gender on green consumption

7/31/2019 Gender on Green Consumption

1/29

1

Green Consumption and Peer effect:

An Application to Seafood Products in FranceSterenn Lucasa*, Frdric Salladarrb, Dorothe Brcarda

aUniversit de Nantes, LEMNA, Institut dconomie et de Management de Nantes IAE, Chemin de laCensive du Tertre, BP 52231, 44322 Nantes Cedex 3, France.

bUniversit-IUT de Rennes 1, CREM-CNRS, LEMNA, Campus de Beaulieu, Avenue du Gnral Leclerc,CS 44202, 35042 Rennes Cedex, France.

Abstract:

The consumers are increasingly concerned with the environmental impacts of they

consumption, and some are willing to pay a premium for a more environmentally

friendly product. But what influences the probability to have a positive willingness to

pay (WTP)? Beyond the socio-economic characteristics and the environmental

consciousness, the social norms can also largely influence consumer behavior and thus

consumer WTP a premium for labeled products. In this paper we work on the influence

of the peers behaviors on the probability to have a positive WTP a premium for labeled

seafood products, through a propensity score-matching estimation. We find that the

social influences is significant positive, having pro-eco-label peers increase the

probability to have a positive WTP for labeled seafood products. Under gender

restriction, we find that this effect is consistent for women but does not hold for men.

Keywords: Consumer behavior, social interaction, eco-label, seafood, propensity score

matching.

JEL Classification: D12, Q22, C21

* Corresponding author. E-mail address:[email protected], +33 240 141 731
mailto:[email protected]:[email protected]:[email protected]:[email protected]


2/29

2

1 Introduction

Opinion surveys agree that consumers are increasingly concerned with the

impacts of the products they consume on the environment. According to the

Eurobarometer survey (European Commission, 2009), the French are particularly

sensitive to these impacts: when buying or using products, 78% are generally or fully

aware of their environmental effects. This preoccupation leads 72% of them to be ready

to buy environmentally friendly products even if they cost a little bit more (European

Commission, 2011). Paradoxically, according to the 2011 Green Brands survey1, 78% of

French people state that the biggest challenge to purchasing green products or services

is that they are too expensive. Besides, 36% do not want to purchase these products at

a higher price. Actually, they are only 20% to declare having recently bought

environmentally friendly products marked with an environmental label (European

Commission, 2011). What can explain consumer willingness to pay (WTP) a premium

for a green product? Although environmental consciousness and socio-economic

characteristics of consumers play a major role, social norms can also largely influence

the product choices of consumers. Indeed, the 2011 Green Brands survey show that,after the packaging (for 30% of the respondents), this is the word of mouth (i.e. family,

friends) that has the greatest impact on their likelihood to purchase green products (for

23%). The purpose of our paper is to detect to what extent the social circle influences

the WTP a premium for a specific green product: an eco-labeled seafood product.

The higher prices of green products may act as a brake on green consumption

and partially explain the attitude/behavior gap underlined by Young et al. (2010).

However, such prices are essential to cover their production costs. Indeed, green

products are generally more labor-intensive, produced on a smaller scale and/or involve

more environmentally friendly technologies. Moreover, their higher price can also be

used to signal their environmental friendliness to consumers (Mahenc 2006). Budget

constraints can thus limit green consumption, even that of the ecological aware

1 The 2011 Green Brands study surveys over 9,000 people, including 1,100 French people. It is conductedby WPP companies Cohn & Wolfe, Landor Associates and Penn, Schoen & Berland (seehttp://www.cohnwolfe.com,, accessed 25/01/2012)
http://www.cohnwolfe.com/http://www.cohnwolfe.com/http://www.cohnwolfe.com/


3/29

3

consumers. Accordingly, consumer income plays an important role in the choice

between green and standard products.

Previous literature on green consumer profile highlights a certain consistency in

their socio-economic characteristics2. First at all, gender appears as a major determinant

of consumer preferences. Actually, women seem generally more likely than men to

choose an eco-labeled product and also to pay a premium for green products (Blend and

Van Ravenswaay, 1999; Loureiro 2003; Loureiro et al., 2002; Brcard et al., 2009;

Salladarr et al., 2010). It is worth noting that the men who are ready to buy green

products at higher prices, are also willing to pay a larger premium than women

(Carlsson and Johansson-Stenman, 2000; Dupont, 2004). Furthermore, the importance

women place on the environmental information is larger than men (Bjrner, 2004; Teisl

et al., 2008). Secondly, the age affects consumer attitude towards green products. Many

papers show that the younger the consumers are, the more ecological oriented they are

likely to be (Loureiro and Lotade, 2005; Srinivasan and Blomquist, 2009; Brcard et al.,

2009; Salladarr et al., 2010), whereas Johnston et al. (2001) highlight the reverse

effect. However, paradoxically, older consumers longer trust the label information than

younger ones (Teisl et al., 2008). The third main determinant of green consumption isthe level of education, which, through a better comprehension of environmental issues,

favors confidence in eco-information and green consumption (Teisl et al., 2008;

Wessells et al., 1999). Among other highlighted socio-economic characteristics of green

consumers, we can also cite the presence of children under 18, a low family size,

confidence in certifying organizations, environmental involvement, interest in politics,

regional affinities, faith in people (Torgler and Garcia-Valias, 2007) and being in favor

of equitable sharing.

Beyond socio-economic characteristics, moral motivation and social interaction

influence pro-environmental behavior too. Frey and Stutzer (2006) underline four types

of moral motivations: pure and impure altruism, internalized norms, intrinsic

2 for detailed surveys on main determinants of pro-environmental attitudes, see, for instance, Torglerand Garcia-Valias (2007) and Brcard et al. (2009).


4/29

4

motivations and social norms. Indeed, using an altruism scale, based on Schwartzs

model (1970, 1977), Clark et al. (2003) and Kotchen and Moore (2007) demonstrate the

positive impact of altruistic behavior on environmentally friendly consumption.Internalized norms and intrinsic motivations refer to individuals values that may bring

him/her warm glow and welfare when he/she acts in favor of environment protection.

They could however be in conflict with extrinsic motivations, such as laws or financial

incentives, leading to a crowding-out effect (Frey and Oberholzer-Gee, 1997). The

social norms affect consumer behavior through the opinion of the other members of the

society about his/her behavior. Indeed, social norms could promote green consumption

when the consumer think that such a behavior is well thought-of, as shown in Section 2synthesizing previous literature in this field. This idea is central to our paper, since we

attempt to highlight the role of peer effect in the existence of a positive WTP for an eco-

labeled seafood product.

