase a McDonald's
Chicken Sandwich than another brand's Chicken Sandwich?

a. Yes
b. No

21. How much more are you willing to pay to get a McDonald's Chicken Sandwich
rather than another brand of Chicken Sandwich?

Between 0% and 10% more
Between 10% and 20% more
Between 20% and 40% more
Between 40% and 60% more
Between 60% and 80% more
Between 80% and 100% more
At least 100% more

(102-”9999‘!”

Information Treatment 1
Please read this further piece of information about McDonald's.
Havin’ fun!!!

McDonald’s is one of life’s many small pleasures that millions of people around
the world enjoy every day. Great food. Fun to eat. Casual environment. Local and
familiar. And always something new!

You want the very best for your kids, and so do we at McDonald’s. That’s why
we’ve made quality a top priority:

a. McDonald’s coffee is made with 100% pure Arabica coffee beans.

b. McDonald’s burger patties are cooked straight on the grill with no added
fat or oil.

c. McDonald’s Premium Chicken Sandwiches are made with all white meat
real chicken.

d. McDonald’s premium salads contain no preservatives, and are assembled
fresh in the restaurant daily.

154

e. McDonald’s Happy Meal Milk jugs contain real 1% low fat white or
chocolate milk.

f. McDonald’s Apple Dippers are made with farm-fresh apples selected for
their crispness, color and texture.

A picture with a group of McDonald’s prodcuts is placed here.

Now please answer the following questions about McDonald's.

A McDonald’s logo is placed here.

22. How would you describe your attitude towards McDonald's? (Seven-point scale,
from 1. Very Negative to 7. Very Positive)

Please rate to what extent you agree or disagree with the following statement about
McDonald's.

23. I believe that McDonald's takes effective measures to provide proper animal
welfare to chickens and hens raised, transported, and processed for production of
food products (e.g., chicken nuggets and eggs) sold in their restaurants. (Seven-
point scale, from 1. Strongly Disagree to 7. Strongly Agree)

24. Do you believe that McDonald's takes MORE, EQUAL or LESS effective
measures to provide proper animal welfare to chickens and hens raised,
transported, and processed for production of food products (e. g., chicken nuggets
and eggs) sold in their restaurants relative to its competitors?

a. More
b. Equal
c. Less
d. Idon’t know

25. Would you be willing to pay a premium if it costs more to purchase a McDonald's
Chicken Sandwich than another brand's Chicken Sandwich?

a. Yes
b. No

155

26. How much more are you willing to pay to get a McDonald's Chicken Sandwich
rather than another brand of Chicken Sandwich?

Between 0% and 10% more
Between 10% and 20% more
Between 20% and 40% more
Between 40% and 60% more
Between 60% and 80% more
Between 80% and 100% more
At least 100% more

(rowan-99's»

Now please answer the following questions about Burger King.

A Burger King logo is placed here.

27. How would you describe your attitude towards Burger King? (Seven-point scale,
from 1. Very Negative to 7. Very Positive)

Please rate to what extent you now agree or disagree with the following statement about
Burger King.

28. I believe that Burger King takes effective measures to provide proper animal
welfare to chickens and hens raised, transported, and processed for production of
food products (e.g., chicken nuggets and eggs) sold in their restaurants. (Seven-
point scale, from 1. Strongly Disagree to 7. Strongly Agree)

29. Do you believe that Burger King takes MORE, EQUAL or LESS effective
measures to provide proper animal welfare to chickens and hens raised,
transported, and processed for production of food products (e.g., chicken nuggets
and eggs) sold in their restaurants relative to its competitors?

a. More
b. Equal
c. Less
(1. Idon’t know

30. If the price of a Boneless Chicken Sandwich were the same across the following
brands, which brand would you choose?

156

 

a. McDonald’s

b. Burger King

0. Kentucky Fried Chicken
d. Wendy’s

e. Others

f. None

Information Treatment 2
Please read this further piece of information about production practices at McDonald's.
PETA’s “McCruelty — I’m hatin’ it” campaign message:

“McDonald’s chicken suppliers in the United States kill birds with cruel
methods. Chickens typically suffer broken limbs, they have their throats out while
they are still conscious and are often scalded to death in defeathering tanks.

