Comparative analysis of consumer attitudes in Croatia and Serbia according to domestic and foreign brands

Abstract Brand management is an effective tool in today's market conditions. In some markets there are numerous brands that can be seen as domestic and foreign. Domestic brands are a broad category and they can represent brands which own domestic products, thus representing brands that are present on some part of the national market, or they may be national brands that are present throughout the entire national market. Successful brand management implies a necessary focus on the appropriate market segment of the consumer and that the brand has an appropriate exchange value (viewed through the functional, emotional and symbolic component of that value). The assumption of seeing what value brands have and what value consumers demand refers to the perception of consumer attitudes about it. This paper examines consumer attitudes by looking at brands through two generic groups (domestic and foreign brands) as well as covering two markets (Croatia and Serbia). Primary research which has been carried out on a sample of 555 consumers (n = 291 in Croatia, n = 264 in Serbia) shows that there are numerous similarities as well as some differences regarding consumer attitudes about domestic and foreign brands.


Introduction
Brands are a superior category of products and services since they involve establishing and maintaining emotional and symbolic values, while products and services are primarily focused on functionality or functional value. Certainly this functionality is necessary for the successful establishment and management of a brand and, thus, it also implies its functional value, but it is insufficient for the brand to be successful because a successful brand implies a synergistic effect of functional, emotional and symbolic values. In order to try to explain this concept, Aaker (1991) defines it as a set of assets (and liabilities) linked to a brand's name and symbol that add to (or subtract from) the value provided by a product or service to a firm and/or that firm's customers. It represents the difference between overall brand preference and multiattributed preference based on objectively measured attribute levels (Park & Srinivasan, 1994) and overall quality and customers' choice intention (Angrawal & Rao, 1996).
Customer-based brand equity is defined from the perspective of the customer and is based on consumer knowledge, familiarity, and associations with the brand (Dollatabady & Amirusefi, 2011;Washburn & Plank, 2002). Proponents contend that, for a brand to have value, it must be valued by a customer. Regarding this, Keller (1993) defines customer-based brand equity as 'the differential effect that brand knowledge about the consumer or how customers respond to the marketing of that brand' and when customers respond more favourably to a product whose brand is identified. The importance of the product brand shall be seen primarily in its impact on consumers' choice and their loyalty through identifying and differentiating quality and origin, as well as creating additional values (Vrane sevi c & Stan cec, 2003). Certainly, the brand's origin has an important role as the more positive the origin of the brand, the easier it is to establish and maintain the emotional and symbolic value of the brand. By analysing relevant literature in this field, it can be concluded that the country of origin of the product and services impact on product appraisal and purchasing behaviour of consumers ( Sapi c & Golo, 2017;Sharma, 2011). The origin affects the emotional and symbolic value of the brand. Taking this into consideration, it is most appropriate, in our opinion, to observe the origin as a 'culture' of a particular brand, i.e., which culture a brand represents, which cultural values it testifies and which are those that it inclines in the long-term. Also, significant differences between regions are the consequence of the differences in social customs and factors such as culture or religion ( -Deri, Armenski, Tesanovi c, Bradi c, & Vukosav, 2013). From the perspective of marketing and economics, the concept of consumer ethnocentrism refers to the attitude of consumers to buy domestic brands. For ethnocentrically-oriented consumers, the purchase of foreign products is an inappropriate, non-patriotic and immoral work that can seriously jeopardise the national economy and create unemployment (Mati c, 2013;Shimp & Sharma, 1987).
In order to create a brand, a lot of investment is needed, as well as expertise and time. Brands are not created overnight, without a clear vision and proper strategy. Companies are becoming more aware that consumers will rather decide to buy a brand that has its own history, than a brand that has been strikingly advertised over the past several years (Peri c, Krasulja, and Radojevi c, 2011). In order to assess the proper brand management and brand positioning it is necessary to learn about consumer attitudes-real and potential-in individual markets, to see their similarities and differences and their ethnocentricity through a comparative analysis in order to take the actions necessary for successful brand management.