The development of eco-labeled seafood products could partially respond to the

overexploitation issue.3 Indeed, 85% of fish stocks are either fully exploited, over-

exploited, depleted or recovering (FAO, 2010). Seafood eco-labels, as the Marine

Stewardship Council (MSC)4

, certify that the fishing activity preserves marineresources, minimizes its environmental impact and practices sustainable management.

By delivering environmental information to consumers, fish eco-labeling could

encourage fish consumers to turn towards eco-friendly consumption and, by this way,

further eco-friendly fisheries. Consumer guidelines can also reinforce the green

consumption incentive. In France, about ten seafood guidelines are published, among

them the WWF, Greenpeace and Nicolas Hulot Foundation ones. It is worth noting that

consumers of eco-labeled seafood products show a similar profile to green consumerspreviously depicted, despite a lack of consensus concerning the age effect. These

consumers pay also attention to species, fresh or frozen form, wild or farmed nature and

3 For an extensive analysis of seafood eco-labeling, see Ward and Phillips (2008).

4 Created by the World Wild Fund for Nature (WWF) and Unilever in 1997, MSC became an

independent non-profit organization in 1999. It, at present, certifies fisheries catching 7% of the total

global capture production for direct human consumption, and will soon increase to 12%. Despite its goal

is to promote sustainable fisheries, the MSC is the subject of a controversy (Jacquet and Jaulry, 2010).


5/29

5

geographical origin of the fish (Wessells et al., 1999; Johnston et al., 2001; Roheim et

al., 2004; Johnston and Roheim, 2006; Brcard et al., 2009; Salladarr et al., 2010).

Interestingly, Johnston and Roheim (2006) show that high overfishing encouragesconsumers to turn towards less exploited species, but that the eco-label alone is not

enough to divert consumers to their most-favored species. Nevertheless, eco-labeling is

an important tool for the fishing industries in order to promote more sustainable

fisheries, to protect endangered species and to benefit from the eco-labeled seafood

premium (Deere, 1999). Although eco-labeled seafood demand has been widely studied

last years, the novelty of our research stems from the attention paid to the role of social

interactions in consumer WTP a premium for eco-labeled seafood products.

In order to analyze the peer effect on WTP for labeled seafood products, we use

a propensity score-matching model, in the line with Rosebaum and Rubin (1983) 5. The

propensity score-matching model is a popular approach used to estimate causal

treatment effects in many research fields, as labor economics (Heckman et al., 1997;

Messe et al. 2009), health economics (Harding, 2003) and education (Dearden et al.,

2005). This method allows highlighting the role played by a treatment on individuals. In

our paper, we focus on the impact of having pro-label peers (the traitement) on theprobability to be willing to pay a premium for a labeled seafood product. The basic

principle consists in splitting individuals into two groups, a treated group and an

untreated one. Obviously, we cannot observe for a same individual his/her WTP with

and without treatment, since he/she belongs to only one group. Consequently, the

method consists in identifying all relevant pretreatment characteristics (age, gender,...)

which are similar between individuals without pro-label peers and individuals with pro-

label peers. In this way, we can attribute the difference in WTP a premium for labeledproducts between treated and control groups to the peer effect. In order to make this

studies more relevant, we are also interested in peer effect under gender restriction,

since gender is an important determinant of green consumption. Furthermore, different

sensitivity of women and men to peer effect has already been shown by Carlsson et al.

(2008). We use original data from a French survey carried out on around 900

5 For a detailed survey, see Caliendo and Kopeinig (2008);


6/29

6

respondents by he RICEP6 in 2010. Beyond respondent attitude toward eco-labeled

seafood products, traditional socio-economic characteristics and fish purchase criteria,

the survey gives information about the peer effect through a question concerningrespondent perception of the circle attitude towards eco-labeled seafood products.

Our results show that being surround by people interested in eco-labeled fish

increases the probability of willing to pay a premium for a labeled seafood product, this

result is constant across several specifications. A deeper analysis of differences

according to gender highlights that women are strongly influenced by her peers

behaviors, while its seems that there is no impact on men probability of willing to pay a

premium for a labeled seafood product.

The remainder of the paper is structured as follows. In section 2, we survey

previous literature dealing with peer effects in green consumption patterns. In section 3

we introduce the database and the econometric method. In section 4, we analyze our

empirical results. Section 5 brings the paper to a conclusion.

2 Peer Effect and green consumption

Brock and Durlauf (2001) define social interaction as the idea that the utility or

payoff an individual received from a given action depends directly on the choices of

others in that individual's reference group (Brock and Durlauf, 2001). Cowan et al.

(1997) distinguish three types of reference groups susceptible to influence a consumer.

The peer group includes similar consumers, regularly interacting among themselves,

with whom the consumer would like to share consumption patterns. The contrast group

encompasses consumers with whom the consumer would not like to interact and from

whom he/she wishes to distinguish himself/herself. The aspirational group is the one

with which the consumer does not regularly interact, but hopes to do so. Accordingly, a

consumer may copy behaviors of consumers inside his/her peer and aspirational groups,

whereas he/she may avoid demanding the same goods as his/her contrast group.

6 The RICEP (Rseau dInformation et de Conseil en Economie des Pches) is the French Network ofInformation and Advice in Economics of Fisheries.


7/29

7

Social interactions were already highlighted in Veblen (1899) and Leibenstein

(1950)s articles, which emphasized that consumers are aware of the consumption

choices of others. Granovetter and Song (1986) formalized this idea in thresholdmodels with interpersonal effects in consumer demand. The influence of others may

lead to bandwagon effect or to snob effect. In the case of bandwagon effect,

individual utility increases with the number of individuals consuming the same good.

This arises from the desire of each consumer to purchase the same good as his/her peer

and aspirational groups in order to conform with these groups, to get into the swim of

things and to be fashionable. The utility could also increase because it depends on joint

consumption by others7

. Finally, consumption of others may facilitate access to the goodand makes it cheaper. In the case of snob effect or reverse bandwagon effect,

individual utility decreases with the number of individuals consuming the same good.

This may be due to the wish of a consumer to purchase a good his/her contrast group

does not consume. This effect mainly concerns luxury goods and stems from the

satisfaction arising from having a rare good. This effect may also arise from congestion

effects for some goods and services8. The same good can be affected by both effects,

according to the consumption volume..

The interactions among consumers have mainly been analyzed in social

psychology (Schultz, 1998) and in education (Steinberg et al., 1996) and applied to

criminal orientation (Glaeser et al., 1996), voting behavior (Gerber and Rogers, 2009),

retirement savings (Beshears et al., 2009) and charitable giving (Frey and Meier, 2004).

