When our senses get tricked: A case about Country of Origin Effect and Ethnocentrism.

I When our senses get tricked: A case about Country of Origin Effect and Ethnocentrism By Pedro Luís Raposo Osório Veiga Master in Marketing Thesis Supervised by Raquel Meneses, PhD 2014 II “The curious task of economics is to demonstrate to men how little they really know about what they imagine they can design.” ― Friedrich von Hayek III Author’s Biography Born in 1990, Pedro Luís Raposo Osório Veiga, moved from São João da Madeira to Oporto. Graduated from Faculdade de Economia do Porto (FEP) in 2012, and in September of the same year started his professional career at Optimus, SA as Channel and Product Manager. Along with his professional career he started a Marketing Master’s Degree at FEP, having finished the curricular year in 2013. In September of 2013, under orientation of Raquel Meneses, PhD, started his Master’s thesis. Currently, and since 2013, Pedro Veiga is a Stock Manager at Parfois, SA. IV Acknowledgments I wish to thank my Professor, Raquel Meneses PhD, for the overwhelming support during this research. I thank my mother and my girlfriend for all the patience and words of support and kindness. To Rafael and Joana for all the moments of joy, and for being my second family. V Abstract Throughout the world, capital and means of production have flown from developed countries to countries under development, with the first group being unable to reverse this tendency. Regardless of this scenario, there is a type of capital which is not able to move, Country of Origin Effect (COE). A product can be made anywhere in the World, but the perceptions created in each individual, are invaluable and inseparable. Related to this phenomenon is the way Ethnocentrism affects the perceptions of what domestic production and importations are. This study focused on a specific segment, food. In particular, is analyzed the COE and Ethnocentrism effect on Olive Oil and Chocolate. This choice was based on the symmetrical prestige and notoriety of different origins: Portugal and Switzerland. Following this approached, it is portrayed as opportunistic to study, not only what characteristics of both products that are valued most by the costumers, but also if those valuations vary with the COE of each product and the level of Ethnocentrism of the individual. The bibliographic review showed several limitations regarding the design of the studies done in this field, which were, for the most part, improved in this research. A research model was developed and applied, in order to place a series of tests on the effect that subjective variables such as COE and Ethnocentrism had on objective variables regarding the product’s characteristics. Either kind of variable came from the literature review. The data collection was performed using a questionnaire and after each individual experienced each product, which resulted in 341 valid answers. The results were computed using a Structural Equation Modeling (PLS), for the Ethnocentrism effect and a Paired Sample t-test for the Country of Origin effect. Highlighted are the strong effects of COE and Ethnocentrism regarding Olive Oil. Key Words: Country of Origin, Ethnocentrism, Olive Oil, Chocolate, PLS VI Resumo Por todo o mundo o capital e os meios de produção têm sido transferidos de países desenvolvidos para aqueles em desenvolvimento, com o primeiro grupo a não ter capacidade de inverter esta tendência. Apesar deste cenário, há um tipo de capital que não migra, o Efeito de Pais de Origem (EPO). Um produto pode ser feito em qualquer parte do Mundo, mas as perceções criadas em cada individuo são valiosas e inseparáveis. Relativo a este fenómeno é a forma como o Etnocentrismo afeta as perceções daquilo que é produzido domesticamente e do que é importado. Este estudo foca-se no segmento da comida. Em particular, é analisado o EPO e o efeito do Etnocentrismo no Azeite e no Chocolate. Esta escolha teve por base o simétrico prestígio e notoriedade relativos a diferentes origens: Portugal e Suíça Seguindo esta premissa, é oportuno estudar, não só as características mais valorizadas de ambos os produtos, mas também se as avaliações dessas mesmas características variam com o EPO, e com o nível de Etnocentrismo de cada individuo. A revisão da bibliografia revelou algumas falhas no design de estudos nesta área, que na sua maioria, foram melhorados nesta analise. Para criar e aplicar um modelo de pesquisa, e efetuar uma serie de testes ao efeito que variáveis subjetivas como o EPO e o Etnocentrismo têm em variáveis objetivas como as características físicas dos produtos. Ambos os tipos de variáveis advêm da revisão da literatura. A recolha de dados foi feita através de questionário, apos cada individuo ter experimentado cada produto, tendo sido recolhidas 341 respostas validas. Os dados foram processados através de um Sistema de Equações Estruturais, para o efeito do Etnocentrismo e um Teste t para amostras emparelhadas para o EPO. Dos resultados sobressaem o EPO e o Etnocentrismo relativo o Azeite. Palavras-chave: Efeito Pais de Origem, Etnocentrismo, Azeite, Chocolate VII Contents Tables Index.................................................................................................................................... VIII Figures Index .................................................................................................................................... IX Graphics Index .................................................................................................................................. X 1Introduction ..................................................................................................................................... 1 2-Literature Review ............................................................................................................................ 4 2.1-Country of Origin Effect ........................................................................................................ 4 2.2-Ethnocentrism ......................................................................................................................... 8 2.3-Price-Quality ........................................................................................................................... 9 2.4-Hypothesis Framework ........................................................................................................ 10 3-Methodology .................................................................................................................................. 17 3.1-Empirical Framework .......................................................................................................... 17 3.2-Questionnaire ........................................................................................................................ 21 3.3 Sample and Data Collection ................................................................................................. 24 3.4 Data Analysis ......................................................................................................................... 24 4-Results and Discussion .................................................................................................................. 36 4.1 Descriptive ............................................................................................................................. 36 4.2-COE ........................................................................................................................................ 38 4.3-Ethnocentrism ....................................................................................................................... 49 5-Conclusion ..................................................................................................................................... 61 6-Bibliography .................................................................................................................................. 63 7-Attachments................................................................................................................................... 69 7.1-Questionnaire – Adults ........................................................................................................... 69 7.2-Questionnaire-Children .......................................................................................................... 77 7.3-Moderation Spreadsheet ........................................................................................................ 86 7.4-Mediation Spreadsheet .......................................................................................................... 86 VIII Tables Index Table 1Senses analyzed with the evaluation of each characteristic……………………………………..……..23 Table 2Symbols and respective meanings used in the graphical representation………………..……….28 Table 3-Adult’s Paired Sample T-test p-values ................................................................................. 38 Table 4Adults Means greatness ...................................................................................................... 43 Table 5-Adults Price confidence interval .......................................................................................... 44 Table 6-Children’s Paired Sample T-test p-values ............................................................................. 44 Table 7-Children Means greatness.................................................................................................... 48 Table 8-Price's confidence interval ................................................................................................... 49 Table 9-Cronbach's Alpha Quality ..................................................................................................... 50 Table 10-Adults’ CETSCALE Validation .............................................................................................. 50 Table 11-Adults' Path Coefficient ...................................................................................................... 51 Table 12-Adults' Moderation's t-statistics ........................................................................................ 52 Table 13-Adults' Mediation's t-statistics ........................................................................................... 54 Table 14-Children’s CETSCALE validation .......................................................................................... 55 Table 15-Childrens' Path Coefficient ................................................................................................. 56 Table 16-Childrens' Moderation's t-statistics ................................................................................... 57 Table 17-Children' Mediation's t-statistics ....................................................................................... 