It would cost McDonald’s NOTHING to demand that its chicken suppliers switch
to a far less cruel slaughter method. But McDonald’s refuses.

Tell McDonald’s to stop the cruelty.”

A “McCruelty: I’m hatin it” logo by PETA is placed here.

Now please answer the following questions about McDonald's.

A McDonald’s logo is placed here.

31. How would you describe your attitude towards McDonald's? (Seven-point scale,
from 1. Very Negative to 7. Very Positive)

Please rate to what extent you now agree or disagree with the following statement about
McDonald's.

32. I believe that McDonald’s takes effective measures to provide proper animal
welfare to chickens and hens raised, transported, and processed for production of

157

food products (e.g., chicken nuggets and eggs) sold in their restaurants. (Seven-
point scale, from 1. Strongly Disagree to 7. Strongly Agree)

33. Do you believe that McDonald's takes MORE, EQUAL or LESS effective
measures to provide proper animal welfare to chickens and hens raised,
transported, and processed for production of food products (e.g., chicken nuggets
and eggs) sold in their restaurants relative to its competitors?

a. More
b. Equal
c. Less
d. Idon’t know

34. Would you be willing to pay a premium if it costs more to purchase a McDonald's
Chicken Sandwich than another brand's Chicken Sandwich?

a. Yes
b. No

35. How much more are you willing to pay to get a McDonald's Chicken Sandwich
rather than another brand of Chicken Sandwich?

Between 0% and 10% more
Between 10% and 20% more
Between 20% and 40% more
Between 40% and 60% more
Between 60% and 80% more
Between 80% and 100% more
At least 100% more

carbon-99‘.»

A Burger King logo is placed here.

36. How would you describe your attitude towards Burger King? (Seven-point scale,
from 1. Very Negative to 7. Very Positive)

158

Please rate to what extent you now agree or disagree with the following statement about
Burger King.

37. I believe that Burger King takes effective measures to provide proper animal
welfare to chickens and hens raised, transported, and processed for production of
food products (e.g., chicken nuggets and eggs) sold in their restaurants. (Seven-
point scale, from 1. Strongly Disagree to 7. Strongly Agree)

38. If the price of a Boneless Chicken Sandwich were the same across the following
brands, which brand would you choose?

a. McDonald’s

b. Burger King

c. Kentucky Fried Chicken
d. Wendy’s

e. Others

I. None

Please rate to what extent you agree or disagree with the following statement.

39. When I buy meat products, I like to receive detailed information about product
quality. I am not particularly bothered by receiving too much information on the
product. (Seven-point scale, from 1. Strongly Disagree to 7. Strongly Agree)

40. My annual pre-tax, household income is:

a. Less than $ 20,000
b. $ 20,000-S 39,999

j. 180,000 $ or more

159

41. When you buy a beef steak for your consumption, which one of this two products
would you choose assuming that they have the same price:
a. A USDA-certified beef steak which is produced with animal welfare,
environment friendly practices, from grass-fed animals.
b. A beef steak which is “simply a beef steak”.
c. None of the two.

 

160

Appendix F

OTHER TREATMENT — CHAPTER 3

 

In this appendix, 1 provide the information treatments that did not appear in the survey
instrument for Group 1 but that were administered to either Group 2, 3 or 4. Each

treatment was presented to respondents within the same page.

Positive Information Treatment Related to Animal Welfare
Please read these brief pieces of information about McDonald's production practices.
0 John Sauven, Campaign Director of Greenpeace International, claims:

“McDonald’s moved very swiftly to support the Amazon Rainforest campaign. It has
played a key role in bringing the US-based multinational soybean traders to the
negotiating table and this is a significant breakthrough. The soy which is fed to our
pigs, chickens and cows to make meat products is one of the main drivers of
deforestation.”

A Greenpeace logo is placed here.

0 Excerpt from “McDonalds’ Guiding Principles on Animal Welfare”:

“McDonald’s commitment to animal welfare is global and guided by the following
principles. These principles apply to all the countries in which McDonald’s does
business.

161

Quality: McDonald’s believes treating animals with care and respect is an integral
part of an overall quality assurance program that makes good business sense.