Recently, social interactions have also been studied in the field of green consumption.

Bandwagon effects in green consumption can lead to new social norms, in which

consuming green products is well thought of others. Consumer behavior can then beoriented towards green consumption, through green nudges9.

Green nudges consist in promoting environmental social norms and, by this

7 Granovetter and Soong (1986) give the vivid example of the utility derived from eating in an emptyrestaurant against the one arising from eating in a lively restaurant.

8 Granovetter and Soong (1986) carry on their example underlining the disutility arising from a jam-packed restaurant.

9 For a survey, see Centre dAnalyse Stratgique (2011).


8/29

8

way, inciting consumers to ecological behavior, without being prescriptive or guilt

inducing. Implementation of green nudges simply requires informing consumers

about their peer ecological behavior (Oullier and Sauneron, 2011). We can then fillconsumer attitude/behavior gap (Young et al. 2010) by including ecological behavior

in social norms.

Several experiments highlighted the effectiveness of green nudges, especially

in energy consumption field. Allcott (2011) analyzes a natural field experiment

conducted in the United State consisting in sending Home Energy Report letters, which

inform costumers about the energy use of their similar neighbors. He shows that this

resulted in a fall in energy consumption by 2%, although the same effect would be

reached with a rise in short-run electricity price by 11 to 20% or a 5% long run price

increase. Nevertheless, according to Schultz et al. (2007), social comparison does favor

green consumptions only in the case where household energy consumption is initially

below the average level of energy consumption of their neighbors. In the reverse case,

providing information triggers a boomerang effect leading the most energy efficient

households to raise their energy consumption. In order to avoid this adverse

phenomenon, Schultz et al. (2007) show that it is sufficient to add to the descriptivenorm an injunctive norm (Cialdini et al., 1990), as a smiley. Conversely, using a

Swedish survey dealing with the choice between green and non-green electricity, Ek and

Sderholm (2008) do not succeed in supporting the influence of social interaction on the

green electricity choice. This can be explained by the absence of visibility of others

choices.

Several studies confirm the efficiency of green nudges not only for energy use

but also for other ecological behaviors. Using a survey conducted in Germany, Welschand Khling (2009) highlight the role played by routines and reference group behavior

in the choice of solar energy equipment, green electricity programs and organic foods.

In the survey, they ask to respondents whether people around them behave ecologically

for each studied product10. The authors find that consumption patterns of reference

groups are significant covariates of all three kinds of green products, especially for

10For example, concerning the organic food respondents were ask Do many of your friends,neighbors and relatives buy food that is labeled as organic food?.


9/29

9

organic food and, to a lesser extent for green electricity. They also underline that

possible motives for installing solar thermal systems are a desire for autarchy or for

status (Mercedes-Benz on the rooftop). Through a choice experiment, Carlsson et al.(2008) analyze the effect of the number of consumers choosing environmentally

friendly coffee over standard coffee on the WTP for green coffee. The respondents were

asked to choose among three kinds of coffees, which vary with respect to their content

of ecological and fair trade beans and their prices. Three treatments were used, differing

only in the information given about the choices made by others consumers: 10%, 50%

or 90% of others were supposed to consume 100% of ecological beans. Using a random

parameter logit model, the authors cannot support the bandwagon effect at the aggregatelevel. However, focusing on gender differences, they demonstrate that the bandwagon

effect significantly increases only women willingness to choose and to pay a premium

for the ecological coffee, although men are more inclined to consume ecological coffee.

Schultz (1998) implemented a field experiment on waste recycling in California.

During four weeks, 120 households of a district were informed of the recycling

behavior of the neighbors. This green nudge led to a 19% increase in the volume of

recycling. According to the author, handwritten notation positively influenced theresults, strengthening the household feeling of closeness. In the same way, an

experiment conducted by Goldstein et al. (2008) concerning the motivation of

environmental conservation in hotels, shows that an indication as the majority of

guests reuses their towels is more effective than a simple call to environmental

protection. Furthermore, it is also more efficient when the comparison group is close to

the individual.

Alpizar et al. (2008) conducted an experiment dealing with donation for anational park in Costa Rica. They emphasize that informing on a typical donation raises

household contribution when the reference amount is high and increases the probability

of contribution when the reference amount is low. In the same way, using a Dutch

survey, Pieters et al. (1998) show that behavior and ability attributed to other

households have a significant positive effect on consumer pro-environmental behavior.

Finally, Janssen and Jager (2002) investigate the role of social norms in diffusion of


10/29

10

green products. They assume that either firms, being previous providers of green or

non-green products, keep the same product despite diffusion of green products or that

half of the non green-firms can decide to innovate in order to switch to green products.Consumers seek to satisfy both personal and social needs. Accordingly, consumer

choices rest on four behaviors: repetition of his/her usual consumption behavior,

deliberation on his/her need satisfaction according to each product, imitation of his/her

neighbor behavior and social comparison revealing the product its neighbors consume

the most. Deliberation, inducing rational choices, characterizes Homo economicus.

Repetition is a routine behavior, while imitation and social comparison both require an

attention paid to the reference group behavior. Hence, four types of cognitive processingcharacterize Homo psychologicus. Using simulations, Janssen and Jager (2002) show

that, when a tax is progressively introduced on the non-green product, the diffusion of

green products is faster for Homo economicus than for Homo psychologicus when firms

do not change their product designs. However, when non-green firms innovate in order

to respond to the increased demand for green products, the diffusion of green products

is faster for Homo psychologicus than for Homo economicus. Hence, green nudges

associated with fiscal incentives may favor diffusion of green products when firms

continuously adapt their product designs.

Beyond social norms, the bandwagon effect in green consumption can also be

explained by consumer consciousness that a collective action is more efficient than an

isolated action in order to reach a better environment. Indeed, 30% the 85% of

Europeans who claim to make an effort to protect the environment do not believe that

their efforts have an impact as long as others do not do the same (European

Commission, 2005). Accordingly, consumer considers his/her green consumption as

his/her own contribution to global effort to the benefit of the environment. The extent to

which he/she believes that his/her action make a difference in solving environmental

problem is the Perceived Consumer Effectiveness (Ellen et al., 1991). This argument

holds only if consumers are environmentally oriented and if they believe that

consumption and environmental damage are strongly linked.


11/29

11

3 The database and the econometric model

3.1 Data

The data used for our empirical analysis come from a survey investigating

French consumer perception of and purchase intentions for labeled fresh seafood

products. This survey is a part of a larger project on Sustainable Development of the

Artisanal fisheries in the Atlantic Area (PRESPO) coordinated by the RICEP. The

PRESPO project is structured in six activities, including the commercial optimization,

which focuses on the commercial potential of eco-labeling applied to artisanal fisheries

in the Atlantic area, as mechanism of adding value and commercial optimization.

