measuring the impacts of dierent messengers on consumer ... · on consumer preferences for products...
TRANSCRIPT
1
Measuring the Impacts of Di�erent Messengerson Consumer Preferences for Products Irrigated
with Recycled Water: A Field Experiment
Maik Kecinski, University of Alberta
Coauthors:Alix T. Schmidt, University of Alberta
Tongzhe Li, University of WindsorKent D. Messer, University of Delaware
Julia Parker, University of Delaware
2
Introduction
2
Introduction
2
Introduction
2
Introduction
3
IntroductionLiterature
I E�ective communication not only depends on the message butoften times more importantly on the messenger.
I Individuals are more receptive to information when it iscommunicated by a trusted source.
I But they may also entirely reject a message if they do nottrust the source that conveys it.
I Understanding the impacts di�erent messengers have on anindividual’s behavior holds important policy implications -policy communication but also nudging consumers towardsaccepting new technologies and policies.
3
IntroductionLiterature
I E�ective communication not only depends on the message butoften times more importantly on the messenger.
I Individuals are more receptive to information when it iscommunicated by a trusted source.
I But they may also entirely reject a message if they do nottrust the source that conveys it.
I Understanding the impacts di�erent messengers have on anindividual’s behavior holds important policy implications -policy communication but also nudging consumers towardsaccepting new technologies and policies.
3
IntroductionLiterature
I E�ective communication not only depends on the message butoften times more importantly on the messenger.
I Individuals are more receptive to information when it iscommunicated by a trusted source.
I But they may also entirely reject a message if they do nottrust the source that conveys it.
I Understanding the impacts di�erent messengers have on anindividual’s behavior holds important policy implications -policy communication but also nudging consumers towardsaccepting new technologies and policies.
4
IntroductionLiterature
I In the recent past, the most trusted source of information werescientists while trust in media and governments was oftenranked significantly lower (Hunt and Frewer 2001, Haynes etal. 2007, Dolnicar and Hurlimann 2010, Arbuckle et al. 2015).
I In the more recent present, research suggests that the trustindividuals place in scientists may be declining (Bubelba et al.2009, Leiserowitz et al. 2012, Higgins 2016, Makri 2017).
5
IntroductionLiterature
I These studies suggest that the reason why public trust inscientists may be declining is due to increasinginterdisciplinary work, global research focus and privatefunding from industry (Higgins 2016, Makri 2017).
I Leiserowitz et al. (2012) attribute some of this decline to the2009 ’Climategate’ ”scandal.”
I Previous studies have mostly relied on stated preferencesurvey approaches...leaving opportunities to gain valuableinsight using experimental approaches.
5
IntroductionLiterature
I These studies suggest that the reason why public trust inscientists may be declining is due to increasinginterdisciplinary work, global research focus and privatefunding from industry (Higgins 2016, Makri 2017).
I Leiserowitz et al. (2012) attribute some of this decline to the2009 ’Climategate’ ”scandal.”
I Previous studies have mostly relied on stated preferencesurvey approaches...leaving opportunities to gain valuableinsight using experimental approaches.
5
IntroductionLiterature
I These studies suggest that the reason why public trust inscientists may be declining is due to increasinginterdisciplinary work, global research focus and privatefunding from industry (Higgins 2016, Makri 2017).
I Leiserowitz et al. (2012) attribute some of this decline to the2009 ’Climategate’ ”scandal.”
I Previous studies have mostly relied on stated preferencesurvey approaches...leaving opportunities to gain valuableinsight using experimental approaches.
6
IntroductionResearch Questions
I We used artefactual field experiments to explore the e�ects ofdi�erent messengers on consumer behavior when makingpurchasing decisions concerning sustainable, albeit, potentiallyrepulsive products irrigated with recycled water. Specifically:
(1) Does providing information on the benefits of sustainablegrowing techniques a�ect consumers’ acceptance of their use?
(2) What are the impacts of messenger communication onindividual behavior?
(3) How e�ective are di�erent types of messengers (newspapers,scientists, government agencies, non-profit organizations) inincreasing the acceptability of recycled water products?
6
IntroductionResearch Questions
I We used artefactual field experiments to explore the e�ects ofdi�erent messengers on consumer behavior when makingpurchasing decisions concerning sustainable, albeit, potentiallyrepulsive products irrigated with recycled water. Specifically:
(1) Does providing information on the benefits of sustainablegrowing techniques a�ect consumers’ acceptance of their use?
(2) What are the impacts of messenger communication onindividual behavior?
