Affective and Cognitive Bases of Implicit and Explicit Attitudes towards Domestic and Foreign Food Brands: Brand Preferences in Ethnic Primary and Minority Groups

Abstract

The aim of the research is to investigate implicit and explicit attitudes towards domestic and foreign food brands and their affective and cognitive bases in Ethnic Primary and Ethnic Minority groups. Measurements: methodologically balanced procedures of implicit association tests (IAT): Self-Concept IAT, Affective IAT, Cognitive IAT; affective and cognitive explicit procedures. In the Primary Ethnic Group, affective-cognitive independence was revealed, which made it possible to develop two constructs using factor analysis: the affective and cognitive bases of ‘overall’ attitude towards food brands. The structure of the variability of these constructs contains both associations and propositions. Consumer preferences formed on affective and cognitive bases were stronger for domestic food brands than for foreign ones. This has some analogies with both explicit and implicit consumer ethnocentrisms. In the Ethnic Minority Group, implicit-explicit independence was revealed, and implicit and explicit ‘overall’ attitudes towards food brands were built. The structure of the variability of these ‘overall’ attitudes contains affective and cognitive components. Consumer preferences based on implicit ‘overall’ attitude towards food brands are stronger for domestic food brands than for foreign ones. This has some analogies with implicit consumer ethnocentrism. Consumer preferences based on ‘overall’ explicit attitude towards food brands are ambivalent or unpronounced.

1. Introduction

Study of socio-psychological patterns of consumer attitudes and consumer behaviour has gained in importance of a scientific problem related to the factors that determine consumer preferences and the process of making decisions. Decision making is a complex psychological process involving the motivation, emotions and cognitions of the consumer. Studies by scientists working in the field of consumer behaviour psychology substantiate that it is not enough to consider only conscious, intentional processing of information in order to make a consumer choice, but also indicate the need to bear in mind automatic information processing (e.g. [1]). In psychology, consumer behaviour is understood as a special type of social behaviour, when the interest of researchers is aimed at understanding the socio-psychological regulators of this domain: values, norms, attitudes, social stereotypes, attributive processes. An analysis of publications (e.g. [2]) provides an insight into the intensity of research on consumer behaviour.

1.1 Brand

Brand name is “the unique identifier that characterises a product as belonging to a particular manufacturer or distributor”. “Brand names enable consumers to differentiate between many products that serve the same function. A particular brand name may carry expectations of value, reliability, service, and perhaps status for the user”, “brands are designed to enhance consumer appeal for a product and to boost the manufacturer’s image” [3]. Brand studies combine knowledge from cognitive psychology, social psychology, neuroscience, sociology, and the field of marketing and consumer behaviour ([4], p. 9). In social psychology, the brand is considered as one of the forms of social communication and as an object of social cognition. The peculiarity of the brand as the object of social cognition is manifested in the fact that the brand can be considered, firstly, as a materialised object (e.g. ad, logo), and secondly, as an image that is the result of social perception ([5], p. 45). This approach deserves attention, since the image as a result of the perception process is the basis for specific behaviour, and the consumer, when deciding, relies directly on the image embodied in the brand; moreover, the perception of these images may be different for everyone.

1.2 Some approaches to understanding brands

Kevin Lane Keller [6] developed an approach based on the achievements of cognitive psychology, where the brand is viewed as a cognitive construct in the form of an image of the world. When perceiving a brand, the consumer experiences a stream of associations that influence the final image of the brand. Jennifer Aaker [7] developed a personality approach based on advances in personality psychology. This approach proposes the concept of “brand personality”, which is defined as a set of personality traits associated with a brand. The relational approach was developed based on the achievements of existential psychology. This approach continues to develop an understanding of brand-consumer relationships that are analogous to interpersonal interactions [8]. At the beginning of the twenty-first century, due to socio-cultural, political, economic changes, there is a paradigm shift in the understanding of the brand. A societal approach is emerging that analyses issues related to the anthropology of consumption, socio-cultural influences and consumer empowerment [9]. The cultural approach, on the other hand, reflects the extent to which a brand can “embody” in a culture, become meaningful in it, and transform from a brand into a “brand icon” [10]. Currently, the most important socio-psychological mechanisms of brand perception, in the context of the formation of its attractiveness, are personification, identification and categorisation. Several researchers recognise that the presence of a brand significantly reduces the time for choosing a product, as it serves as a signal to the consumer to retrieve enough information from memory to decide about the product.