Animal Treatment: McDonald’s supports that animals should be free from cruelty,
abuse and neglect while embracing the proper treatment of animals and addressing
animal welfare issues.

Partnership: McDonald’s works continuously with our suppliers to audit animal
welfare practices, ensuring compliance and continuous improvement.”

A McDonald’s logo is placed here.

0 Temple Grandin, a recognized expert and leader in the field of animal welfare
from Colorado State University, claims:

“I have been working for almost 25 years in designing facilities before I started
doing work with McDonald's. When I worked with implementing the McDonald's’
operations in 1999, it just made a huge, huge change. It was a massive tipping point
where the whole culture of the meat industry changed to where animal welfare is
important. I am very proud of that".

A picture of a livestock expert with cows in background is placed here.

162

Appendix G

METHODOLOGICAL NOTE — CHAPTER 3

This methodological note provides a detailed report of the analysis conducted in Chapter
3. Results obtained from the analysis described within the chapter are derived after
undertaking the following intermediate steps:

0 Simple Piecewise LGM

o Associative LGM

0 Curve-of-F actors LGM

o Multi-group Associative LGM

o Predictive LGM with WTPP

0 Predictive LGM with Attitudes

The entire analysis has been performed with the structural equation program EQS,

copyright by PM. Bentler, Multivariate Software, Inc., Version 6.1, l985-2006 (B91).

163

Simple Piecewise LGM

Piecewise LGM represent a specific case of LGM that describes structural changes in
observed measures over time (Duncan et al., 1999). Therefore, in this study piecewise
LGM is used to describe structural changes in consumers’ beliefs, attitudes and WTPP
created by the sequence of positive and negative information treatments. When building
the models, the difference between piecewise LGMs and general LGMs is only in the
arbitrary choice of the values of the fixed parameters (i.e., loadings) linking the factors to
the observed variables. In general LGMs, the values of these loading is linearly

dependent for all factors, such as:

V1=1*F1+0*F2+0*F3+e1; (1)
v2 =1*F1 +1*F2 + 2*F3 + e2; (2)
v3 =1*F1 + 2*F2 + 4*F3 + e3; (3)
Fl=a1Ml+b1D1; (4)
F2 = a2M2 + b2D2; (5)
F3 = a3M3 + b3D3; (6)

where the loadings of the linear growth F2 are 0, 1, 2 and the loadings of the quadratic
growth factor are 0, 2, 4 (Duncan et al., 1999). The interpretation of the parameters is the
same as in the text of the chapter. In a piecewise model describing a structural change the
fixed parameters of the loadings are not necessarily linearly dependent and can be of
opposite directions among factors. For example, in the piecewise LGM described in

Figure 2, the loadings of F2 are 0, 0.5, 0, while the loadings of F3 are 0, 0, -1. Then, in

164

this case F2 can be interpreted as an increase factor, while F3 as a decrease factor after

the structural change (i.e., the negative information treatment) occurs.

A simple piecewise LGM model is first built for each measure individually. This
provides information about the individual significance of coefficients describing growth
and decrease after the shocks (Mi), as well as a measure of each factor variance (Di).
Results of the piecewise LGM for attitudes of respondents included in Group 1 of the

experiment are reported in Table 17.

Table 17 - Simple Piecewise LGM with Consumer Attitudes in Group 1

 

 

 

 

 

 

 

 

 

 

 

Mean Std. Dev. Mi Di
V1 4.07 1.64 F1 407* 036*
V2 4.46 1.56 F2 078* 039*
V3 3.36 1.68 F3 070* 3.56*
Chi-Square 0.000 with -3 d.f.
CFI 0.987

 

 

 

Legend: V1 to V3 indicate observed measures of attitudes fiom Time 0 to Time 2. F1 = Intercept Factor of
Attitudes; F2 = Increase Factor of Attitudes; F3 = Decrease Factor of Attitudes.
Note: the asterisk (*) indicates significance at the 95% level.

Results provide evidence that the growth and decrease trends are significant when
the information treatment is given and that variance is significantly large. The model is
under-identified because the number of free parameters to be estimated is higher than the
number of known parameters; therefore we add parameters in the following steps of
building a LGM. A similar piecewise LGM model has been run for the measures of
animal welfare beliefs and WTPP of respondents in Group 1 and for all respondents’

measures in Groups 2, 3 and 4.