(3) How e�ective are di�erent types of messengers (newspapers,scientists, government agencies, non-profit organizations) inincreasing the acceptability of recycled water products?
7
Experimental DesignBecker Degroot Marschak (BDM) Mechanism, Becker et al. (1964)
Willingness-to-PayNon-hypothetical purchasesParticipants were paid $15
No Information Recycled Water Conventional Waterabout Irrigation Irrigation Irrigation(1) Strawberries (2) Strawberries (3) Strawberries
(4) Cotton T-shirts (5) Cotton T-shirts (6) Cotton T-shirts
I Each participant made 6 bids (Bij) in random order.I Random draw (Rij): [0, 15] uniform distribution.
Purchase =
I1, if Bij Ø Rij (WTP)0, if Bij < Rij (WTP).
7
Experimental DesignBecker Degroot Marschak (BDM) Mechanism, Becker et al. (1964)
Willingness-to-PayNon-hypothetical purchasesParticipants were paid $15
No Information Recycled Water Conventional Waterabout Irrigation Irrigation Irrigation(1) Strawberries (2) Strawberries (3) Strawberries
(4) Cotton T-shirts (5) Cotton T-shirts (6) Cotton T-shirts
I Each participant made 6 bids (Bij) in random order.I Random draw (Rij): [0, 15] uniform distribution.
Purchase =
I1, if Bij Ø Rij (WTP)0, if Bij < Rij (WTP).
7
Experimental DesignBecker Degroot Marschak (BDM) Mechanism, Becker et al. (1964)
Willingness-to-PayNon-hypothetical purchasesParticipants were paid $15
No Information Recycled Water Conventional Waterabout Irrigation Irrigation Irrigation(1) Strawberries (2) Strawberries (3) Strawberries
(4) Cotton T-shirts (5) Cotton T-shirts (6) Cotton T-shirts
I Each participant made 6 bids (Bij) in random order.I Random draw (Rij): [0, 15] uniform distribution.
Purchase =
I1, if Bij Ø Rij (WTP)0, if Bij < Rij (WTP).
8
Experimental DesignStep 1: Water Definitions
Recycled WaterI ”Recycled water is highly treated wastewater from various
sources, such as domestic sewage, industrial wastewater andstorm water runo�.”
Conventional WaterI ”Typical sources of conventional water include: surface water,
groundwater from wells, rainwater, impounded water (ponds,reservoirs, and lakes), open canals, rivers, streams, andirrigation ditches.”
9
Experimental DesignStep 2: Treatment Information
10
Experimental DesignStep 3: Placing Bids
11
Experimental DesignStep 3 cont.: Placing Bids
12
ResultsSample
I Experiments were conducted at the Department of MotorVehicles and outside a dairy and ice cream shop in Delaware.
I 201 total participants.
I 59% of participants were female.
I The average age of the participants was 44 years.
I 20% were conservative, 30% liberal, 40% moderate, and 10%had no political a�liation.
13
Results - Descriptive
$3.00
$3.50
$4.00
$4.50
$5.00
$5.50
$6.00
Newspaper Science Government Non-profit Control
Mea
n B
id
Treatment (Information Source)
Mean Bids by Treatment and Product
Unspecified T-shirt Bid Conventional T-shirt Bid Recycled T-shirt Bid
Unspecified Strawberry Bid Conventional Strawberry Bid Recycled Strawberry Bid
14
Results - Hurdle Model - Zero Inflated (Cragg 1971)
Coe�cient Std. Error p-ValueSelection Model(More likely to bid)Newspapers 0.2853 0.1652 0.084Scientists -0.4965 0.1474 0.001Government Agencies 0.1449 0.1651 0.380Non-profit Organizations 0.3676 0.1772 0.038Control (Baseline)Constant 1.2381 0.1172 0.000
Outcome Model (Value)Newspapers 1.2034 0.3939 0.002Scientists 0.9386 0.4261 0.028Government Agencies 0.8467 0.4110 0.039Non-profit Organizations 0.7837 0.4088 0.055Control (Baseline)Constant 3.2197 0.3340 0.000
N = 1206. Lower limit set to zero. Regression is not product specific.