Theoretical framework
Brands are equally important for companies (businesses) and for consumers (clients) (Vrane sevi c, 2016). The importance of branding for companies is recognised by the fact that they are a means of identification, a means of legal protection, a signal of product quality, a means of enriching products with added value, a source of competitive advantages and a source of financial revenue. On the other hand, for customers (clients) brands are: the identification of origin, determination of product and/or service liability, reduction of risk by purchase and use, reduction of costs and time needed for selection of products and/or services due to diversified offer, indicating the responsible individual for products and services owned by the brand, a means of expressing symbolic aspirations and a sign and quality certificate.
When it comes to managing brands in a larger number of markets and the relation of domestic and foreign brands, this is the case when it is important to study consumer attitudes because attitudes can be opposed and may affect management activities related to the necessary adaptation or standardisation of global and/or foreign brands to these receptive markets, whereby the more diverse market environment in the receptive markets the greater brand adaptation is required, and vice versa (Krupka, Ozreti c-Do sen, & Previ si c, 2017).
This adaptation and standardisation can be viewed through the elements of the marketing web (basic and extended), brand identity elements, but also through all marketing activities related to brand management. Furthermore, it is also important to see the advantages and disadvantages of approaching the taken or desired position of local brands in relation to existing brands on the market (regardless of the fact that these brands are global, i.e., foreign or domestic, local brands). Analysis revealed significant differences in the image of the socialist era brands between their home and foreign markets. Variations are linked to a cultural context of newly (re)emerging states with high ethnocentrism, animosity and uneasy political relationships. Transnational brands, however, appear immune to much consumer ethnocentrism (Bre ci c et al., 2013).
The brand's origin may also be significant for consumers while assessing their exchange value, consumers (clients) should not be expected to evaluate the exchange value of the brand solely on the basis of the origin of a certain country, since this estimate should be the result of a number of factors (Krajnovi c, Jadre sko, & Bosna, 2015). There are numerous global brands in all industries on both Croatian and Serbian markets, but in both countries there are also numerous historically 'common' brands which are active and which exist from the time when both countries were part of the same state union (until 1991 in the SFR Yugoslavia). Also, these brands were considered in today's independent markets (Serbia and/or Croatia) and in some situations still are, to be common 'domestic' brands (Babi c, 2012). A question which may be asked is to what extent these brands are experienced as foreign ones or even still domestic by regular customers in relation to new potential consumers. Also, there are brands that are considered exclusively local brands and domestic (national) brands on both of these markets. So far conducted research has shown certain principles-in Serbia, a country with extremely high value of the dimension of avoidance of uncertainty ( Sapi c & Golo, 2017), numerous consumers believe that word-of-mouth reduces the risk of purchasing and this is rather seen as a characteristic of older consumers (Mari c, Grubor, & -Duranovi c, 2016). Numerous research confirms that the high degree of avoidance of uncertainty affects negative attitudes of consumers from the aspect of acceptance of foreign products (Griffith, Yalcinkaya, & Rubera, 2014;Tellis, Stremersch, & Yin, 2003;van Everdingen, Fok, and Fok, 2009;Yeniyurt & Townsend, 2003). Consumers in Croatia prefer brands which they have positive experiences with (Stan cin & Gregurec, 2017), just as they have emphasised consumer ethnocentrism (Renko, Crnjak Karanovi c, & Mati c, 2012). Mostly, separate surveys on the observed markets indicate more or less determined consumer attitudes towards global, national and local brands. With the intention of deepening and discovering new insights, we have carried out a survey involving two countries (Croatia and Serbia) and we have also tried to look at some specific aspects (for example, the attitude of actions such as 'Let's buy domestic', the attitude towards informing through virtual reality-www, social networks; the attitude towards (non) identity of product quality of domestic and foreign brands that are being produced for the markets of the Eastern Europe countries and the Balkans), and thus we have made a comparative analysis of all the obtained results.

The aim of the research
The aim of this research is to examine the relationship between socio-demographic variables (gender, age, working status, degree of education, monthly personal income of respondents and monthly personal income of respondent households, number of household members and number of household members who earn income) and attitudes of respondents from Serbia and Croatia towards branded products. In particular, the main aims of the research were: To determine the attitude of consumers in Croatia and Serbia toward branded goods in general and the purchase of such. To investigate the consumers opinion about the quality of branded goods (whether brands guarantee the quality and on what level they do this). To research whether consumers in Croatia and Serbia take into account the country of brand's origin.