59 IX Figures Index Figure 1-Hypothesis Framework ........................................................................................................10 Figure 2Graphical representation of Structural Equations ..............................................................28 Figure 3-Recursive and non-Recursive Models graphical representation .........................................29 Figure 4-Reflective and Formative Indicators graphical representation ...........................................30 Figure 5-T-Statistics Formula .............................................................................................................32 Figure 6-Age Cluster Diagram ............................................................................................................34 Figure 7-Strutural Model ....................................................................................................................35 6 developed country to generate prestige. To access the power of the COE (Russel, 2006) studied the effect of positive and negative prompts towards cinema consumers in the United States and France. Taking into account the history between the two countries and the hegemony of Hollywood over the global film industry the authors classify a negative prompt as showing French consumers that the film is American and vice-versa. The results show that the COE is more powerful among French consumers, in this industry. Also taking part on the COE there are three ranges of elements: the country’s economics and physical resources and industrial capabilities, its cultural values and institutional norms, and its national government policies (Sethi S. Prakash, 1999). These elements create competitive advantages or disadvantages to corporations, and manifest themselves on the strategy adopted in the internationalization process: as portrayed by (Luciano Ciravegna, 2013) small and medium enterprises (SME) in emerging economies tend to use school based networking to reach international markets, when compared with SMEs from developed economies that value serendipitous events to expand their network and their business. The COE is also influenced by the technology fit between a country and the product produced, thus low technological capabilities perceived from a country impact the quality interpreted of a product with high technological attributes produced in that country. Although when high technological skills perceived from a country do not improve the fit between that country and a lesser technological attributed product, it (the product) sees its credibility improved in the market place (Story, 2005). The distinction between these three COE components, COA, COP an COD, and their correlation has been widely studied with Chao (1993) concluding that design and quality are two different dimensions and Insch (1995) going further to say that COD evaluation depends of the level of wealth of the country of the buyer. The latter author also suggest that product familiarity enables consumers to incorporate better the Country of Origin cue in their quality evaluation. 7 After Dichter (1962) made a case about the “made in” effect, this cue, plays a more enduring role than a Brand (Tse, 1993), potentially damaging Brand Image (Myung Soo Jo, 2003), and molding the Brand Personality (Thakor MV, 1996). Strategic Brand Alliances (SBA) are a mean for weaker brands to leverage on a stronger Brand’s Equity when entering in a new market (cross boarder SBA) (Christian Bluemelhuber, 2007). Key factors for a successful SBA are, not only product fit, class proximity between the promoted products (Bernard L. Simonin, 1998) and brand fit, similarity between brand image, but also Country of Origin fit, alignment of Product-Country Image (Christian Bluemelhuber, 2007). Further research (Jin K. Lee, 2013) have shown that when both Brands involved in a SBA have a positive COE, both of them see the attitudes towards them enhanced. Adding to this conclusion, the same authors postulate that when the hosting Brand has a stronger COE than the partner Brand, the latter absorbs this effect. Consumer involvement into advertising claims are also influenced by the country of origin of the product, this being that highly benefiting claims about product with a COE with a poor Product-Country Image produces a fragile consumer involvement and vice-versa (Peeter W.J. Verlegh, 2005) Several authors have proposed that the referred effect exists and that consumers tend to use Country of Origin to access product quality (Liefeld, 1993) .Brands are told to try to leverage their brand equity on COE, and command premium prices, since consumers tend to be very sensitive to it (Keller, 1993), but several publications provide evidence that although this might work under laboratory conditions, it would not in the marketplace, excepting hedonic products such as fragrances and wine (Kamakura, 1999). In recent years, several researchers have found that, using single extrinsic cued models generate many limitations in analyzing the Country of Origin effect. This realization provided the need to develop and apply more sophisticated models using multiple cues (Helop, 1993). The use of multiple cues, usually diminishes the power of the COE relative to other cues, being extrinsic and intrinsic. Another aspect of concern in this field belongs to the fact that COE tends to become weaker as the consumer transits from product 8 perception to attitude formation and behavioral intention (Kamakura, 1999). Multiple cued models are also being used to infer real objective quality of product or a bias towards it, based on a COE. Several other extrinsic cues were found to be moderators of the Country of Origin Effect of products, i.e. warranty, brand or store image. Related to this matter, Chao (1998) while manipulating the Country of Assembly, Country of Design and Country of Parts, of a television set demonstrated that negative Country of Assembly effect could not be compensated by a positive Country of Design. Non sensory attributes of a product are major influencers of the hedonic experience of consumption. 2.2-Ethnocentrism Ethnocentrism is mainly described as being a range of attitudes either towards a belonging group or community, or towards a non-belonging group or community (Hammond, 2003). The root of Ethnocentrism is not a closed discussion: it can be educational, cultural, genetics or an aggregate of all of these factors. Ethnocentrism is known to enable individuals to contribute to the group’s common welfare at their own cost and is wise to know that in-group preference is not the same as out-group hate. In their thesis Hammond (2003) postulate that individuals engage in ethnocentric behavior even when there is no artificial incentive for it, revealing that ethnocentricity is organic. In the same experience, the authors pay close attention to the dynamics towards those who belonging to a group, by having a common characteristic, do not participate in the group’s common welfare. Bender (2013) took a deep look into how guidebooks are created and especially into how the message in them differs when made by a foreign publisher. Analyzing Swiss guidebooks, semi-optically, the authors verify that when done by a foreign author, the use of stereotypes increases even when they are no longer true. Group membership can be based on a variety of characteristics such as language or religion. Being such a strong attitude, Ethnocentrism, has led to innumerous conflicts but also influencing consumer choice. 9 The economics term “Transaction Cost”, suggests that, the more specific the investment more attention should be paid by the organization into safeguarding this investment against any opportunistic behavior (Williamson, 1985). Svendsen (2011) reveals that the more ethnocentric the market is the less foreign-specific investment it attracts and more homespecific investment is deployed. Summarizing, Ethnocentrism, in the marketing literature (Shimp, 1984) is define as the “consumer beliefs in the superiority of their own country’s products, rooted in morality, consumer ethnocentrism is intended to capture the notion that some consumers believe that it is somehow wrong to purchase foreign made products, because it will hurt the domestic economy, cause the loss of jobs, and it is plainly unpatriotic”. Usually the lower level of income, and the lower level of education of individuals tend to promote consumer ethnocentrism, which is suggested to be based on fear of job losses from foreign competition (Shimp, 1984). In a recent study, Philippidis (2011) found that ethnocentricity and neo-phobia play a significant role in the consumption of ethnic food in Spanish restaurants. 2.3-Price-Quality Behavior theory suggests that consumers tend to prefer the lowest price for any given pool of similar products. Even so, empirical evidence (Burton, 1989), infer that consumers also extrapolate information from price, i.e. quality. Being this stated, higher prices decrease consumer’s utility, because they pay more for the products, but also increase utility since it induces higher quality (Krishman, 1985). Min Ding (2010), proposes a model based upon the following premises: consumers infer quality from a products price and consumers have a reference price from a given product. Although this price-quality relationship is not univocal, integrative research finds that most consumers find this relationship positive. This pointed, is important to understand what causes a consumer to perceive value. Andreas Herrmann (2007) defines perceived value as the ratio of perceived benefits and perceived sacrifices. This definition is validated by Grewal (1998), who stated that perceived price downwards, and perceived quality upwards, 10 influenced perceived value. Moreover, Kime (2014), by manipulating catalog messages, Grewal (1998) also found that consumers tend to be more sensible to effective price changes, than to effective quality changes. Zeithaml (1988), after several company interviews and focus groups, defines perceived quality, as being different from objective quality, mechanic and humanistic, since it represents the consumer’s judgment of a product or service. The author also suggest that objective quality does not since quality is always a perception made by someone. Min Ding (2010), proposes a model based upon the following premises: consumers infer quality from a products price and consumers have a reference price from a given product. 2.4-Hypothesis Framework Figure 1-Hypothesis Framework Source: the Author Ethnocentrism COE Price Physical Perception H1A H5 H2A H2B H3A/B/C Age Gender Education Quality H4 H6A H2C H6B H1B H1C 11 Hypothesis under study After having portrayed the theory regarding this field of investigation, it will be, in this chapter, selected several variables in order to build a model that will be the basis for this study. By selecting these variables it will be enounced the hypothesis that are going to be tested. Country of Origin Effect (COE) The Country of Origin of a product is taken into consideration by the consumer into his purchase decision process. It serves as a valid form of extrinsic information, influencing positively or negative the purchase intention (Cameron, 1994), (Thorelli, 1989). Playing a role similar to the Brand name, the “Made in” effects are usually stronger in time, than the Brand’s, according to Tse (1993).Taking into consideration the stated above, it is considered that the Country of Origin, being it what is the labeled product by/with, has a significant impact on the price attributed to it, which leads us to the first hypothesis of our model Consumer’s perception of a specific product can be greatly influenced by its Country of Origin (Johny K. Johansson, 1985). Also Liefeld (1993) suggests that the Country of Origin of a product is used by the consumer to infer the quality of that same product. Using the theoretical base stated above, and believing that the Country of Origin has a significant effect on Perceived Quality, it is proposed the following hypothesis: Country of Origin direct Effect H1A: The Country of Origin has an effect on consumer’s perceived physical characteristics of each product H1B: The Country of Origin has an effect on consumer’s perceived quality of each product H1C: The Country of Origin has an effect on consumer’s price attribution for each product 12 COE-Moderation The above stated hypothesis (H1A, H1B and H1C) are not stable and the theory states that they are influenced by several other independent variables. These variables are known as moderator variables, they affect the strength or direction of the stated hypothesis, according to their level. Moderator variables may be natural occurrences of the experiment, such as age or gender, or may be artificial/manipulated, like bad illumination or loud noises, (Ro, 2012). The moderator effect is useful for the researcher to understand the inconsistency levels of the relationships that the moderator variable moderates, but it is not the main objective of the study (Kenny, 1986). The range of factors affecting the above stated hypothesis H1A, H1B and H1C are beyond human calculation and consideration. Even so, the effect of the following will be moderated: -Age, since it is long acclaimed that age has a different effect on the way the individual interprets and evaluates a foreign product. According to Schooler (1971), younger consumers tend to prefer better imported goods, in contrast with older consumers who, generally value better domestic produced