1.2.1 Attitudes towards brand

In recent years, the attention of researchers has been focused not only on the study of factors that affect consumer awareness, but also on the study of unconscious processes that affect consumer preferences [1, 11, 12, 13, 14]. Consumer behaviour is influenced by beliefs and consumer attitudes towards brands or brand attitudes, which are formed throughout life. Consumer attitudes are associated both with automatic, unconscious affective and cognitive processes, and with controlled, conscious ones. Considering the processes underlying consumer attitudes will contribute to a better understanding of how overall attitudes are formed.

1.2.1.1 Beliefs

“Beliefs may be based on knowledge, opinion, or faith, and they may or may not carry an emotional charge.” … “Manufacturers are very interested in the beliefs people have about their products and services” ([4], p. 95). These beliefs shape the product and brand image, and people respond on that image [4].

1.2.1.2 Brand origin country

What is important to marketers is that consumers often have different beliefs about brands or products or product brands depending on their country of origin. The influence of the country of origin varies depending on the type of product. In addition, attitudes towards country of origin may change over time ([4], p. 95).

1.2.1.3 Attitudes

There are many definitions of attitude in psychology. They all involve evaluative judgements about the object of the attitude, which recapitulate decisions about liking or disliking, summarising various types of information about a problem, object or person. In these evaluative judgements, thoughts, feelings and past experiences are brought together into a single evaluative summary. Attitudes vary in valence and strength [15]. For example, Russell Fazio ([16], p. 247) defines attitude as “associations in memory between a given object and a given summary evaluation of the object—associations that can vary in strength and, hence, in their accessibility from memory”. According to this definition, “an attitude can be viewed as a simple two-node semantic network, with one node representing the object, the second node the global evaluation of the object, and the link between the two nodes the strength of the association” ([17], p. 80). Such an understanding of attitude points to association as a mechanism for the formation of attitudes, and to memory as a storage system for mental representations of attitudes.

1.2.1.4 Affective base, cognitive base and behavioural base of attitude

The earliest theory considered that attitudes consist of three components: affect, cognition and behaviour, which have been used respectively (e.g. [18]) to describe:

• Positive and negative feelings that one holds towards an attitude object.

• Beliefs that one holds about the attitude object.

• Overt actions and responses to the attitude object.

Contemporary researchers of attitudes argue that attitude does not consist of affect, behaviour and cognition, but it is a general evaluative summary of the information derived from its affective base, cognitive base and behavioural base (e.g. [17], p. 82). The modern conceptualisation of the attitude is that it represents an overall assessment of an object based on cognitive, affective and behavioural information [15, 19].

1.2.1.5 Implicit and explicit attitudes

Consumer behaviour is highly dependent on both automatic, unconscious cognitive processes and controlled, conscious ones [20, 21]. Accounting of the automatic affective and semantic (cognitive) processes underlying consumer attitudes will contribute to a deeper understanding of how the general representation of attitudes is formed. Theoretical models present implicit and explicit attitudes as two qualitatively different ways of processing cognitive information. One of them is associative, based on associations of similarity, contiguity and space, time, while the other, a propositional process, operates based on logical analysis and reasoning [22, 23]. Both explicit and implicit attitudes can be seen as the result of a learning process whose mechanisms are of a dual nature. Both mechanisms correspond to two independent and competing systems: associative and propositional. Associations and propositions, as forms of knowledge, differ in the way they are mentally represented and in the conditions under which they are activated and guide behaviour [24]. Implicit attitudes are predominantly seen as the result of associative processes, while explicit attitudes are seen as propositional [22, 25]. The process of actualisation of attitudes may require volitional efforts and be conscious, controlled, or, on the contrary, be spontaneous, unconscious, automatic. In the first case, we are talking about explicit attitudes, and in the second about implicit ones.

1.2.1.6 Memory system model (MSM)

Research [12] devoted to the study of attitudes towards domestic and foreign food brands highlights the contribution of the multisystem memory model (MSM) [26] to the development of implicit social cognition. MSM differs in the way it implicitly processes information from traditional two-process models of implicit social cognition based on automatic and controlled processes. According to MSM, implicit processes represent multiple systemic representations of associative knowledge, while two-process models reflect single-system representations of associative knowledge [26, 27]. The emergence of the MSM model, which includes many forms of learning and memory associated with different neural substrates, and perceptual and behavioural systems, has influenced a different understanding of implicit social processes [26, 27, 28]. Exploring intergroup implicit attitudes using the MSM approach [29, 30, 31], it was found that affective forms of implicit bias correspond to affective forms of learning and memory, which are supported by the amygdala and related subcortical contours. In contrast, implicit stereotypes reflect semantic associations that correspond to conceptual forms of learning and memory and to corresponding areas of the neocortex such as the left prefrontal cortex ([28], as cited in Ref. [12]). According to MSM, implicit attitude can “reflect a combination of affective and semantic (i.e. cognitive) associations”. This statement made it possible to study implicit consumer attitudes, considering how they are based not only on implicit, automatic, affective associations, but also on automatic, conceptual, cognitive associations [12]. These provisions of the MSM theory were the theoretical basis for the creation of various versions of affective and cognitive implicit association tests (IAT).