165

if!

 

Associative LGM

The associative LGM is one large model that describes the change factors for several

measures at the same time to analyze if there is covariance among the change across the

measures (Duncan et al., 1999). An associative LGM is built where the increase and

decrease factors load to measures of beliefs, attitudes and WTPP simultaneously, where

the co-variances among each of the nine factors (three factors for each measure) are

estimated. The factor loadings are the same as in the simple piecewise LGM for each of

the three variables. The co-variance matrix from the associative LGM is reported in Table

18.

Table 18 - Co-variance Matrix of the Associative LGM with Consumer Attitudes in

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Groupl

F1 F2 F3 F4 F5 F6 F7 F8 F9
F1 268*
F2 136* 189*
F3 0.02 0.00 001*
F4 -0.86* -0.03 0.03* 243*
F5 079* -0.71* 0.05"“ 0.26 487*
F6 0.02 0.01 0.00 0.04* 0.08* 002*
F7 097* 038* 0.00 -0.57* 0.21 0.00 207*
F8 0.24 0.80* -0.02 0.20 -1.21* 0.00 1.31* 236*
F9 0.00 -0.01 0.00 0.02 0.00 -0.01 0.03* 0.02 0.01*

 

Legend: Fl = Intercept Factor of Attitudes; F2 = Intercept Factor of Beliefs; F3 = Intercept Factor of
WTPP; F4 = Increase Factor of Attitudes; F5= Increase Factor of Beliefs; F6 = Increase Factor of WTPP;
F7 = Decrease Factor of Attitudes; F8= Decrease Factor of Beliefs; F9 = Decrease Factor of WTPP.

Note: values on the diagonal are factor variances Di; the asterisk (*) indicates significance at the 95% level.

166

 

Results provide evidence that there is covariance among the increase and decrease factors
across the three measures of beliefs, attitudes and WTPP. The associative LGM model

has also been run with data of the measures from respondents in Groups 2, 3 and 4.

Multi-Group Associative LGM

The multi-group associative LGM is used to analyze if there are differences across the
parameters from respondents’ data in Group 1 and Group 3, which provide evidence also
to test the stated hypotheses in Chapter 3. In particular, a control has been performed to
establish if there are differences across factor means and factor variances across Group 1
and Group 3, where respondents in Group 1 received a positive information which is
unrelated to animal welfare and respondents in Group 3 received a positive information

related to animal welfare.

To control for these differences across parameters in the two groups, an equality
constraint is imposed to the model. Therefore, the LM test is performed to explore which
constraints have to be released in order to obtain a significant fit improvement. Results

are presented in Table 13 in the Chapter.

The same procedure has been used to compare differences in parameters across
Group 2 and Group 4. An interpretation of these results is provided in the text of the

Chapter.

Curve-of-Factors LGM
The curve-of-factor LGM describes the change of several measures with only one set of

factors to analyze if the same pace of change is the same across several measures or not

167

(Duncan et al., 1999). In this case, a curve-of-factors LGM is built to analyze if a unique
set of factors can describe the change occurring across beliefs, attitudes and WTPP.
When running the model with data from respondents in group 1, as the overall fit of the
model with data is low (chi-square=248.68 with 30 d.f. and p-value<0.001; CFI=0.697;
RMSEA= 0.285), results show that the changes in the three measures cannot be
effectively described by only one set of factors and so that there are differences in the
pace of change across beliefs, attitudes and WTPP. The same curve-of-factors LGM is
also run with only two out of the three variables and repeated the same analysis with
measures of respondents in Group 2, 3 and 4. In each evaluated case, the curve-of-factors

LGM failed to provide an adequate fit.

Predictive LGM with WTPP

As the curve-of-factors LGM suggests that no unique change factor can effectively
describe the change in beliefs, attitudes and WTPP simultaneously, an analysis of what
are the predictors of the change factor for each measure independently has been done.
First, a predictive LGM is run with the WTPP measures by adding all the expected
predictive variables (i.e., demographics, chicken consumption habits, food values) to the
simple piecewise WTPP model and estimating the impact of each of these variables on

the intercept, increase and decrease factors.

