Amsterdam School of Economics
THE INFLUENCE OF THE GREEN INFORMATION SYSTEM AND
SOCIAL NORMS ON CUSTOMER’S DECISION TO CHOOSE THE
ECO-LABELED PRODUCT
Author: Aida Žukauskaitė, 10828699Supervisor: Theo Offerman
Master’s in Economics
Behavioral Economics & Game Theory
April, 2016
Abstract
Regarding the increasing interest into fostering a sustainable consumption, the experiment was implemented with the aim to find the most effective technique in swaying the decision making towards the eco labeled products. It was checked if the green information systems can affect the behavior more effectively than social norms. The online shopping for the home cleaning detergents was modeled in four treatment conditions. Results did not contradict the previous findings that the social norms are one of the best techniques in nurturing sustainable consumption. Although the information given by the green mobile application had a significant effect on the decisions to choose eco labeled products. Results differ for male and female participants. Social norm was more effective in female behavior change while the application information improved the behavior for male participant more than social norm. This leads to the conclusion that green mobile applications are the effective technique for shaping people’s behavior.
Content
Abstract ... 2
1 Introduction ... 1
2 Related literature ... 3
2.1 Experiments related with pro-environmental behavior and social norms ... 4
2.2 Green IS studies ... 5
2.3 Social norm and mobile application ... 7
3 Methodology ... 7
4 Results ... 13
4.1 Main statistics ... 13
4.2 General results ... 13
4.3 Analysis of perceptions towards environmental problems ... 17
4.4 ‘More information’ effect on choices ... 20
4.5 Regression analysis... 23
4.6 Limitations ... 26
5 Discussion and conclusion ... 27
6 References ... 29
7 Apendix ... 32
7.1 Instructions ... 32
7.1.1 Social norm treatment ... 32
1
1 Introduction
Global warming and anthropogenic impact on the environment are common concerns nowadays. Parallel to growing environmental problems, the use of technologies is conquering human lives progressively. Gaining benefit from the latter, technologies should be applied to invoke conscious and sustainable behavior. The main focus of my research is to test whether the information, provided by the green information system (IS), creates the added-value towards the eco-shopping and works as effectively as social norms.
In the age of technology marketers, business and individuals widely use the advantages provided by smart technologies, mobile applications and internet. There should be more attention brought to the possibilities to shape the pro-environmental behavior of individuals with the help of technology as well. Fortunately, regarding the escalation of the climate change, there is a growing number of green information technologies and systems (green IT/IS). There are various types of it – some help in saving the energy, by regulating the electricity use, other optimizes the resource use by managing it more effectively or simply by providing additional information for the customers (Loeser et al., 2011).
In this study the main attention is on mobile applications which provide information about the sustainability of the products offered in the supermarket. There is a number of such applications, to mention few – Question Mark, EWG Food Scores, Orange Harp, Open Label and Good Guide. After scanning a barcode the applications provide the health related issues, environmental information, tell if a product is fair-trade, its ingredients etc. Usually the applications have their own ranking system for the products or use other common methodology to evaluate the sustainability of a product. Such applications can be effective in shaping the behavior of society because more and more people are engaged into using smart phones and mobile applications in their daily lives. Research (SessionM, 2015) has shown that more than 90% of Americans stated that they use smartphones when retail shopping. 53.8% use it for comparing prices, 48.4% are searching for the product information and 42% are checking online reviews. So the mobile tagging, which is providing additional information after scanning a barcode of the product, is seen as a promising tool for marketing as well as for engaging more people into the mindful consumption (McKenzie-Mohr & Schultz, 2014).
2 The information that is provided by the green application can stimulate the customer to move her attention towards sustainable brands, helping the brands to grow and consumers to behave pro-environmentally. Though there is a lack of scientific research on how to employ digital tools to affect the buying process.
However a number of studies of Behavioral Economics analyzed people’s behavior and it plays a big role in enhancing people to consume wisely, to behave environmentally friendly and to be more conscious about their decisions. Despite the great number of behavioral techniques, social norms are found to be one of the best to influence people to behave pro-environmentally (Bolsen et al. , 2014; Smith et al., 2012; Nolan et al., 2008; Dolan & Metcalfe, 2013).
My research focuses on both – social norms and the green IS appliance into fostering individuals to consume environmentally friendly products.
Motivation of choosing this topic is the focus on sustainable development by mostly all future strategies of the main world’s organizations like the European Union, the United Nations, the European Comision, the Organisation for Economic Co-operation and Development (OECD), the World Economic Forum and governments. A lot of exertion is being put by the latter into fostering the environmentally friendly consumer. The private sector and organizations also initiate green movements and ecological product design to help sustainable consumption to become more present in the daily routines and people are becoming more environmentally concerned. As already mentioned, it is still not clear however what behavior techniques make the biggest changes in consumption. Recently the most common programs to nurture sustainability are based on information-intensive campaigns (McKenzie-Mohr & Schultz, 2014; Schultz & Wesley, 2014). However it is seen that they are not as efficient as expected. Certainly, Hughner et al. (2007) had shown that regardless of the facts that 46-67% of population have positive attitudes towards organic food, actual ecological purchase behavior forms only 4-10% of all purchases. Rresearch by Pedersen & Neergaard (2006) found that only 50% of those who view environmentally friendly clothes as important, actually buy the environmentally friendly clothes themselves. These facts suggest that there is a need for other behavior changing techniques. Additionally there is criticism with respect to the traditional information-based campaigns pointing out that knowledge and understanding of the environmental issues increase often do not lead to a change in behavior (Linden, 2015).
3 The experiment was implemented by testing the choices during the shopping of detergents. This type of products was chosen because they cause damage not only for the consumers themselves but also for nature and society in terms of pollution. The participants of the study were expected to think more about the environment while buying such kind of products, compared to buying food, drinks or cosmetics.
The results of the experiment show that social norms and application information affects the behavior in a positive way. The results are inconsistent for different genders – male are more affected by the application information while female participants respond more to social norm information.
The remainder of the thesis is structured as follows: in the next section I discuss the related literature and the experiments that have been implemented. Then follows the methodology of the experiment where the design and setup is described in detail. Next, I provide the results and the analysis of the differences between the treatments which is followed by the discussion, the limitations and the conclusion.
2 Related literature
With the environmental problems being a big issue nowadays there is a considerable number of studies made to clarify how to convince people to consume consciously and to engage them into the long term sustainable consumption (McKenzie-Mohr & Schultz, 2014). There is a great focus on the role of social norms in behavior design (Goldstein et al., 2008; Schultz et al., 2007; Smith, et al., 2012) and particularly on designing pro-environmental behavior. However there are contradicting findings regarding social norms. The studies try to understand the characteristics and behavior intentions of sustainable consumer. After interviewing people about sustainable consumption, (Young et al., 2010) found out, that the main barriers for people not to buy environmentally harmless products are the lack of time and information available about the companies. They also found that customers have to trust the information source and labels offering shortcuts to choosing greener products, supporting the findings of Vermeir & Verbeke (2006), Sammer & Wüstenhagen (2006), Solér (1996). That would be a reasonable argument to promote the use of mobile applications, which provide the structured third party information customers want to know in a structured way.
