The relationship between ecological knowledge, ecological concern, ecological affection, subjective norms and the green purchase behavior in Brazil

1 Faculdade Reges de Ribeirão Preto and Assistant Professor at FIA, Rua Garibaldi, 715, apto. 21 Higienópolis – Zip Code: 14010-170 – Ribeirão Preto-SP. 2 Faculdade de Economia, Administração e Contabilidade –FEA-USP, Av. Prof. Luciano Gualberto, 908, sala G173 Cidade Universitária – Zip Code 05508-010 São Paulo – SP. 3 Graduate Program in Business Administration, FEA-RP/USP Av. dos Bandeirantes 3900 – Zip Code: 14.040-900 Ribeirão Preto-SP. 4 Faculdade Dom Bosco de Piracicaba, Rua Luiz Razera 300, apto. 54F, Zip Code: 13417-530 Piracicaba-SP. ‘ Accepted 30 August, 2013


INTRODUCTION
The current model of economic growth based on maximizing production and consumption have generated simultaneously both imbalances in economics and social situation and environmental cenarium as well, depending on the environmental impacts seen in recent decades.This problem in turn, was reflected in the increased concern about environmental issues -specially from the 60s and 70s, when ecological movements started-the emergence of environmental marketing in academic research, focused on consumer behavior whose main *Corresponding author.E-mail: hrstamashiro@ibest.com.br.Tel: 55 + 16 3625-2912 focus has been investigating the impact of these issues on the green or ecological behavior and attitudes of consumers (Straughan and Roberts, 1999;D'souza et al., 2007).
In fact, with this attempt to reduce environmental impacts, environmentally responsible behavior (ERB), defined by Lee et al. (2013) as the effort made by people to reduce environmental impacts and contribute to environmental preservation, has received special attention in order to understand how human behavior affects the sustainable development of ecological environments.
According to Rasmussen (2011), as we move into this new decade, the need for a transition to an environmentally friendly future becomes more evident every day.Thus, the search for meeting the needs of a society that grows and develops more and more, especially with regard to technologies and new patterns of production, trade and consumption of products and services, has generated direct effects on the environment.
The understanding of all these aspects appears as a key issue for the sustainability of marketing activities, whose challenge is to develop and launch new products that meet the consumers' needs, given that, in the opinion of Costa and Jongen (2006), it is considered the starting point for the survival of the companies.Thogersen and Olander (2002), Amine (2003) and Gilg et al. (2005) advocate the idea that the adoption of such attitude by the companies is not just an adjustment to the issues of specific legislation, but also a marketing response to the pressures that consumer groups began to exercise since the early 1990s, due to a greater concern with the aspects focused on the environment.According to Firat (2009), pressures arising from environmental organizations, competition between firms and academic discussions are factors that contributed to the appearance of the environmental awareness since 1980, especially due to the climate changes, droughts, and other events as a result of global warming.
Given the context explained here, the concern to understand who consumes green products and their characteristics has been the keynote of marketing studies (Gonçalves- Dias et al., 2009).This concern has motivated the implementation of various discussions within the corporate and academic world, in an attempt to define the green or eco-consumer (D'souza et al., 2007).Udo (2007) states that, due to the complexity in understanding consumer behavior, a number of studies tried to identify the determinants of behavior in the process of green purchases, as well as the profile of the green consumer.According to this author, the fundamental question lies in the understanding of green consumers and their characteristics, as well as the purchase criteria adopted by them.Therefore, many researchers have sought to channel their efforts in search for answers that indicate the determinants of green purchases, understood as those that show concern for the physical environment, such as water, air, land (Shrum et al., 1995).
The first study, empirically tested, which sought to examine the relationship between ecological knowledge, ecological affection, ecological concern and the consumer buying behavior was developed by Maloney and Ward (1973).Despite the tenuous conclusion of the study, not proving the existence of a strong correlation between the buying behavior of the respondents with ecological knowledge, affection and concern, the model proposed and reviewed by Maloney and Ward (1973), revised in 1975, served as the basis for future research investigating the consumer buying behavior, trying to establish a relationship between the buying behavior and the environmental issues, such as ecological affection, ecological concern and ecological knowledge (Kennear and Taylor, 1973;Mackenzie, 1990;Martin and Simintiras, 1995;Chan, 1996;Zanoli and Naspetti, 2001;Kleef et al., 2005;Chan et al., 2008).
In addition to the environmental issues that may affect the buying behavior of consumers, there is the influence of family, friends and neighbors, which may interfere with this process.Therefore, the subjective norms were also studied by several articles, whose aim was to seek to measure the influence of others with regard to consumer behavior.Among the studies identified in this regard, we highlight those conducted by Ajzen (1985Ajzen ( , 1991Ajzen ( , 2001)), Ajzen and Fishbein (2004) and Mahon et al. (2006).
It is within this perspective that this paper fits into the debate on the relationship between environmental issues and the purchase behavior of ecological or green products (Ottman, 1994), named by Layrargues (2000) as one whose power of choice of the product covers, besides the price/quality issue, the environment.Thus, the purpose of this study was to determine whether there is a correlation between the constructs: ecological knowledge, ecological affection, ecological concern, subjective norms and green purchase behavior.
The fact that in Brazil, studies focused on the relationship between environmental factors and consumer behavior are still embryonic in the country, represents an opportunity for studies correlating the purchase behavior of consumers with environmental issues.Since the environmental damage is not an exclusive result of the irresponsibility of companies, but of the whole society, it is believed that extending the discussions involving the position of consumers in view of these issues proves to be relevant, given that the environmental degradation is the result of a crisis of bad behaviors of people (Maloney and Ward, 1973;Green-Demers et al., 1997).The following section provides a description about the material and method of study.