Methods
The conducted research is quantitative by nature and it was carried out through a structured questionnaire as an instrument for collecting primary data into research that included questions related to socio-demographic information about the respondent and a group of questions that, at the same time, represented dependent variables. The collected responses depicted personal assessments of the respondents regarding the extent they agreed to the claims that we used to examine their attitude towards brands. The Likert scale of agreement with established assertions was being used and it had the intensities of agreeing from 1-5, where 1 represents the answer (personal assessment) I completely disagree; 2 ¼ I disagree; 3 ¼ I am not sure; 4 ¼ I agree; and 5 ¼ I completely agree.
The collection of primary data was carried out during March and April 2017 through on-line testing by using the software solution Qualtrics (https://www.qualtrics.com/).
We have used the following for data analysis: descriptive analysis (percentage, arithmetic mean), Hi-square independence test, t-test, multiple linear regression and Pearson correlation coefficient. The level of statistical significance was set to p < 0.05 and all the obtained data were processed in the SPSS program, version 19.

Sample description
The sample may be claimed to be large enough if the obtained results can be considered indicative. Our intention was not to obtain a statistically representative sample since this was impossible due to the selected method of collecting primary data, i.e., on-line testing. A total of 264 respondents participated in the survey in Serbia, of which 124 were male (47%) and 140 female (53%). The survey in Croatia covered 291 respondents, of which 103 were male (35.4%) and 188 female (64.6%).

Independent variables
The researched and presented sample characteristics, described in the sample (subsection 3.3), which at the same time represented the independent variables in this study are: gender of the respondents, age, working status, level of education, monthly personal income of respondents, monthly income of the household, number of members living in the household and the number of household members who earn income.

Dependent variables
Dependent variables were used to examine the attitude of the respondents towards branded products. This was examined through two closed-type questions and ten claims where the respondents assessed to what extent they agreed with them on a five-step Likert type scale, where 1 signified I completely disagree and 5 meant I completely agree.
The questions of the closed type are: (a) I buy almost exclusively branded goods (possible answers: yes and no) and (b) When it comes to the influence of the brand on the product quality I find that: (three responses are offered, with the option of choosing only one). The brand has no impact on the quality of the product because there are too many copies; Despite all the copies the brand is still able to guarantee the quality of the product; The brand has no impact on the quality of the product, but only increases its price).
The claims to which respondents should have expressed their (dis)agreement on the Likert scale are: I prefer foreign brands; When buying, I take care of the country of brands' origin; For the same category of product, I would rather pay more for a foreign brand; I support the campaign 'Let's buy domestic'; I am actively being informed about favourite brands through the manufacturer or seller's website, social networks or consumer forums; I believe that the brand guarantees me enough product quality; I believe that large companies send goods of lower quality to the countries of Eastern Europe and the Balkans; I like to try the new home brand when it appears on the market; I like to try a new foreign brand when it appears on the market; I usually communicate positive or negative brand impressions to my friends, colleagues and relatives. In total, there were 10 claims on which the respondents should have expressed their (dis)agreement.

Descriptive analysis (percentages, arithmetic mean)
The majority of respondents in Serbia, as much as 92.4%, stated that they did not buy exclusively branded goods. A similar situation is typical for Croatia, where 83.5% of the respondents said they did not buy exclusively branded goods.
The Serbian sample showed that most of the respondents (37.9%) pointed out their agreement with the claim 'I prefer foreign brands', while 45.4% of them agreed with the statement 'When buying, I take care of the country of brands' origin'. Almost every other respondent (54.5%) disagrees with the claim 'For the same category of product, I would rather pay more for a foreign brand', and 60.6% of the respondents support the campaign 'Let's buy domestic', while 43.9% of those surveyed are actively being informed about favourite brands through the manufacturer or sellers' website, social networks or consumer forums. In addition, 42.4% of the respondents believe that the brand guarantees them enough product quality; however, seven out of 10 respondents (69.7%) believe that large companies send goods of lower quality to the countries of Eastern Europe and the Balkans. When it comes to new domestic and foreign brands on the market, 62.1% of respondents like to try out a new home brand when it appears on the market, while 48.5% do it when a new foreign brand appears. Also, 71.2% of those surveyed stated that they usually communicate positive or negative brand impressions to my friends, colleagues and relatives. These results are presented in Table 1.
The majority of respondents (56.3%) in the Croatian sample pointed out their disagreement with the claim 'I prefer foreign brands'. When it comes to the country of brand's origin, the results show that the majority of respondents (47.1%) in Croatia took into account the country of origin of brands when buying, but most of the respondents (69.1%) would prefer to pay more for the same product category for the foreign brand. The Croatian sample showed that 79.7% of respondents supported the campaign 'Let's buy domestic' and 42% reported that they were not actively informed about favourite brands through the manufacturer or sellers' site, social networks or consumer forums. The results show that the largest number of respondents (86.3%) believes that large companies send goods of lower quality to the countries of Eastern Europe and the Balkans. When it comes to new domestic and foreign brands on the market, 68.4% of them like to try out a new domestic brand when it appears on the market, while 45.4% do it when a new foreign brand appears. Positive or negative impressions of the brand are usually communicated to friends, colleagues and relatives by 83.5% of those surveyed. These results are presented in Table 2.
The results show that respondents in Serbia mostly agree with the claim 'I usually communicate positive or negative brand impressions to my friends, colleagues and relatives' (3.86), and then with the assertion 'I believe that large companies send goods of lower quality to the countries of Eastern Europe and the Balkans' (3.82), while the least of those surveyed agree with the statement 'For the same category of product, I would rather pay more for a foreign brand' (2.57) (Figure 1).
The respondents in Croatia mostly agree with the claim 'I believe that large companies send goods of lower quality to the countries of Eastern Europe and the Balkans' (4.24) and then with the statement 'I usually communicate positive or negative brand impressions to my friends, colleagues and relatives' (4.05), while the least  of those surveyed agree with the statement 'For the same category of product, I would rather pay more for a foreign brand' (2.22) (Figure 2). When it comes to the brand's impact on product quality, the largest number of those surveyed on the Serbian sample (60.6%) believe that, despite all the copies, the brand is still able to guarantee the quality of the product.
As noted on the Serbian sample, the majority of those surveyed in Croatia (52.2%) believe that, despite all the copies the brand is still able to guarantee the quality of the product.