The aim of research is to investigate the affective and cognitive bases of implicit and explicit attitudes towards domestic and foreign food brands, as well as brand preferences in ethnic primary and ethnic minority groups.

Research questions

1. Are there any relationships between the results of implicit and explicit measurements of attitudes towards domestic or foreign food brands?

2. What common factors underlie the relationships between the results of measuring attitudes towards domestic and foreign food brands using Affective IAT, Cognitive IAT, Self-Concept IAT, Explicit Affective Procedure and Explicit Cognitive Procedure?

3. How are consumers distributed according to the preferences for domestic and foreign food brands?

2. Method

2.1 Participants

The number of participants is 203. Two ethnic groups were formed by the self-determination of the participants. All participants live in the same main city of the research country for almost their entire lives.

“Ethnic Minority Group”—124 were aged 17–57 years old (Mdn = 30.5, M = 31.3, SD =9.70). Male—30%, female—70%. Education: general—45%, bachelor’s degree—40%, master’s degree—15%. Family status: unmarried—52%, married—45%. Occupation: unemployed—40%, employed—60%. Monthly income (EUR) per person: 12%—less than 500, 69%—from 500 to 1000 EUR, 19% more than 1000 EUR.

“Primary Ethnic Group”—79 were aged 16–60 years old (Mdn = 36.0, M = 34.9, SD = 12.6). Male—48%, female—47%. Education: general—46%, bachelor’s degree—35%, master’s degree—19%. Family status: unmarried—48%, married—51%. Occupation: unemployed—11%, employed—87%. Monthly income (EUR) per person: 17%—less than 500, 48%—from 500 to 1000 EUR, 35% more than 1000 EUR.

2.2 Measurements

The attitudes towards domestic and foreign food brands were measured using specially designed and methodologically balanced five procedures [5, 12].

2.2.1 The procedures of affective, cognitive and self-concept implicit association tests (IATs)

The affective and cognitive IAT procedures [5, 12] were developed in accordance with the classic Two-Category IAT [32].

“Affective IAT measures the automatic affective associations of domestic and foreign food brands with attributes having positive or negative valence” ([12], p. 219).

“Cognitive IAT measures the automatic semantic (cognitive) associations of domestic and foreign food brand with attributes that characterise the price and quality of foods” ([12], p. 219).

The procedure for the Two-Category Self-Concept IAT [5, 12] was developed based on the IAT reviewed in Ref. [33].

Self-Concept IAT measures automatic associations between categories: “Self” and “Others” and attributes characterising domestic or foreign food brands ([12], p. 220).

2.2.2 D-scores

D-scores were defined for the affective, the cognitive, and the Self-Concept IAT procedures: D(affective), D(cognitive) and D(Self), respectively. When $D\ge 0.15$, the explicit preferences of domestic food brands occur, when $D\le -0.15$, explicit brand preferences are for foreign foods. When $-0.15<D<0.15$, “no implicit effect is found”, that is two situations are possible: ambivalence of implicit associations or very weak implicit associations [12, 34].¹

2.2.3 Affective and cognitive explicit procedures

The self-assessment procedures were based on some aspects of the Osgood’s semantic differential [5, 12, 35].

“Affective Explicit Procedure measures the explicit preferences of domestic or foreign food brands, which are based on the assessment of bipolar adjectives characterising the affective properties of perception of these food brands” ([12], p. 220). Cronbach’s α = .91: the internal consistency is excellent.

“Cognitive Explicit Procedure measures the explicit preferences of domestic or foreign food brands based on the assessment of bipolar adjectives characterising the perception of the price and quality of the brands of these foods” ([12], p. 220). Cronbach’s α = .88: the internal consistency is good.

2.2.4 E-scores

E-scores were defined for the explicit affective and the cognitive procedures: E(affective) and E(cognitive), respectively. When E ≥ 4.5, the explicit preferences of domestic food brands occur, when E ≤ 3.5, explicit brand preferences are for foreign foods [12].

2.2.5 Questionnaires

A specially designed preliminary questionnaire aimed at finding out how familiar the research participants are with the logos of domestic and foreign food brands [5, 12].

A special demographic questionnaire, including gender, age, family status, education, occupation, income per family member per month, and attitudes towards domestic and foreign food products that participants use on a daily basis, was also used [5, 12].