goods. -Gender has already been described as a major influencer for the preference of acquiring, or not, imported goods. Dornoff (1974) described women as being more permeable to the value represented by foreign goods, while men, are portrayed as being more patriotic and has valuing more domestic produced goods. -Formal Education has not been left out of the equation when talking about influencers of the effect of the country of origin of a product. Cunningham (1972) has long described a scenario, where formal education has a significant role in moderating the country of origin effect. Formal education is a continuum effect: the greater the education, the greater the tendency for the individual to better value goods from another country. Using the theoretical base stated above, that the Country of Origin has a significant effect on Perceived Quality, the following hypothesis are proposed: 13 Country of Origin Effect Moderators H3A1: Age profile is a significant moderator of effect between Country of Origin and the Perceived Physical Characteristics of each product H3A2: Age profile is a significant moderator of effect between Country of Origin and the Quality of each product H3A3: Age profile is a significant moderator of effect between Country of Origin and the Price of each product H3B1: Gender profile is a significant moderator of effect between Country of Origin and the Perceived Physical Characteristics of each product H3B2: Gender profile is a significant moderator of effect between Country of Origin and the Quality of each product H3B3: Gender profile is a significant moderator of effect between Country of Origin and the Price of each product H3C1: Formal Education profile is a significant moderator of effect between Country of Origin and the Perceived Physical Characteristics of each product H3C2: Formal Education profile is a significant moderator of effect between Country of Origin and the Quality of each product H3C3: Formal Education profile is a significant moderator of effect between Country of Origin and the Price of each product Although being inwards preference different from outwards hate, ethnocentrism is described by Hammond (2003) as the amplitude of attitudes either towards a belonging group or towards a non-belonging group, this including either people, places or products. Ethnocentrism direct effect H2A: Ethnocentrism has an effect on consumer’s Perceived Physical Characteristics of a given product H2B: Ethnocentrism has an effect on the Price attributed to a given product H2C: Ethnocentrism has an effect on the Quality attributed to a given product 14 Ethnocentrism-Moderator Starting from what the previous research stated above, it is opportune to test the moderator effect of Age, Gender and Formal Education in all other hypothesis H3A4: Age profile is a significant moderator of effect between Ethnocentrism and the Perceived Physical Characteristics of each product H3A5: Age profile is a significant moderator of effect between Ethnocentrism and the Quality attributed to a given product H3A6: Age profile is a significant moderator of effect between Ethnocentrism and the Price attributed to a given product H3B4: Gender profile is a significant moderator of effect between Ethnocentrism and the Perceived Physical Characteristics of each product H3B5: Gender profile is a significant moderator of effect between Ethnocentrism and the Quality attributed to a given product H3B6: Gender profile is a significant moderator of effect between Ethnocentrism and the Price attributed to a given product H3C4: Formal Education profile is a significant moderator of effect between Ethnocentrism and the Perceived Physical Characteristics of each product H3C5: Formal Education profile is a significant moderator of effect between Ethnocentrism and the Quality attributed to a given product H3C6: Formal Education profile is a significant moderator of effect between Ethnocentrism and the Price attributed to a given product 15 Ethnocentrism-Mediator Understanding that Price is a function, and a quantification, of quality (value), it is important to this study to represent the relationship between Product Perception and Price, and since Product Perception is a function of the Ethnocentricity level of the consumer, it plays a mediator role in the relationship between Ethnocentricity and Price. Mediator variables, such as these, add information and are portrayed as an active organism to the form of how the association, between the independent variable and the outcome variable, occurs (Ro, 2012). The mediator variable provides the study and the researcher a causal chain of effects (Kenny, 1986) and an explanation of how physical events impact internal psychological significance. In the present study, the way Ethnocentricity (independent variable) influences the proposed Price (outcome variable) for either Olive Oil or Chocolate through the Product Perception (mediator variable). Accounting for the theoretical base above, it is found to be relevant to study the mediator effect of Perception on Price and on Quality through the following hypothesis: H4: Perceived Physical Characteristics has a mediator effect on the Quality attributed to a given product, from Ethnocentrism H5: Perceived Physical Characteristics has a mediator effect on the Price attributed to a given product, from Ethnocentrism Conclusion The chapter that now ends was dedicated to present the main concepts and definitions related to the Country of Origin Effect, Ethnocentrism and Price in order to establish a base of comprehension of the field of study. It was presented the ways in which Country of Origin gets divided in to, from Country of Design to Country of Assembly through Country of Production, and how each one of these affects consumer behavior. 22 information regarding the products, and not only the Country of Origin. Next it was clearly stated the Country of Origin of all products, and not just the geographical area, and by doing this it can robustly be defined the influence that each Country has on the perceptions of its products, in this case Olive Oil and Chocolate. Lastly, and to get the most truthfully information out of each individual, the study was presented to the population not as a market research but as a help request from a local store, to not give the respondents the feeling of judgment, since the theory states that what is stated under questionnaire is usually not what the individuals really do (Kamakura, 1999). All questions present in the questionnaire were developed to be a reflection of the variables of the model created, since these will be the inputs that will be the measuring of those same variables. Although most of the questions were elaborated by the author, there are several that were based upon already tested scales. The questionnaire is formed by 76 questions, either multiple choice or Likert scaled questions and divided into five different sections. Siskos (1995) found out that perceived quality, for both consumers and distributors, was linked to sensory properties such as taste, aroma, color, appearance and texture. Region of Origin, composed by local agronomic conditions, traditional human know how and raw product characteristics, but also influenced by PDO (Protected Denomination Origin) were a property linked to perceived quality found by Dekhili (2009). The same author found that price and the size of the container were major influencers of consumer choice. The first and second section, were related to, respectively, olive oil and chocolate. Individuals were asked to answer several scale questions related to physical features of olive oil and chocolate, which, as stated by Cameron (1994), are categorized as intrinsic characteristics. These questions were about several physical characteristics which are considered to be objective, sensory based, when evaluation each products. In order to access the quality perceived from each feature, scale questions varied from 1 to 5 (1Very Low; 2-Low; 3Indeferent; 4-Sigificantly; 5Very Significantly). 23 Table 1Senses analyzed with the evaluation of each characteristic Source: the Author After the scale questions, given a reference price, the individuals were asked to attribute a price to the Portuguese sample of each product with an open ended question. To finalize each section several multiple choice questions regarding Branding and Packaging were asked. Although these final questions were not taken into consideration in our model, we found it to be interesting and related to our topic of study. In these sections of the questionnaire all the questions were elaborated by the author. The fourth section is composed strictly by the Consumer Ethnocentric Tendencies Scale (CETSCALE) (Sharma, 1987), an internationally validated scale. In order to apply this scale in Portuguese and to rightly perform this measurement, the scale was translated by the author to Portuguese and then asked to a native speaker to translate it back to English, with the latter being very similar to the original version. Along with the translation, the CETSCALE, as did the whole questionnaire, was also adapted when performed on the Children sample. This section Composed by 17 Likert questions varying as the previous section, from 1 to 5 (1Completely Disagree; 2-Desagree; 3Indifferent; 4-Agree; 5-Complety Agree), serves as a way to access Ethnocentrism of an individual. Characteristic Sense Color Sight Smell Smell Acidity Bitterness Spiciness Brightness Sight Cocoa Aroma Smell Cocoa Flavor Milk Flavor Thikness Melt Crispiness Taste Touch Chocolate Olive Oil Taste 24 Ethnocentrism is impossible to measure, but the CETSCALE measures, in different dimensions, reflections of this attribute. Since this study hypothesizes an effect of Ethnocentrism on consumer quality perception and price evaluation, it was found this to be an appropriate scale to use. 3.3 Sample and Data Collection Since our study requested the personal participation from each respondent, it was tried to reach a population with an eased access for the author. Although the sample is heterogeneous, the preference gave to students is justified only by the facilitated access: exploring the author’s academic and personal network the study was performed at São João da Madeira’s Senior College, Faculdade de Economia do Porto, at an Elementary and High school of Santa Maria da Feira and on a Free Fair. No pre-requisites were imposed to the individuals in order to participate in our study. Regarding under aged individuals, the permission from the schools involved was requested and all the procedure took place under the respective classes’ teacher presence. Being sensitive to the fact that the experiment would be performed with under aged children, was decided to adapt the questionnaire to this target. It was done so by, alongside with the teachers from the elementary school, simplifying some of the vocabulary of the questionnaire. Structurally, the only change made when simplifying the questionnaire was related to the question about the price, of both products: instead of asking the respondent of the experiment for a specific price, it was asked if the compared product should be more expensive, cheaper or of equal price. When the tasting was conducted with this target, this took place in the classroom with the respective teacher always present. 3.4 Data Analysis The data collected through the questionnaires was treated differently regarding the method employed: while the effect caused by the Country of Origin on each product’s physical characteristics perceptions and quality was analyzed using the Paired Sample T-test, the effect of Ethnocentrism on physical characteristics’ perception, quality and price was researched using a Structural Equation Model (SEM). 