ENVIRONMENTAL AWARENESS AND THE GREEN CONSUMER
The current economic growth model based on the maximization of production and consumption has generated, simultaneously, huge economic, social and environmental imbalances, as a result of the environmental impacts observed in recent decades.In this regard, the importance of understanding and applying the concepts of ideas related to the environment and the consumer awareness of individuals has proven to be necessary and urgent.According to Green-Demers et al. (1997), people also contribute to the degradation of the environment in various ways, for example: careless waste disposal, poor consumer habits, etc.
Environmental awareness, as supported by Bedante and Slongo (2004), refers to the tendency of an individual to stand with respect to issues related to the environment, proving to be for or against it.According to these authors, higher levels of environmental awareness tend to get them to make decisions that consider the environmental impact of their attitudes and actions.In this approach, Sheth et al. (2001) show that the environmentally aware consumer is conceived as the individual who seeks products perceived as those that generate the least impact on the environment.
While Elkington and Hailes (1989) admit that everyone is increasingly aware of the threats to the environment, Goleman (2009), in turn, believes that almost all of us go out for shopping without being aware of their true impact and the impact of our habits.According to this author, our inability to instinctively recognize the connections between our actions and the problems resulting therefrom leaves us wide open to create threats that we censor.
Whereas the consumer not always make a critical reflection before starting the buying activity (Kleef et al., 2005), Gilg et al. (2005) draw attention to the fact that the green purchase needs to be understood as the development of sustainable lifestyles that incorporate other environmental actions in a more holistic conceptualization.The solution to the problems of pollution and environmental degradation depends not only on the actions of companies and governments, but also on changing the human behavior (Maloney and Ward, 1973).A similar view can be observed in Elkington and Hailes (1989) and Green-Demers et al. (1997), who claim that, if on the one hand people contribute to environmental degradation (for example, by disposing waste without proper treatment, keeping poor consumer habits, etc.), on the other hand, part of the solution lies in the hands of people.Bhattacharya et al. (2009) showed that consumers expressed a preference for products from companies that invest in social-environmental responsibility.In Brazil, Serpa and Ávila (2006) argue that Brazilian consumers are increasingly assigning the companies a responsibility to generate changes in society, a responsibility that goes beyond the generation of profits, jobs and paying taxes.Kalafatis et al. (1999) assume that there is social awareness in consumers willing to contribute to the environment protection through the choices of products and services they buy, and traders, in turn, are responding to these demands by developing environmentally friendly products.In this regard, Mohr and Webb (2005) believe that fewer and fewer people are judging a company just for economic reasons.Tanner (1999) concluded in his studies that, although in recent decades there has been an increased awareness regarding the environmental threat affecting the planet, the environmental degradation continues.Reitman (1992) argues that there is a wide range of green products offered by companies (biodegradable detergent, unbleached coffee filters, recycled paper and products), but most of these products do not reach the level of success expected in the market, despite the increased consumption responsible for the environmental preservation.
In general, Ottman (1999) classifies green products as durable, non-toxic, made from recycled materials and with minimal packaging.It means that green products can not be analyzed only from the viewpoint of the final results of the production process, but also in light of the entire production chain.
An environmentally friendly or green product, according to Ottman (1994), should have the following characteristics: a) be manufactured with a minimum amount of raw materials and with renewable or recyclable raw materials that conserve the natural resources in the extraction process; b) be manufactured with the utmost energy efficiency and with the least use of water; c) be packed in lighter packaging; d) provide a higher durability and serve multiple purposes; e) be reusable; f) be biodegradable.This author draws attention, however, to the fact that there are no products completely green or environmentally friendly, because when they are produced, they all consume energy, resources and generate emissions into the atmosphere.
The expressions, environmentally friendly and green consumer, little known until recently, reached high proportions and connotations, particularly for those involved with the production process (Mackenzie, 1990).For being considered a relatively new area, there are few studies that bring the definition of "green consumer" (Shrum et al., 1995;Gilg et al., 2005).Within this approach, Hopwood et al. (2005) add that although the phrase "sustainable development" (emphasis added) have different meanings and generate different views, deep inside, its concept appears as an attempt to combine questions on socioeconomic and environmental issues.
By analyzing the environmental issue and the corporate concern with this aspect, Zimmer et al. (1994) make a chronology that follows this order: the 1970s are considered the ones that make the importance of ecology, due to the environmental pollution and the consumption of energy.The 1980s bring to light the importance of discussing environmental issues due to the worsening of air pollution.Finally, the 1990s show the emergence of

Authors Concepts Ottman (1999)
It is the one that seeks to consume products that cause little impact to the environment -the so-called green or environmentally friendly products.

Layrargues (2000)
It is the individual who makes their buying choices taking into account not only price and quality, but also the fact that it is environmentally friendly.

Portilho (2004)
It is the one who in addition to the variable quality/price, includes in their "power of choice", the environmental variable preferring products that do not harm the environment.

Dias (2007)
It is the one who is concerned with the environment and adopts a behavior consistent with these values.

True Green
They hold strong environmental beliefs and experience them; they are almost three times more likely to avoid buying products from a company with questionable environmental reputation.They are twice as likely to buy products more "green."They believe they can make a difference in solving environmental problems.

Green Money
They support environmental issues much more through the donation of money than time or action.They are more likely to spend more for "green" products: a 15% increase in the price is acceptable.

Almost Green
They want more environmental legislation, but they do not believe they can do something for a positive environmental impact.They regularly engage in proenvironmental activities.but they do not want to pay more for pro-environmental products; they are uncertain when they are forced to choose between the environment and the economy.

Whiners
They perform few environmental actions, but believe that other consumers are also not doing their job, and that the company should solve the problems.They believe that.besides being very expensive among the "non-green" products available, the "green" products do not work very well.They feel confused and uninformed about environmental issues, believing that everything is someone else's problem and that a third party should solve it.