2.3 Procedure

The researchers followed guidelines for confidentiality, voluntary participation and psychological ethics. Self-reported procedures were followed by all participants. Initially, all 203 participants filled out the preliminary and demographic questionnaires. Firstly they performed IATs, secondly performed Affective and Cognitive Explicit Procedures. The number of participants who performed IATs in each group was as follows. In Ethnic Minority Group: Affective and Cognitive IAT—88; Self-Concept IAT—95, all three IATs were completed by 60 participants. In Primary Ethnic Group Affective IAT—71; Cognitive IAT—72; Self-Concept IAT—71; all three IATs were completed by 70 participants.

2.4 Apparatus

Certified licenced software E-Prime 2®.

2.5 Statistical methods

The study of the data showed that the data distributions allow the use of parametrical statistical methods. The methods of statistical data processing were chosen in accordance with the research questions.

3. Results

3.1 The first research question

The relationships between the results of the measurements using the experimental procedures IATs and self-report procedures were established using Pearson correlation coefficients. The effect size r interpretation is: .10—small, .30—medium, .50—large [36].

In each ethnic group, some relationships between the results of measuring attitudes towards domestic and foreign food brands were found. No inconsistency of measurement results was observed. But the relationships between the results of the measurements in the Primary Ethnic Group and in the Ethnic Minority Group were different (Table 1, Figure 1).

Variables	Pearson r and 95% Confidence Intervals
	Primary Ethnic Group	Ethnic Minority Group
Implicit-Explicit Relationships
D(affective)-E(affective)	r(71) = .15, p = .22, ns; [−.11; .35]	r(88) = .15, p = .17, ns; [−.061; 35]
D(cognitive)-E(cognitive)	r(72) = .21, p = .083, [−.023; .42]	r(88) = .37***, p < .001; [.17; .54]
D(self)-E(affective)	r(71) = .065, p = .59, ns; [−.17; .29]	r(95) = .26*, p = .010; [.065; .44]
D(self)-E(cognitive)	r(71) = .24*, p = .041; [.010; .45]	r(95) = .31**, p = .002; [.11; .47]
Affective-Cognitive Relationships
D(affective)-D(cognitive)	r(71) = .063, p = .60, ns; [−.17; .29]	r(88) = .61***, p < .001; [.46; .73]
E(affective)-E(cognitive)	r(79) = .092, p = .42, ns; [−.13; .31]	r(124) = .87***, p < .001; [.82; .91]
D(cognitive)-E(affective)	r(72) = −.029, p = .81, ns;	r(88) = .45***, p < .001; [.27; .60]
D(affective)-E(cognitive)	r(71) = .028, p = .82, ns;[−.21; .26]	r(88) = .070, p = .52, ns; [−.14; .28]
Self-Concept IAT and Affective IAT, Cognitive IAT Relationships
D(self)-D(affective)	r(70) = .35**, p = .003; [.13; .54]	r(60) = .30*, p = .020; [.050; .51]
D(self)-D(cognitive)	r(71) = .46**, p < .001; [.25; .63]	r(60) = .30*, p = .019; [.053; .52]

Table 1.

The correspondences between brand attitudes measurements results: Correlations, effect sizes and 95% confidence intervals for them.

Figure 1.

Relationships between measurements in two ethnic groups.

3.2 The second research question

To understand what the common factors underlying the relationships between implicit and explicit attitudes towards brands and its affective and cognitive bases are the factor analysis for each ethnic group was done: two factors (Kaiser criterion), Principal Component Method, Varimax Rotation with Kaiser Normalisation, Rotation converged in 3 iterations, Scores based on method “Regression” (s. Tables 2 and 3). For each group, the cumulative percent for two factors of total variance explained, Kaiser–Meyer–Olkin (KMO) and Bartlett’s Tests of Sphericity results are as follows.

Variables	Rotated Component Matrix		Component Score Coefficient Matrix: β-coefficients
	Components		Components
	1	2	1	2
E(cognitive)	.945	.036	.495	−.036
E(affective)	.942	.061	.492	−.021
D(affective)	−.209	.834	−.161	.498
D(cognitive)	.307	.783	.114	.437
D(self)	.066	.652	−.005	.376

Table 2.

Rotated component matrix and component score coefficient matrix. Ethnic minority group.

Variables	Rotated Component Matrix		Component Score Coefficient Matrix: β-coefficients
	Components		Components
	1	2	1	2
D(cognitive)	.813	−.156	.523	−.237
D(self)	.787	.289	.451	.150
E(cognitive)	.556	.076	.335	−.005
E(affective)	−.103	.767	−.161	.674
D(affective)	.256	.703	.069	.573

Table 3.

Rotated component and component score coefficient matrices. Primary ethnic group.