Hi-square test
Hi-square test revealed a statistically significant difference (at the level of 0.01) between the respondents from the Serbian and Croatian samples x2 (1, 551) ¼ 13.314, regarding the issue whether the respondents buy exclusively branded goods. Although the majority of respondents in both countries say that they do not exclusively buy branded goods, the results show that this percentage is higher within respondents in Serbia (93.8%) than in Croatia (83,5%).
A statistically significant difference (at the level of 0.05) between respondents from Serbia and Croatia was found also when it comes to the influence of the brand on product quality x2 (2, 555) ¼ 6.400. The results show that the majority of respondents from the mentioned countries believe that, despite all the copies, the brand manages to guarantee the quality of products, but the number of those who consider it is higher in Serbia (60.6%) than in Croatia (52.2%). The opinion that the brand does not have an impact on the quality of the product, but only increases its price is expressed more by respondents in Croatia (40.5%) than those in Serbia, while a larger number of respondents from Serbia (9.1%) compared to those in Croatia (7.2%) considers that the brand has no effect on the quality of the product because there are too many copies (Figures 3 and 4).

T-test
The T-test examined whether there was any difference in responses of the respondents of different sexes (on Serbian and Croatian sample) regarding the dependent variables which examined the attitude of the respondents towards the branded products.  When it comes to the differences in the responses of the respondents from the Serbian and Croatian samples, the t-test showed that respondents differ in terms of preferences for foreign brands t(553) ¼ -5.85, p < 0.01, and the results show that respondents from Serbia (3.03) preferred foreign brands more than respondents from Croatia (2.53).
The results show that there are also differences between the respondents when it comes to their willingness to pay more for the foreign brand of the same category of product (499.835) ¼ -3.74, p < 0.01, and it has been noted that respondents in Croatia (2.22) are less willing to do so in comparison to respondents in Serbia (2.56).
Significant differences between the respondents are also obtained when it comes to supporting the campaign 'Let's buy domestic', t(496,130) ¼ 6.38, p < 0.01, and the results show that the mentioned campaign is more supported by respondents in Croatia (4.02) than in Serbia (3.47).
The T-test revealed the difference between the respondents and their belief that large companies send goods of lower quality to the countries of Eastern Europe and the Balkans, t(469,703) ¼ 5.35, p < 0.01, while the results show that respondents in Croatia (4.24) believe in this to a greater extent than those in Serbia (3.80).
A statistically significant difference between the respondents also exists in their habit of communicating positive or negative brand impressions to their friends, colleagues and relatives, t(529.593) ¼ 2.21, p < 0.05, and the results show that respondents in Croatia (4.05) do this to a greater extent than respondents in Serbia (3.88). When other variables are considered there were no statistically significant differences between the respondents from the Serbian and Croatian samples.