Basic Brown
They are less involved with environmental issues and their basic belief is that there is not much that individuals can do to solve the problem.
Source: Authors based on the information of Ottman (1994).
the "green trend" when marketing professionals realized that consumers would respond to efforts in relation to the environment.Based on these discussions, the concept of green consumer emerges, as shown in Table 1.
The concepts presented in Table 1 refer to the idea of people who, when buying their products, are concerned with issues such as the environmental degradation and that, because of this, are willing to take action to resolve or at least alleviate the problem, including modifying their personal consumption habits.It should be noted that the sustainable consumption as a form of consumption that uses natural resources to meet the needs, without necessarily compromising the needs and aspirations of future generations, comes from the changes in the attitudes of consumers and society in general.Based on this approach, we assume the idea that the formation of the green awareness can be understood as part of a set of actions aimed at environmental preservation.Green consumers are classified according to their degree of environmental concern, as shown in Table 2.
The descriptions of the consumer groups shown in the table above allow them to be regrouped into four sub-categories proposed by Gilg et al. (2005): the committed environmentalists, supporter environmentalists, occasional environmentalists and non-environmentalists.

DETERMINING FACTORS OF THE ECOLOGICAL BEHAVIOR (GREEN) OF THE CONSUMER
The term ecological behavior is used in a positive sense, meaning the same as eco-friendly, that is, one acting in favor of the environment.Among the commitments made by consumers, the "environmental behavior" is comprised by the commitments that will be verbally assumed based on the "environmental intentions" (Chan, 1999;2001).For Chan and Lau (2000), the ecological behavior of an individual is highly dependent on their ecological knowledge, affection and intention.In this regard, Kollmuss and Agyeman (2002) argue that educating people about environmental questions leads to a greater proenvironmental behavior.Given the recent nature of this construct in the literature, Pato and Tomayo (2006) assume the possibility of the existence of several measures of this construct, probably reflecting the lack of an accurate definition of what is the ecological behavior.
According to Steg and Vlek (2009), the effectiveness of behavioural interventions generally increases when they are aimed at important antecedents of the relevant behaviour and removing barriers for change.Therefore, it is important to understand which determining factors of the environmental behavior.The descriptions about each one of these constructs are presented below.

Ecological knowledge (EK)
The literature review showed that, with respect to the variable "ecological knowledge," named by Schahn and Holzer (1990) as "factual knowledge," it tends to affect the action of people even when the focus does not refer exclusively to environmental issues.Byrne (2001) assumes that the individual level of environmental knowledge, as well as the level of environmental orientation or commitment, can affect the consumer behavior, leading the consumer to environmental behaviors.The question, however, lies in knowing what the issues related to "environmentally friendly" products really mean to the consumer (Wagner, 2003).Kollmuss and Agyeman (2002) state that, in most cases, the increased knowledge and awareness do not necessarily lead to a pro-environmental behavior.In this sense, Tanner and Kast (2003) suggest that, although the environmental knowledge has been the subject of many studies, some empirical tests show no conclusive uniformity that indicates the direct relationship between environmental knowledge and environmental behavior.In Brazil, the characteristic is of a society still poorly informed and that have not shown enough strength to threaten organizations who are not committed to environmental degradation (Tagliaferro, 2000).
The results of the studies conducted by Dispoto (1977) showed that the knowledge about ecology, environment and pollution was the main predictor of the environmental behavior, but they found a tenuous relationship between the two variables.These results also suggest that environmental education and planned conservation may be the basis for changes related to the behavior of individuals.Thus, the generalized belief is that knowledge positively influences people's behavior (Chan, 1999;2001).The assumptions presented above shows that scholars are still not fully convinced that the knowledge about environmental issues necessarily leads a consumer to adopt a pro-environmental buying behavior.Maloney and Ward (1973) and Maloney et al. (1975) agree with this view, by emphasizing that the environmental behavior of an individual is highly dependent on their ecological knowledge, affection and intention.D 'Souza et al. (2007) argue that consumers who are well informed about environmental problems are more likely to purchase green products.
In a recent study, Barber et al. (2010) show that increased information leads to a greater knowledge about the environment and this is a prerequisite for changing attitudes.The results of a survey conducted by Xiao and Dunlap (2007) suggest that, in general, the attitudes of respondents are relatively related to the "concern for the environment."Barber et al. (2010) claim that knowledge and attitudes are important factors, as they serve as a basis for promoting changes in the human actions towards the environment.These claims, in turn, led to the formulation of the first research hypothesis of this study.

Hypothesis 1: the level of ecological knowledge of consumers on environmental issues influences their green purchase behavior.
In addition, Maloney and Ward (1973) argue that it is important to determine what people "know" about ecology, environment and pollution, as well as how individuals feel in the midst of these issues and what commitments they have made or are willing to assume for them.Concentrating the efforts of this study to make this research is critical because, according to Chen and Chai (2010), the knowledge gap on the use and values of green products prevents consumers from committing to a purchase decision.

Ecological concern (EC)
The environmental concern towards the environment, called ecological or green concern (Ottman, 1992), has been the subject of several studies abroad (Chan and Lau, 2000;Chan et al., 2006;Chan et al., 2008).Zimmer et al. (1994) point out that understanding the dimensions related to the concern for the environment within the consumer market is the first step for the development of green marketing strategies.Ling Yee (1997) adds that the environmental concern has been measured in different ways.Citing data from Gallup's Health of the Planet Survey (1992), this author reports that 67% of Germans rated environmental problems as very serious, compared with 51% in the United States and 42% in Japan.On the other hand, green consumers in the United States account for 37% of the population, while in European countries such as Switzerland, Germany and England, they account for 50% already (Andrade et al., 2002).
Complementing this line of reasoning, Brehm et al. (2006) suggest that a few studies indicate low levels of environmental concern by American consumers, but such conclusions are inaccurate, as they have indicated a conscious environmental behavior when compared to the world population in general.Results of a survey conducted by Xiao and Dunlap (2007) suggest that, overall; the attitudes of respondents are relatively related to the "concern for the environment." The results of a study conducted in China by Chan and Lau (2000) that sought to investigate the influence of cultural values, ecological affection and ecological knowledge on the green purchase behavior of Chinese consumers show that there is a positive relationship between these variables, but despite this, the level of ecological knowledge of the Chinese people is low and they have a minimum green purchase behavior.Kim and Choi (2005) argue that the consumer concern about environmental issues may not be translated into pro-environmental behaviors, but consumers with a strong concern for the environment are more likely to purchase environmentally friendly products than those less concerned with these issues; for them, this concern contributes to affect the operational aspects of the companies.Based on these assumptions, it was possible to propose the second research hypothesis.