Ethnic Minority Group:

73.2% = 38.4% + 34.8%; KMO = .56; χ²(10) = 95.3, p < .001; N = 60

Primary Ethnic Group:

57.2% = 33.3% + 23.9%; KMO = .55; χ²(10) = 32.8, p < .001; N = 70

3.2.1 Ethnic minority group

It was found that in the Ethnic Minority Group Component 1 is described mainly by explicit variables and Component 2 is described mainly by implicit variables (Table 2). Component 1 can be called by “Explicit ‘Overall’ Brand Attitude”, and Component 2 by “Implicit ‘Overall’ Brand Attitude”.

From Table 2 it follows that the greatest contribution to the Implicit “Overall” Brand Attitude is made by D(affective): ${\mathrm{\beta }}_1=.50$, followed by D(cognitive): ${\mathrm{\beta }}_2=.44$; D(self): The largest contributor to the Explicit “Overall” Brand Attitude is E(cognitive): ${\mathrm{\beta }}_1=.50$, followed by E(affective): ${\mathrm{\beta }}_2=.49$.

Figure 2 shows the percentage of variability of the variables that make up the Implicit “Overall” Brand Attitude and Explicit “Overall” Brand Attitude. These percentages are obtained using a rotated component matrix (Table 2) and for a fixed component (column of the matrix) represent the ratio of the square of the factor loading of a variable to the sum of the squares of all factor loadings of that component.

Figure 2.

Variability percentage of implicit and explicit “overall” attitudes towards brands and their constituent variables.

The variability of Implicit “Overall” Brand Attitude consists of 100% associations and 0% propositions. The variability of Explicit “Overall” Brand Attitude composition contains 7% associations and 93% propositions (Figure 2).

“Cut Points”. Using factor analysis, two independent variables were constructed: Explicit “Overall” Brand Attitude and Implicit “Overall” Brand Attitude. The large values of both “Overall” Brand Attitudes correspond to the preference for the domestic brand, small values—to the preference for a foreign brand. The “cut points” for Explicit “Overall” Brand Attitude are −0.83 and 0.31, for the Implicit “Overall” Brand Attitude are −1.75 and −0.69. The calculation of “cut points” for variables obtained by factor analysis is described in detail in the study [37].

3.2.2 Primary ethnic group

It was found that in the Primary Ethnic Group Component 1 is described mainly by cognitive variables, and Component 2 is described mainly by affective variables. Component 1 can be called by “Cognitive Base of “Overall” Brand Attitude”, and Component 2 by “Affective Base of ‘Overall’ Brand Attitude” (Table 3).

The largest contribution to the Cognitive Base of “Overall” Brand Attitude comes from D(cognitive): ${\mathrm{\beta }}_1=.52$, followed by D(Self): ${\mathrm{\beta }}_2=.45$; E(cognitive): ${\mathrm{\beta }}_3=.33$. The largest contribution to the Affective Base of “Overall” Brand Attitude is by E(affective): ${\mathrm{\beta }}_1=.67$, followed by D(affective): ${\mathrm{\beta }}_2=.58$ (Table 3).

Figure 3 shows the percentage of variability of the variables that make up the Cognitive Base of “Overall” Brand Attitude and Affective Base of “Overall” Brand Attitude. These percentages are also obtained using a rotated component matrix (Table 3) and for a fixed component (column of the matrix) represent the ratio of the square of the factor loading of a variable to the sum of the squares of all factor loadings of that component.

Figure 3.

Variability percentage of affective and cognitive bases of “overall” attitudes towards brands and their constituent variables.

The “Cognitive Base of ‘Overall’ Brand Attitude” variability is due to 37% of implicit Self-associations; 58% of cognitive variables (18% propositions and 40% associations); 5% of affective variables. So, there are 81% of associations and 19% of propositions. Pertinently, this is also implicit attitude, but with a cognitive “strand”.

The variability of “Affective Base of ‘Overall’ Brand Attitude” composition contains 7% of implicit Self-associations; 2% of implicit cognitive; 91% of affective: 42% affective associations and 49% of affective propositions. Therefore, there are 51% of associations and 49% of affective propositions.

“Cut Points”. Using factor analysis, two independent variables were constructed: Cognitive Base of “Overall” Brand Attitude and Affective Base of “Overall” Brand Attitude. The large values of both Bases of “Overall” Brand Attitudes correspond to the preference for the domestic brand, small values—to the preference for a foreign brand. The “cut points” of the Affective Base of ‘Overall’ Brand Attitude are −1.51 and −0.038; for the Cognitive Base of “Overall” Brand Attitude are −1.64 and −0.46. The calculation of “cut points” for variables obtained by factor analysis is described in detail in the study [37].