4 This section is structured as follows. Firstly the recent studies that are related with the social norms and forming sustainable consumption are discussed. Secondly the studies about the green IS and mobile use are presented. At the end of the section I discuss the relevance of the literature for the research I did.
2.1 Experiments related with pro-environmental behavior and social norms
Social norm is the group-based standards or rules regarding appropriate attitudes and behaviors. There are two types of them: descriptive norms and injunctive norms. Injunctive norms involve perceptions of which behaviors are typically approved or disapproved. Descriptive norms involve perceptions of what most people do (Cialdini et al., 1991). Descriptive norms are used in this research, because they are more commonly used and found to be more effective in the number of researches (Schultz et al., 2008; Bolsen et al., 2014). The experiments where injunctive norm is as powerful as descriptive (Schultz et al., 2007) are present however as well. The supportive fact for the use of descriptive norm is that individuals tend to cooperate on the collective action problem only when they believe others to cooperate on it (Bolsen et al., 2014).
Regarding the role of social norms for the pro-environmental behavior stimulation, there are a lot of studies made, to test the effectiveness in different markets and types of consumption (Bolsen et al. , 2014; Cialdini et al., 1991; Goldstein et al., 2008; Linden, 2015; Nolanet al., 2008; Schultz et al., 2007; Smith, et al., 2012; van der Linden, 2015). Nolan et al. (2008) made experiments in the field regarding people’s beliefs about what influences their energy consumption. The participants received five experimental messages: descriptive norm, self-interest, environmental, social responsibility related, or information-only. The results show that the descriptive normative message had the biggest influence in reducing the energy consumption. Social norms were found to be more effective than an industry standard messages in the famous experiment at the hotel where it was proposed to guests to reuse the towels (Goldstein et al., 2008). Here the provincial descriptive norms were used, activating the social norm in the very small community - hotel room guests, so people unconsciously were effected by close similarity to other guests. Bolsen et al. (2014) research results in the same conclusions, that social norms stating supportive behavior of others had more influence than science-based communications. Although their results are not really consistent, because in one experiment they found little support for the norm based
5 behavioral intentions to act pro-environmentally while in the other the effect was stronger. The influence of science based information had little evidence at the experiment. Based on not consistent results because of partisan people beliefs towards global warming, the real effect of social norms and science based information should be tested in the wider experimental setting. Overall these studies suggest that social norm is the effective technique to engage people in sustainable consumption.
However, it is known that social norms can have a boomerang effect and work in the opposite way than expected. This is because individuals tend to overestimate their behavior compared to the given social norm and the norm can work in two ways – those whose behavior is below the average can improve it, but those performing behavior above the average can be demotivated and start to behave in an undesirable way (Schultz et al., 2007).
A recent study about water bottle consumption intentions (Linden, 2015) showed that people are most keen to change their behavior when they get information about the water consumption combined with social norm. In this case the information about the bottled water production makes more influence than the social norm alone. The replication of the towel reuse in the hotel room experiment resulted in showing that social norm improved the pro-environmental behavior although it was not more effective than the standard message which is in line with few more replications (Bohner & Schlüter, 2014).
Taking into account the latter studies, it cannot be stated that the descriptive norm is the most reliable behavior design technique and that is why I can question if the mobile application information has the bigger impact on customer behavior.
2.2 Green IS studies
The importance of the green IS has been increasing during the past years, because there are a lot of opportunities to use IS in sustainable development, although there is not so much research done in this field (Mohrenfels & Klapper, 2012; Loeser et al., 2011; Jenkin et al., 2011). The term green IS refers to the intelligent use of information systems (ISs) to facilitate sustainable behavior within organizations and society and thereby establishes how the transformative power of the ISs can help to create an ecologically sustainable society (Watson et al., 2010).
Regarding the green IS and their effect on people’s behavior, Mohrenfels & Klapper (2012) ran laboratory and field experiments to test the scale of the influence of mobile information,
6 transmitted via the extended packaging (the ability to get more information about the product after scanning the barcode), on the perception and consumption of green brands. They found that the usage of the extended packaging can positively influence the consumer choices to buy the green brand product and increase the willingness to pay (WTP). The down side of their research is that they gave mobile phones with the designed application to the participants and asked to use it, while in the real life people choose themselves to use such technology or not and sometimes disregard it because of a lack of time or interest. Atkinson (2013) tested the eagerness to use QR codes1 while shopping for sustainable products. QR
code is the commonly used technique to provide additional information about the product or service for the consumer, with a purpose to ease the information search on the smartphone. Atkinson (2013) tried to clarify the conditions when people do not trust eco labels the most and are eager to get more information. His results indicate that consumers use the QR code scanning to check the credibility of the eco labels more often when they are not convenient with the government trust, have intention for boycotting and are inclined with market mavenism. The study indicates that people are interested in having more information before making the purchase decision. Barnes (2002) claims that extensive packaging is particularly effective at providing timely information about the product, taken into account that through the QR code scanning one can reach customers with relevant information at the time of shopping. This suggests that the QR code online content could be used to shape the behavior of consumers towards the sustainable consumption.
The studies regarding the mobile application use tested the perceptions and beliefs about the brand, price, willingness to pay, but there is no study that tests the real intentions to use the applications voluntarily. Instead participants were asked to use the barcode scanning or were given the information by explicitly using the applications. My study is novel in a way that participants are concentrated on choosing the products in the experiment and they are not forced to see the information from the application – it is their own choice. So the willingness to use the application for scanning the bar code is controlled as well. Besides, the real
1 QR code (abbreviated from Quick Response Code) is two-dimensional bar code that can be converted to content, including URLs, phone numbers and text, when scanned by a camera-enabled smartphone with a QR reader, invented by Denso Wave in Japan in 1994 originally to track parts in car manufacturing (Atkinson, 2013).
7 behavior, not the perceptions towards the intentions is tested, because the experiment is designed to reflect the real shopping process.
Using the mobile application to get the information about products could be perceived from the other point of view as well – it can be seen as pull advertising. Pull advertising is defined as the communication sent by the advertisers upon request of the consumer (Atkinson, 2013; Barnes, 2002). It refers mostly to the invitations to use QR scanning to get more information about the product, promotions or coupons at the presence of the shopping. The similarity between the pull advertising and the green application can be seen in terms of the message sent by the advertiser and the message given by the mobile application. In the case of the mobile application, the application owners could be seen as the sustainable consumption advertisers. Pull advertising is seen as more relevant and useful than usual push advertising because the information from pull advertising is searched and chosen to know by the consumer, it is less irritating and perceived as more convenient (Atkinson, 2013). This adds to the credibility of the mobile application effectiveness in shaping the behavior.