The output indicates that parameters are linearly dependent, and so that the output
of this model cannot be trusted. From the EQS 6.1 output, results indicate that linearly
dependent parameters are the errors of the three WTPP measures over time (e1, e2 and e3

in the generic piecewise LGM). This is due to the fact that the majority of WTPP values

168

are zero (around 85%), as only few respondents are WTP a premium price for
McDonald’s chicken sandwiches, no matter their demographics and the information
treatments they receive. Output is similar when the same predictive LGM with WTPP
from respondents’ data in Group 2, 3 and 4 is run. Therefore, data collected do not allow
analyzing predictors of WTPP changes over time. The same predictive LGM is then

repeated with respondents’ attitudes.

Predictive LGM with Attitudes

Results of final predictive LGM are presented in Tables 15 and 16 in the Chapter. To
build the final predictive LGM illustrated in these tables, a first preliminary predictive
LGM is run with only demographic and chicken consumption habit predictors. A second
preliminary predictive LGM with only food value predictors is also-run. As overall
goodness-to-fit with the data was bad, a Wald Test is performed to drop the independent
variables that bring the least contribution in explaining the dependent variables and those
that create serious problems of multi-collinearity. Therefore, in the predictive LGM with
attitudes measures from respondents in Group 1, respondents’ education (which has high
co-variance with income), chicken consumption frequency and value for food

sustainability and origin (as suggested by the Wald test) are dropped.

Therefore, a third predictive LGM is run with all the predictors but the variables
dropped previously, and then evaluated the model looking again at the overall goodness-
to-fit, the Wald test and the co-variance among independent variables. At this stage, the
respondents’ value for taste variable is also dropped, as suggested by the Wald test.

Therefore, a fourth and final predictive LGM is built with the remaining variables, which

169

are respondents’ gender, income and age, and obtained the results in Table 15 in the
Chapter. The same procedure has been used to come up with the final predictive LGM

with attitudes of respondents in group 2, 3 and 4.

170

REFERENCES

Aaker, D. (1991). Managing Brand Equity: Capitalizing on the Value of a Brand Name,
The Free Press, New York.

Ahluwalia, R., Burnkrant, R. E. & Unnava. H. R. (2000). “Consumer response to
negative publicity: The moderating role of commitment.” Journal of Marketing Research
37, 203-214.

American Veterinary Medical Association. (2006). “American Veterinary Policy
Statements.” Information from the Association Web Site: www.AVMA.org.

Carlson, K.A. and Pearo, L.K. (2004). “Limiting predecisional distortion by prior
valuation of attribute components.” Organizational Behavior and Human Decision
Processes 94, 48-59.

Carlson, K.A., Meloy, MG. and Russo, J.E. (2006). “Leader-Driven Primacy: Using
Attribute Order to Affect Consumer Choice.” Journal of Consumer Research 32, 513-518.

Carlsson, F ., F lykblom P. and Lagerkvist, C.J. (2007). “Consumer willingness to pay for
farm animal welfare: mobile abattoirs versus transportation to slaughter.” European
Review of Agricultural Economics 34 (3), 321-344.

Chen, Y.J. (2008). “Consumer Preferences for W001 Production Attributes.” PhD
Dissertation, Kansas State University, Kansas.

Darby, M. and Karni, E. (1973). “Free competition and the optimal amount of fraud.”
Journal of Law and Economics 16 (1), 67-88.

Dawar, N. & Pillutla, M. M. (2000). “Impact of product-harm crises on brand equity: The
moderating role of consumer expectations.” Journal of Marketing Research 37, 215-227.

171

 

Duncan, T. E., Duncan, S. C., Strycker, L. A., Li, R, & Alpert, A. (1999). An
introduction to latent variable growth curve modeling. Mahwah, NJ: Erlbaum.

Eagly, A. & Chaiken, S. (1993). The psychology of attitudes, Fort Worth, TX: Harcourt
Brace Jovanovich.

The Economist (1999), "Coca Cola: Bad For You," (June 19), 62.