3.3 The third research question

To find out how consumers are distributed according to their preferences for domestic or foreign food brands, pie charts were plotted, as well as Fisher’s angular transformation test and Cohen’s effect sizes h (0.2 for small effect, 0.5 for medium effect, 0.8 for large effect) [38] were used to compare the percentages.

Figure 4 illustrates consumers’ distributions demonstrating preferences for domestic and foreign food brands, as well as the uncertain preferences, which are without the revealed preference effect, that is ambivalent or unexpressed.

Figure 4.

Distributions of consumers’ preferences in two groups.

Preferences distributions were found for the results of

• the Affective IAT—D(affective), Cognitive IAT—D(cognitive), Self-Concept IAT—D(self),

• Affective Explicit Procedure—E(affective), Cognitive Explicit Procedure—E(cognitive),

• as well as for the Implicit “Overall” Brand Attitude, the Explicit “Overall” Brand Attitude, Cognitive Base of “Overall” Brand Attitude and Affective Base of “Overall” Brand Attitude retrieved using factor analysis.

3.3.1 Ethnic minority group

In Ethnic Minority Group, the distribution of explicit preferences for domestic and foreign food brands was determined by Explicit “Overall” Brand Attitude; that is, its variability contains 7% associations and 93% propositions. The distribution of implicit brand preferences was determined with the help of Implicit “Overall” Brand Attitude, because its variability consists of 100% associations and 0% propositions (Figure 2). Implicit preferences for the domestic food brand dominate over other implicit preferences: from 63.3 to 78.3%. In the “worst” case, Ethnic Minority Group, D(cognitive) (N = 60), comparing the percentage of implicit “admirers” of domestic brands (63.3%) with participants who prefer foreign brands, or with undetected effects (100%—63.3% = 36.7%), we get a statistically significant result: φ^* = 2.95, p = .003, the effect size Cohen’s h = 0.54 is medium.

In the same group, Explicit “Overall” Brand Attitude contains only 31.7% of domestic brand explicit preferences, and 56.7% of uncertain preferences. This is already a statistically significant dominance: φ^* = 2.79, p = .005, the effect size Cohen’s h = 0.51 is medium.

3.3.2 Primary ethnic group

In Primary Ethnic Group the variability of Cognitive Base of “Overall” Brand, contains 81% associations and 19% propositions (Figure 3). In essence, this is also implicit attitude. Therefore, 65.7% of consumers showed implicit preference for domestic brands. In the “Affective Base of ‘Overall’ Brand Attitude”, 58.6% of consumers showed domestic brand affective preferences. This is statistically significantly more than the other part of consumers (35.7%): φ^* = 2.54, p = .011, the effect size Cohen’s h = 0.46 is almost medium.

4. Discussion

4.1 The first research question: relationships

Bertram Gawronski and Skylar Brannon ([39], p. 7) wrote about a widespread assumption in attitude research, that is “that measurements of implicit evaluation provide valuable information that cannot be gained from measurements of explicit evaluations”. An analysis of numerous empirical studies conducted in recent years indicates conflicting opinions regarding the understanding of relationships between implicit and explicit dimensions of the same psychological construct [40]. In 2014, Bar-Anan and Nosek expressed surprise that, despite advances in various areas of implicit social cognition, very little is known about the relationship between implicit measurement results. “Little is known about the relations among indirect measurements. The fact that two measurements are indirect does not itself guarantee that they are influenced by similar psychological processes, predict similar behaviours, measure the same construct, or even correlate with one another” ([41], p. 699). Analysis of empirical research in recent years has also revealed a lack of empirical research using several implicit measurements simultaneously for different psychological constructs. Particularly, this lack concerns the study of consumer attitudes. Therefore, the formulation of the first research question is appropriate.

The effect size of the relationship effect depends on the type of attitude, on the participants in the research, and on the measurement conditions [41].

One of the reasons for the discrepancy between the results of implicit and explicit measurements may be related not only to the problem of the unreliability of self-assessment procedures, but also to the need to consider those variables that may affect the consistency of the measurement results.

Therefore, it is necessary to consider different variables and their associated different subgroups of participants, as well as experimental conditions that can affect the correspondence between implicit and explicit measurements [12, 25, 42, 43, 44, 45, 46, 47, 48].

All found relationships indicate the presence of convergent validity of the respective measurement procedures.

When considering the correspondence between implicit and explicit measurements, it is necessary to pay attention to the procedural features of constructing implicit and explicit measurements [49]. For example, if the target categories in an implicit association test are visual stimuli (e.g. a visual stimulus is a brand logo), then when constructing explicit methods, identical stimuli must be used. This study also used the principle of methodological balance [12]. Many researchers believe that if this feature is not considered, this can lead to an erroneous conclusion about genuine correspondences between implicit and explicit measurements [39, 40, 50].