Summing up the literature related with the mobile application use there is a proof that the green IS can be used to sway people to purchase greener products.
2.3 Social norm and mobile application
All the discussed studies tested whether the information about the environment or the information systems influence behavior. None of them however has a design where the information, provided in the mobile applications would be compared with other behavioral techniques. The information provided by the applications could be perceived as the informational messages given in the other studies, but it is different from the latter, because the information from the application is structured, usually rates the products by the common system and people choose by themselves to know that information, so they approach it differently. To conclude, the consideration of the effectiveness of the mobile information and its comparison with social norms is one more novel feature of my study.
3 Methodology
The study is based on the online experiment about decision making, which is modeled as shopping for the cleaning and laundry detergents. The participant is asked to imagine that
8 she is shopping for the mentioned products and gets six different product sets, each consisting of three items. She has to choose one product which she intends to buy in each set. The six sets are: laundry detergents, universal cleaner, fabric softener, laundry powder, dish washing liquid and toilet bowl cleaner. Each set is provided in a separate question. In each question the participant is given a certain amount of money which she can spend for buying the products given in that question. The amount is such that all the products in the set would be affordable and there would be a change left. To encourage participants to behave as in a real life situation as an incentive the lottery was used in the experiment. At the beginning each participant is told that after the research is over, two persons will be randomly chosen and provided with the products that they have chosen in one randomly selected question. The change that is left from the given amount of money after buying the particular product was sent to the participants as well. So the participant is induced to think if she wants to save money and buy a cheap product, or to buy an expensive one.
The articles are chosen in a way, that the information about it would be stored in the ‘Good Guide’ application and that it would be possible to buy them in Lithuania or in the online shop within Europe. This is an additional reason why detergents were chosen for the experiment. It is hard to find food or cosmetics that are sold in Europe and would be stored in the application, which is more applied for the American market. One more reason is that unlike shower products or cosmetics, choices of detergents are not affected by preferences for the brand, smell, allergies etc. as much as the preferences for the latter.
The prices of the products that are provided in the experiment are the real prices at the moment of the research and are calculated averages from the sources where the products are sold (supermarkets and online shops).
Products within the sets have three attributes that vary among them: price, eco label and brand popularity. The variations of attributes are presented in Table 3.1.
Attribute Variation
Price High, medium, low Eco label Eco, non-eco
Brand Popular, not popular
Table 3.1. Attributes of the products
The sets are modeled in a way, that the attributes of the products create different variations amid questions, to capture the effect of each attribute on the buying decision. For example,
9 in the first question the product Woolite Complete has the following attributes: low price, non - eco label and is not popular, Ecover Laundry Liquid has the high price, eco label and is popular, Tide Liquid Detergent is medium priced, non – eco labeled and popular. Popularity is decided according to the supply of the product in the supermarkets. If it cannot be bought in most of the supermarkets, it is said to be not popular and popular otherwise. The eco label is clearly recognizable from the package of the product.
The experiment consists of the four conditions: the control and three treatments. In the control condition the participant is only asked to choose a product from a given set. She is not provided with any additional information except the price of the item. Other attributes are expected to be captured by the participant from the picture of the product.
The first treatment condition is the Application treatment, where the participant is given the opportunity to choose to get the additional information by pressing the option ‘More information’ in each question for each product (Figure 3.1). If one chooses to see ‘More information’, in the next step he is shown the information that would be provided by the mobile application ‘Good Guide’ (Figure 3.2) after scanning the barcode of the product. Such design is chosen to transmit the barcode scanning procedure to the screen as accurately as possible - ‘More information’ has to be pressed for every product, while one should scan every barcode separately. The application rates the products according to the common system for all products and provides the basic information about the composition of the
10 product. The participant has the opportunity to check the methodology of the ranking by pressing the provided link. As it can be seen, the rankings (range from 0 to 10) of the effect on health, environment and society are given, along with the ingredients and their effect on health.
The participant of the second treatment condition – the Social Norm treatment (later Social treatment) – is given the descriptive norm in each question. Two types of descriptive norms are used – a norm about the behavior of others within the most popular supermarket in the country (Figure 3.3) and a norm about the behavior of citizens of the country the person has been living in at the time of research. The norms are excessive and the numbers representing the engaged behavior are augmented. According to the previous studies (Kahneman, 2011; Nolan et al., 2008) if the descriptive norm states what a small number of people behave in a desired way, it is demotivating. Such norm is perceived as the acceptable behavior. Regarding these findings, the number of participants referred to in the norm is chosen to be bigger than 50%.
For the norms to work properly, at the beginning of the questionnaire the participant is asked what is her current living location. The questionnaire is fitted for the citizens from Lithuania and the Netherlands. If the participants lived somewhere else at the time of the experiment, he was not shown any social norm and such responses were treated as the responses from the control
Figure 3.2. The application information about the product
11 condition (there were 3 such responses), because
then the setup was the same as in the control – no social norm was shown nor any additional information.
The last treatment joins the Application and the Social Norm treatments (later Social and Application treatment) and tests how both techniques work together. So in this treatment condition participant is given a social norm, the same that is used in the Social Norm treatment and has the opportunity to check additional information about the product, provided by
the application as in the Application treatment condition. Again, if the participant lived in the other country than Lithuania and the Netherlands, he was not shown the social norm and the response was treated as from the Application treatment (there were 4 such cases).
Summing up, all experiment conditions includes both – Lithuanians living in the country and those, who live abroad. Only conditions, where the social norm is included, contains participants living abroad who are residing in the Netherlands but still have Lithuanian origins. After six questions, where the participant had chosen the products, he was asked several questions to control for his attitude towards environmental problems and conscious consumption, income level, age, nationality and the country where she has been living for the last two years. The questions about the environment ask if the participant thought about the environment at all while choosing the products and how often she buys environmentally friendly commodities. She was also asked what influenced his decision the most, where the possible answers are: price, brand, other, information about the behavior of others (in the Social and Social and Application treatments) and ratings - information from the application (in the Application treatment and Social and Application treatment). In the following question the participant is asked to grade the statements “I am interested in the environmental problems and know what are the biggest issues of it”, “I think that the global warming is a big problem” and “I am concerned about the impact my personal consumption makes on the environment”. The Likert scale for Level of Agreement is used, ranging from 1 to 7, where 1 is ‘Strongly disagree’ and 7 is ‘Strongly agree’.
Figure 3.3. The social norm within the popular Lithuanian grocery store
12 Participants of the experiment are only Lithuanians, but living in different countries. However, those who lived not in Lithuania at the time of the experiment, take 19% of the respondents. Although some of the respondents were not living in Lithuania, all of them grew up in Lithuania, so the values and attitudes are formed by the same government and society. The questionnaires were distributed online, randomly assigning the condition of the experiment for each participant. The sample was taken from Facebook friends and friends of the friends, they hold different education levels, professions and are from all over Lithuania. The other part of the participants were the young members of the Lithuanian Association of Families with Deaf and Hearing Impaired Children families. Most of the participants do not know me in person and none of the participants knew the purpose of the experiment. The aim was to focus on millennials – those who are aged around 21-30, because they are seen as the buyers of the future and will take the biggest share of the shoppers in upcoming years. So the biggest part of participants belong to the mentioned age group.