Elliot, K.A. and Freeman. RB. (2003). Can Labor Standards Improve Under
Globalization? Washington, DC: Institute for International Economics.

Erdem, T. and Swait, J. (1998). “Brand Equity as a Signaling Phenomenon.” Journal of
Consumer Psychology 7 (2), 131-157.

EU Commission (2009). Information from EU Commission Web Site:
http://ec.europa.eu/.

Farm Foundation. 2006. Future ofAnimal Agriculture in North America. Oak Brook, IL:
Farm Foundation. available online.

Fishbein, M. (1967). “A Behaviour Theory Approach to the Relations Between Beliefs
about an Object and the Attitude Toward the Object.” Readings in Attitude Theory and
Measurement, John Wiley, New York, 389-400.

Fishbein, M. and Ajzen, I. (1975). Belief Attitude, Intention and Behavior: An
Introduction to Theory and Research. Reading, Massachusetts: Addison-Wesley.

Floydl, D.L., Prentice-Dunn, S. and Rogers, R.W. (2000). “A Meta-Analysis of Research
on Protection Motivation Theory.” Journal of Applied Social Psychology 30 (2), 407-429.

Fox, J .A., Hayes, DJ. and Shogren, J .F. (2002). “Consumers Preferences for Food
Irradiation: How favorable and unfavorable descriptions affect preferences for irradiated
pork in experimental auctions.” Journal of risk and Uncertainty 24 (1), 75-95.

Harper, GO and Makatouni, A. (2002). “Consumer perception of organic food
production and farm animal welfare.” British Food Journal 104 (3-5), 287-299.

Hayes, D.J., Fox, J.A. and Shogren, J.F. (2002). “Experts and activists: how information
affects the demand for food irradiation.” Food Policy 27 (2), 185-193.

Henson, S., and Traill, W. (2000). “Measuring perceived performance of the food system
and consumer food-related welfare.” Journal of Agricultural Economics 51, 388—404.

172

Herr, P. M., Kardes, F. R., & Kim, J. (1991). Effects of word-of-mouth and product-
attribute information on persuasion: An accessibility-diagnosticity perspective Journal of
Consumer Research. 17, 454-463.

Hudson, D. and Lusk, J. (2004). “Activists and Corporate Behavior in Food Processing
and Retailing: A Sequential Bargaining Game.” Journal of Agricultural and Resource
Economics 29 (1), 79-94.

Kamins, M.A., F olkes, V.S. and Perner, L. (1997). “Consumer Responses to Rumors:
Good News, Bad News.” Journal of Consumer Psychology 6 (2), 165-187.

Keller, K. L. (1993). “Conceptualizing, Measuring, and Managing Customer-Based
Brand Equity.” Journal of Marketing 57 (1), 1-22

Klein, J. and Dawar, N. (2004). “Corporate social responsibility and consumer’s
attributions and brand evaluations in a product harm crisis.” International Journal of
Research in Marketing 21, 203-217.

Klein, J.G., Smith, NC. and John, A. (2004). “Why We Boycott: Consumer Motivations
for Boycott Participation.” Journal of Marketing 68, 92-109.

Lagerkvist, C.J., Carlsson, F. and Viske, D. (2006). “Swedish Consumer Preferences for
Animal Welfare and Biotech: A Choice Experiment.” Journal of A grobiotechnology
Management and Economics 9(1), 51-58.

Liljenstolpe, C. (2008). “Evaluating Animal Welfare with Choice Experiments: An
Application to Swedish Pig Production.” Agribusiness 24 (1), 67-84.

Lusk, J .L., House, L.O., Valli, C., Jaeger, S.R., Moore, M., Morrow, J .L. and Traill, W.B.
(2004). “Effect of information about benefits of biotechnology on consumer acceptance
of genetically modified food: evidence from experimental auctions in the United States,
England, and France.” Eur0pean Review of Agricultural Economics 31(2), 179-204.

Martin, A. (2007). “Burger King Shifts Policy on Animals.” The New York Times, March
28th, 2007.

McDonald’s (2009). Information from Corporate Web Site: www.mcdonalds.com.