In this research, we used implicit and explicit measurements, specially developed by us, related to the affect, cognitive basis of attitude towards domestic and foreign product brands [5, 12]. Therefore, we hope that our research will reduce the lack of existing implicit and explicit measurements to measure attitudes towards brands of products of domestic and foreign manufacturers.

4.1.1 Implicit-explicit relationships

The issue of correspondence of the results of implicit and explicit measurements is one of the important methodological issues in implicit social cognition [40].

For the research of implicit attitudes vs. explicit attitudes in each of the groups, we examined the relationship between the results of the following measurements:

• Affective IAT and Affective Explicit Procedures,

• Cognitive IAT and Cognitive Explicit Procedures,

• Self-Concept IAT and Affective Explicit Procedures,

• Self-Concept IAT and Cognitive Explicit Procedures.

The results are presented in Table 1, in Figure 1 and in paragraph 4.1.1.

The correlation between the two types of assessments may be low for various reasons. Let us consider some conditions under which the results of implicit and explicit measurements may or may not agree with each other.

4.1.1.1 Introspective inaccessibility of implicit representations

According to Ref. ([51], p. 8), implicit attitudes are “introspectively unidentified (or inaccurately identified) traces of past experience that mediate favourable or unfavourable feeling, thought, or action”. Based on this conceptualisation, it is often assumed that self-report measurements reflect conscious attitudes that are introspectively accessible, while implicit measurements reflect unconscious attitudes to which people do not have introspective access. Introspection can increase awareness of previously unconscious (sub) implicit attitudes (e.g. [52]). Therefore, the correspondence between explicit and implicit measurements may increase the more time people spend on introspection. One explanation for the results of the research may be that too many research participants may not devote enough time to analysing their internal attitudes towards domestic and foreign brands [17, 39].

The discrepancy between the results of using implicit and explicit measurements can also be explained by using two-process attitude models. Let us consider some of them.

4.1.1.2 Consistency of implicit and explicit measurements and the MODE model. Motivational influences

Implicit and explicit attitudes have a single underlying mental representation. But their differences are determined by the level of mental processes that determine behaviour. If the motivation or ability to participate in labour-intensive processes is low, then the subsequent behaviour is influenced by an automatically activated attitude, and if it is high, then the behaviour is determined by “thoughtful attitudes” [53, 54]. The implicit-explicit correspondence of the measurement results can be explained in terms of the MODE model. Implicit measurements, as a rule, are not subject to motivational influences. Explicit self-reports are often influenced by the social desirability of research participants. It follows from Fazio’s MODE model of attitude that explicit and implicit measurements should be highly correlated for participants who are unmotivated and unable to control their explicit self-reports [55]. Explicit-implicit correspondence can be high for relatively “utilitarian” research topics, for example, when studying consumer attitudes.

4.1.1.3 Consistency of implicit and explicit measurements and Wilson’s dual attitude model

Correlations can also be explained in terms of the Dual Attitude Model [55], which uses two attitudes in relation to the same attitude object. The Dual Attitude Model assumes that there are two mental representations that are separate and do not interact with each other. Within the framework of the model under consideration, the implicit attitude is associated with stable memory structures, while the explicit attitude is associated with a new human experience that has not yet entered deeper memory structures but is already relevant and used. The discrepancy in the measurement results is due to the influence of deep (stable) memory structures. The results of the ratio of direct and indirect measurements depend on the level of cognitive effort and interest of the subjects in the results of measurements [55] ([25], p. 19). The Dual Attitude Model is susceptible that when attitudes change, the old implicit attitude remains in memory, influencing people’s judgement and behaviour when they are unable or unmotivated to participate in the laborious process of retrieving their new explicit attitude from memory.

Suppose that in response to new information a person changes his/her explicit attitude to a new explicit attitude. The research participant makes a certain cognitive effort to retrieve this new attitude from memory. This new explicit attitude is reflected in self-reported measurements. Wilson uses the term “implicit” for perfectly learned old attitudes that cannot be erased from memory. Thus, measurements based on the results of implicit measurements “capture” well-studied old attitudes (which have already become implicit), which are automatically activated as a response to the object of attitude [39].

For example, some domestic brand M of dairy products has been explicitly and already implicitly associated with positive incentives for decades. The research participant hears from acquaintances or learns from the press that new technologies have been used to make the brand M, which has made this brand worse. The explicit attitude of this participant to the brand M changes a lot.

Implicit Measures reflect old ideas about the domestic brand, and they are most likely to be automatically activated when confronted with a relevant stimulus as a response to the object of attitude. Most likely, these are positive implicit associations of the domestic food brand M (affective basis, cognitive basis, associations with oneself).