25 Paired Sample T-test As it was stated in the introduction, this study is focused on analyzing the impact of the Country of Origin on the physical perceptions of two different products: olive oil and chocolate. The reason of choice, regarding this two products was the symmetrical Country of Origin Effect when related to Portugal: olive oil with a positive Country of Origin Effect and chocolate with a negative Country of Origin Effect. The Country of Origin Effect of this products is also symmetrical when related to Switzerland: olive oil is negative and chocolate is positive. In order to understand the role played by the Country of Origin, is key to analyze, for every characteristic of each product, if the distribution of answers is independent from the Country of Origin. To execute this analysis it will be used the Paired Sample T-Test. This test’s output enables the researcher to access if two paired samples have similar means (Maria Helena Pestana, 2008). Usually each observation in analyzed twice, forming paired samples, which differences are tested in order to understand if the result is equal to zero. This is the scenario portrayed in this study, since each observation, individual, tasted both the chocolate and the olive oil twice, but with different stimulus, regarding the Country of Origin. The difference between each pair of observations, in this case the pair constituted by the score gave, to the same characteristics regarding different origins of the same product, by each individual, is calculated by d = X1i-X2i . This stated, the test hypothesis came as: H0: μd = 0 H1: μd ≠ 0 26 In order to calculate the t-value, the formula to employ is: T = ( - μd)/(s’d / √n) ∩ T(n-1) Where: = (∑di)/n ; sd = (∑(di- )2/(n-1) Being d’s mean and d’s variance s’2 Ethnocentrism – (SEM) After collecting all the data throughout the several institutions that cooperated with this study, it was necessary to estimate the model that would enable the analysis and process all the data. Below it will be discussed the estimation method and why it is appropriate for the model in study. Next, it will described several Structural Equation Models (SEM), and the reason why it was chosen PLS (Partial Least Square) Estimation Models of 1st and 2nd Generation Structural Equation Models (SEM) is a generalized statistical technique used in order to measure and test the validity of theoretical models which is defined by hypothetical and causal effects between variables. (Marôco, 2014) Regarding the 1st Generation models, Haenlein (2004) was the pioneer in referring that it has several limitations. The authors state that this kind of models are only applicable when the researcher have a research model with a simple structure, this being a model with only one dependent variable, and several independent variables. The same authors present another limitation of 1st Generation models, which is due to the fact that this kind of models assume that all the variables are measurable. This fact restricts any research model containing reflexive variables, since these are not measured directly and which measurement is made through reflections, manifestations. Also referred by the authors, is the lack of error measurement, when using 1st Generation models. d d d d 27 The 2nd Generation Models, especially Structural Equation Models (SEMs), present themselves as being superior when compared to 1st Generation ones, since they suppress the limitations point out above. Firstly they can incorporate simultaneously several dependent and independent variables, they are able to consider non-directly observed variables, and lastly consider and measure the estimation errors for the observed variables. Distinction between observed and non-observed variables was presented by Marôco (2014) , being observed variables the ones with direct measurement, and non-observed, or latent variables that manifest themselves through indicators. The research model of this study is constituted by both observed variables and latent variables, and being so, 1st Generation Models can be excluded. Structural Equation Models LISREL (Linear Structural Relationship) is a Structural Equation Model (Karl G. Jöreskog, 1982), which analyzes the effects between both observed and latent variables (Marôco, 2014). As stated before, one difference between 1st and 2nd Generation Models is the capacity that the later has to enable the researcher to include latent variables in his model. Latent variables are measured by a scheme of relationships of cause and effect. These relationships build a pool of both direct and indirect connections called paths, where each direct path is an equation. Structural Equation Models is formed by the pool of equations present in the research model and divided between the measurement model and the structural model (Figure 3). The structural model is composed by the effects between the exogenous (ξ) and endogenous (η) latent variables, enabling to measure and identify the effects between them. The measurement models provide the effects between the latent variables and their indicators. 28 Figure 2Graphical representation of Structural Equations Source: the Author Table 2-Symbols and respective meanings used in the graphical representation Source: the Author Structural Equation Models are estimated in two different ways, with two different algorithms:  Covariance Structure Model – CSM - are run upon algorithms which describe the variance and covariance of each variable  Partial Least Square – PLS – uses a formula which is focused on error minimization Symbol Meaning ξvector of latent exogenous variables xvector of measures of predictor variables δvector of errors of measurement of x λx matrix of coefficients or loadings of x on the latent exogenous variables ηvector of latent endogenous variables yvector of measures of dependent variables εvector of errors of measurement of y λy matrix of coefficients or loadings of y on the latent endogenous variables ϒregression coeficient from ξ to η →relation (from cause to effect) 11 1 Measurement Model Measurement Model Structural Model λx11 λx21 221 λx31 221 ϒ11 λy11 λy21 δ1 λx11 δ2 λx11 δ3 λx11 X1 λx11 X2 λx11 X3 λx11 ξ1 η1 Y1 Y2 ε1 ε2 29 Regarding these two ways for model estimation, there are four major differences. Firstly, Partial Least Square (Nobre, 2006) uses an interactive sequence of ordinary least squares (OLS), and analyzes one variable at a time. Using this process, it enables the minimization of the residual variance of all dependent variables by applying multiple linear regressions to the latent variables estimates (Tobias, 1995). The type of model that CSM and PLS can estimate is also different, recursive and non-recursive (Figure 4). According to Marôco (2014) in non-recursive models a variable can be both cause and effect and get effect of another variable, on the other hand, in recursive models a variable can either be cause or get effect by another variable, but never both. Regarding this matter, Partial Least Square, has a limitation since it only can be used in order to estimate recursive models. Figure 3-Recursive and non-Recursive Models graphical representation Source: the Author 11 1 λx11 λx21 221 λx31 221 ϒ11 λy11 δ1 λx11 δ2 λx11 δ3 λx11 X1 λx11 X2 λx11 ξ1 η1 Y1 ε1 ε2 11 1 No Recursive Models λx11 λx21 221 λx31 221 ϒ11 λy11 λy21 δ1 λx11 δ2 λx11 δ3 λx11 X1 λx11 X2 λx11 X3 λx11 ξ1 η1 Y1 Y2 ε1 ε2 X3 λx11 Recursive Models Y1 λy21 30 The third difference between CSM and PLS is related to the effects supported by each type of variable, latent and manifest variables. Regarding the first type of effect, formative, latent variables are formed by manifest variables. These later variables can be positively or negatively correlated and do not have the need to be in the same conceptual dimension (Marôco, 2014). Another type of effect, reflective, latent variables manifest themselves on manifest variables. In reflective effects, a pool of manifest variables, being the manifestation of a latent variable, their correlation must always be positive, and is always in the same conceptual dimension. Figure 4-Reflective and Formative Indicators graphical representation Source: the Author Regarding this features, PLS presents one advantage over CSM since it allows easily both types of effects, contrasting with CSM which formative effects are harder to represent. One fourth difference between CSM and PLS is related to the statistical tests, because while CSM requires parametrical assumptions, Partial Least Square does not. This means that there is no need for hypothesizing for the distribution of the observed variables. All statistical inferences are provided via Bootstrapping. Choosing Estimation Model In order to choose one type of model, there was a need to take into account the characteristics of the variables present in the research model. Since all variables present in the model are latent, immediately was chosen a 2nd Generation Model. After this, it was needed to understand how each algorithm works, to access which one would be more suited for this study. 11 1 Reflective Models λx11 λx21 221 λx31 221 λy11 λy21 δ1 λx11 δ2 λx11 δ3 λx11 X1 λx11 X2 λx11 X3 λx11 ξ1 η1 Y1 Y2 δ1 λx11 Formative Models 31 Taking into account all the considerations provided previously, where it was provided the main characteristics and differences of both CSM and PLS. As matter of statistical rigor it was decided to choose a PLS model, since its algorithm takes just one variable latent into consideration at a time, which provides dependent variables with minimal residual variance when applying the multiple linear regressions to the latent variables. Moderation The field of study’s literature is rich in pointing out several moderators to the effect of Ethnocentrism and Country of Origin, ranging from age (Schooler, 1971), to education (Tsiros, 1995), through gender (Nes, 1982), which is worth to mention, contradicts the position took by Hammond (2003), who states that Ethnocentrism is innate. This state of the art enabled this study to seek moderation effects to several relationships of the County of Origin effect and Ethnocentrism model proposed in figure 2. With this purpose, this study will not only seek for significant moderation effects but also for validation of the hypothesis stated in chapter 2.4 for different sub groups in the Adults and Children’s samples . To access the significance of each moderator on the created model it was used Smart PLS and ran the model via bootstrapping, for each sub-sample in order to get the Path Coefficients. After knowing the results, are verified the statistical significance of the moderators by calculating the t-statistic, meaning this that there is rejection for the null hypothesis if the pvalue is lesser than 0.05 for a significance level of 5%. The t-statistic is calculated using the formula below: 38 In this part of the chapter will be presented the results of our model regarding the major effects impacting quality perception and price of olive oil and chocolate. The whole sample was divided into two: adults and children. This division was made at 16 years old, being , anyone older than 16 years old considered and measured as an Adult, and anyone youger considered and measured as a Child. It was choosen 16 years old as spliting point since, in Portugal, this is considered to be the age where and individual is elegeble to work (Chapter IV, 2012), which inferres greater maturity and cognition and imposes the need to analize differently. From this point forward, all analysis will be divided between adults and children. 