The survey consisted of around fifty questions dealing with consumer behavior,

expectation and willingness to pay for a labeled fresh seafood product. The database

includes 91111 questionnaires completed in face-to-face interviews during April-June

2010 in neutral place in regard to seafood consumption. Face-to-face interviews allow

maximizing the number of respondents and avoiding poor completed questionnaire.

However, they may produce several biases. They are prone to social desirability biases,

which is the tendency of respondents to reply in the manner that will be viewed

favorably by others. However, the answers show that seafood quality has priority over

environmental criterion in seafood choice of respondents (see Table A.1). This allows us

to consider that the social desirability biases are limited.

The survey was divided in four parts: perception of professional fishing,

perception of seafood products, consumption and purchase of seafood products and

socio-economic characteristics (given in Table A.1 in the appendix). Beforehand, a pre-

survey have been conducted in face-to-face from fifty individuals in order to refine label

definitions. It resulted in definitions of three labels: a health label guaranteeing the

products do not contain toxic substances, an eco-label guaranteeing the products were

caught in an environmentally friendly way and a fair trade label guaranteeing that

production conditions meet minimum standards such as decent working conditions.

11Only 337 are exploitable in this context, due to a lake of answer concerning the willingness to pay morequestion and the missing values for the other variables.


12/29

12

In this survey, we asked to individuals: If you choose a labeled product (health,

eco or fair trade) would you be willing to pay more for the guarantees it provides?

('yes' , 'no', or 'do not know'). Only 655 persons answer to this question. Around 70% of

respondents are agreed to pay a premium for a labeled product. For this latter question,

descriptive statistics exhibit some differences according to the marital status (around

67% for singe persons vs around 73% for couple), to the education level (around 65%

for primary and secondary vs around 75% for tertiary), and to the financial satisfaction

(Would you say that your income allows you a gratifying consumption? -yes, no, or

do not know - around 65% when financial satisfaction is low versus around 85% when

financial satisfaction is high) but not by gender, age, and number of child(ren). We

asked also to the individuals: Do you think that people from your circle would agree to

buy eco-labeled seafood products? (yes, no, or do not know). Only 610 persons

answer to this question. Approximately 90% gave a positive answer. For this latter

question, the percentage of answers is very similar according to the gender, the marital

status, and the number of child(ren) but it differs for age (around 96% for less than 30

years vs 81% for 55 years and more), education level (around 85% for primary andsecondary vs around 95% for tertiary) and to the financial satisfaction (around 87%

when financial satisfaction is low versus around 92% when financial satisfaction is

high). People who answer to the question If you choose a labeled product (health, eco

or fair trade) would you be willing to pay more for the guarantees it provides? are not

necessary the same that those who answer to the question Do you think that people

from your circle would agree to buy eco-labeled seafood products? (both questions

have missing values). Thus, only 337 questionnaires are exploitable when all variablesare taken into account.

This question allows us to estimate the peer effect. Indeed, it is not possible to

ask directly whether the individual is influenced by others behavior, since a so direct

question would bring to an answering bias. Therefore, we assume that a positive answer

about circle behavior is a sign of a peer effect in green consumption. As we analyze the

labeling of seafood product, one might argue that the lack of 'salience and visibility' in


13/29

13

labeled food could limit peer effect (Janssen and Jager, 2002). However, in green-

fashion context, we can underline that labeled food consumption is a way to show his

green consumption without a lot of expenses. Besides, Welsch and Kuhling (2009) showthat the influence of the reference group consumption pattern is greater for organic food

than for electricity and solar panel.

In order to study the peer effect on the willingness to pay a premium for a

labeled seafood product we examine two kinds of characteristics. First, we consider the

socio-economic characteristics: the gender, the age, the familial situation (couple or

single), the household size (number of children), the education level (Primary,

Secondary and Tertiary) and the financial satisfaction.

Secondly, we consider purchase criteria on seafood product, this being common

in the literature (Wessells et al., 1999 ; Jaffry et al., 2004 ; Bernus et al., 2003). Several

criteria have been used: the attention paid to price, geographical origin, wild or farmed

origin, fishing technique, degree of seafood product exploitation, appearance, product

reputation, vendor advice and nutritive quality of the product (the exact question if :

when you buy seafood products, you pay attention to , the respondents have to

choose a point on a scale 0-10). In order to test the reliability of the answers, a

correlation matrix was calculated for all purchasing criteria. Apart from the price

variable, all the variables are positively related: this may be due to underlying factors

which could be revealed through a factor analysis12. Two factors appeared relevant: a

production process one, which encompasses four variables (the geographical origin,

wild versus farmed origin, the fishing technique, and the degree of exploitation of the

product) and a product characteristic one, which encompasses four variables too (the

appearance, the reputation of the product, vendor advice, and the nutritive quality of theproduct). Furthermore, we consider the person responsible of the shopping in the

households (Who does the shopping in your household? myself or others) as the

12 The Bartlett test of sphericity concludes that a factor analysis is relevant and the Kaiser-Meyer-Olkin measure of sampling adequacy increases when the price variable is omitted. To pick the rightnumber of factors, we used the Kaiser and the Cattell criteria, and Horns parallel analysis. For eachfactor, Cronbachs alpha statistic, which determines the internal consistency of items in a surveyinstruments to determine its reliability, was computed, equal to 0.70 for the first (process) factor and 0.64for the second (product). According to Nunnaly (1978), a score of 0.7 obtained on a substantial sample isan acceptably reliable coefficient but lower thresholds are sometimes used in the literature.


14/29

14

interview has been made outside any shopping place.

In order to work on the role played by the peer effect on the willingness to pay

more for a labeled seafood product, we first made a Probit model which shows thatnotably circle behavior has an positive and significant impact on the probability to have

a positive willingness to pay. Therefore, we can underline several factors influencing the

probability to have a positive WTP a premium for labeled seafood products.

3.2 Determinants of the willingness to pay

The issue of willingness to pay more for labeled seafood products is analyzed

through aProbitmodel. The probability to have a positive WTP is linked with severalexplanatory variables, like socio-economics features (age, gender, marital status,

presence of child(ren), and education), the purchase criteria usually take into account

when choosing fresh fish (product, process and price) and financial parameters

(financial satisfaction). As we have particular interest for peer effect influence on WTP,

this parameter will be included in all specifications.