The experiment was anonymous. Participants were asked to provide the address at the end of the questionnaire, if they wanted to participate in a lottery. The questionnaires were created online using the Qualtrics tool.
The expected result of the study is that green IS information has a positive effect on pro-environmental decisions and that it works better than the social norm, although the latter assumption is ambiguous. The joined treatment is expected to result in the most environmentally friendly choices. The hypotheses of the experiment are:
H1. The green IS information affects the buying of the eco-labeled products in the
positive way.
H2a. The green IS information works better than the social norms in enhancing a
pro-environmental shopping decisions.
H2b. The green IS information works worse than the social norms in enhancing
pro-environmental shopping decisions.
H3a. The green IS information joined with the social norms evokes pro-environmental
behavior more effectively than each technique alone.
H3b. The green IS information joined with the social norms evokes the
pro-environmental behavior as effectively as each technique alone. The following section presents the results from the experiment.
13
4 Results
4.1 Main statistics
In this section the results of the experiment are analyzed. 232 Lithuanians participated in the experiment, although 179 had finished the questionnaires. There are 45 responses in the Control condition, 54 in the Application treatment, 39 in the Social treatment and 41 in the Social and Application treatment conditions.
85% of the participants fall into the 21-30 year age group, 91% are younger than 30 years. So as it is stated before, the experiment was concentrated on analyzing the future buyers – millennials. The biggest part of the participants earn around the average salary in Lithuania. The average annual income is 6624 Euros (Statistics Lithuania, n.d.) and 31% of the participants stated their annual income is 4.000-9.999 Euros. 66% earn less than 15k Euros per year (17% under 4k per year, 19% earns 10k to 15k Euros per year) and 15% did not reveal their income level. The income levels on average did not differ among the treatment conditions (𝒳2(27, N=179) = 24.21, p=0.618). All participants grew up in Lithuania, 19% of them have been living outside of Lithuania for the last two years at the time of the experiment. There are 37% male and 62% female participants, one participant did not want to tell the gender. The Chi-squared test showed that there is no significant difference among the distribution of genders for the four conditions (𝒳2(6, N=179) = 5.65, p=0.464).
4.2 General results
The results are analyzed in terms of the number of the eco labeled products chosen among the six questions by participant of the study. Table 4.1 represents the percentages of participants who chose a certain number of the eco labeled products among each condition. It is worth noticing that only 13 participants, which is 7% out of all, did not choose any eco labeled product. But it can be seen that in the Social treatment condition none of the participants chose all six products to be eco labeled, but in this condition almost 50% of the participants chose four or five ecological products. In the Control condition the reverse tendency can be seen, while almost 50% of the participants chose only one or two ecological
14
Number of eco products chosen
Control App Social Social
and App Total 0 13% 4% 8% 5% 7% 1 24% 26% 10% 5% 17% 2 24% 22% 15% 17% 20% 3 20% 11% 21% 24% 18% 4 11% 15% 23% 29% 19% 5 2% 11% 23% 12% 12% 6 4% 11% 0% 7% 6%
Table 4.1. Percentages of participants who chose eco labeled products
products. The Application treatment condition result is similar to the Control condition with almost 50% of the participants choosing only one or two ecological items. Whereas in the joined treatment condition almost a half (49%) of the participants chose more than three ecological items. The Chi-squared test yields a significant difference between the conditions (𝒳𝟐(18, N=179) = 32.87, p=0.017). Comparing treatments with the Control condition, there are significant differences at 5% significance level among the Control and Social (𝒳𝟐(6, N=84) = 14.97, p=0.020) and Control and Social and Application (𝒳𝟐(6, N=85) = 15.26, p=0.018) treatments, but not among the Control and Application treatment (𝒳𝟐(6, N=99) = 8.52, p=0.202). Comparing other treatments pairwise, there are no significant differences at 5 % significance level between Application and Social and application (𝒳𝟐(6, N=94) = 12.07, p=0.060), Social and Social and application (𝒳𝟐(6, N=79) = 5.65, p=0.464), but there is between Application and Social (𝒳𝟐(6, N=93) = 12.61, p=0.050).
Table 4.2. Means and st. deviations of the number of chosen ecological products per conditions
The differences are seen in the average number of ecological items (Table 4.2) chosen as well. The average is the lowest in the Control condition (M=2.16) and the highest in the Social and Application treatment condition along with the lowest standard deviation (M=3.38, SD=1.48). The second highest average number of the ecological products is in the Social treatment condition (M=3.10, SD=1.57). The Application condition average (M=2.85, SD=1.80) is between the Control and the Social treatment conditions. In line with the Chi-squared test
Treatment condition Mean Std. Dev. N
Control 2.16 1.54 45
Application 2.85 1.80 54
Social 3.10 1.57 39
15
Control Application Social Norm
Control - - -
Application -2.05(97)* - -
Social -2.79(82)** -0.69(91) -
Social + Application -3.72(83)*** -1.50(92) -0.79(77)
Table 4.3. t statistics for pairwise comparisons of the means of the number of chosen ecological products between conditions (degrees of freedom in parentheses). * p<0.05, ** p<0.01, *** p<0.001
results, t-test shows significant differences between the Control condition and other treatments (Table 4.3). The highest significance level is between Control and Social and Application treatment (p= 0.0004), the second is Social treatment (p= 0.006). Pairwise comparison of treatment conditions resulted the insignificant differences in means.
Summarizing the results of the tests it can be concluded, that all techniques make the statistically significant positive effect towards choosing the eco labeled products. Although it cannot be said which is better than the others when comparing the behavior techniques itself.
Table 4.4. Eco purchases among the conditions and separate questions
To analyze the data a bit deeper I checked percentages of participants who had chosen ecological products in every question (Table 4.4). The patterns can be seen where the bigger
Questions / Eco-label Control Application Social Social+App.
Q1. Laundry detergent Non eco 67% 89% 95% 63% Eco 33% 11% 5% 37% Q2.Universal cleaner Non eco 54% 71% 38% 44% Eco 46% 29% 62% 56% Q3.Fabric softener Non eco 56% 73% 51% 56% Eco 44% 27% 49% 44% Q4.Laundry powder Non eco 54% 67% 44% 39% Eco 46% 33% 56% 61%
Q5.Dish washing liquid
Non eco 44% 44% 28% 29%
Eco 56% 56% 72% 71%
Q6.Toilet bowl cleaner
Non eco 41% 40% 33% 34%
16
Figure 4.1. Eco purchases among the conditions and separate questions
part of participants chose the ecological products in certain questions independently of the treatment condition, while in the other questions the ecological purchases are the minority. It can clearly be seen in Figure 4.1 that in the first and the third questions the non-eco labeled products were chosen more often than the eco labeled independently of the treatment condition. This suggests, that only the eco label does not imply the purchase. There are other factors that influence the decision as well. In these two cases (first and third questions), the ecological products were more expensive than the non-eco labeled ones. So if the price of the ecological product is too high compared to the other products, the eco label does not have a strong effect on choosing the product.