4.2-COE The main purpose of this thesis was to evaluate in which way the Country of Origin effect was going to affect the perceptions of the participants in the questionnaire. For this analysis it was divided the sample in two groups: Adults and Children. After this first division, it will be analyzing the data between males and females, age cluster and between levels of education. Adults-COE Table 3-Adult’s Paired Sample T-test p-values Source: the Author Adults Males Females College Graduates non-College Graduates Yound Adults Middle Aged Adults Old Aged Adults OliveOilQuality .000 .000 .000 .000 .000 .000 .000 .080 OliveOilAcidity .840 .820 .291 .292 .171 .443 .774 .333 OliveOilBitterness .000 .218 .602 .023 .003 .001 .043 .453 OliveOilSpiciness .664 .241 .370 .649 .310 .145 .030 .323 OliveOilDensity .003 .221 .023 .000 .567 .006 .038 .800 OliveOilColor .000 .092 .659 .000 .000 .000 .000 .030 OliveOilSmell .000 .479 .791 .000 .383 .002 .000 .897 ChocolateQuality .677 .084 .364 .355 .676 .800 1.000 .701 ChocolateBrightness .002 .381 .194 .007 .141 .011 .165 .323 ChocolateMelts .411 .929 .257 .864 .163 .196 .895 .762 ChocolateCocoaTaste .854 .339 .579 .935 .715 .366 .536 .040 ChocolateCocoaAroma .143 .178 .134 .842 .052 .923 .722 .002 ChocolateMilkTaste .037 .207 .959 .010 1.000 .022 .698 .872 ChocolateDensity .681 .297 .376 .755 .791 .331 .626 .772 ChocolateCrispiness .245 .025 .145 .241 .635 .056 .676 1.000 39 When analyzing the adults sample as a whole for the effect of the Country of Origin on the perceptions of physical characteristics regarding the two products, it was found, for olive oil, that only Acidity and Spiciness registered a p-value greater than our significance level of 0.05, which accept the null hypothesis of no difference between the means of the two variables related to this characteristic. Quality, with a p-value of 0.00, gets the null hypothesis rejected for the absence of difference between the means of the two related variables, and validates H1B. The other variables, Bitterness, Density, Color and Smell, registered p-values of, respectively, 0.00, 0.03, 0.00 and 0.00, these results partially validate hypothesis H1A for Olive Oil. Splitting the analysis in gender, and starting with males, is found a severe significance regarding the absence of difference between variable’s mean , with all variables, excepting Quality registering p-values greater than our significance level of 0.05, which makes a case for the acceptance of the null hypothesis, in this case H1A gets rejected. Quality registered a p-value of 0.000, resulting in the rejection of the null hypothesis, and a validation of H1B. The case for female individuals is similar to that of males, but Density also got the null hypothesis rejected , with a p-value of 0.023 , meaning that the difference between the variable’s mean is different than 0. For this sub group all other variables regarding olive oil registered p-values greater than 0.05 which is the base for accepting the null hypothesis of no difference between the variable’s mean. Quality, with a p-value of 0.000, also gets the null hypothesis rejected, representing a difference between the related variable’s mean. These results validate hypothesis H1B These results make a case for the rejection of H1A for both genders. 40 Analyzing the sub-groups of college graduates and non-college graduates, both sub-groups reject the null hypothesis for Bitterness and Color, with, respectively, p-values of 0.023 and 0.003, 0.00 and 0.00. These results mean that there is a difference between the means of the variables regarding the stated characteristics. Adding to the previous stated results, for college graduates was found a base for rejecting the null hypothesis regarding Density and Smell, with p-values of 0.00 and 0.00 respectively, which makes Acidity and Spiciness, for this group, the only characteristics with no difference between the variable’s mean. Quality, for both sub-groups got a p-value of 0.00, which makes a rejection out of the null hypothesis and a validation of hypothesis H1B The stated reject hypothesis H1A for non-college graduates and partially validate hypothesis H1A for college graduates. When the adults sample is divided by age, Young Adults found p-values smaller than the significance level of 0.05 for Bitterness, Density, Color and Smell, respectively, 0.001, 0006, 0.000 and 0.002. On the other hand, for the same group with p-values of 0.443 and 0.145, respectively Acidity and Spiciness, got the null hypothesis validated for the absence of difference between the means of the respective variables. Regarding Middle Aged Adults, the only characteristic with the null hypothesis validated was Acidity, with a p-value of 0.774. All other characteristics, Bitterness, Spiciness, Density, Color and Smell, with p-values, of 0.043, 0.030, 0.038, 0.00 and 0.00, have the null hypothesis rejected which reflects the difference between the variables related to those characteristics. Finally, Old Aged Adults, have symmetrical portray, when compared to Middle Aged Adults, having just one characteristic, Color, rejecting the null hypothesis, with a p-value of 0.030, smaller than the significance level of 0.05. Acidity, Bitterness, Spiciness, Density and Smell, registered pvalues of 0.333, 0.453, 0.323, 0.8 and 0.897, respectively, making the null hypothesis valid of no difference between the variables’ mean. The presented results validate hypothesis H1A for Middle Aged Adults, partially validate for Young Adults and reject for Old Aged Adults. For Young Adults, Middle Aged Adults and Old Aged Adults, the null hypothesis got validated for the latter group, with a p-value of 0.080. The first two groups, both with a p- 41 value of 0.000 regarding Quality, reject the null hypothesis of absence of difference between the related variable’s mean. Focusing on the chocolate’s characteristics the adults sample all together, registered only two variables, Brightness and MilkTaste, with p-values smaller than the significance level of 0.05, rejecting the null hypothesis of no mean’s difference between the variables related to those characteristics. All other variables, Melts, CocoaTaste, CocoaAroma, Density and Crispiness, registered p-values of, respectively, 0.411, 0.854, 0.143, 0.681 and 0.245, all greater than 0.05, our significance level, validating this way the null hypothesis of no difference between the means of the respective variables. These results completely reject hypothesis H1A for the adult sample, regarding chocolate. Quality, with a p-value of 0.677, gets the null hypothesis validated for the absence of difference between the related variables, which makes rejection out of the hypothesis H1B. Splitting the adults sample between gender, as was did before, males registered only one characteristic with a p-value smaller than the significance level of 0.05, Crispiness, this meaning a rejection of the null hypothesis of no difference between the variables’ mean related to this characteristic. All other characteristics, registered p-values greater than the significance level: Brightness, Melts, CocoaTaste, CocoaAroma, MilkTaste and Density with, 0.381, 0.929, 0.339, 0.178, 0.207 and 0.297 which makes the null hypothesis valid for the variables of these characteristics. Regarding the female sub-group all variables registered p-values greater than 0.05, validating for all of them the null hypothesis of no difference between the variable’s mean. Taking into consideration these results the hypothesis H1A is rejected for both genders. Quality with a p-value of 0.084 for males, and 0.364 for females, validates the null hypothesis for no difference between the related variables, which lays the base do reject hypothesis H1B. Taking a closer look into the adults sample and dividing the sample between college graduates and non-college graduates, the first group behaved similar to the adults sample as a whole rejecting only Brightness and MilkTaste with p-values of 0.007 and 0.010. All other 42 characteristics, Brightness, Melts, CocoaTaste, CocoaAroma, MilkTaste and Density, with p-values of respectively, 0.864, 0.935, 0.842, 0.755 and 0.241 validated the null hypothesis of no difference between the related variables’ mean. Non-college graduates, registered a similar scenario as did females, with p-values, for all characteristics greater than the significance level, 0.05, meaning this, the validation of the null hypothesis of no difference between the related variables’ mean. The results presented make the base for rejection of hypothesis H1A for both college and non-college graduates. Quality, for both college and non-college graduates, with respectively p-values of 0.335 and 0.676, validated the null hypothesis for absence of difference between the related variable’s mean. These results reject hypothesis H1B. Dividing the adults sample by age cluster, the young adults, regarding chocolate characteristics, present two characteristics with p-values smaller than the significance level, Brightness and Milk Taste, as did college graduates and the sample as a whole. Brightness with p-value of 0.11 and Milk Taste with p-value of 0.022 reject the null hypothesis, meaning a significance in the difference between the related variable’s mean. All other characteristics, Brightness, Melts, CocoaTaste, CocoaAroma, MilkTaste and Density, with p-values of 0.196, 0.366, 0.923, 0.331 and 0.56, respectively, validate the null hypothesis. Middle aged adults, show a similar behavior as females and non-college graduates, with all characteristics having a p-value greater than 0.05 and by this validating the null hypothesis. Old Aged Adults, portray two variables with p-values smaller than the significance level of 0.05, CocoaTaste and CocoaAroma, respectively 0.040 and 0.02, rejecting the null hypothesis of no difference between the related variables means. All other variables, Brightness, Melts, MilkTaste, Density and Crispiness, with p-values of 0.323, 0.762, 0.872, 0.772 and 1, all greater than our significance level validate the null hypothesis. Attending to these results H1A gets rejected for all three age clusters. Regarding Quality all age clusters, Young Adults, Middle Aged Adults and Old Aged Adults validated the null hypothesis, with p-values of respectively, 0.8, 1 and 0.701. With these results H1B gets rejected. 43 Adults-COE-Price and Quality There is no way to understand if more of one characteristic is necessarily better than more of another. The characteristics regarding our products are simply that, characteristics, neither good nor bad, that is why the objective of this study is to understand if the Country of Origin affects the perceptions regarding these characteristics. The only exception is Quality, which more Quality is necessarily better, and less Quality is necessarily worst. Regarding Quality, the Paired Sample T-Test demonstrates the that there is a significant mean difference between the variables related to olive oil quality , for all sub-groups excepting Old Aged Adults, but does not show which is greater, if the Portuguese or the Swiss. Since olive oil was the product with more characteristics significantly different, our study shows on table 5 which mean is greater, using the difference between the samples’ mean, of both Portuguese olive oil and Swiss olive oil. The results show that for all cross sections, including Old Aged, although not significantly, Portuguese olive oil registers the greatest quality evaluation. Table 4Adults Means greatness Source: the Author Since our values from the Paired Sample T-test regarding chocolate quality revealed no significant difference between the two samples the same analysis will not be performed. The last paragraphs are the perfect transition to analyze the variable price. The reference price present in the questionnaire regarding Swiss olive oil was 5€. The table 6 shows the confidence interval with a confidence level of 95%, for the distribution of prices attributed to Portuguese olive oil, and portrays that, for all sub-groups, this interval contains the price for Swiss olive oil. This fact, contrasting with the table 5 shows that, although recognizing superior quality for Portuguese olive oil, individuals are not willing to pay more. This scenario is aligned with Kamakura (1999), when the individuals are affected by the COE but this effect is not materialized in different behaviors. Adults Males Females College Graduates non-College Graduates Yound Adults Middle Aged Adults Old Aged Adults OliveOilQuality PT PT PT PT PT PT PT PT 44 Table 5-Adults Price confidence interval Source: the Author As already stated and showed before, there was no significant difference between Portuguese and Swiss chocolate quality, and since the analysis regarding quality was done, it is wise to relate those values with the price attributed to Portuguese chocolate. The absence of difference between chocolate’s evaluations contrasts with the difference on the price attributed to Portuguese chocolate. The reference price for Swiss chocolate, was 5.6€, and for a significance level of 95% the confidence interval for the whole sample, for College and non-College graduates, Young Adults and Middle Aged Adults, was always below the reference price, which once again might be extrapolated as an inability to significantly recognize