Multiple regressions
Multiple regressions examined how gender, age, working status, level of education, monthly personal income, household monthly income, number of household members and the number of members in the household who earn income are related as a linear combination of predictors with a group of dependent variables. We investigated whether the ratio of subjects to branded products could be predicted depending on the mentioned linear combination of the predictors.
The results show that the preference of the respondents to foreign brands on the Serbian sample can be significantly statistically predicted by this combination of the predictors R 2 ¼ 0.151, F(8.255) ¼ 5.647, p < 0.01, and the number of household members is emphasised as a single significant indicator, b ¼ 0.366, t ¼ 4.440, p < 0.01, which means that respondents living in a household with a larger number of members prefer foreign brands. A significant prediction was obtained for the Croatian sample, R 2 ¼ 0.112, F(8,282) ¼ 4.446, p < 0.01, and as individual significant predictors the following are stressed: gender, b ¼ -0.138, t ¼ -2.378, p < 0.05, and the age of subjects, b ¼ -0.258, t ¼ -4.178, p < 0.01, thus based on the results it can be concluded that younger men prefer foreign brands more than others.
When it comes to the influence of the country of brand's origin on the respondents during the purchase, the results from the Serbian sample show that this combination of predictors can foresee the aforementioned, R 2 ¼ 0.254, F(8.255) ¼ 10.838, p < 0.01, and as individual significant predictors the following are singled out: gender, b ¼ -0.156, t ¼ -2.555, p < 0.05, age of the respondents, b ¼ 0.151, t ¼ 2.098, p < 0.05, working status, b ¼ -0.226, t ¼ -3.574, p < 0.01, the number of household members, b ¼ 0.239, t ¼ 3.101, p < 0.01, and the number of household members who earn income, b ¼ -0.351, t ¼ -5.488, p < 0.01. Based on the results, it can be concluded that elderly employed men with a larger number of household members and a smaller number of household members who earn income are more likely to take into account the country of brand's origin when buying. A significant prediction was obtained in the Croatian sample, R 2 ¼ 0.107, F(8.282) ¼ 4.243, p < 0.01, and the age of the respondents was emphasised as an important predictor, b ¼ 0.199, t ¼ 3.209, p < 0.01, which means that the older respondents take more account of the country of brand's origin while shopping.
A statistically significant prediction on the Serbian sample was also obtained when it comes to respondents' willingness to pay more for a foreign brand of the same product category, R 2 ¼ 0.069, F(8.255) ¼ 2.372, p < 0.05. As individual significant predictors, the monthly income of the household, b ¼ -199, t ¼ -1.974, p < 0.05, the number of household members, b ¼ 0.320, t ¼ 3.718, p < 0.01, and the number of household members who earn income, b ¼ -0.183, t ¼ -2.565, p < 0.05 were emphasised, which means that respondents with a lower monthly household income, fewer household members and a larger number of members who earn income in the household are the least willing to pay more for a foreign brand of the same category. On the Croatian sample, the regression model did not prove to be significant.
When it comes to respondents' active informing about their favourite brands through the manufacturer or seller's site, social networks or consumer forums, the obtained results for the Serbian sample show that it can also be predicted by this combination of predictors, R 2 ¼ 0.122, F(8.255) ¼ 4.438, p < 0.01. As individual significant predictors the following have been emphasised: gender, b ¼ 0.133, t ¼ 2.012, p < 0.05, the working status, b ¼ 0.218, t ¼ 3.185, p < 0.01, monthly personal income, b ¼ 322, t ¼ 3.056, p < 0.01, and monthly household income, b ¼ -0.268, t ¼ -2.746, p < 0.01, which means that women with higher personal income and lower household income who are not employed are more actively being informed about their favourite brands through these sources. On the Croatian sample, the regression model did not prove to be significant.
It is shown on the Serbian sample that the opinion of the respondents who believe the brand guarantees sufficient product quality, R 2 ¼ 0.116, F(8.255) ¼ 4.178, p < 0.01 can be statistically predicted to a great extent, and the following have been emphasised as individual significant predictors: gender, b ¼ -0.226, t ¼ -3.402, p < 0.01, working status, b ¼ 0.168, t ¼ -2.447, p < 0.05, monthly household income, b ¼ -0.278, t ¼ -2.835, p < 0.01, and the number of household members, b ¼ 0.198, t ¼ 2.353, p < 0.05. Based on the obtained results, it can be concluded that employed men with a lower monthly income of the household and a larger number of household members consider, to a greater extent, that the brand guarantees sufficient product quality. A significant prediction was obtained in the Croatian sample, R 2 ¼ 0.136, F(8.282) ¼ 5.546, p < 0.01, and the following have been emphasised as individual significant predictors: gender, b ¼ -0.181, t ¼ -3.157, p < 0.01, and the age of the examinees, b ¼ -0.246, t ¼ -4.034, p < 0.01, which means that younger men are more likely to believe that the brand guarantees a sufficient product quality.
A statistically significant prediction for the Serbian sample was also obtained for the belief that large companies send goods of lower quality to the countries of Eastern Europe and the Balkans, R 2 ¼ 0.161, F(8.255) ¼ 6.131, p < 0.01, and the following have been emphasised as individual significant predictors: age of subjects, b ¼ 0.299, t ¼ 3.932, p < 0.01, and the number of household members, b ¼ 0.169, t ¼ 2.067, p < 0.05, which means that the older respondents living in households with a larger number of members mostly hold this opinion. On the Croatian sample, no predictor was shown to be statistically significant.
This combination of predictors on the Serbian sample also proved to be statistically significant for predicting the respondents' desire to try out a new domestic brand when it appears on the market, R 2 ¼ 0.105, F(8.255) ¼ 3.676, p < 0.01, and the following have been emphasised as individual significant predictors: gender, R 2 ¼ 0.182, t ¼ 2.694, p < 0.01, and the number of household members, b ¼ -0.213, t ¼ -2.496, p < 0.05. Based on the results we may conclude that most likely women living in households with fewer members are willing to try the new domestic brand when it appears on the market. A significant prediction was also obtained in the Croatian sample, R 2 ¼ 0.061, F(8.282) ¼ 2.288, p < 0.05, and the following have been emphasised as individual significant predictors: gender, b ¼ 0.126, t ¼ 2.104, p < 0.05, and the degree of professional education of the respondents, b ¼ 0.142, t ¼ 2.194, p < 0.05, which means that most likely highly educated women are willing to try the new domestic brand when it appears on the market.
When it comes to the respondents' desire to try out a new foreign brand when it appears on the market, it has been shown that on the Serbian sample it can be predicted by this combination of predictors, R 2 ¼ 0.164, F(8.255) ¼ 6.233, p < 0.01, and the following have been emphasised as individual significant predictors: the age of the examinees, b ¼ 0.165, t ¼ -2.174, p < 0.05, working status, b ¼ 0.228, t ¼ -3.416, p < 0.01, the degree of professional education, b ¼ -0.228, t ¼ -2.846, p < 0.01, and monthly personal income, b ¼ 0.228, t ¼ 2.217, p < 0.05. The obtained results show that younger employees of lower education and higher monthly personal income are more likely to try a new foreign brand when it appears on the market. A significant prediction was also obtained on the Croatian sample, R 2 ¼ 0.081, F(8,282) ¼ 3.118, p < 0.01, and the age of the respondents was emphasised as a significant individual predictor, b ¼ -0.240, t ¼ -3.812, p < 0.01, which means that the new foreign brand is most likely to be tried by younger respondents when it appears on the market.
This combination of predictors on the Serbian sample was also shown to be statistically significant when it came to the habits of the respondents to communicate positive or negative impressions of the brand to their friends, colleagues and relatives, R 2 ¼ 0.134, F(8.255) ¼ 4.927, p < 0.01; the following have been emphasised as individual significant predictors: monthly personal income, b ¼ 0.262, t ¼ -2.499, p < 0.05, number of household members, b ¼ -2.210, t ¼ -2.522, p < 0.05, and the number of household members who earn income, b ¼ -0.207, t ¼ -2.999, p < 0.01, which means 80 T. VRANE SEVI C AND N. PERI C those respondents with lower personal income, who live in households with fewer members and fewer members earn revenue most often communicate their impressions of the brand to their closest ones. On the Croatian sample, the regression model proved to be statistically significant, R 2 ¼ 0.085, F(8.282) ¼ 3.255, p < 0.01, and the following have been emphasised as individual significant predictors: the gender of examinees, b ¼ 0.225, t ¼ 3.820, p < 0.01, and monthly household income, b ¼ 0.154, t ¼ 2.321, p < 0.05, which means that women with higher household incomes most often communicate their positive or negative impressions of the brand to their friends, colleagues and relatives. Based on the obtained results, it is concluded that this linear combination of predictors proves to be significant for predicting almost all dependent variables which we used to examine the attitude of the respondents towards the branded products.