The results are presented in Table 1 and we can see the importance of the peer

effect. Having peers in favor of eco-label increases the probability to be willing to paymore for label seafood products, this result is strongly significant and holds for any

specifications. In the first model, the socio-economics characteristics who significantly

influences the probability to have a positive WTP for labeled seafood products is the

presence of child(ren) in the household. This result is in line with previous results in the

literature on WTP for labeled products: households with children are more likely to

have a positive WTP.

In order to improve the estimation, we add purchase criteria characteristics to themodel 1. The purchase criteria are regrouped into two factors as previously underlined:

the process factor and the product factor, in addition to the price variable. Being in

charge of the shopping task is also taken into account. Peer effect is still strongly

significant as regards to the probability to have a positive WTP but that the presence of

child(ren) is no longer significant. Nonetheless, being attentive to the process factor

positively influences the probability to have a positive WTP while being attentive to the

price is negatively link with the WTP more for labeled seafood products. Thus, more an


15/29

15

individual is attentive to the fish production or fishing issues more he/she agrees on the

idea to pay a premium for labeled seafood products, but this does not hold when the

individual is attentive to the price, for which the reverse is observed. Both effects aresignificant. Being in charge of the grocery does not influence the WTP.

Finally, in model 3 we add the financial satisfaction which is positively linked to

the probability of having a positive WTP. The peer effect and being attentive to process

factors are still linked to the probability to have a positive WTP. Nevertheless, being

attentive to the price is no longer significant and having child(ren) becomes again

lightly significant. The influence of age becomes also lightly significant, old people

have a weaker probability to have a positive WTP than the young (the link between ageand the probability to have a positive WTP reaches its minimum at 52 years old), which

is in line with the previous literature (Loureiro and Lotade, 2005; Srinivasan and

Blomquist, 2009; Brcard et al., 2009; Salladarr et al., 2010).

Inasmuch as gender is often considered as an important determinant of green

consumption we decompose the model 3 into gender effect. The results show that men

and women are not identically influenced concerning the determinants of the WTP for

label seafood products. First, for men, only purchase criteria significantly influence the

positive WTP. Being attentive to the process factor increase the men probability to be

willing to pay more for label seafood products and being attentive to the price decrease

this probability. The peer effect less important for men than for previous scenario but

most important it is not significant. Concerning the women, the peer effect is strong and

significant. This result is in line with previous literature concerning the influence of

peer, generally women are more influence be others behavior than men. Beyond, age

negatively influence women WTP, and process factor, presence of children and financial

satisfaction positively influence the probability to be willing to pay more for labeled

seafood product.


16/29

16

Table 1 The determinants of pay more variable

Model (1) Model (2) Model (3)Gender

Men WomenCircle effect*

No Ref. Ref. Ref. Ref. Ref.Yes 0.846*** 0.976*** 0.921*** 0.300 1.132***

(3.92) (3.58) (3.31) (0.50) (3.41)Gender

Men Ref. Ref. Ref.Women -0.003 -0.044 -0.072

(0.02) (0.26) (0.39)Age 0.002 -0.059 -0.069* 0.028 -0.114**

(0.06) (1.48) (1.67) (0.38) (2.13)Agesquared (/100) 0.012 0.068 0.075* -0.022 0.117**

(0.34) (1.55) (1.66) (0.27) (2.00)Marital status

Single Ref. Ref. Ref. Ref. Ref.Couple 0.024 0.062 -0.005 -0.201 -0.087(0.16) (0.32) (0,02) (0.55) (0.31)

Child(ren) at homeNo Ref. Ref. Ref. Ref. Ref.Yes 0.349*** 0.307 0.355* 0.274 0.455*

(2.14) (1.55) (1.69) (0.75) (1.67)Education

Primary/Secondary Ref. Ref. Ref. Ref. Ref.Tertiary 0.138 0.005 -0.035 0.362 -0.180

(0.95) (0.03) (0.18) (1.03) (0.72)Seafood characteristics

Process 0.191*** 0.179*** 0.249** 0.173***(4.13) (3.60) (2.42) (2.90)

Product -0.038 -0.059 -0.127 -0.037(0.77) (1.06) (1.24) (0.52)Price -0.121*** -0.075 -0.167* -0.066

(2.60) (1.50) (1.74) (1.04)Do the shopping

No Ref. Ref. Ref. Ref.Yes 0.257 0.411 -0.002 0.707

(0.80) (1.26) (0.00) (1.60)Financial Satisfaction

No Ref. Ref. Ref.Yes 0.650*** 0.172 0.822***

(3.29) (0.45) (3.35)

Constant -0.582 1.243 0.832 0.699 1.068

(0.94) (1.34) (0.85) (0.39 (0.86)

Number of observations 480 374 337 121 216

Pseudo-R2 0.05 0.12 0.15 0.14 0.20Log-Likelihood -246.65 -164.25 -141.58 -52.64 -89.93

Note: Absolute value of t statistics are in parentheses, * significant at 10%; ** significant at 5%; ***significant at 1%.Source: RICEP (2010), data from survey PRESPO.

Overall, peer effect have a strong and almost always significant influence on

WTP, expect for men. Nevertheless, this effect may be overestimated by aProbitmodel,


17/29

17

as the strong coefficient of peer effect in the previous models can confirmed it. Thus, in

order to have a better estimation of this particular effect on the WTP for labeled seafood

product we use a propensity score-matching model, in the line with Rosenbaum andRubin (1983).

4 The propensity score matching estimates

In order to evaluate the robustness of peer effect influence that are based on

Probitregression, we estimate propensity score matching models13 As we study the role

played by the peer effect on the probability to be willing to pay more for labeled

seafood products, we highlight the difference in probability to be willing to pay more

between the two kinds of individual: the one with peers interested in labellisation, and

the one without. The key idea of propensity score matching is to find among individual

with no peers in favor of label those individuals who are similar to the individual with

pro-label peers in all relevant pretreatment characteristics (set of explicative variable).

That being done, the outcome difference between the control group and the treated

group can be attributed to the neighborhood effect. Nevertheless, in order to estimate

this average treatment effect, two assumptions need to be made: the unconfoundedness

assumption14 and the overlap assumption. First the unconfoundedness assumption, i.e.

that differences in outcomes between treated and control with the same pretreatments

characteristic are attributable to the treatments. This implies that all variable that

influence treatment assignment as well as potential outcomes have to be observed. The

second assumption concerns the perfect predictability of the treatment (overlap or

common support assumption). Corresponding assumptions apply when the treatment is

becoming employed.

Based on the previousProbitestimation, we use a propensity score with nearest-

neighbour matching and kernel (Gaussian and Epanechnikov kernel) methods when

13See Caliendo and Kopeinig (2008) for a survey on the propensity score matching models, andBockerman and Ilmakunnas (2009) or Heckman et al. (1997) for example.