the quality of the foreign product, but being willing to pay more for it. Children-COE Table 6-Children’s Paired Sample T-test p-values Source: the Author Adults Males Females College Graduates non-College Graduates Yound Adults Middle Aged Adults Old Aged Adults Upper OliveOilPrice 5.1735 5.5435 7.3017 5.2126 5.4117 5.1991 5.7267 5.5698 Lower OliveOilPrice 4.6343 4.5355 3.9737 4.5984 4.3911 4.5131 4.2595 4.4046 Upper ChocolatePrice 4.6249 7.9891 6.8875 4.5820 4.8997 4.4223 4.7676 5.6807 Lower ChocolatePrice 4.2345 3.3005 3.5634 4.1005 4.2319 3.9733 3.6600 4.8240 Children Males Females 4thGraders 6thGraders 9thGraders OliveOilQuality .000 0.013 0.00 .000 0.693 0.024 OliveOilAcidity .000 0.051 0.00 .000 0.162 0.012 OliveOilBitterness .044 0.745 0.01 .000 0.480 0.045 OliveOilSpiciness .001 0.001 0.146 .012 0.052 0.129 OliveOilDensity .641 0.425 0.84 .918 0.781 0.259 OliveOilColor .000 0.005 0.00 .000 0.071 0.003 OliveOilSmell .297 0.455 0.52 .000 0.022 0.247 ChocolateQuality .003 0.002 0.24 .008 0.000 0.009 ChocolateBrightness .000 0.000 0.01 .000 0.012 0.404 ChocolateMelts .221 0.308 0.49 .448 0.021 0.083 ChocolateCocoaTaste .097 0.398 0.11 .478 0.031 0.558 ChocolateCocoaAroma .362 0.559 0.47 .229 0.002 0.503 ChocolateMilkTaste .295 0.041 0.65 .140 0.208 0.001 ChocolateDensity .362 0.934 0.16 .302 0.004 0.129 ChocolateCrispiness .044 0.594 0.02 .508 0.045 0.504 45 As done with the adult’s sample, this study will also analyze the results regarding the difference between the variables related to all the characteristics of the two products used, olive oil and chocolate. Considered for this study as Children, were every individual younger than 16 years old, (Chapter IV, 2012). The data was collected at different education establishments: at an elementary and middle school, due to facilitated accessibility. From this work resulted 134 responses but with 3 being withdrawal for not having responded to more than 10% of the questionnaire Starting with olive oil, table 11 shows that the children’s sample as a whole registered two characteristics with p-values greater than 0.05. In fact, Density and Smell, with p-values of 0.641 and 0.297, respectively validated the null hypothesis of no difference between the related variables’ mean. All other variables, Acidity, Bitterness, Spiciness and Color, registered p-values smaller than 0.05 which rejects the null hypothesis. These results make a case for a partial validation of hypothesis H1A. Quality, with a p-value of 0.000 reject the null hypothesis for absence of difference between the related variable’s mean. These results plant the base for validation of hypothesis H1B Dividing the sample between genders, male children, registered for Spiciness and Color, pvalues smaller than the significance level, respectively, 0.001 and 0.005, which makes a case for rejection of the null hypothesis and the absence of difference between the related variables’ mean. Acidity, Bitterness, Density and Smell, with p-values of 0.051, 0.745, 0.425 and 0.455 validate the null hypothesis. The female sub-group, within the children sample registered for Acidity, Bitterness and Color, p-values smaller than the significance level, rejecting this way the null hypothesis for the absence of difference between related variables’ mean. Spiciness, Density and Smell, with p-values of 0.146, 0.084 and 0.52 validate the null hypothesis. These values result in a rejection of hypothesis H1A for male children and a partial validation of the same hypothesis for females. 46 Quality with a p-value of 0.013 for males and 0.00 for females, reject the null hypothesis of absence of difference between related variable’s mean, which validates hypothesis H1B Within the children’s sample, this study focused on different levels of education. Firstly the 4th graders, showed a p-value greater than the significance level for just one characteristic, Density. The stated variable, had a p-value of 0.918, which validates the null hypothesis. All other variables, Acidity, Bitterness, Spiciness, Color and Smell, with p-values of, respectively 0.000, 0.000, 0.012, 0.000 and 0.000, all lesser than 0.05, reject the null hypothesis of the absence of difference between the related variables’ mean. The 6th graders sub group, behaved symmetrically to the 4th graders with only one characteristic with pvalue smaller than 0.05, Smell. With p-value of 0.022, Smell gets the null hypothesis rejected, which indicates a difference between the related variables’ mean. All other variables, Acidity, Bitterness, Spiciness, Density and Color, registered p-values greater than the significance level of 0.05, and by this validating the null hypothesis. Lastly 9th graders, showed a p-value greater than 0.05, the significance level, for Spiciness, Density and Smell, making the null hypothesis valid and the difference between the related variables’ mean absent. Acidity, Bitterness, and Color, with p-values of 0.012, 0.045 and 0.003, respectively, reject the null hypothesis. These results validate the hypothesis H1A for 4th graders, partially validate the same hypothesis for 9th graders and reject for 6th graders. Regarding Quality, 4th graders and 9th graders, with p-values of 0.00 and 0.024 respectively, reject the null hypothesis which makes the difference between the related variable’s mean different than 0. Analyzing 6th graders, with a p-value of 0.693, gets the null hypothesis validated. These results partially validate H1B Focusing this analysis on chocolate, regarding the children’s sample as a whole, Brightness and Crispiness are the only variables with p-values smaller than the significance level of 0.05, respectively, 0.000 and 0.044. These results reject the null hypothesis. All other variables, Melts, CocoaTaste, CocoaAroma, MilkTaste and Density, with 0.221, 0.097, 0.362, 0.295 and 0.362 being their respective p-value, validate the null hypothesis of no difference between the related variable’s mean. These results make a rejection out of hypothesis H1A. 47 Quality with a p-value of 0.003 rejects the null hypothesis for no difference between the related variables, which validates hypothesis H1B. Within the children’s sample, and as was done with olive oil, the analysis will be focused on the sub-groups regarding gender. Starting with males, and as the sample all together, Brightness, with p-values of 0.000 got rejection for the null hypothesis. Along with the latter one, males also reject the null hypothesis for the absence of difference between the variables’ mean related to MilkTaste, with a p-value of 0.041. All other variables, with p-values greater than the significance level of 0.005, validate the null hypothesis. Female Children, with pvalues of 0.01 and 0.02, reject the null hypothesis for Brightness and Crispiness. For Melts, CocoaTaste, CocoaAroma, MilkTaste and Density, female children reported a p-value of, respectively 0.49, 0.11, 0.47, 0.65 and 0.16, which by being greater than 0.05 validate the null hypothesis for the absence of difference between the related variables’ mean. These results reject hypothesis H1A for both males and females. Related to quality, the behavior of both genders, was very different, while males with a pvalue of 0.002 demonstrated absence of difference between the related variables mean, females with 0.24 as a p-value did not. These results, validate H1B for males, but not for females. Addressing now the results related to the sub groups constituted by the 4th graders, only Brightness, with a p-value of 0.000 got rejection for the null hypothesis. All other characteristics, with p-values greater than the significance level of 0.05, Melts, CocoaTaste, CocoaAroma, MilkTaste, Density and Crispiness, validate the null hypothesis for the absence of difference between the related variables’ mean. On the other hand 6th graders showed a symmetrical scenario, by rejecting the null hypothesis for all variables, excepting MilkTaste. With a p-value of 0.208, greater than the significance level, the null hypothesis is validated, which states an absence of difference between the related variables’ mean. Lastly, 9th graders show a similar pattern to those observed already in other sub-groups by rejecting the null hypothesis for only one characteristic, MilkTaste. With a p-value of 0.001, smaller than the significance level, the null hypothesis is rejected, making different the related variables’ 54 Young Adults and Old Aged Adults Between these two groups there was no significant moderation validated, since all effects validated in both sub-groups did not present a t-statistic greater than 1.64, regarding the moderator. Since it was already stated the validated effects for Young Adults, it will be skipped this parte and analyzed just Old Aged Adults Validated effects were detected in this sub-group of Adults. In fact Old Aged Adults registered, positive, greater than Young Adults, with t-statistics over 1.64, on the effect between Ethnocentrism and Portuguese Olive Oil Quality, Ethnocentrism and Swiss Olive Oil Quality, Portuguese Olive Oil Perception, and Portuguese Olive Oil Price, with, respectively, 2.14, 2.77, 2.07 and 2.27 to what t-statistics is concerned. Middle Aged and Old Aged Adults There was no effect with significant moderation between these two sub-groups and since they were already analyzed above. Mediation Table 13 shows, after performing the Sobel test for Adults, the for each mediation variable of our model. Table 13-Adults' Mediation's t-statistics Source: the Author After performing the Sobel Test for all paths of the created model with mediation variables, the results show, that there is no significant mediation for all mediation variables, since the Mediator Effect Adults PT Olive Oil Perception Ethnocentrism -> Price Olive Oil PT -0.1515 PT Chocolate Perception Ethnocentrism -> Price Chocolate PT -0.3695 PT Olive Oil Perception Ethnocentrism -> Quality Olive Oil PT -0.2831 SW Olive Oil Perception Ethnocentrism -> Quality Olive Oil SW -0.4037 PT Chocolate Perception Ethnocentrism -> Quality Chocolate PT 0.3489 SW Chocolate Perception Ethnocentrism -> Quality Chocolate SW 0.4703 55 absolute values were always under 1.64. With this being, it will not validate neither for chocolate nor olive oil Hypothesis H4 and H5. With this result in mind it can be suggested that any valid effect between Ethnocentrism on either Price or Quality is based mainly on pre conceived ideas, and never by actual perception of the product. Children - Ethnocentrism Model – Validation By the theory reported before and as performed with the Adults sample, for measuring internal consistency and reliability was used Cronbach’s Alpha (α), Composite Reliability (CR), as well as Average Variance Extracted (AVE) will be used to measure the variance shared by the set of items that are part of our scale. In this study, the model studying children presents a Cronbach’s Alpha of 0.8496, which as presented before, is a “Good” value. Regarding Composite Reliability (CR), the present model has a value of 0.8769, which as presented earlier qualifies as robust value. The Average Variance Extracted (AVE) of this model is 0.3329, which infers that only 33, 29% of the variance of the set of items that composes the CETSCALE is error free. As was done with the Adults sample, and although the AVE value was not proper, the CETSCALE will be used, without any modification. Table 14-Children’s CETSCALE validation Source: SmartPLS output Children -Regression Weights and Statistical Tests – Structural Model Following, it will be presented the results of the estimation of our model, for Children, using PLS algorithm. As well as with the Adults Sample, on table 15, on the first column right, is the t-statistic for each effect present on the Children’s model. Taking into consideration a significance level of 0.1 it must, as was already done with the Adults model, be worked with values above 1.64. AVE Composite Reliability Cronbrach's Alpha CETSCALE 0.3329 0.8769 0.8496 56 Table 15 shows that the effect of Ethnocentrism and Quality Portuguese Olive Oil, Perception Portuguese Olive Oil and Price Portuguese Olive Oil, and Perception Portuguese Olive Oil and Quality Portuguese Olive Oil are all significant effects. These