Nayga, R.M., Aiew, W. and Nichols, JP. (2005). “Information Effects on Consumers’
Willingness to Purchase Irradiated Food Products.” Review of Agricultural Economics 27
(1), 37-48. '

173

Nelson, P. (1970). “Information and Consumer Behavior.” Journal of Political Economy
78 (2), 311-329.

Nestle’ (2009). Information from Corporate Web Site: www.nestle.com.

Nilsson, T. Foster, K. and Lusk, J.L. (2006). “Marketing Opportunities for Certified Pork
Chops.” Canadian Journal of Agricultural Economics 54 (4), 567—583.

Okada, E. M. & Reibstein, D. J. (1998). When !@#? (bad stuff) happens: Effects of
related and unrelated positive associations on the influence of negative secondary
associations. In J. W. Alba & J. W. Hutchinson (eds.) Advances in Consumer Research
(Volume 25, pp. 349-356.). Provo, UT: Association for Consumer Research.

Pullig, C., Netemeyer, R. G., & Biswas, A.(2006). Attitude basis, certainty, and challenge
alignment: A case of negative brand publicity. Journal of the Academy of Marketing
Science, 34, 528-543.

Reardon, T., Henson, S. and Gulati, A. (2010). “Links Between Supermarkets and Food
Prices, Diet Diversity and Food Safety in Developing Countries” in Hawkes, C., Blouin,
C. Henson, 8., Drager, N., Dube’, L. (2010), Trade, Food, Diet and Health: Perspectives
and Policy Options, Blackwell Publishing Ltd.

Roehm, ML. and Tybout, AL. (2006). “When Will a Brand Scandal Spill Over, and
How Should Competitors Respond?” Journal of Marketing Research 43, 366-373.

Rogers, R.W. (1975). “A Protection Motivation Theory of Fear Appeals and Attitude
Change.” Journal of Psychology 91, 93-114.

Rousu, Huffman, Shogren, Tegene (2002). “The value of verifiable information in a
controversial product market: evidence from lab auctions of genetically modified foods.”
Staff Working Paper 3, Iowa State University, IA.

Russo, J.E., Meloy, MG. and Medvec, V.H. (1998). “Predecisional Distortion of Product
Information.” Journal of Marketing Research 35 (4), 438-452.

SAI Platform (2009). Information from the Sustainable Agriculture Initiative Platform
Web Site at: http://www.saip1atform.org/.

Schroder, M.J.A. and McEachem, MG. (2004). “Consumer value conflicts surrounding
ethical food purchase decisions: a focus on animal welfare.” International Journal of
Consumer Studies 28 (2), 168—177.

174

Scott, CA. and Tybout, A.M. (1981). “Theoretical Perspectives on the Impact of
Negative Information: Does Valence Matter?” Advances in Consumer Research 8, Kent
B. Monroe, ed. Ann Arbor, MI: Association for Consumer Research, 408—409.

Shocker, A.D, Srivastava, R.K. and Ruekert, R.W. (1994). “Challenges and opportunities
facing brand management.” Journal of Marketing Research 31 (2), 149.

Smith, R. E. (1993). “Integrating information from advertising and trial: Processes and
effects on consumer response to product information.” Journal of Marketing Research 30,
204-219.

Smith, R. E. and Vogt, C. A. (1995). “The effects of integrating advertising and negative
word-of-mouth communication on message processing and response.” Journal of
Consumer Psychology 4, 133-151.

Stott, A.W., Milne, C.E., Goddard, P.J., and Waterhouse, A. (2005). “Projected effect of
alternative management strategies on profit and animal welfare in extensive sheep
production systems in Great Britain.” Livestock Production Science 97, 161—171.

Tanner, J.F., Hunt, J.B. and Eppright, DR. (1991). “The Protection Motivation Model: A
Normative Model of Fear Appeals.” The Journal of Marketing 55 (3), 36-45.

Teegen, H., Doh, J .P. and Vachani, S. (2004). “The importance of nongovernmental
organizations (N GOs) in global governance and value creation: an international business
research agenda.” Journal of International Business Studies35, 463-483.

Tonsor, G.T., Olynk, N. and Wolf, C. (2009a). “Consumer Preferences for Animal
Welfare Attributes: The Case of Gestation Crates.” Journal of Agricultural and Applied
Economics, Forthcoming.