15
Results - Hurdle Model - Zero Inflated (Cragg 1971)
Coe�cient Std. Error p-ValueSelection Model(More likely to bid)T-shirt 0.0572 0.1685 0.734Newspapers 0.6859 0.2759 0.013Scientists -0.3432 0.2413 0.155Government Agencies 0.5206 0.2781 0.061Non-profit Organizations 0.6905 0.9916 0.021Control (Baseline)Female 0.0781 0.1751 0.656Age 0.0040 0.0046 0.386Political A�liationConservative 0.4003 0.2008 0.046Moderate 0.2407 0.2348 0.305Other 0.8099 0.3903 0.038Liberal (Baseline)
Constant 0.3372 0.3485 0.333
16
Results - Hurdle Model - Zero Inflated (Cragg 1971)
Coe�cient Std. Error p-ValueOutcome Model (Value)T-shirt 2.0744 0.4322 0.000Newspapers 1.3290 0.7154 0.063Scientists 1.5474 0.7730 0.045Government Agencies 0.8300 0.7446 0.265Non-profit Organizations 0.7121 0.7428 0.338Control (Baseline)Female -0.6441 0.4343 0.138Age -0.0230 0.0115 0.047Political A�liationConservative -0.7782 0.5119 0.128Moderate -1.3362 0.6323 0.035Other -0.2360 0.7729 0.760Liberal (Baseline)
Constant 4.0786 1.0069 0.000N = 338. Lower limit set to zero. Regression for products irrigated with recyced water.
17
Results - Di�erence-in-Di�erence Model
Di�-in-Di� Coe�cient Std. Error p-ValueNewspapers 0.6453 0.3834 0.092Scientists 0.3528 0.3975 0.375Government Agencies 0.3262 0.3969 0.411Non-profit Organizations 0.1806 0.3993 0.651Gender 0.1131 0.2453 0.645Age 0.0047 0.0065 0.467T-Shirt 0.4329 0.1686 0.010Political A�liationConservative -1.2544 0.3422 0.000Moderate -0.5358 0.2882 0.063Other -0.7459 0.4610 0.106Liberal (Baseline)
Constant -0.4245 0.5145 0.409N = 338.
18
Summary
I E�ective communication is complex.
I Discrepancies in behavior exist between placing bids versuspaying more money.
I Messengers can significantly impact this behavior - mosthaving a positive coe�cient except for the Science messenger,which had a negative coe�cient in the selection model.
I We find that the Newspaper treatment had the strongestoverall impact on individual behavior, leading participantstowards placing more bids and also to increase their WTP forrecycled water products.
19
Summary
I Water recycling is a sustainable and cost-e�ectivemethod to ensure water availability now and in thefuture.
I Understanding how to increase or decreaseacceptability of new technologies or policies impactsthe success of such programs.
I The right messenger can provide a nudge in the rightdirection; the wrong messenger can have the oppositee�ect.
20
References
I Leiserowitz AA, Maibach EW, Roser-Renouf C, Smith N, Dawson E (2012) Climategate, public opinion,and the loss of trust. Am Behav Sci 57(6):818-837.
I Makri A (2017) Give the public the tools to trust scientists. Nature 541(7637):261.I Higgins K (2016) Post-truth: a guide for the perplexed. Nature 540(7631):9.I Arbuckle JG Jr, Morton LW, Hobbs J (2015) Understanding farmer perspectives on climate change
adaptation and mitigation: the roles of trust in sources of climate information, climate change beliefs, andperceived risk. Environ Behav 47(2):205-234.
I Haynes K, Barclay J, Pidgeon N (2007) The issue of trust and its influence on risk communication during avolcanic crisis. Bull Volcanol 70(5):605-621.
I Hunt S, Frewer LJ (2001) Trust in sources of information about genetically modified food risks in the UK.Brit Food J 103(1):46-62.
I Bubela T, Nisbet MC, Borchelt R, Brunger F, Critchley C, Einsiedel E, Geller G, Gupta A, Hampel J,Hyde-Lay R, Jandciu E, Jones S, Kolopack P, Lane S, Lougheed T, Nerlich B, Ogbogu U, O’Riordan K,Ouellette C, Spear M, Strauss S, Thavaratnam T, Willemse L, Caulfield T (2009) Science communicationreconsidered. Nat Biotechnol 27(6):514-518.
I Becker G, DeGroot M, Marschak J (1964) Measuring utility by a single-response sequential method. SystRes Behav Sci 9:226-236.
21
Results
05
1015
05
1015
0 5 10 15 0 5 10 15 0 5 10 15
Shirt- Unspecified Shirt- Conventional Shirt- Recycled
Strawberry- Unspecified Strawberry- Conventional Strawberry-Recycled
Den
sity
Bid in US DollarsGraphs by ProductDummy
Histograms of Bids by Product
22
Experimental DesignLink