results, validate hypothesis H2C, H6B and H6A. Table 15-Childrens' Path Coefficient Source: SmartPLS output Children – Moderation - Gender and Education Level To what the Children’s Sample is concerned it was found relevant to use Gender and Educational Level as moderators, as was done regarding the analysis towards accessing presence of COE, with the significance levels showed in table 16. Original Sample (O) Sample Mean (M) Standard Deviation (STDEV) Standard Error (STERR) T Statistics (|O/STERR|) Ethnocentrism -> PerceptionOliveOILSW 0.3578 0.1464 0.3223 0.3223 1.11 Ethnocentrism -> PerceptionOliveOILPT 0.2498 0.2013 0.1972 0.1972 1.2668 Ethnocentrism -> PerceptionCHOCPT 0.513 0.2801 0.4454 0.4454 1.1519 Ethnocentrism -> PerceptionCHOCSW 0.31 0.2987 0.1676 0.1676 1.549 Ethnocentrism -> PriceOliveOILPT -0.0443 -0.0351 0.115 0.115 0.3854 Ethnocentrism -> PriceCHOCPT 0.1733 0.177 0.1325 0.1325 1.3081 Ethnocentrism -> QualityOliveOILPT 0.2811 0.2356 0.0915 0.0915 3.0718 Ethnocentrism -> QualityOliveOILSW 0.0503 0.0361 0.1115 0.1115 0.451 Ethnocentrism -> QualityCHOCSW 0.0317 0.0141 0.0899 0.0899 0.3522 Ethnocentrism -> QualityCHOCPT 0.0433 -0.0043 0.1331 0.1331 0.3254 PerceptionOliveOILSW -> QualityOliveOILSW -0.5281 0.0039 0.6179 0.6179 0.8547 PerceptionOliveOILpt -> PriceOliveOILPT 0.3333 0.3957 0.1177 0.1177 2.8308 PerceptionOliveOILpt -> QualityOliveOILPT -0.6454 -0.6368 0.1603 0.1603 4.0264 PerceptionCHOCpt -> PriceCHOCPT -0.2195 -0.111 0.2285 0.2285 0.9609 PerceptionCHOCpt -> QualityCHOCPT -0.0779 0.0216 0.2116 0.2116 0.3681 PerceptionCHOCSW -> QualityCHOCSW -0.0529 0.1096 0.2668 0.2668 0.1984 57 Table 16-Childrens' Moderation's t-statistics Source: the Author Gender As was done with the general model, it was validated the effect between Ethnocentrism and Portuguese olive oil Quality, Portuguese olive oil Perception and Portuguese olive oil Price and Portuguese olive oil Perception and Portuguese olive oil Quality, for both genders, with t-statistics greater than 1.64. To what the two last effects is regarded the moderator effect was also significant, with the t-statistics being respectively 4.01 and 3.54, and with females having grater regression weights than males. When is taken into consideration the first effect, Ethnocentrism and Portuguese olive oil Quality although being for both genders an effect statistically significant, there is no validation for an effect of moderation. Related to males only, it was also validated the Effect between Ethnocentrism and Swiss chocolate Perception, Ethnocentrism and Swiss chocolate Quality and Ethnocentrism and Portuguese chocolate Quality, with t-statistics greater than 1.64. Considering Females alone, it got validation of the effects between Ethnocentrism and Portuguese chocolate Perception, Ethnocentrism and Portuguese chocolate Price and Swiss olive oil Perception and Swiss olive oil Perception. All of these effects registered t-statistics above 1.65 and the effects were both positive and greater than male’s. 4th Graders and 6th Graders As it was done with Age at the Adults sample, this study will be analyzing the three moderators related to education comparing the effect between each other. ChildrenMale ChildrenFemale Moderator 4th Grade 6th Grade Moderator 4th Grade 9th Grade Moderator 6th Grade 9th Grade Moderator Ethnocentrism -> PerceptionOliveOILSU 1.57 0.73 1.16 0.92 1.16 2.65 0.92 2.65 Ethnocentrism -> PerceptionOliveOILPT 0.77 0.59 1.30 0.96 1.30 3.16 0.96 3.16 Ethnocentrism -> PerceptionCHOCPT 1.03 3.24 1.11 0.33 1.11 1.43 0.33 1.43 Ethnocentrism -> PerceptionCHOCSU 2.05 1.18 1.94 1.06 1.94 4.33 0.15 1.06 4.33 Ethnocentrism -> PriceOliveOILPT 0.76 0.33 0.57 1.19 0.57 1.90 1.19 1.90 Ethnocentrism -> PriceCHOCPT 0.42 1.92 1.46 0.21 1.46 2.94 0.21 2.94 Ethnocentrism -> QualityOliveOILPT 1.84 2.06 0.24 0.98 1.20 0.98 3.76 2.09 1.20 3.76 Ethnocentrism -> QualityOliveOILSU 0.11 0.18 0.58 0.40 0.58 1.50 0.40 1.50 Ethnocentrism -> QualityCHOCSU 2.02 1.01 2.51 0.82 0.14 0.82 0.95 0.14 0.95 Ethnocentrism -> QualityCHOCPT 2.69 1.11 2.35 1.60 0.89 1.60 1.43 0.89 1.43 PerceptionOliveOILSU -> QualityOliveOILSU 1.30 1.81 1.40 0.96 1.40 0.19 0.96 0.19 PerceptionOliveOILpt -> PriceOliveOILPT 2.23 7.29 4.01 1.09 0.62 1.09 4.30 2.73 0.62 4.30 2.08 PerceptionOliveOILpt -> QualityOliveOILPT 4.66 3.46 3.54 2.86 0.99 2.86 7.59 4.96 0.99 7.59 PerceptionCHOCpt -> PriceCHOCPT 0.89 0.84 0.98 3.36 2.04 0.98 7.75 1.98 3.36 7.75 5.23 PerceptionCHOCpt -> QualityCHOCPT 0.79 0.55 1.05 3.89 1.05 0.78 3.89 0.78 3.07 PerceptionCHOCsu -> QualityCHOCSU 1.60 0.31 1.79 3.75 0.45 1.79 0.30 0.51 3.75 0.30 Gender Education 58 Starting with the moderation between 4th graders and 6th graders, there was no significant value of moderation regarding the validated hypothesis of the general model, as to neither one of the other hypothesis. Although the previous stated, it was found a base to validate, regarding 4th graders, the effect between Ethnocentrism and Swiss chocolate Perception, Portuguese olive oil Perception and Portuguese olive oil Quality, with t-statistics of 1.94, and 2.86, respectively. The first is a case for a positive effect, greater for 4th graders than 6th graders, the second is a negative effect, being the 4th graders the greater in absolute terms. Towards 6th graders solo, it got validation for two effects, being them Portuguese chocolate Perception and Portuguese chocolate Price and Portuguese chocolate Perception and Portuguese chocolate Quality. For both, the effect is greater for 6th graders than for 4th graders and are both positive. Relating 4th and 6th graders, both validate the effect between Swiss Chocolate Perception and Swiss Chocolate Quality, with t-statistics greater than 1.64 and for both groups this effect is positive. Even though this effect was validated for both groups, there was no evidence of moderation. 4th Graders and 9th Graders Moderating the effect between 4th and 9th graders, shows significance, with a t-statistic of 4.86, for one of the effects validated in the general model, Portuguese olive oil Perception and Portuguese olive oil Quality, with 4th grader with a t-statistic of 2.86 and 9th grader with a t-statistic of 7.59, and contrasting them 4th graders having a negative weight between the two variables and 9th graders having a positive one. Since this analysis has already discussed the effects validated for 4th graders above, it will focus on 9th graders from now on. The sub-group formed by all the 9th graders of the initial Children’s sample validated all the effects of the general model, and also the effect between Ethnocentrism and Swiss olive oil Perception, with a t-statistic of 2.65, Ethnocentrism and Portuguese olive oil Perception, with a t-statistic of 3.16, Ethnocentrism and Swiss chocolate Perception, with a t-statistic of 4.33, Ethnocentrism and Portuguese olive oil Price, with a tstatistic of 1.90, Ethnocentrism and Portuguese chocolate Price, and Portuguese chocolate 59 Perception and Portuguese chocolate Price, with a t-statistic of 7.75. Both effects were positive and in both the weight of 9th graders was greater than 4th graders. 6th Graders and 9th Graders To finalize the analysis of this moderator regarding Children, and since both groups’ significant effects were already stated, 6th graders and 9th graders registered significance moderation for the effect between Portuguese Chocolate Perception and Portuguese Chocolate Price, with the moderation having a t-statistic of 5.23. The weights of each subgroup have opposite directions, being the 6th graders’ positive and the 9th graders negative. The same effect, Portuguese Chocolate Perception and Portuguese Chocolate Price, was statistically moderated between 4th graders and 9th graders, being greater for 9th graders. Regarding the moderation between 6th graders and 9th graders, there was no effect significantly moderated between these two groups, and both have already been discussed separately above. Mediation The research model, only have one kind of mediator variable, although it is present in several paths of model: Perception. Related to both chocolate and olive oil, and to the Origin of the products, Portugal and Switzerland, Perception mediates all the relationships between Ethnocentrism and Price and Quality. Table 17-Children' Mediation's t-statistics Source: the Author After performing the Sobel Test for all paths of the created model with mediation variables, the results show, that there is no significant mediation for all mediation variables, regarding Mediator Effect Children PT Olive Oil Perception Ethnocentrism -> Price Olive Oil PT 1.0938 PT Chocolate Perception Ethnocentrism -> Price Chocolate PT -0.7252 PT Olive Oil Perception Ethnocentrism -> Quality Olive Oil PT -1.2146 SW Olive Oil Perception Ethnocentrism -> Quality Olive Oil SW -0.709 PT Chocolate Perception Ethnocentrism -> Quality Chocolate PT -0.353 SW Chocolate Perception Ethnocentrism -> Quality Chocolate SW -0.3075 60 the Children sample, since the absolute values - table 17 - were always under 1.64. With this being, it will not validate neither for chocolate nor olive oil Hypothesis H4 and H5. As hypothesized regarding Adults, it can be suggested that any valid effect between Ethnocentrism on either Price or Quality is based mainly on pre conceived ideas, and never by actual perception of the product. Conclusion The now ending chapter was dedicated to present the collected data through a descriptive analysis followed by the analysis of the proposed hypothesis. The data collected provided comprehension of the dynamics of the Country of Origin effect between several groups and sub-groups that composed our sample, and both for Chocolate and Olive Oil. In fact, it was concluded that the more effect significant regarding Children, when compared with Adults. As to what the product is concerned, this effect was, generally, more significant for Olive Oil than for Chocolate. Our general models, of both Adults and Children, although with some validation, showed to not provide significant values for our hypothesis: for the Adults sample H2A of Portuguese Olive Oil and H6A of Swiss Chocolate. The Children Sample validated H2C, H6B and H6A of Portuguese Olive Oil. The scenario above, justified an exhaustive comprehension of the results regarding our moderators. By doing this it enabled the validation for several hypothesis of different subgroups, which provide a portrait of distinct realities between genders, ages and education levels. On one hand some of these results were aligned with the state of the art of our field of studies, but on the other hand, a fraction of our results were not validated and require further research. 