Correlation
The inter-relation between the dependent variables, which referred to the relation of the respondents to the branded products, was examined by the Pearson linear correlation coefficient.
The obtained results of the correlation matrix, which are listed in Table 3, show that the highest degree of dependence in the Serbian sample was found between the claims 'I prefer foreign brands' and 'For the same category of product, I would rather pay more for a foreign brand' (r ¼ þ0.520, p < 0.01), which means that the respondents prefer foreign brands more and they are more willing to pay more for the foreign brand of the same category of product and vice versa. A low correlation between the mentioned variables (r ¼ þ0.367, p < 0.01) was found on the Croatian sample.
The results from the Croatian sample show that the highest degree of dependency was found between the following variables: 'I like to try out a new domestic brand when it appears on the market' and 'I like to try out a new foreign brand when it appears on the market' (r ¼ þ0.472, p < 0.01). In addition, such a moderate correlation between these variables was also found on the Serbian sample (r ¼ þ0.476, p < 0.01) ( Table 4).
A moderate level of correlation on the Serbian sample was also found among the following variables: 'I am actively being informed about favourite brands through the manufacturer or seller's website, social networks or consumer forums' and 'I usually communicate positive or negative brand impressions to my friends, colleagues and relatives' (r ¼ þ0.447, p < 0.01), which means that the respondents are more actively informed about their favourite brands through the above sources and that they are more likely to share their positive or negative impressions about the brand with the nearest ones. Within other variables with present dependence, low correlations were found in both samples (Tables 3 and 4).