14This assumption is know as the selection on observables (Heckman and Robb, 1985) or conditionalindependence assumption (Lechner, 1999) too.


18/29

18

calculating the average treatment effect on the treated (ATT)15. The ATT allow us to

highlight the effect of the treatment for the individual who are under the treatments, i.e.

we look at the estimation of the peer influence in case of pro-label peers on theprobability to be willing to pay more for labeled seafood products. The covariate used to

estimate the ATT are similar to the variable used in the Probitspecification. Model 1 is

composed only of socio-economics characteristics, in model 2 we take purchase criteria

in addition and model 3 take also financial dimensions. The matrix of correlation

between all the covariates, the outcome and the treatment show no strong correlation.

We choose to use propensity score in the common support condition16. The common

support condition improve the quality of the matches used to estimated ATT despite italso reduce the sample and so on high quality matches may be lost at the boundaries of

the common support17 If the proportion of lost individuals is small this presents few

problems (Bryson et al., 2002). In our case, for each model only few observations are

drop by the common support definition, thus we assume there is no failure of the

common support.

For the models 1, 2 and 3, propensity score is performed using the region of

common support for the propensity scores, which included 551, 421 and 372 individuals

with pro-label peers, respectively, and 59, 45 and 40 control cases respectively. The

validity of the matching is tested through the balancing property18 for all the

specifications. For all the variables matching succeeds in making the means of the

covariates close to each other for the treated and controls in each of the 5 blocks19.

First column of the Table 2 reported the estimated treatment of the treated for the

model 1. When nearest-neighbor (NN) matching is used the average treatment of having

15In our analysis we use the pscore programs written by Becker and Ichino (2002)

16We follow the Becker-Ichino (2002) definition of commun support: all treated with a propensityscore superior to the untreated maximum propensity score and all the untreated with a propensity scoreinferior to the treated minimum propensity score are excluded in order to obtain the common support.

17See Lechner (2001) for a discussion on the interest of the common support conditions

18See Ichino and Becker (2002) for details

19This number of blocks ensures that the mean propensity score is not different for treated and controls in eachblocks


19/29

19

pro-labels peer is 0.443, which is statistically significant at the 1% level. To check the

robustness of the result, kernel matching is also used. With Gaussian and Epanechnikov

with a high bandwidth parameters (0.1), Kernel results show a lower impact of peereffect, around 0.32 (significant at the 1% level) If we use a lower bandwidth parameters

(0.01)20 increase the strength of the peer effect, with an impact of 0.488 (significant at

the 1% level). In all case, despite a range of values between 0.32 and 0.488, having pro-

label peers positively influence the probability to be willing to pay more for labeled

seafood products. The result is constant across the others specifications since models 2

and 3 lead to the same conclusion, despite some ATT values variations.

Indeed, when we take into account the purchase criteria of an individual as

further matching variables, the impact of peer effect is lightly lower in average but still

significant The nearest-neighbor matching still shows up high values of average

treatment on the treated such as the Epanechnikov Kernel with low bandwidth

parameters. Those matching estimate the average treatment of having pro-label peers

around 0.354 and 0.339 respectively. Meanwhile, the matching with Gaussian and

Epanechnikov Kernel with a high bandwidth values estimate the average treatment of

having pro-label peers around 0.29. In all case, having pro-label peers positively

influence the probability to be willing to pay more for labeled seafood products with a

significant at the 1% level. Concerning the third model, the result is lower for the NN

matching, around 0.28 point, and quite similar for the others specifications (between

0.33 and 0.28). All results are significants at the 1% level.

20 Using a small values for the bandwidth parameters decrease the bias but increase the variance. SeeCaliendo and Kopeinig (2008) for more details.


20/29

20

Table 2 Average treatment on the treated for having pro-ecolabel peers

Outcome

Model (1) Model (2) Model (4)

Model (4)

Gender

Matching method Men Women

Nearest Neighbor 0.443*** 0.354*** 0.284** 0.394 0.368***(0.072) (0.117) (0.117) (0.282) (0.108)

Kernel Gaussien 0.320*** 0.293*** 0.283*** 0.273 0.291**(0.086) (0.107) (0.110) (0.253) (0.143)

Kernel Epan 0,1 0.324*** 0.288*** 0.290** 0.409 0.280*(0.090) (0.105) (0.116) (0.285) (0.149)

Kernel Epan 0,01 0.488*** 0.339*** 0.330*** 0.663 0.327**(0.072) (0.117) (0.107) (0.489) (0.131)

Note: Bootstrap standard errors (1000 replications) in parentheses, * significant at 10%; ** significant at 5%; *** significant at 1%.Source: RICEP (2010), data from survey PRESPO.

The last model, the model 3 is decomposed under gender restriction. The

propensity score under the common support included 228 women with pro-label peers

and 24 control cases, and 144 men with pro-label peers and 16 control cases. The

validity of the matching is tested through the balancing property. For all the variables

matching succeeds in making the means of the covariates close to each other for the

treated and controls in each of the 5 block. Concerning the gender specification, we can

underline that none of the men results are significant and the distance between the

different results is strong (between 0.66 and 0.27). Nevertheless, the results for women

show a strong impact of having pro-label peers on the probability to be willing to pay

more for labeled seafood products. According to the NN matching the average treatment

effect of having pro-label peers increases of 0.368 points the probability to be willing to

pay more for labeled seafood products (significant at the 1% level). The Epanechnikov

Kernel 0.01 gives some similar result, with an average treatments effect around 0.32,

while the Gaussian and Epanechnikov Kernel 0.1 give an average treatments effect

around 0.29. Those results are in line with the previous literature on peer effect under

gender restriction (Carlsson et al., 2008), showing that women are significantly

influence by peer behavior, when men are not. Nonetheless, the non significant effect

for men could be link with the little number of observations that we can used in case of

men matching.


21/29

21

5 Conclusion

The green consumption is a way for consumers to show that they are

increasingly concerned with the impacts of the products they consume on the

environment. The socio-economics characteristics, but also the moral motivation and

social interaction influence pro-environmental behavior. In this paper we have attempt

to underline the role played by the social interaction on the probability to have a

positive WTP in case of labeled seafood products. For this propose we us a Probit

estimation to emphasize the determinant of the probability to have a positive WTP and a

Propensity score-matching to highlight the effect of having pro-eco-label peers.