Hypothesis 2: the level of ecological concern of individuals influences the adoption of a green purchase behavior.
Even with advances still far from the desirable, Kim and Choi (2005) consider the environmental concern as fundamental in studies related to the ecological behavior of individuals, as it can be considered a variable predictor of the environmentally friendly behavior.

Ecological affection (EA)
The affection is related to emotions and feelings that an object causes to a person.The affective component, according to Karsaklian (2008), corresponds to the evaluation of the image formed.He summarizes the positive or negative feelings and the consequent emotions, which means that it is stimulated by the emotions used in advertising and marketing arguments.The emotional factor, according to Schouten (1991), is one of the most interesting components of human behavior and, particularly, the consumer behavior.For this author, the emotions follow consumers at all times and it would not be precisely when they make decisions about their purchases (their moment to express their identity) that they would give up being humans and not express their emotions.
Arnold and Reynolds ( 2009) reinforce the idea that conducting studies in this regard has contributed to a better understanding of the effects of affection, showing that mood, feelings and emotions are related to almost all aspects of consumer behavior.This is particularly evident in retail environments, where mood has been the main subject in studies addressing the environmental influences on consumption, such as the atmosphere of retail.Therefore, the authors suggest that understanding how people deal with affection is important for the study of consumer behavior in retail.
Studies indicate that even with little knowledge about the environment, people have a strong emotional connection with it (Maloney and Ward, 1973;Dispoto, 1977), but consumers with high levels of ecological concern and affection are more likely to exhibit positive attitudes towards the environment than those who have a low level of concern and affection (Laskova, 2007).According to Zanoli and Naspetti (2001), although the relationship between affection and cognition is not yet fully understood, the cognitive interpretations (interpretations learned) of consumers influence and are influenced by affective reactions.Therefore, we assume the third hypothesis.
Hypothesis 3 -the level of ecological affection leads individuals to adopt a green purchase behavior.Martin and Simintiras (1995) suggest that the influences of ecological knowledge and ecological affection on the ecological behavior of individuals are independent variables and, therefore, both influence the behavioral responses of consumers.

Subjective or social norms (SN)
For a long time, the relationship between attitudes and behavior has been studied in the literature of social psychology (Ajzen and Fishbein, 2004).The reasoned action theory, developed by these authors, seeks to relate attitude and behavior through the mediation of behavior intention, considering the social norms, which consist of the individual's belief about how other people expect them to behave.
The subjective norms are the basis of the operation of society (Karsaklian, 2008).This author assumes that the impact of the influence group corresponds to all compartments of life, including with respect to the purchase and consumption behavior of individuals.This author also believes that this social interaction, conceived as a process that occurs between two or more individuals, where the action of an individual leads to the action of the other, is at the same time, a response to another individual and stimulus to his actions.These norms are related to predictions of how their desires will be perceived by the groups of which they are part and because of that, people are more or less willing to obey.The other consists in the motivation to comply with such behavior (Thogersen, 1999).
In the reasoned action theory or controlled behavior of Ajzen (1985Ajzen ( , 1991Ajzen ( , 2001)), Ajzen et al. (1996), Ajzen and Fishbein (2004), the behavior is determined by the intention to perform a pro-environmental behavior (or any other behavior).That intent, in turn, is determined by the attitude on the behavior (attitudinal component), by the subjective norm (normative component) and by the perceived control of the subject over the behavior.
The theory of planned behaviour (TPB) has been widely adopted as one of the most powerful tools to test the behavioral intention of consumers.According to Ajzen (2011) since its introduction 26 years ago, the theory of planned behavior became one of the most frequently cited models in the research about for the prediction of human social behaviour.Its popularity is revealed by conducting a Google Scholar search for the keyword 'theory of planned behavior or theory of planned behaviour.'From 22 citations in 1985, the number of citations per year has grown steadily to a total of 4550 in 2010 (Ajzen, 2011).Kim and Choi (2005) report that subjective norms may influence the formation of the purchasing attitudes of green products, as they may be related to the welfare of the group members.Along this same line of thought, Mahon et al. (2006) assume the idea that the subjective norms measure the influence of others with regard to the behavior of the individual.Thogersen (1999) shows that most Danish consumers developed personal standards over the choice of "environmentally friendly" packagings, revealing that personal standards are a significant predictor of their propensity to choose green products in supermarkets.
Recent results of a study conducted by Chen and Chai (2010) revealed that the personal standard was the most important contributor to the attitude of respondents regarding green products.Based on these findings, we made up the fourth research hypothesis.

Hypothesis 4: subjective norms exert influence on the ecological purchase behavior of individuals.
The assumptions formulated for each one of the variables are summarized in Figure 1.

METHOD Population, sampling plan and data collection techniques
In this study was considered women (living in São Paulo State) who are consumers of cosmetics, aged over 18 years old and living in the urban areas.The choice of this age group is justified by the fact that we found no accurate data on the female population in other age groups.The population selected is based on the fact that in Ribeirão Preto, 65,9% of total population are 18 years old or more; in addition, if considering just this age group (18 year old or more people), 53,1% are female.Thus, this study was based on convenience sampling to determine the participants of the sample.
A total of 800 questionnaires were sent and 500 questionnaires were considered valid.The discard of some questionnaires received was necessary for the following reasons: a) incompletely filled; b) returned blank and c) not returned.Data collection was performed through the personal application of self-administered questionnaires, as suggested by Cooper and Schindler (2003).One of the strategies adopted in this modality was to approach the respondents in places of large circulation.