Results analysis
The results of the survey show that most of the respondents in Serbia and Croatia do not buy exclusively branded goods, while the percentage of those who do is higher in Serbia than in Croatia. When it comes to differences in terms of demographic characteristics, the Hi-square test has shown that there are no significant differences between the respondents of the Croatian sample who purchase exclusively branded goods. The Serbian sample showed that respondents over 40 years old, pensioners and the unemployed, highly educated, as well as respondents who live alone, or in a two-member or five-member household are not buying only branded goods, while among other categories of respondents there are groups that by only branded products, mostly among the youngest (20% of those age up to 20).
When it comes to the brand's impact on product quality, the results show that respondents from both samples, in most cases, consider that, despite all the copies the brand manages to guarantee product quality and a Hi-square test in both samples found statistically significant differences between respondents regarding their demographic characteristics. The results in Serbia show that, although the respondents of both sexes believe that in spite of all the copies the brand manages to guarantee the quality of products, men hold this opinion rather than women. Unlike the respondents under the age of 20, pupils, those who graduated only from elementary school and respondents living in a five-member household, who consider that the brand only increases the price of the product but does not have an impact on its quality, other respondents mostly consider that despite all the copies the brand still manages to guarantee the quality of the product. The results of the Croatian sample show that respondents under the age of 20 and those aged 51-65, middle-aged, those with high school education, respondents who do not earn income and those with an income

82
T. VRANE SEVI C AND N. PERI C between 501-1,000 euros believe to a higher percentage that the brand has no impact on product quality but only increases the price, while other categories of respondents consider that the brand, despite all the copies, manages to guarantee the quality of the product.
The T-test showed that respondents of both samples could differ significantly in terms of their attitude to branded products. The results in Serbia show that, during a purchase, men are more likely to take into account the country of the brand's origin than women and that they believe that the brand guarantees sufficient product quality. On the other hand, women rather than men are more likely to try a new domestic brand when it appears on the market. The results of the Croatian sample show that women prefer foreign brands less than men, but also they are more willing to communicate their brand impressions to friends, colleagues and relatives. Men in Croatia, as well as in Serbia, believe to a greater extent than women that the brand guarantees them enough product quality.
The regression analysis was used to examine whether the respondents' attitude to the branded products could be predicted on the basis of independent variables. The analysis showed that this linear combination of the predictors in both samples proved to be significant for predicting most of the dependent variables that were used to examine the respondents' attitude to the branded products.
Pearson's correlation coefficient examined the inter-relation between the dependent variables which were used to examine the respondents' attitude to the branded products, and the results showed that the highest level of correlation regarding the Serbian sample was found within the following variables: I prefer foreign brands and For the same category of products I would rather pay more for foreign brand. On the other hand, the highest level of correlation regarding the Croatian sample was found between the following variables: I like to try the new domestic brand when it appears on the market and I like to try out a new foreign brand when it appears on the market.