The Probitestimation underlines the determinants of the probability to have a

positive willingness to pay a premium for labeled seafood products. In all specifications

included purchase criteria being attentive to the process factor, i.e. the geographical

origin, wild or farmed origin, fishing technique and degree of seafood product

exploitation, increase the probability to have a positive willingness to pay a premium

for labeled seafood products. This result is in line with the previous literature who

underlines the importance of purchase criteria in order to achieve more sustainable fish

consumption ( Wessells et al., 1999; Jaffry et al., 2004; Brcard et al., 2009; Salladarr

et al., 2010). It is worth noting that financial satisfaction increases the probability to

have a positive WTP. This result highlight the budgetary consideration that consumer

can face when choosing more ecological products. Nonetheless, despite a likely

overestimation, the importance of the peer effect on those models conducts us to a first

idea on the importance of circle behaviors on consumer probability to have a positive

WTP. This peers effect is strongly significant in all specification, except for men.

Using the propensity score matching allow us to underline more precisely therole played by peers behavior on the probability to be willing to pay more for labeled

seafood products. In all the specification used, having pro-eco-label peers increases

significantly the probability to be more willing to pay for labeled seafood products, the

range of the values being between 0.488 and 0.283. This phenomenon is confirmed for

women, who are strongly influenced by having pro-eco-label peers. Women with pro-

eco-label peers increase their probability to have a positive willingness between 0.368


22/29

22

and 0.28. Meanwhile, men do not seems to be influenced by the peers behavior, as none

of the estimates is significant. This result is in line with the result of Carlsson et al.

(2008) who find women more willing to buy ecological coffee when they were told thatecological coffee was a common choice across the population.

Overall, this work allows us to underline than beyond socio-economic

characteristics, the social interactions have an influence on consumer behavior. Thus,

based on this influence, consumer behavior can then be oriented towards a more

sustainable fish consumption.


23/29

23

Appendix

Table A.1 Descriptives statistics

Descriptive statistics Pay more Gender

Obs Mean Min MaxYes

(N=271)

No

(N=66)

Men

(N=121)

Women

(N=216)

Circle effect*

Yes 610 0.903 0 1 0.946 0.800 0.941 0.903No 610 0.097 0 1 0.054 0.200 0.059 0.097

Pay more

Yes 655 0.705 0 1 0.816 0.788

No 655 0.295 0 1 0.184 0.212

Gender

Men 911 0.561 0 1 0.374 0.333

Women 911 0.439 0 1 0.626 0.667AgeLess than 30 years 911 0.127 0 1 0.188 0.240 0.206 0.19530-39 years 911 0.453 0 1 0.273 0.267 0.338 0.23340-54 years 911 0.333 0 1 0.340 0.333 0.316 0.35255 years and more 911 0.087 0 1 0.199 0.160 0.140 0.220

Marital status

Couple 911 0.648 0 1 0.653 0.627 0.669 0.636Single 911 0.352 0 1 0.347 0.373 0.331 0.364

Child(ren) at home

Yes 911 0.420 0 1 0.505 0.440 0.537 0.460No 911 0.580 0 1 0.495 0.560 0.463 0.540

Education

Primary and Secondary 911 0.406 0 1 0.299 0.333 0.294 0.314Tertiary 911 0.594 0 1 0.701 0.667 0.706 0.686

Seafood characteristics

Appearance 701 7.838 0 10 7.949 8.045 8.012 7.942Price 701 7.832 0 10 7.675 8.295 7.537 7.942Nutritional Quality 701 5.391 0 10 5.834 5.462 5.126 6.122Wild or farmed origin 701 4.429 0 10 5.232 4.000 5.158 4.899Geographic origin 701 3.767 0 10 4.411 3.151 4.300 4.090Vendor advice 701 3.549 0 10 3.976 3.030 3.528 3.942Product reputation/prestige 701 3.153 0 10 3.560 3.439 3.382 3.624Resource availability 701 3.123 0 10 3.648 2.636 3.435 3.461Fishing Technique 701 2.249 0 10 2.553 1.819 2.321 2.403

Financial satisfaction

Yes 800 0.725 0 1 0.806 0.560 0.780 0.746No 800 0.275 0 1 0.194 0.440 0.220 0.254

Do the shopping

Yes 907 0.867 0 1 0.946 0.893 0.911 0.949No 907 0.133 0 1 0.054 0.107 0.089 0.051

* The exact question is People from circle would agreed to buy eco-labeled seafood products.Source: RICEP (2010), data from survey PRESPO


24/29

24

References:

Becker, S.O. & Ichino, A., 2002. Estimation of average treatment effects based on

propensity scores. The Stata Journal, 2(4), pp.358377.

Bernus, A., Olaizola, A. & Corcoran, K., 2003. Labelling information demanded by

European consumers and relationships with purchasing motives, quality and safety

of meat.Meat Science, 65(3), pp.10951106.

Bjrner, T., 2004. Environmental labeling and consumers choice - an empirical analysis

of the effect of the Nordic Swan. Journal of Environmental Economics and

Management, 47(3), pp.411-434.

Blend, J.R. & van Ravenswaay, E.O., 1999. Measuring consumer demand for

ecolabeled apples. American Journal of Agricultural Economics, 81(5), pp.1072

1077.

Brock, W.A. & Durlauf, S.N., 2001. Discrete choice with social interactions. The

Review of Economic Studies, 68(2), p.235.

Brcard, D. et al., 2009. Determinants of demand for green products: An application to

eco-label demand for fish in Europe.Ecological Economics, 69(1), pp.115-125.

Caliendo, M. & Kopeinig, S., 2008. Some Practical Guidance for the Implementation of

Propensity Score Matching. Journal of Economic Surveys, 22(1), pp.31-72.

Carlsson, F. & Johansson-Stenman, O., 2000. Willingness to pay for improved air

quality in Sweden.Applied Economics, 32(6), pp.661-669.

Carlsson, F., Garca, J.H. & Lfgren, ., 2008. Conformity and the demand for

environmental goods. , 2465(286), pp.1-20.


25/29

25

Centre dAnalyse Stratgique (2011), Green nudges: new incentives for ecological

behavior. Note 216.

Clark, C.F., Kotchen, M.J. & Moore, M.R., 2003. Internal and external influences on

pro-environmental behavior: Participation in a green electricity program. Journal

of Environmental Psychology, 23(3), pp.237246.

Cowan, R., Cowan, W. & Swann, P., 1997. A model of demand with interactions among

consumers.International Journal of Industrial Organization, 15(6), pp.711732.

Dearden, L. et al., 2005. Education subsidies and school drop-out rates. The Institute for

fiscal Studies.

Deere, C., 1999. Eco-labeling and sustainable fisheries. The World Conservation Union

(UICN) and the Food and Agricultural Organization of the United Nations (FAO),

Washington and Rome.