Tonsor, G.T., Olynk, N. and Wolf, C. (2009b). “Media Coverage of Animal Handling
and Welfare: Influence on Meat Demand.” Paper prepared for presentation at the
American Agricultural Economics Association Annual Meetings, Milwaukee, WI, July

2009.

Tonsor, G.T., Wolf, C. and Olynk, N. (2009). “Consumer voting and demand behavior
regarding swine gestation crates.” Food Policy 34, 492—498.

Tybout, A.M., Calder, BJ. and Stemthal, B. (1981). “Using Information Processing
Theory to Design Marketing Strategies.” Journal of Marketing Research 18 (1), 73-79.

175

 

Verbeke, V.A.J. and Viaene, J. (2000). “Ethical Challenges for Livestock Prodution:
Meeting Consumer Concerns about Meat Safety and Animal Welfare.” Journal of
Agricultural and Environmental Ethics 12, 141—151.

Videras, J. (2006). “Religion and animal welfare: Evidence from voting data.” Journal of
Socio-Economics 35(4), 652-659.

Wachenheim, OJ. and VanWechel, T. (2004). “The Influence of Environmental-Impact
Information on Consumer Willingness to Pay for Products Labeled as Free of Genetically
Modified Ingredients.” Journal of Food Distribution Research 35 (2), 1-13.

Weinberger, M. G., Allen, C. T. and Dillion, W. R. (1981). “The impact of negative
marketing communications: The consumers’ union/Chrysler controversy.” Journal of
Advertising 10, 20-28.

176

CONCLUSION

This research work represents an initial attempt to build a theory of branding products
with credence attributes. This aims at analyzing which credence attribute information can
differentiate a brand from its competitors and so make a firm gain a benefit advantage

over its competitors.

This research attempts to provide both a conceptual and methodological
contribution to the agri-food marketing literature. From a conceptual standpoint, this
research integrates agricultural economics literature with a literature built upon the theory
of attitude formation, developed in psychology and applied to marketing. From a
methodological point of view, this research introduces different uses of structural
equation modeling and path modeling (Hair et al., 2006) to tackle research questions on
the impact of credence attribute information on consumers’ attitudes and buying

intentions.

In the first essay, results provide evidence that credence attribute information on
Michigan locally-grown has both a direct and indirect effect on consumers’ attitudes
towards apples. This suggests that consumers like “locally-grown” mainly because it
creates inferences with other desired credence attributes, such as pesticide-free, and

experience attributes, such as good flavor or firmness. This has implications for managers

177

 

seeking to obtain a benefit advantage by communicating the “locally-grown” attributes of

their brand or other attributes that consumers infer from “locally-grown”.

In the second essay, results suggest that credence attribute information on Liguria
extra-virgin olive oil differentiates the brand of the firm providing the information in
terms of attitudes and willingness-to-pay, but that credence attribute information on
Southern Louisiana cream cheese does not. However, in both these cases, individual
brand information related to the place-of-origin creates a higher level of brand
differentiation relative to generic attribute information. This has major implications for

managers seeking brand differentiation by adding place-of-origin to their products.

Finally, results from the third essay show that brand information'either related or
unrelated to animal welfare issues has the positive effect of mitigating the impact of
negative information shocks on consumers’ attitudes and intentions to buy McDonald’s.
However, we found that McDonald’s information related to animal welfare has a more
positive effect on some consumer segments, such as elder individuals with higher
education. On the other hand, females with higher income are more sensitive to

information unrelated to animal welfare which aims at mitigating a negative shock.

Although introducing a conceptual framework based on the theory of attitude
formation as well as a structural equation and path modeling approach to the context of
agri-food marketing, this research does not provide results that are highly generalizable,
as they remain limited to specific products, attributes and inforrnation treatments. Future
research should first test if these results hold across different products, either food or non-
food, and across different credence attributes. Moreover, future research should test how

results vary across various information sources, as well as across various information

178

 

contents. By undertaking this direction, future research would establish the necessary and
sufficient conditions that complete the development of a theory of branding products with

credence attributes.

179

 

Mllllllllllllllllllllllllllljlllll“