Data collection instrument
For measuring the independent variables (ecological knowledge, ecological affection and ecological concern, subjective norms) and the dependent variable (ecological purchase behavior), the scales used in the study were based on previous studies, but duly adapted to the object of study of this research.Therefore, we prepared the survey instrument based on the adaptations of the scales proposed in the models of Maloney et al. (1975) and Ajzen (1985;1991;2001), Ajzen and Fishbein (2004).To measure the dependent variable (green purchase behavior), the questions were based on the model proposed by Kim and Choi (2005).The metric scales used were the interval, since the goal at this stage was to measure the degree of ecological knowledge, concern and affection of the respondents, and check whether these variables and subjective norms influence the adoption of a green purchase behavior, also called the Likert scale of 11 points, which were 10-strongly agree, 8-agree 5-not sure, 3-disagree, 1-strongly disagree.
Defined the scales, the research instrument was divided into distinct blocks: the first addressed issues related to "ecological knowledge."The second, had questions about the "ecological affection".The third, in turn, brought issues regarding "ecological concern".Then questions regarding "subjective norms" were posted.The fifth group presented some specific questions about "purchase behavior of ecological cosmetics".Finally, the last block consisted of questions relating to sociodemographic characteristics, age, occupation, education level and income level.
First of all, we developed a pre-test with 53 university professors.According to them there were some ambiguous questions that the researcher rephrased so as to come with the final questionnaire.The questionarie was applied in different districts to avoid income biases problems, as shopping centers and beauty salon.The data was collected during three months during 2011 in Ribeirão Preto city, São Paulo State.
The scales were evaluated for validity and reliability.Pearson's correlations were calculated between independent variables (ecological knowledge, ecological affection and ecological concern, subjective norms) and the dependent variable (ecological purchase behavior).Reliability was measured using cronbach alpha.A theoretical review is presented as follows.

Respondents description
The 500 respondents that constituted the sample of the survey indicated the following sociodemographic characteristics: most of the respondents are in the age group between 21 and 30 years old (42.6%), single (47.2%), work (63%), completed high school or is attending higher education (50.0%) and is pr ate that there are no respondents belonging to the social class "E".Table 3 brings the crossover frequency regarding the classification (1 st and 2 nd places) of the features that the respondents value in beauty products, personal care and perfumery.
Firstly, these data seem to indicate that consumers highly value the quality issue when buying cosmetics (feature "better products", chosen in the first place), and that it seems to be associated with branded products.Secondly, it is possible to observe that the feature "products with refill" is also equally valued by the respondents: Approximately 17.4% of them chose this as the most valued feature.It is also noted that products with recyclable packages (C2) also represent another feature valued by the respondents.Although this feature was mentioned by only 7.4% of them as the most prioritized feature, it was cited as to the third most valued feature.The data of prioritized features in third place are shown in Table 4. Cheaper products, as well as fashion trends have been declared as the least edominantly in classes "B" and "C".The statements indic valued features between those analyzed.The reason why "products with refill" have more frequency than "better products" and "branded products" can be probably explained because these last two features are closely related, as previously discussed.With respect to the companies in the cosmetics sector most cited under the ecological point of view, the results show that Natura, Boticário and Avon were by far the most mentioned companies by the 500 respondents (together accounting for almost 90% of the valid citations).Table 5 shows the results related to the block of questions that sought to analyze how much the respondents know about ecological cosmetics.
There is a consensus on the fact that the cosmetic companies should disclose more products about the green products they produce, since 63.6% of the respondents marked the maximum value on the scale for this question.Therefore, the suggestion verified in the scales highlighted is that these companies do more advertising campaigns that show the benefits of consuming green products and more awareness campaigns regarding the consumption of these products (95.8% marked values between 7 and 10 on the level of agreement to these questions.Thus, there is a strong consensus of the respondents regarding the need to invest more in the dissemination of green products, informing their benefits and making people aware about their consumption.The following section presents the results concerning the determination of the level of ecological knowledge, ecological concern, ecological affection, the influence of subjective norms and the purchase behavior of green products.