The limitations of research and recommendations
A relatively small sample of respondents (n ¼ 264 subjects in Serbia, n ¼ 291 subjects in Croatia) could be considered as the main constraints of the conducted research. The samples are not completely comparable according to their characteristics, e.g., the Croatian sample, unlike the Serbian one, has a higher proportion of women compared to men; and this is not a result of the present population in Croatia). The Croatian sample illustrated a similar situation with the age structure, therefore it cannot be claimed that the sample is consistent regarding the characteristics of the population.
The very selection of on-line testing as a way of collecting primary data, whereby the participants themselves decided whether to participate in the research or not, is a kind of restriction which indicates that statistically neither its representativeness of the population nor the possible level of this (observed through reliability and a possible range of deviations) can be implied. Although this is a significant limitation, we believe that, due to the number of respondents (a total of 555 in both samples; 264 in Serbia and 291 in Croatia); the obtained results can be considered highly indicative for the reached conclusions.
An additional limitation refers to the fact that the brands were not observed according to the branches they represent (e.g., food brands, technical brands, service brands, etc.), but merely as a generic concept: domestic and foreign brands.
Also, an additional limitation refers to the methodology and statistical treatment (descriptive analysis, multiple regression and correlations) which can be considered as basic, but it could be treated as a preliminary investigation of the phenomena.
For future research we recommend a bigger sample in total and per country and at the same time in additional countries besides Croatia and Serbia. Additionally, we recommend usage of a more complex approach, i.e., Structural Equation Modelling for providing more detailed insight into the phenomena.
The limitations are evident, but we feel that they could not significantly devalue the obtained results and the conclusions we have reached. The presented results and conclusions are in our opinion a good basis for further research in the field.

Conclusion
The concluding remark is that the differences are not so evident-at the overall level they are not large, but not irrelevant for brand management on the Croatian and Serbian markets, since they are statistically significant. The majority of respondents (92.4% in Serbia and 83.5% in Croatia) said they did not buy exclusively branded goods. In general, respondents neither agreed nor disagreed with the statement 'I preferred foreign brands to domestic' (in Croatia mean is 2.53 vs 3.05 in Serbia, mean on a 1-5 Likert scale; where 1 represents the answer (personal assessment) I completely disagree; 2 ¼ I disagree; 3 ¼ I am not sure; 4 ¼ I agree; and 5 ¼ I completely agree). At the same time respondents mainly disagree with the statement 'I would rather pay more for foreign brand in the same category of product'; M ¼ 2.57 in Serbia and M ¼ 2.22 in Croatia (mean on a 1-5 Likert scale). At the same time the highest agreement is achieved with the statements 'I usually communicate positive or negative brand impressions to friends, colleagues and relatives' (in Serbia, M ¼ 3.86 and M ¼ 4.05 in Croatia) and 'I believe that large companies send goods of lower quality to the countries of Eastern Europe and the Balkans' (in Serbia, M ¼ 3.82 and M ¼ 4.24 in Croatia).
In our opinion, the overall results and all the identified differences are 'operational' in the marketing sense, i.e., they are susceptible to marketing strategies and activities through the segmentation process, the selection of target segments and the positioning of brands using the elements of the marketing mix (the basic: 4 P and the extended: 8 P). Still, the degree of standardisation and brand adaptation on individual markets should be taken into account. Also, we should take into account that the brand management is a dynamic category and that in the present circumstances products are merely elements of a brand identity, while brands are an effective tool to achieve market success. Brands can be understood as global and domestic (national) or foreign and local, but it should be taken into account that their functional, emotional and symbolic values have to be harmonised in order to be successful. Among all the brand classifications we believe the best is the one that divides them into successful or unsuccessful on the markets where they are present. Finally, brands are chosen by people and not vice versa. The consumer is still in first place and he will agree on a trade relationship only if he receives more or at least the same of what has been invested, whereby he is the one who evaluates the value of both invested and obtained which is expressed on a functional, financial, cultural and emotional level.
The research gives a better understanding of consumers preferences to domestic vs foreign products and, accordingly, a product selection may be made and define the marketing strategy of the company. Also, the obtained results can be of use to experts and scientists from the state authorities and businesses, since they provide insight into purchasing consumer habits that have a significant impact on the national economy. We hope that this research will be seen as a stimulus for new similar research, while the abovementioned constraints (sample selection, sample representativeness, branding by activity, methodology and statistical treatment, etc.) will be taken into consideration.

Disclosure statement
The authors report no conflicts of interest.