Ek, K. & Soderholm, P., 2008. Norms and economic motivation in the Swedish greenelectricity market.Ecological Economics, 68(1-2), pp.169-182.

European Commission (2005) Attitudes of Europeans citizens towards the environment.

Special Eurobarometer 217.

European Commission (2009). Europeans attitudes towards the issue of sustainable

consumption and production. Flash Eurobarometer 256.

European Commission (2011) Attitudes of Europeans citizens towards the environment.

Special Eurobarometer 365.

Frey, B.S. & Oberholzer-Gee, F., 1997. The cost of price incentives: An empirical

analysis of motivation crowding-out. The American economic review, 87(4),

pp.746755.


26/29

26

Frey, Bruno and Stephen Meier (2004), Social Comparisons and Pro-Social Behavior:

Testing Conditional Cooperation in a Field Experiment, American Economic

Review, 94, pp: 1717 1722.

Frey, B.S. & Stutzer, A., 2006. Environmental morale and motivation. Research in

Economics, (288).

Goldstein N.J. et al. (2008), A room with a viewpoint: Using social norms to motivate

environmental conservation in hotels, Journal of Consumer Research, vol. 35(3).

Granovetter, M. & Soong, R., 1986. Threshold models of interpersonal effects in

consumer demand 1. Journal of Economic Behavior & Organization, 7(1), pp.83

99.

Harding, D.J., 2003. Counterfactual models of neighborhood effects: The effect of

neighborhood poverty on dropping out and teenage pregnancy. American Journal

of Sociology, 109(3), pp.676719.

Heckman, J.J. & Robb, R.J., 1985. ALTERNATIVE METHODS FOR EVALUATINGTHE IMPACT OF INTERVENTIONS An Overview. Journal of Econometrics, 30,

pp.239-267.

Heckman, J.J., Ichimura, H. & Tood, P.E., 1997. Matching as a Econometric Evaluation

Estimator: Evidence from Evaluating a job Training Programme. The Review of

Economic Studies, 64(4), pp.605-654.

Jacquet, J., Pauly, D., 2010. Seafood stewardship in crisis. Nature 467, 28-29.

Jaffry, S. et al., 2004. Consumer choices for quality and sustainability labelled seafood

products in the UK.Food Policy, 29(3), pp.215228.

Janssen, M. a & Jager, W., 2002. Stimulating diffusion of green products. Journal of

Evolutionary Economics, 12, pp.283-306.


27/29

27

Johnston, R.J. & Roheim, C.A., 2006. A battle of taste and environmental convictions

for ecolabeled seafood: a contingent ranking experiment. Journal of Agricultural

and Resource Economics, 31(2), pp.283-300.

Johnston, R.J. et al., 2000.Measuring consumer preferences for ecolabeled seafood: an

international comparison.

Kotchen, M.J. & Moore, M.R., 2007. Private provision of environmental public goods:

Household participation in green-electricity programs. Journal of Environmental

Economics and Management, 53(1), pp.116.

Lechner, M., 2001. A note on the common support problem in applied evaluation

studies.Discussion paper no.2001-01.

Lechner, M., 1999. Earnings and employment effects of continuous off-the-job training

in East Germany after unification. Journal of Business & Economic Statistics,

17(1), pp.74-90.

Leibenstein H (1950) Bandwagon, Snob, and Veblen Effects in the Theory of Con-

sumers Demand, Quart. J. Econ. 64:183-207.

Loureiro, M., 2003. Rethinking new wines: implications of local and environmentally

friendly labels.Food Policy, 28(5-6), pp.547-560.

Loureiro, M. & Lotade, J, 2005. Do fair trade and eco-labels in coffee wake up the

consumer conscience?Ecological Economics, 53(1), pp.129138.

Loureiro, M., McCluskey, J.J. & Mittelhammer, R.C., 2002. Will Consumers Pay a

Premium for Eco-labeled Apples? Journal of Consumer Affairs, 36(2), pp.203

219.

Mahenc, P., (2006) Signaling the environmental performance of polluting products to

green consumers. International Journal of Industrial Organization, 26(1), pp.59-


28/29

28

68.

Pieters, R. et al., 1998. Consumers attributions of proenvironmental behavior,motivation, and ability to self and others. Journal of Public Policy & Marketing,

17(2), pp.215225.

Roheim, C.A. et al., 2004. Consumer Preferences for Ecolabeled Seafood: Results of a

Connecticut Survey.

Rosenbaum, P.R. & Rubin, D.B., 1983. The central role of the propensity score in

observational studies for causal effects. Biometrika, 70(1), p.41.

Salladarr, F. et al., 2010. The Demand for Seafood Eco-Labels in France. Journal of

Agricultural & Food Industrial Organization, 8(1).

Schultz P.W. (1998), Changing behavior with normative feedback interventions: A field

experiment on curbside recycling, Basic and Applied Social Psychology, vol.

21(1).

Schultz P.W. et al. (2007), The constructive, destructive, and reconstructive power of

social norms, Psychological Science, vol. 18.

Srinivasan, A.K. & Blomquist, G.C., 2009. Ecolabeled paper towels: consumer

valuation and expenditure analysis. Journal of environmental management, 90(1),

pp.314-20.

Steinberg L. with Brown, B. and Dornbusch, S. (1996) Beyond the Classroom (New

York: Simon and Schuster).

Teisl, M.F., Rubin, J. & Noblet, C.L., 2008. Non-dirty dancing? Interactions between

eco-labels and consumers. Journal of Economic Psychology, 29(2), pp.140159.

Torgler, B. & Garcia-Valias, M.A., 2007. The determinants of individuals attitudes

towards preventing environmental damage. Ecological economics, 63(2-3),


29/29

29

pp.536552.

Veblen T (1899) The Theory of the Leisure Class. An economic Study of Institutions.

Macmillan, London.

Ward T., Phillips B. (2008), Seafood ecolabelling. Principles and practice. Wiley-Black-

well.

Welsch, H. & Kuhling, J., 2009. Determinants of pro-environmental consumption: The

role of reference groups and routine behavior. Ecological Economics, 69(1),

pp.166176.

Wessells, C.R., Donath, H. & Johnston, R.J., 1999. US consumer preferences for

ecolabeled seafood: results of a consumer survey. Kingston: University of Rhode

Island, Department of Environmental and Natural Resource Economics.

Unpublished Report.

Young, W. et al., 2010. Sustainable consumption: green consumer behaviour when

purchasing products. Sustainable Development, 18(1), pp.2031.

gender on green consumption

Documents