Determination of the level of ecological knowledge, ecological concern, ecological affection, the influence of subjective norms and the purchase behavior of green products
Initially data was examined for the presence of missing values and outliers.For each dimension, missing values were checked and excluded from the final data set.After eliminating cases which had missing values, Mahalanobis Distance was calculated for each group of variables and a t-test was ran in order to identify (=0.01%)outliers for each dimension, according to Hair et. al. (1998).Data was analysed as follows: -Purchase behavior of ecological cosmetics (PBEC): this construct owns 10 items and totalized 500 valid cases, 16 outliers were found and deleted from the data set.
-Subjective norms (SN): this construct owns 5 items and totalized 500 valid cases, 13 outliers were found and deleted from the data set.
-Ecological affection (EA): this construct owns 10 items and totalized 500 valid cases, 28 outliers were found and excluded from the data set.
-Ecological concern (EC): this construct owns 10 items and totalizes 498 valid cases, 4 outliers were found and deleted from the data set.
-Ecological knowledge (EK): this construct owns 10 items and totalizes 498 valid cases, 11 outliers were found and excluded from the data set.
After this data purifying procedure, reliability was assessed by Cronbach's Alpha coefficient.In these analysis items which reduced dimension reliability were deleted.Table 6 summarizes results.Exploratory factor analysis was conducted for each group of variables.This technique was conducted in order to recognize data structure and assess factors reliability.Principal Component Analysis was applied considering Varimax Rotation and Eigenvalue higher than 1. Results are presented on Table 7.In general, results obtained from the factor analysis are appropriate, since adherence indexes such as KMO, extracted variance, measures of sampling adequacy (MSA), comunalities and Cronbach's alpha, showed adequate numbers which indicates scales used in the present study own satisfactory adequacy and reliability.
In order to proceed to additional analysis, new variables named factors were created by calculating the avarage of all the variables which belong to that dimension, according to the results found on Table 7.For example, factor PBEC was obtained by the average of its variables: PBEC1, PBEC2, PBEC3, PBEC4, PBEC5, PBEC7, PBEC9, PBEC10.
It was found that the construct PBEC has a unidimensional structure, characterized by a single factor (composed of items PBEC1, PBEC2, PBEC3, PBEC4, PBEC5, PBEC7, and PBEC9 PBEC10).The final construct has an average of 6.70 and a standard deviation of 2.35, and the coefficient of variation equal to 35.1%.These results show that most respondents (52%) have a high degree of ecological purchase behavior (score in the composite scale greater than or equal to 7).
The construct SN has a unidimensional structure, characterized by a single factor (composed of items SN20, SN21, SN22, SN23, SN24 and SN25).The final construct has an average of 5.70 and a standard deviation of 2.94, and the coefficient of variation equal to 51.5%.These results show the existence of a large dispersion of data, where 38.6% of respondents stood on a high level on the scale (score in the composite scale greater than or equal to 7).Approximately 35.4% of the respondents were below the midpoint of the scale (score lower than 5).
The construct EA has a unidimensional structure, characterized by a single factor (composed of items EA25, EA26, EA27, EA29, EA30 and EA31).The final construct has an average of 9.33 and a standard deviation of 0.897, and the coefficient of variation equal to 9.6%.These results show a low dispersion of scores, the vast majority of respondents (95.2%) has a high degree of ecological affection (score in the composite scale greater than or equal to 7), 74.4% of whom had scores above 9 and 33.6% with maximum score (value equal to 10 on the scale).
The construct ecological concern (EC), in turn, has a structure composed of two factors: factor 1 (EC1), composed of items EC37, EC39, EC40, EC41, EC43 and EC44 and factor 2 (EC2), composed of items EC35 and EC36.It is noteworthy that Factor 1 has an average of 7.20, a standard deviation of 2.112 and coefficient of variation equal to 29.3%.These results show that the scores are relatively dispersed on the scale, but the majority of respondents (61.5%) has a high degree of ecological concern/Factor 2 (score in the composite scale greater than or equal to 7).Factor 2, in turn, has an average of 8.77, a standard deviation of 1.628 and coefficient of variation equal to 18.6%.These results show a low dispersion of scores; the vast majority of respondents (89.3%) has a high degree of ecological concern/Factor 2 (score in the composite scale greater than or equal to 7), 66.8% of those with scores higher than 9 and 42.9% with maximum score (value equal to 10 on the scale).
The construct ecological knowledge (EK) also showed a structure composed of two factors: factor 1 (EK1), composed of items 51, 52, 53 and 54, and factor 2 (EK2), composed of items 46 and 47.Factor 1 has an average of 7.39, standard deviation of 1.69 and coefficient of variation equal to 22.9%.These results show a relative dispersion of scores; the vast majority of the respondents (60.2%) has a high degree of ecological knowledge/ Factor 1 (score in the composite scale greater than or equal to 7), 21.6% of those with scores higher than 9 and 11.4% with maximum score (value equal to 10 on the scale).
Factor 2 has an average of 8.20, standard deviation of 1.92 and coefficient of variation equal to 23.4%.These results show that scores on the scale are relatively dispersed, but the majority of respondents (81.3%) has a high degree of ecological knowledge/Factor 2 (score in the composite scale greater than or equal to 7), 49.4% of them with scores higher than 9 and 29.1% with maximum score (value equal to 10 on the scale).In the following section, we present the hypothesis test on the influence of the independent variables (SN, EA, EC, EK) on the dependent variable (PBEC).

Test of hypotheses on the influence of the independent variables (SN, EA, EC, EK) on the dependent variable (PBEC)
Factors described on the item above were used as input to a regression analysis.Since this article's objective is to identify relations among purchase behavior and social norms, ecological affect, ecological concern and ecological knowledge, we propose a multiple regression model that seeks to ascertain how these constructs (SN, EA, EC and EK) impact the purchase behavior of ecological cosmetics (PBECdependent variable).Inittialy Pearson correlation (r) was calculated as presented on Table 8.The Pearson correlation r of 0.689, regarding the correlation between subjective norms (SN) and the purchase behavior of ecological cosmetics (PBEC), shows that the two-tailed significance level of 0.01, is statistically significant (sig.=0.000) , thus leading to the rejection of the null hypothesis of no correlation between the two variables and the acceptance of the hypothesis 1 formulated.
The correlation between ecological affect (EA) and the purchase behavior of ecological cosmetics (Pearson correlation r) was 0.276.Although this correlation is not strong, at the two-tailed significance level of 0.01 it is statistically significant (sig.=0.000), thus leading to the rejection of the null hypothesis of no correlation between the two variables.
For the correlation between ecological concern (EC factor 2) and PBEC, the Pearson correlation r was 0.540.This correlation, on a two-tailed significance level of 0.01, is statistically significant (sig.=0.000), thus leading to the rejection of the null hypothesis of no correlation between the two variables and to the acceptance of the third hypothesis formulated above.A second test of linear correlation of Pearson was conducted between the scores of the composite scale of ecological concern/ dimension 2 (EC factor 1) and the scores of the composite scale of the purchase behavior of ecological cosmetics (PBEC), in the same sample of respondents.
These results indicate that the Pearson correlation r was 0.229.Although small, this correlation, at the twotailed significance level of 0.01, is considered statistically significant (sig.=0.000), thus leading to the rejection of the null hypothesis of no correlation between the two variables and the acceptance of hypothesis 4 formulated.These are results that, if analyzed from the standpoint of the arithmetic mean of the level of ecological concern (7.20 and 8.77 for EC2 and EC1, respectively), can be considered as very positive because, on average, the purchase behavior of ecological cosmetics of the sample surveyed is lower (6.70).
As to the correlation between Ecological Knowledge (EK factor 2) and PBEC, the Pearson correlation r was 0.301.This correlation, on a two-tailed significance level of 0.01, is statistically significant (sig.=0.000), thus leading to the rejection of the null hypothesis of no correlation between the two variables and the acceptance of the hypothesis 5 formulated above.A second linear correlation test of Pearson between the scores of the composite scale of ecological concern/dimension1 (EK factor 1) and the scores of the composite scale of the purchase behavior of ecological cosmetics (PBEC) was performed on the same sample of respondents.The results indicate that the Pearson correlation r was only 0.170.Although small, this correlation, on the two-tailed significance level of 0.01, was considered statistically significant (sig.=0.000), thus leading to the rejection of the null hypothesis of no correlation between the two variables and the acceptance of the hypothesis 6 formulated above.
Regression analysis was, then, applied.A stepwise model was chosen in order to include in the final model just those constructs which had significant influence on purchase behavior (dependent variable).Final model (model 4) obtained R 2 =55.2%, which can be considered appropriate, for a social sciences study.Table 9 exposes coeffients for the models obtained by the regression analysis.
We observe some constructs were not kept in final model (EC factor 1 and EK factor 1), since they do not influence purchase behavior.It is noticed that social norms (SN) has the strongest influence on purchase behavior, followed by ecological concern (EC factor 2), ecological knowledge (EK factor 2) and ecological affect.In this sense, we can consider all the constructs studied have significant influence on purchase behavior.
Collinearity was assessed by tolerance and VIF Residual normality was also examined through a Kolmogorov-Smirnov (K-S) test, which found residuals follow a normal distribution (p-value=0,083).These findings guarantee model's adequacy.