Treatment condition Gender t-test statistics
(degrees of freedom in parenthesis) Male Female Control Mean 1.64 2.39 -1.53(43) St.Dev 1.22 1.63 N 14 31 Application Mean 2.68 2.97 -0.57(52) St.Dev 1.91 1.73 N 22 32 Social Mean 2.31 3.5 -2.37(37)** St.Dev 1.97 1.17 N 13 26
Social + Application Mean 3.39 3.36 0.05(38) St.Dev 1.24 1.68
N 18 22
Table 4.5. Differences in the means of chosen acological products among the genders. * p<0.05, ** p<0.01, *** p<0.001
0% 20% 40% 60% 80% 100% Non eco Eco Non eco Eco Non eco Eco Non eco Eco Non eco Eco Non eco Eco Q1. Laundry detergent Q2.Universal cleaner Q3.Fabric softener Q4.Laundry powder Q5.Dish washing liquid Q6.Toilet bowl cleaner % o f th e p ar ticip an ts
Questions and product sets. Type of product chosen - eco/non-eco
App Control Social Social+App
17 The questions about the personal characteristics – age and income had no significant difference (at 5% significance level) neither between the treatments nor for the number of ecological products chosen.
Checking the effect of gender differences, there are no significant differences among treatments in terms of distribution of male and female participants (𝒳𝟐(3, N=178) = 2.27, p=0.518). Testing every treatment separately (Table 4.5), there is a statistically significant difference at 5 % level in the mean of chosen ecological products only in the Social norm treatment condition. However, there was the biggest difference in the number of participants, so the result can be biased.
The following section analyzes the personality and other factors that could be related with the decision to choose ecological product.
4.3 Analysis of perceptions towards environmental problems
Further analysis is concerning the additional general questions about the perception towards environmental problems and personal characteristics which were asked at the end of the questionnaires.
Firstly, participants were asked what influenced their decisions the most. The results are presented in Table 4.6. The biggest part of individuals state, that the price was the main influence for making the decision (over 30% in all the conditions). The brand was also perceived as decision shaping, around 20% of the participants chose this answer. It is worth noticing that while the social norm was the only information participants got (Social treatment), they saw it as a bigger influence (even 18% of participants stated it had an effect on their decision), while the norm is seen as a mere influencer when it comes with the
Row Labels Application Social Social and
Application Price 31% 44% 30% Ratings 28% - 28% Other 22% 21% 10% Brand 19% 18% 28% Norm - 18% 5%
18 opportunity to check the ratings (Social and Application treatment), then only 5% state they were influenced by the norm. In the later condition even 28% state that the main influencer was the information from the application.
The results from the question if the participants were thinking about the environment while choosing the products (Table 4.7) show, that the least part of the participants answered ‘Yes’ in the Control condition (22%), while over 24% answered positively in the Social and Application treatment, 36% in Social and 37% in Application condition. The most common answer in all conditions were ‘In some questions’ and most participants (almost 38%) declared not thinking about the environment at all in Control condition.
App Control Social Social and App
Yes 37% 22% 36% 24%
No 17% 38% 18% 17%
In some questions 46% 40% 46% 59%
Table 4.7. Answers to the question whether participant thought about environment while choosing a product
Those who stated they thought about the environment, actually chose the eco labeled products around 50% of the time in the Control treatment condition, 71% in the Social, 65% in the Application and 60% in the Social and Application treatment condition. Those who said they thought of the environment in some questions chose eco labeled products 37%, 54%, 39% and 67% of the time respectively. So it can be concluded that despite the fact that people think about the environment they do not fully indulge into the sustainable consumption. The later results are followed by the same conclusions from the average number of ecological products chosen among the conditions for particular answers (Table 4.8). It is clearly seen that those who stated they thought of the environment while choosing actually chose on average more ecological products than those who stated they did not think about the environment.
Answers Control Application Social Social and App
Yes 3.00 (2.36) 3.95 (1.85) 3.64 (1.50) 3.60 (1.58) No 1.59 (1.12) 1.78 (1.09) 1.71 (1.25) 2.71 (1.80) In some questions 2.22 (1.11) 2.36 (1.52) 3.22 (1.48) 3.42 (1.35)
Table 4.8. The average number (stand. deviation) of eco labeled products chosen among the answers to the question if participant thought about environment while choosing.
19
Level of agreement
(1-strongly disagree, 7-strongly agree)
1 2 3 4 5 6 7
I am interested in environmental problems and know what are the biggest issues of it.
5% 7% 9% 13% 32% 25% 9%
I think that global warming is a big problem. 3% 4% 4% 5% 21% 35% 27% I am concerned about the impact my personal
consumption makes on the environment.
6% 6% 9% 20% 29% 18% 13%
Table 4.9. Personal interests and concerns about the global warming
Analyzing the answers for the graded statements (Table 4.9), it is clearly seen that most of the participants are informed about global warming and the harm of consumption for the environment. In the first statement 66% agree that they are interested in the environmental problems and know the biggest issues of it. Even 84% of the participants agree that global warming is a big problem and 60% are concerned about their consumption’s impact on nature. Analyzing every condition separately, the same tendency is seen. None of the conditions differ significantly from the overall results in the first and second statements (𝒳2(21, N=179) = 28.50, p=0.127, 𝒳2(21, N=179) = 19.66, p=0.547 accordingly). For the third statement the Chi-squared test for the differences among the conditions in terms of agreement on the statement resulted that there is a difference at 5% significance level (𝒳2(21, N=179) = 35.68, p=0.024). The percentages of those who are concerned about the impact of their personal consumption on the environment can be seen in Figure 4.2. Certainly, in the Application treatment condition 67% of participants agree that they are concerned about their personal impact, 62% agreed in the Social treatment condition and 70% in the Social and Application treatment condition. Whereas the Control condition had only 40% participants who agreed on this statement. There can be multiple conclusions, either the participants in different conditions differ in terms of their beliefs about their personal impact on environmental problems or additional information encourages the participants to think more about the effect of their behavior on the environment.