61 5-Conclusion The present study, based on the Country of Origin effect, granted a general characterization of the international Olive Oil and Chocolate market, as of the Portuguese market. Focusing on Olive Oil, although far from the top producers, Portugal, with a privileged location, is a main global producer, exporter and consumer of Olive Oil. Along with a previous stated, also, identification was found of several characteristics of each product which are impacted by the Country of Origin. Lastly it was identified and analyzed the impact of Ethnocentrism on Quality, Price and Perception on both products. In general terms it was observed that the Country of Origin is important in the way that individuals perceive the product’s physical characteristics, but on those products that have a positive COE. This fact might justify, and should incentivize the use of a “made in” label since it adds value to the product, but at the same time, might place the product in a way, in the consumer’s mind that unable international trade. For those products with a negative COE, the previously reported effect, dilutes. This research challenges that the idea that both quality and physical features of a product are intrinsic, since the way that each individual sees, smells, tastes and touches a product are influenced by both COE and Ethnocentrism. On the other hand, this research strengthens the idea postulated by Kamakura (1999), that COE has a strong effect on physical perceptions, but that those perceptions not always are manifested on a higher price. The results from this study, might have been biased due to the crises period, where individuals tend to protect their own, either individuals or products. This idea manifests itself on when a higher ethnocentric level leads the individual to purpose a smaller price, for both being well adjusted to demand and to enable trade. This study’s sample was heterogeneous, but slightly biased towards college students. Is important to point out though, that in Portugal the schooling is now mandatory until the 12th 62 grade. Even though there are studies that reveal that the standard of living moderate both the effect of Country of Origin and Ethnocentrism. This study, the children’s sample was gathered in a private school, which might have a biased effect on the results. Also with a biasing effect on the results is the geographical area where the data was collected at the north of Portugal, (Porto, Santa Maria da Feira, São João da Madeira and Paredes) Another limitation is related to the type of products, since they were presented without neither brand nor package. On one hand, this fact enabled this study to refine the COE, but on the other hand creates an unrealistic scenario, where COE gains unrealistic strength. Regarding future research, it would be interesting to repeat this study, with the same design using other types of products that incorporate more technology. All in all, the design of this study, that proposed to create a more realistic situation as possible, where the respondents had to experience the products that they were rating, implies that they would have to test and manipulate the products of future research, which mandates that the products used must be of easy manipulation, intuitive and must not raise ethical concerns. On future research the author also finds it interesting to adding to the data analysis moderated mediation, in order to analyze mediation within sub-groups. 63 6-Bibliography Alberts, J. L. (2006). Constructing quality: The multinational histories of chocolate. Geoforum, 9991007. Alexander Josiassen, B. A. (2008). Country-of-origin contingencies: Competing perspectives on product familiarity and product involvement. International Marketing Review, 423-440. Andersson, N. H. (1971). Integration Theory and attitude change. Psychol Rev , 171-206. Andersson, N. H. (1981). Methods of information integration Theory. Academic Press. Andersson, N. H. (1991). Contribution to information integration theory. Lawrence Erlbaum Associates. Andreas Herrmann, L. X. (2007). The influence of price fairness on customer satisfaction: an empirical test in the context of automobile purchases. Journal of Product & Brand Management, 49–58. Andrews, E. L. (1997, October 24). Great Chocolate War Reveals Dark Side of Europe. The New York Times. Bernard L. Simonin, a. J. (1998). Is a company known by the company it keeps? Assessing the spillover effects of brand alliances on consumer brand attitudes. Journal of Marketing Research, 30-42. Bluemelhuber C, C. L. (2007). Extending the view of brand alliances effects. International Marketing Review, 427-43. Bruno Godey, D. P. (2012). Brand and Country of Origin Effect on consumer's decision to purchase luxury products. Journal of Business Research, 1461-1470. Burton, D. R. (1989). The Relationship between Perceived and Objective Price-Quality. Journal of Marketing Research, 429-443. Cameron, E. a. (1994). Consumer Perception of Product Quality and Country-of-Origin Effect. Journal of International Marketing, 49-62. Caobisco. (2012). Caobisco. Retrieved from Caobisco Chocolate, Biscuits and Confectionery of Europe: www.caobisco.eu Castellan, S. S. (1988). Nonparametric tests for the behavioral sciences. New York: McGraw Hill. Chao, P. (1993). Partitioning country of origin effects: consumer evaluations of a hybrid product. Journal of International Business Studies, 291-306. Chao, P. (1998). Impact of Coutry-of-Origin Dimensons on Product Quality and Design Quality Perceptions. Journal of Business Research, 1-6. Chapter IV. (2012, August 29). Diário da República, p. 4826. 70 Sabor: amargura Sabor: picante Quanto à densidade Definição da cor Quanto ao cheiro frutado  O preço deste azeite biológico suíço é de 5€/0,5L. Por quanto é que acha que deveria ser vendido este azeite biológico português? _____________________  Aconselha a que o nome da marca de azeite português tenha um significado em português? Sim Não  Aconselha a que a marca do azeite português tenha referências patrióticas? Sim Não  Aconselha a que a marca do azeite português deva estar escrita numa das línguas oficiais suíças (italiano/alemão/francês)? Sim Não  Aconselha a que a marca de azeite português deva estar escrita em inglês? Sim Não  Aconselha a que a embalagem de azeite português tenha imagens alusivas a Portugal? Sim Não  Aconselha a que a embalagem do azeite português tenha imagens que, de alguma forma, lembrem a Suíça? Sim Não 71 CHOCOLATE POR FAVOR OBSERVE, TOQUE, CHEIRE E PROVE AS DUAS AMOSTRAS DE CHOCOLATE.  Como classificaria o chocolate português? 1. Muito Pouco 2. Pouco 3. Indiferente 4. Significativo 5. Muito Quanto à qualidade Aspeto: brilhante Toque: derrete-se facilmente Sabor: Intensidade do Cacau Aroma: Intensidade do cacau Sabor: Intensidade do Leite Densidade Estaladiço / Crocante  Como classificaria o chocolate suíço? 1. Muito Pouco 2.Pouco 3.Indiferente 4.Significativo 5.Muito Quanto à qualidade Aspeto: brilhante Toque: derrete-se facilmente Sabor: Intensidade do Cacau Aroma: Intensidade do Cacau 72 Sabor: Intensidade do Leite Densidade Estaladiço / Crocante  O preço deste chocolate suíço é de 5,60 € (valor excluído de transportes). Por quanto é que acha que deveria ser vendido o chocolate Português? ________  Aconselha a que o nome da marca de chocolate português tenha um significado em português? Sim Não  Aconselha a que a marca do chocolate português tenha referências patrióticas? Sim Não  Aconselha a que a marca do chocolate português esteja escrita numa das línguas oficiais suíças (italiano/alemão/francês)? Sim Não  Aconselha a que a marca do chocolate português deva estar escrita em inglês? Sim Não  Aconselha a que a embalagem do chocolate português tenha imagens alusivas a Portugal? Sim Não  Aconselha a que a embalagem do chocolate português tenha imagens que, de alguma forma, lembrem a Suíça? Sim Não 73 Questionário Os dados fornecidos ao longo deste questionário são considerados confidenciais e em nenhum momento serão partilhados com nenhuma entidade externa à Faculdade de Economia do Porto e à organização que o promove. O questionário terá uma duração esperada inferior a 5 minutos. Por favor responda de forma consciente e verdadeira. Responda às seguintes perguntas consoante o seu grau de concordância, em que 1- “Discordo totalmente”; 2-“Discordo”, 3-“Indiferente” 4- “Concordo” e 5- “Concordo totalmente” Produção 1.Discordo Totalmente 2.Discordo 3.Indiferente 4.Concordo 5.Concordo Totalmente Saber que o produto foi produzido integralmente numa unidade fabril portuguesa é um fator importante que influencia positivamente a minha decisão compra. Saber que todos os trabalhadores que produziram um produto são Portugueses influência positivamente a minha decisão de compra. Saber que apenas matérias-primas portuguesas entram na composição de um produto influencia positivamente a minha decisão compra. 74 Embora sejam usadas apenas matérias-primas e trabalhadores portugueses na produção de um produto, o facto de a organização ser maioritariamente detida por estrangeiros influencia negativamente a minha compra Design 1.Discordo Totalmente 2.Discordo 3.Indiferente 4.Concordo 5.Concordo Totalmente O design de um produto é um fator importante que influencia positivamente a minha compra. O facto de o design de um produto ter sido desenvolvido por portugueses influencia positivamente a minha decisão de compra. Saber que o design foi produzido por uma organização portuguesa, mesmo que a atuar fora do país, influência positivamente a minha decisão de compra. Marca 1.Discordo Totalmente 2.Discordo 3.Indiferente 4.Concordo 5.Concordo Totalmente O facto de a Marca estar escrita em português influencia positivamente a minha compra. As Marcas portuguesas devem fazer alusão a Portugal. 75 As Marcas portuguesas devem estar escritas em português Se o produto for apenas para exportar, a marca não deve estar escrita em Português. Por favor, demonstre o seu grau de concordância com as frases abaixo, numa escala de 1 a 5, em que 1-“Discordo totalmente”; 2-“Discordo”, 3-“Indiferente” 4- “Concordo” e 5- “Concordo totalmente” 1.Discordo Totalmente 2.Discordo 3.Indiferente 4.Concordo 5.Concordo Totalmente A população portuguesa deve comprar sempre produção nacional em vez de recorrer a importações. Apenas os produtos que não são produzidos em Portugal devem ser importados Compre produtos Portugueses. Mantenha Portugal a trabalhar. Produtos Portugueses sempre. Comprar produtos a estrangeiros é antiportuguês. Não é correto comprar produtos ao estrangeiro porque isso põe a população portuguesa sem emprego. Um verdadeiro português deve sempre comprar produtos portugueses. Nós devemos comprar produtos produzidos em Portugal em vez de deixar que outros países fiquem ricos à nossa custa. É sempre melhor comprar produtos Portugueses. Deve haver muito pouco comércio ou compra de bens de outros países, a não ser por necessidade. Deveriam ser impostos limites a todas as importações. Portugal não deve comprar produtos estrangeiros, porque isto fere as empresas Portuguesas e causa desemprego 76 Pode-me custar no longo prazo, mas eu prefiro apoiar os produtos portugueses. Os produtos estrangeiros deviam ser taxados de forma pesada para reduzir a sua entrada em Portugal. Os estrangeiros não deveriam ter permissão para colocar os seus produtos nos nossos mercados. Apenas devíamos comprar a países estrangeiros os produtos que nós não conseguimos obter dentro do nosso próprio pais. Consumidores Portugueses que comprem apenas produtos que são feitos noutros países são responsáveis por pôr outros portugueses no desemprego. Idade:___________ Sexo: Masculino Feminino Nível Educacional: ___________________ Nacionalidade: __________________ 77 7.2-Questionnaire-Children QUESTIONNAIRE CHILDREN AZEITE POR FAVOR OBSERVA, CHEIRA E PROVA AS DUAS AMOSTRAS DE AZEITE.  Como classificarias o azeite português? 1. Muito Pouco 2. Pouco 3. Indiferente 4. Significativo 5. Muito Quanto à qualidade Sabor: grau de acidez Sabor: amargura Sabor: picante Quanto à grossura Definição da cor Quanto ao cheiro à azeitona 78  Como classificarias o azeite suíço? 1. Muito Pouco 2. Pouco 3.Indiferente 4.Significativo 5. Muito Quanto à qualidade Sabor: grau de acidez Sabor: amargura Sabor: picante Quanto à grossura Definição da cor Quanto ao cheiro à azeitona  Na tua opinião o azeite biológico português deveria ser mais caro ou mais barato que o Suíço? Mais Caro Mais Barato Igual  Aconselhas a que o nome da marca de azeite português tenha um significado em português? Sim Não  Aconselhas a que a marca do azeite português tenha referências ao teu país? Sim Não  Aconselhas a que a marca de azeite português deva estar escrita em inglês? Sim Não  Aconselhas a que a embalagem de azeite português tenha imagens alusivas a Portugal? Sim Não 79  Aconselha a que a embalagem do azeite português tenha imagens que, de alguma forma, lembrem a Suíça? Sim Não CHOCOLATE POR FAVOR OBSERVA, TOCA, CHEIRA E PROVA AS DUAS AMOSTRAS DE CHOCOLATE.  Como classificarias o chocolate português? 1. Muito Pouco 2. Pouco 3. Indiferente 4. Significativo 5. Muito Quanto à qualidade Aspeto: brilhante Toque: derrete-se facilmente Sabor: Intensidade do Cacau Aroma: Intensidade do cacau Sabor: Intensidade do Leite Cremosidade Estaladiço / Crocante 86 7.3-Moderation Spreadsheet Source: the author adapted from http://statwiki.kolobkreations.com/wiki/Main_Page 7.4-Mediation Spreadsheet Source: the author adapted from http://www.unc.edu/~preacher/sobel/sobel.htm Comentario Sample A Sample B Sample Size <-- Enter Data Here Regression Weight Standard Error (S.E.) t-statistic <-- View Results Here p-value (2-tailed) (m-1)^2 1 (m+n-2) -2 (n-1)^2 1 sqrt(1/m+1/n) #DIV/0! 1st half denom 0 2nd half denom 0 sqrt(1st half + 2nd half) 0 Full denom #DIV/0! numerator 0 Effect