DISCUSSION
In general, we can say thay there is a strong consensus of the respondents regarding the need to invest more in the dissemination of green products, informing their benefits and making people aware about their consumption.The feedback from the respondents is consistent with the study of Goleman (2009), whose idea lies in the fact that it is possible to find some products with ecolabels, but that; in general, there is still a lack of wellfounded indicators that alert buyers on the hidden impacts of what they buy.Thus, consumers will continue to sporadically or insignificantly reward the environmentally-friendly products (Golemn, 2009) and the habits that intensify the environmental threats will continue.
The correlation between subjective norms (SN) and the purchase behavior of ecological cosmetics (PBEC), showed the two-tailed significance level.Our findings are supported by the theories of controlled behavior of Ajzen (1985Ajzen ( , 1991Ajzen ( , 2001)), Ajzen and Fishbein (2004); Ajzen et al. (1996).The assumptions regarding the subjective norms discussed by these authors focus on the idea that people are influenced by perceived social pressure (not equal to social norms).
The correlation between ecological affect (EA) and the purchase behavior of ecological cosmetics is not strong but it is statistically significant.These results are in line with the assumptions of Dispoto (1977), Maloney and Ward (1973), as explained in the theoretical framework.According to these authors, even with little knowledge about the environment, people have a strong emotional connection with it.
The congruence of this study also extend to the premises of Schouten (1991), who conceives the emotional factor as one of the most interesting components of consumer behavior.As the emotions follow the consumer at all times, it would not be precisely when they make decisions about their purchases that they would fail to express them.In this regard, Sauerbronn and Barroso (2005) emphasize that consumption decisions are partly a result from the affection of people, as indicated in this study.
It is noteworthy, however, according to Laskova (2007), consumers with high levels of ecological concern and affection are more prone to positive attitudes towards the environment than those with a low level of concern and affection.Within the same approach, the findings of Martin and Simintiras (1995) suggest that the influences of ecological knowledge and ecological affection on the ecological behavior of individuals are variables that influence the behavioral responses of consumers.
For the correlation between ecological concern (EC factor 2) and PBEC, the Pearson correlation was small, but considered statistically significant.It is noteworthy, however, that these results do not eliminate the ambiguity of the dimensionality of the "environmental concern", pointed by Xiao and Dunlap (2007).These are results that, however, do not agree with the fact that the concern about environmental issues does not necessarily translate into a pro-environmental behavior (Kim and Choi, 2005;Hines et al., 1987).
Similar inconsistencies were also found by Salgado and Pretto (2007) who, at the end of their studies with 324 Brazilian consumers on the relation between social and environmental responsibility, concluded that most consumers have some knowledge on the subject, but does not take this aspect into account upon purchase.Perhaps this is due to our inability to recognize the connections between our actions and the problems resulting therefrom, which leaves us wide open to create the threats that we censor (Goleman, 2009).It is believed that this disconnection between what we do with the consequences of our actions is not necessarily the result of "the lack of knowledge", but the real "commitment" to the environment.Elkington and Hailes (1989) seem to share that thought, as according to them, people are increasingly aware of the threats to the environment, although they keep buying products for simple necessities or luxury goods, making choices that affect the environmental quality of the world where they live.These assumptions are consistent with Tanner (1999), whose study concluded that, although in recent decades there has been an increased awareness regarding the environmental threat experienced by the planet, the environmental degradation still continue.
To the correlation between ecological knowledge (EK factor 2) and PBEC indicate that the Pearson correlation was small, but at the same time, considered statistically significant.Although Tanner and Kast (2003), Kollmuss and Agyeman (2002) have suggested that empirical tests have not conclusively shown a uniformity that indicates a direct relationship between knowledge on environmental issues and pro-environmental behavior, the findings in this study contradict these assumptions by revealing the existence of correlation between each other.
Similar results, however, were also found in the studies of Dispoto (1977), whose findings showed that the knowledge about ecology, environment and pollution is the main predictor of the environmental behavior.And still, they showed a small correlation between the two variables.Similar conclusions can be verified in the studies of D 'Souza et al. (2007).These authors found that consumers better informed about environmental problems are more likely to purchase products.
It is believed that this correlation is associated with the high level of knowledge of the respondents about the environmental issues, whose arithmetic mean was 7.39 (factor 2) and 8.20 (factor 1), which opposes to the assumptions of Tagliaferro (2000), whose characteristic attributed to Brazil is of a society still poorly informed and that has not shown enough strength to threaten organizations uncommitted to environmental degradation.The findings of Maloney and Ward (1973) showed, for example, that most people indicated a low knowledge about ecological issues, but still, they believe that the ecological behavior of an individual is highly dependent on their ecological knowledge, affection and intention.
A similar study, conducted in China by Chan and Lau (2000) also proved that there is a positive relationship between these variables, but despite that, the level of ecological knowledge of the Chinese people is low and indicates a minimum green purchase behavior.The general belief for Chan (1999) is that knowledge positively influences people's behavior.
Another aspect to note is that, although the respondents know little about green products, more than 95% have a high degree of ecological affection and also a high degree of ecological concern.It is believed that this may have influenced the fact that 52% have a high degree of ecological buying behavior.These results are similar to studies Maloney and Ward (1973) who, despite showing a low knowledge of most people about ecological issues still believe that the environmental behavior of an individual is highly dependent on their ecological knowledge, affection and purchase intent.
It is also observed that, while Tanner and Kast (2003), Kollmuss and Agyeman (2002) have suggested that empirical tests have not shown a conclusive uniformity that indicates a direct relationship between the knowledge about environmental issues and pro-environmental behavior, it is noted that the results found in this study contradict these assumptions by revealing the existence of correlation between each other.The considerations concerning the formulated hypotheses and the results obtained are presented below.The results regarding the verification of the relationship between the variables analyzed are listed as follows.
Given the results reported and discussed here, it was possible to draw some conclusive analyzes.A finding to be highlighted is that, compared with other independent variables, the variable SN showed the strongest correlation with the dependent variable (r=0.689),followed by the variable EC2 (r=0.540), by the variable EK2 (r=0.301) and the variable EA (r=0.276), which is also supported by regression Beta coefficient.It is believed that this result is due to the assumptions of Ajzen (2006), as the subjective norms lead to the formation of the behavioral intent, since they take the shape of attributes of choice of individuals (Karsaklian, 2008).Since these rules are related to the predictions of how their desires will be perceived by the groups of which they are part, the individuals are more or less willing to obey them (Thogersen, 1999).Corroborating with the results found here, Solomon (2008) notes that someone else may perfectly act as an influencer, by simply making recommendations for or against the purchase of certain products without necessarily having bought or used them before.Table 10 summarizes results about hypotheses developed and tested.
As we can see, all s we can see, all hypotheses were confirmed in this research.The conclusions and limitations are presented in the next section.