Analyzing the perceptions of the participants about their own behavior, it is interesting to see how much in line their answer about the frequency of buying ecological products is with the actual percentage of the ecological products they chose in the questions. So I compared the answer to the question ‘How often do you buy environmentally friendly products?’ where
20
Figure 4.2. Answers among the conditions in question to grade the statement "I am concerned about the impact my personal consumption makes on the environment."
participants had to choose the percentages of the ecological products among their purchases. 29% of the participants were quite accurate and stated that the real shopping basket includes the same percentage of the ecological goods as their chosen set of the products. However, only 26% of participants stated that more than 50% of products they buy are ecological, whereas even 50% of the baskets hold more than three ecological items, which means that more than a half of the products is ecological. Only 13% overestimate their perception of themselves as environmentally friendly and state that in their everyday basket the percentage of ecological products is bigger than the percentage of ecological products they chose. Of course the judgement is a bit too severe because the shopping baskets of this research consists only of detergents, while the everyday basket consists more of other goods. So maybe someone who stated that buys sustainable products very often, buys only the sustainable food and food takes the biggest part of the basket.
4.4 ‘More information’ effect on choices
Focusing on the effect of the possibility to get ‘More information’ and the effect of the information itself the Application and the Social and Application treatment conditions are analyzed. It is worth mentioning how many participants used the opportunity to check the additional information about the product and pressed the ‘More information’ at all. Firstly, the analysis of the Application treatment condition is detailed below.
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Control Application Social Social and Application
% o f p ar ticip an ts Conditions 7 6 5 4 3 2 1 Level of agreement. (0-Strongly disagree, 7- Strongly agree)
21 The Application condition had 54 participants. Only 30 (which is 55%) pressed ‘More information’ at least for one of the products. 13% of participants checked the information for only one product among all the questions and only one participant checked the ratings for all the products in all the questions. On average 12.5 (SD = 2.88) people in every question checked the ecological product’s ranking while on average 11.08 (SD = 1.93) participants checked the non-ecological product’s ranking, so there is almost no difference. The same is concluded from the Mann-Whitney U test for the difference in nonparametric small samples (two samples: more information pressed for the eco labeled products and for the non-eco labeled products). The null hypothesis stating that the samples are not statistically different cannot be rejected (p = 0.341) therefore it can be concluded, that the participants were equally interested in the rankings for both the eco labeled products and the non-eco labeled products.
Analyzing the Social and Application treatment, 28 (out of 47) participants pressed ‘More information’ at least once. Compared to the Application treatment condition, it is a slightly bigger share of the participants, taking 60% of the treatment condition participants compared to 55% in the Application treatment condition. Here four participants checked the rankings in all the questions, five checked it in five questions. On average 11.3 (Std. Dev. 2.58) participants in every question checked the ecological products’ rankings and on average 9 (Std. Dev.2.66) non-ecological products’ rankings. The Mann-Whitney U test for the difference in the samples concludes that there is no statistically significant difference between them (p = 0.0975).
Analyzing the percentages of those who actually chose the eco labeled products after checking the ‘More information’ (Figure 4.3) can be seen that in the Application treatment in
Figure 4.3. Percentages of those who chose eco product among those who checked 'More information' for it
33% 42% 38% 71% 30% 71% 0% 25% 50% 75% 100% Q1 Q2 Q3 Q4 Q5 Q6 Application treatment 42% 53% 31% 91% 56% 100% Q1 Q2 Q3 Q4 Q5 Q6
22 mostly all the questions less than a half of those who saw the ranking of the eco labeled product chose it. Although in the Social and Application treatment in most of the questions more than a half of those who saw the rankings, chose the eco labeled products.
Analyzing those who did not check ‘More information’ for eco products in most of the questions eco labeled products were chosen by less than half of participants.
Figure 4.4. . Percentages of those who chose eco product among those who did not check 'More information' for it
The significant differences between those groups was not found by the Chi-squared test (Table 4.10). Overall, 22% of choices were eco labeled by those who did not click on ‘More information’ for ecological products versus 25% of those who were interested in additional information in Application treatment. In social and Application treatment percentages are 26% and 29% accordingly.
More info' chosen for Eco product
𝒳𝟐 test results
Eco product chosen No Yes
Application treatment
No 78% 22% 𝒳𝟐(1, 324) = 0.385
Yes 75% 25% p-value = 0.535
Social + Application treatment
No 74% 26% 𝒳𝟐(1, 246) = 0.374
Yes 71% 29% p-value = 0.541
Table 4.10. The percentages of participants who clicked 'More information' for labeled products and hen chose eco-labaled products and otherwise.
This suggests that seeing the ranking does not induce to buy the ecological product one hundred percent. Additionally, those participants who did not see the ‘More information’ –
33% 47% 44% 37% 60% 53% 0% 25% 50% 75% 100% Q1 Q2 Q3 Q4 Q5 Q6 Application treatment 33% 56% 47% 48% 73% 56% Q1 Q2 Q3 Q4 Q5 Q6
23 still choose to buy the eco labeled products at the very similar frequency. This suggests, that the more accurate effect should be analyzed in regression.
To sum up, it can be concluded that the participants were not very involved into using the opportunity to see the information from the application. But it is not clear if seeing the ranking had an effect on the choice of the ecological product.
4.5 Regression analysis
Regression analysis is done to capture the effects of the behavior techniques among the treatment conditions controlling other variables. It is clearly seen that neither the social norms nor the application information work perfectly in fostering the environmentally friendly behavior. Therefore regression analysis had been implemented to check if the price, popularity and other variables had a bigger impact on choosing or not choosing the ecological products.
As in the previous subsection, the data is analyzed in terms of the individual answer for every single question. A panel data OLS regression with random effects was used. The random effect model was chosen following the result of The Breusch-Pagan Lagrange multiplier test for existence of a random effect (𝒳2(1) = 50.24, 𝑝 < 0.001). The regression aims to check if the behavior shaping tools have a statistically significant positive effect towards choosing the eco labeled product. Therefore, the dependent variable is binary and equals one if the chosen product in particular question was eco labeled and zero otherwise. Robust standard errors are chosen to correct for the non-constant residuals of the variables.
The results of the regression analysis (Table 4.11) show that only the Social treatment had significant influence (at 1% significance level) on choosing the eco products. The dummy variables of the treatments represent the treatment variable interaction with Female dummy variable. So the results show that Social treatment was the most effective compared to Control condition for female participants, the second most effective was Social and application treatment and the last – Application treatment. Looking to male responsiveness to the treatments, the Social and Application treatment had the most influence – there participants where 28.8% more likely to choose eco labeled products than in the control condition holding other variables constant. The ability to check additional information from the application is expected to sway the decision by 15.9% towards the eco product and the social norm – by 7.6%, compared to the control condition and holding other variables
24
Table 4.11. Regression results. z statistics in parentheses, * p<0.05, ** p<0.01, *** p<0.001
Variables Random Effect Model
Application treatment 0.044
(0.55)
Social treatment 0.162**
(2.58)
Social + App treatment 0.086
(0.70)
Application treatment * Male 0.115
(1.00)
Social treatment * Male -0.086
(-0.71)
Social + App treatment * Male 0.203
(1.15)
Price of eco per liter -0.0697***
(-7.73)
Price per liter of the cheaper alternative -0.004
(-0.31)
Price per liter of the more expensive alternative 0.024*
(2.19)
Info chosen x Eco rank 0.035*
(2.04)
Info chosen x Min Alternative -0.016
(-0.66)
Info chosen x Max Alternative -0.037*
(-2.22)
Info chosen x Eco rank * Male -0.017
(-0.65)
Info chosen x Min Alternative * Male -0.004
(-0.14)
Info chosen x Max Alternative * Male 0.0103
(0.32)
Interest in env. Problems (1-7) 0.025
(1.71) Big problem (1-7) -0.017 (-1.15) Personal affect (1-7) 0.033 (1.90) Gender -0.096 (-1.30) Age -0.006 (-0.39) Constant 0.493*** (4.09) 𝒳2 (degrees of freedom) 187.1 (20) 𝑅2 within 0.118 𝑅2 between 0.151 𝑅2 overall 0.137 N 862
25 constant. So it can be concluded, that female are more likely rely on social norms while male participants were more affected by information of the application. However, the latter coefficients are not statistically significant. But the regression verifies the previous analysis, that regarding female participants, the most effective treatment was Social, the second most effective – Social and Application and the least effective – Application treatment, whereas regarding male participants, the most effective treatment was the Social and Application, then follows the Application treatment and social norm had the least influence on eco decision.
The price did have an influence on the decision. The price increase of the eco labeled product by one unit is expected to decrease the probability of choosing eco product by 0.7% at 0.1% significance level holding other variables constant. Interestingly, the participants were not influenced a lot by the price of the cheap alternatives to the eco product. But the price of the more expensive alternatives among the sets of products had a significant effect at the 5% level. The increase of the price by one unit is expected to increase the probability to choose the eco labeled product by 2.4%, holding other variables constant.
The other significant factor was the ranking given in the application information screen. The variables that captured this effect are interactions of dummy variable of the “More information” choice (1 – “More information” was chosen, 0 – otherwise) with the overall ranking of the product and the gender dummies. Regression showed that the latter variable for eco products improved the probability to choose eco product by 3.5% among female participants and 1.8% among male participants, holding other variables constant. The other factor that had influence on the choice was the ranking of the product which had the higher ranking among the two alternatives for the eco labeled product. The one unit increase in the ranking is expected to sway the decision against the eco labeled product by 3.6% among female and 2.6% among male participants. This makes sense because the closer the non-eco labeled product ranking is to the eco labeled, the more similar is the quality and the eco label loses its power.
Other variables, as in the previous section’s analysis, did not have any significant effect on choosing the eco labeled product.
26
4.6 Limitations
There are few possible shortcomings of the research. In the study by Luchs et al. (2010) researchers found that there can be a negative effect of ethicality of the product for products like tires, detergents and the ones that customers expect to be strong and long-lasting. This suggests that the perceptions of participants towards the ecological cleaning products might be biased and the choices for detergents do not fully represent the behavior towards buying ecological products. Although it would be hard to find the products that would fully represent the behavior and would fulfill all the properties required to be used for the experiment. Secondly, although there are a lot of information about how much smartphones are used in retail shopping in the US (SessionM, 2015), there is not much such data on Lithuanians. TNS.LT (2014) published the report that every second Lithuanian uses mobile applications, but there is no information about the use at the point of sale. Moreover, although mobile tagging is popular in Lithuania, the applications providing additional information about the product’s sustainability are American and not well known in Lithuania. This might have resulted in the low usage of the opportunity to check the additional information during the experiment. It is not clear if the concerns about the impact of personal consumption on the environment was affected by the behavioral techniques or if it was the difference among participants. Those who saw a social norm in every question and checked the additional information about the products are seen to be more concerned about their own effect on the environment. So one more drawback of this experiment is that the beliefs were not tested before the experiment. Although if it would have been tested prior to the questions for choosing the products it might have affected the choices, because then people would have been primed with ideas about the environment and may have thought about it more than regularly while choosing the products.
One more limitation could be seen in the prominence of the eco label. The participants could have understood that this experiment was about the ecological products and it could have resulted in the experimenter demand effect. There is a potential problem if the objectives of misunderstood demand would correlate with the experiment objectives, which is partly true in this case – participants were expected to prefer eco labeled products against other. So there is a risk of biased behavior regarding the experimenter demand effect. But despite the fact that participants might have misunderstood the purpose of the experiment, it was clearly
27 seen that those in treatment conditions behaved differently, so the intention to capture the behavioral techniques effect was fullfiled.
From the latter considerations there is some concern about the external validity of the experiment. But despite the low usage of additional information and the fact that products were detergents, there still can be seen a positive impact of the behavior techniques.
5 Discussion and conclusion
The online experiment has shown that each of the tested behavioral techniques works in a positive way regarding the choice for ecological products. So the first hypothesis that the Green IS works in a positive way is true. Although the increase in environmentally friendly choices is seen for both, the Social and Application treatments, they result in inconsistent effects on different genders. The Social treatment results in a bigger behavior change than the Application treatment for female participants of the experiment and vice versa for male participants. So the hypothesis that Green IS can work better (H2a) in enhancing the sustainable consumption is true for male participants and the hypothesis that the Social norm works better than Green IS (H2b) holds for females. The regression results confirm the latter conclusions. Although the social norm is more effective for females, the ranking scores of the products are seen as the significant variables and as would be expected, increasing rank for the ecological product affects the choice for the eco labeled item in a positive way, whereas increasing the higher score of alternative product has a negative effect. It is logically explainable, because the closer the ranking of sustainability of the other than eco product is, the less affect the eco label has. Here it could be pointed out that the ranking scores have a significant influencers on the choice, whereas people are not very familiar to use the application to get them. That leads to the conclusion that a significant effect can nurtured with the help of the mobile application if people were enhanced to use it more often while shopping.
Furthermore, although the assumption that the joined treatment results in more ecological choices is confirmed by the regression, it is true only for male participants. So the hypothesis H3a stating that the Social norm and Application treatment condition is the most significant among all the treatment variables holds only for males.
28 Although as it was seen in the general result section, not only the behavior techniques affect the decision to buy eco labeled products. The price and other variables differ among the choice of eco labeled and non-eco labeled products and are also the significant variables in terms of choosing the eco labeled products. This is in line with previous findings where the researchers claim that price is the most important barrier of choosing sustainable products in the food market (Vermeir & Verbeke , 2006). It is interesting that the ranking and the price has almost the same effect on the choice, only in different directions – increasing price affects the choice negatively, while the increasing ranking affects it positively.
To conclude, the experiment showed that both techniques are effective regarding the behavior change. However the further research could be done to check how the shopping behavior of other products is affected by the green IS. Also it would be interesting to replicate the experiment in the supermarket to see how people are likely to use the application and if it affects them while they are shopping.