Conclusions and limitations
This research sought to identify the levels of subjective norms (SN), ecological knowledge (EK), ecological concern (EC), ecological affection (EA) to later determine whether there is a correlation between these constructs and the purchase behavior of ecological cosmetics (PBEC).We found consistent results through Pearson Correlation coefficient and regression analysis.Such as we found on the literature review, correlation between the independent variables (ecological knowledge, ecological concern, ecological affection and subjective norms) and the dependent variable (purchase behavior of ecological cosmetics) are significante, it means independent variables strongly influence purchase behavior of environmentally friendly products.
Although there is an association between PBEC (purchase behavior of ecological cosmetics) and EC (ecological knowledge), this relationship is tenuous.This means that people are aware of the environmental problems, but have an ecological behavior that is not consistent with their level of knowledge in this regard.In addition, the dimensions observed in the theory on the subject studied demonstrate theoretical support for the empirical results found in this study, because overall, they indicate similarities with the theoretical framework, given the correlation between the independent variables and the dependent variable.
Given the above, it is concluded at the end of this research, the results presented and discussed are the basis for answering the central questions of the study.In addition, there are evidences that the general and specific objectives proposed were achieved.The contributions envisioned in this research are explained below.In academic terms, an important contribution provided by this study refers to the reference framework.It is worth it to point out, for example, the various concepts presented here that are not yet very familiar in the Brazilian literature, such as: ecological knowledge, ecological affection, ecological concern.
The methodological procedures also offer a number of conceptual aspects that could inspire future research in the area of green marketing, more specifically, strategic planning, consumer behavior, marketing mix management, among others.As the results of this study may contribute to the deepening of the results of new studies directed towards the purchase behavior of the various products of the cosmetics sector, it is assumed that the benefits provided by this study also extend to the scientific field.
Like any scientific research, this study also had limitations that could be overcome in further studies: a) although the indicators used were based on the literature and properly adapted to the specific object of this study, the 11-point scales used, ranging from 0 to 10, were the cause for complaint from some respondents, who felt it was a bit extensive.Scales with many points can better discriminate the respondents, but usually have a lower level of reliability; b) the sample was consistent with the goals, if compared to the size of the group studied.In addition, the sample used is restricted only to women and the city of Ribeirão Preto.The results obtained are attributed only to the female audience who was part of the sample and therefore can not be generalized or extended to the male gender and the state of São Paulo in full.Considering that the respondents of class "E" were not included in the sample studied, it is believed that the results can not be generalized to the city of Ribeirão Preto, given two important aspects: 1) this social group (E) is also a consumer of cosmetics and 2) there was a high rate of refusal of the population to participate in the study (37.5%).
Despite the numerous limitations stated herein, one of the main factors that reveal the scientific rigor used in this research is the ability to recognize the existence of restrictions that the results show.However, it is expected that the limiting aspects discussed here serve as inspiration for other researchers.Therefore, some hypotheses could be taken as a proposal for further studies in order to broaden the academic knowledge and obtain different results other than those found here: 1) establish an analysis that also considers men as consumers, since the literature leaves evidence that the male audience has proven to be a potential group of consumption of cosmetics.Thus, the results of this study will contribute to strengthen studies in this regard; 2) study issues related to the level of environmental awareness of consumers who participate in the production process of ecological goods and services; 3) check the production of green products and the provision of environmentally friendly services by companies in the city of Ribeirão Preto or even in other cities in the State of São Paulo; 4) reapply this research involving consumers from other segments, such as shoes or food; 5) investigate the consumer perception about the benefits of green products.
Finally, we believe that the reapplication of this research, in a broader context that encompasses a comparative analysis between Brazilian consumers and consumers from other countries will largely contribute to the information reported in this study.

Figure 1 .
Figure 1.Interrelationship between the research hypotheses.Source: authors.based on the literature review.

Table 1 .
Concept of green consumer.

Table 3 .
Crosstable between the features prioritized in 1 st × 2 nd place.

Table 4 .
Features of cosmetics prioritized in 3 rd place.

Table 5 .
Frequency percent of the perceptions on the knowledge of green products.
Source: Compiled by the author, based on the results of the field research.

Table 6 .
Summary of the final constructs.

Final sample size Number of Outliers Final Number of items Cronbach's Alpha
SOURCE: Data from field research.

Table 10 .
Statement of hypotheses developed and tested.