29
6 References
Atkinson, L. (2013). Smart shoppers? Using QR codes and ‘green’ smartphone apps to mobilize sustainable consumption in the retail environment. 37, 387-393. Barnes, S. J. (2002). Wireless digital advertising: nature and implications. International
Journal of Advertising, 21(3), 399-420. doi:10.1080/02650487.2002.11104939
Bohner, G., & Schlüter, L. (2014). A Room with a Viewpoint Revisited: Descriptive Norms and Hotel Guests' Towel Reuse Behavior. PLoS ONE, 9(8).
doi:10.1371/journal.pone.0106606
Bolsen, T., Leeper, T., & Shapiro , M. (2014). Doing What Others Do: Norms, Science, and Collective Action on Global Warming. 42(1), 65– 89. doi:10.1177/1532673X13484173 Cialdini, R., Kallgren, C., & Reno, R. (1991). A Focus Theory of Normative Conduct: A
Theoretical Refinement and Reevaluation of the Role of Norms in Human Behavior.
Advances in Experimental Social Psychology, 24, 201–234.
doi:10.1016/S0065-2601(08)60330-5
Delmas, M., & Lessem , N. (2015). Eco-Premium or Eco-Penalty? Eco-Labels and Quality in the Organic Wine Market. 0007650315576119. doi:10.1177/0007650315576119 Dolan, P., & Metcalfe, R. (2013). Neighbors, Knowledge, and Nuggets: Two Natural Field
Experiments on the Role of Incentives on Energy Conservation. Centre for Economic
Performance.
European Comission. (n.d.). Europe 2020. Retrieved from http://ec.europa.eu/europe2020/index_en.htm
Goldstein, N., Cialdini, R., & Griskevicius , V. (2008). A Room with a Viewpoint: Using Social Norms to Motivate Environmental Conservation in Hotels. 35. doi:10.1086/586910 Hughner, R., Stanton, J., McDonagh, P., Prothero, A., & Shultz, C. I. (2007). Who are organic
food consumers? A compilation and review of why people purchase organic food.
Journal of Consumer Behaviour, 6, 94–110. doi:10.1002/cb.210
Jenkin, T., Webster, J., & McShane, L. (2011). An agenda for 'Green' information technology and systems research. 21(1), 17-40. doi:10.1016/j.infoandorg.2010.09.003
30 Linden, S. v. (2015). Exploring Beliefs About Bottled Water and Intentions to Reduce
Consumption: The Dual-Effect of Social Norm Activation and Persuasive Information.
47(5), 526–550. doi:10.1177/0013916513515239
Loeser, F., Erek, K., Schmidt, N.-H., Zarnekow, R., & Kolbe, L. (2011). Aligning Green IT with Environmental Strategies: Development of a Conceptual Framework that Leverages Sustainability and Firm Competitiveness. AMCIS 2011 (p. 222). All Submissions. Retrieved from http://aisel.aisnet.org/amcis2011_submissions/222
Luchs, M., Naylor, R., Irwin, J., & Raghunathan, R. (2010). The Sustainability Liability: Potential Negative Effects of Ethicality on Product Preference. 74(5), 18-31.
McKenzie-Mohr, D., & Schultz, P. (2014). Choosing Effective Behavior Change Tools. 20, 35-46.
Mohrenfels, H. W., & Klapper, D. (2012). The Influence of Mobile Product Information on Brand Perception and Willingness to Pay for Green and Sustainable Products. Orlando: Thirty Third International Conference on Information Systems.
Nolan, J., Schultz, P., Cialdini, R., Goldstein, N., & Griskevicius , V. (2008). Normative Social Influence is Underdetected. 34(7), 913-923. doi:10.1177/0146167208316691 Sammer, K., & Wüstenhagen, R. (2006). The infl uence of eco-labelling on consumer
behaviour – results of a discrete choice analysis for washing machines. Business
Strategy and the Environment, 15, 185–199. doi:10.1002/bse.522
Schultz, & Wesley, P. (2014). Strategies for promoting proenvironmental behavior: Lots of tools but few instructions. European Psychologist, 19(2), 107-117. doi:10.1027/1016-9040/a000163
Schultz, P., Khazian, A., & Zaleski, A. (2008). Using normative social influence to promote conservation among hotel guests. Social Influence, 3(1), 4-23.
doi:10.1080/15534510701755614
Schultz, P., Nolan, J., Cialdini, R., Goldstein, N., & Griskevicius, V. (2007). The Constructive, Destructive, and Reconstructive Power of Social Norms. Psychological Science, 18(5), 429-434. doi:10.1111/j.1467-9280.2007.01917
SessionM. (2015). Retail Shopping, Connecting the Multichannel Shopper. A retail study. Retrieved from
http://cdn2.hubspot.net/hubfs/534866/SessionM_Retail_White_Paper.pdf?t=14532 21238313
31 Smith, J., Winnifred , R., Deborah , J., Katharine , H., Clarke, M., & Chenga, X. (2012).
Congruent or conflicted? The impact of injunctive and descriptive norms on environmental intentions. Journal of Environmental Psychology, 32(4), 353-361. doi:10.1016/j.jenvp.2012.06.001
Solér, C. (1996). Ecologically friendly buying — Theoretical implications of a
phenomenological perspective. Scandinavian Journal of Management, 12(3), 275– 289.
Statistics Lithuania. (n.d.). Main Lithuanian indicators. Retrieved from Official Statistics Portal: http://osp.stat.gov.lt/en/web/guest/home
The United Nations. (n.d.). Sustainable Development. Retrieved from https://sustainabledevelopment.un.org/
TNS.LT. (2014). TNS LT: nemokamas programėles išmaniesiems telefonams naudoja tik kas
antras. TNS.LT. Retrieved from
http://www.tns.lt/lt/news/tns-lt-nemokamas-programeles-ismaniesiems-telefonams-naudoja-tik-kas-antras/ United States Congress. (2013). GREEN Jobs Act of 2013. Retrieved from
https://www.govtrack.us/congress/bills/113/hr2863
van der Linden, S. (2015). Exploring Beliefs About Bottled Water and Intentions to Reduce Consumption: The Dual-Effect of Social Norm Activation and Persuasive Information.
47(5), 526–550.
Vermeir , I., & Verbeke , W. (2006). Sustainable Food Consumption: Exploring the Consumer Attitude - Bevaviour Gap. Journal of Agricultural and, 19(2), 169-194.
doi:10.1007/s10806-005-5485-3
Watson, R., Boudreau, M.-C., & Chen, A. (2010). Information Systems and Environmentally Sustainable Development: Energy Informatics and New Directions for the IS
Community. MIS Quarterly, 34(1), 28-38.
World Economic Forum. (2015). Engaging Tomorrow’s Consumer. Sustainable Consumption initiative. Retrieved from http://www.weforum.org/reports/engaging-tomorrow-s-consumer
Young, W., Hwang, K., McDonald, S., & Oates, C. (2010). Sustainable Consumption: Green Consumer Behaviour when Purchasing Products. 18, 20–31. doi:10.1002/sd.394