Explicit Measures. There are two options.

Option I. The research participant does not use his/her cognitive ability to retrieve recently acquired ideas from memory, such as deterioration in the quality and taste of the domestic product brand M, its price increase. Therefore, when performing self-reporting procedures, this research participant is likely to make his/her judgements spontaneously. Explicit-implicit compliance of measurements can be high.

Option II. The research participant uses cognitive efforts to retrieve recently acquired perceptions from memory, such as the deterioration in the quality and taste of the M brand, its price increase. Therefore, when performing self-reporting procedures, this research participant will evaluate his/her judgements deliberately. His/her explicit attitude towards domestic food brands will never be the same. Explicit-implicit correspondence of measurements is likely to be low.

It is possible that until these implicit associations change, the explicit-implicit discrepancy is likely to remain and lead to bad correlations.

Thus, the Wilson, Lindsey, and Schooler Two Attitude Model relies on implicit measurements reflecting old beliefs that are automatically activated when confronted with a relevant stimulus [55]. In contrast, explicit measurements suggest that they reflect such old beliefs only when people lack the motivation or cognitive ability to retrieve newly acquired beliefs from memory [17].

The Two Attitude Model explains very well the implicit and explicit consumer preferences of domestic and foreign food brands (Figure 4).

4.1.2 Implicit-explicit relationships: main results

The range of correlation coefficients varies depending on the studied psychological construct. A meta-analysis [20] has shown that quite high correlation coefficients are observed in studies of consumer attitudes, a relatively “utilitarian” area compared to a socially sensitive area such as group relationships [56, 57, 58]. The same conclusion was reached in a meta-analysis [49] by Hofmann, Gawronski, Gschwendner, Le, and Schmitt (2005), based on a sample of 126 studies, with the mean effect size of .24.

In the present research, the correlation coefficients of implicit-explicit attitudes towards domestic and foreign food brands in the Ethnic Minority Group were statistically significant and showed (medium) average correspondence of measurements (.27).

In Primary Ethnic Group, the mean effect size of consistency was small (.18). Statistical significance was found only for Self-associations with both cognitive and affective bases of explicit attitudes.

None of the controlled variables (ethnic group, gender, age, education, marital status, profession, income per person, per month) was found that would have an impact on the studied relationships.

Correlations between implicit and explicit measurements are generally rather low. This result may be explained as follows.

4.1.2.1 Introspection

Quite a few participants in the research could not devote enough time to analysing their internal attitudes towards domestic and foreign brands.

4.1.2.2 Distortion of information

Research participants could be fully aware of the attitude that is captured by the implicit measurements, yet they reported a different attitude in the self-reporting procedures [17, 39].

4.1.2.3 Motivational influences

Sufficiently high motivation of participants in the performance of self-reporting procedures influence implicit-explicit relationships.

Dual Attitude Model assumes the simultaneous existence of two attitudes. (1) Attitude whose implicit measurements reflect old beliefs that are automatically activated when confronted with a relevant stimulus. (2) Attitude whose explicit measurements suggest that they reflect old beliefs only when people lack the motivation or cognitive ability to retrieve newly acquired beliefs from memory. Participants in the research most likely either did not change old implicit associations or did not develop sufficient motivation or cognitive ability to retrieve newly acquired representations from memory when performing a self-reporting procedure. Therefore, the revealed correspondence of the results, most likely, showed insufficiently good correlations.

The obtained results and their interpretation are consistent with those of Hofmann et al. [49].

Research [59] has also shown that consumers of generic foods and well-known food brands can also simultaneously have two different attitudes towards an object: an explicit one, which corresponds to thoughtful behaviour, and an implicit one, which corresponds to spontaneous behaviour. Participants whose explicit and implicit preferences were incongruent were more likely to choose their implicit preferred brand when there was little time to choose. But with a sufficiently long time allotted for selection, they were more likely to choose an explicitly preferred brand. Therefore, impulsive behaviour and implicit measurements are important for research in the field of consumer behaviour.

The correspondence between implicit and explicit measurements can also be explained in terms of other attitude models.

4.1.3 Affective-cognitive relationships

Advances in neuroscience have led to the concept of MSM, according to which, implicit processes are multiple systemic representations of associative knowledge, and that implicit attitude can reflect a combination of affective and semantic (i.e. cognitive) associations [28]. In implicit social cognition, it has been argued that affective and cognitive forms of information processing can be performed automatically and are different processes, and that these processes can influence implicit attitudes [26, 27, 28, 30].

To research the relationship between the affective and cognitive foundations of attitudes in each of the groups, we investigated the relationship between the results of the following measurements:

• Affective IAT procedures and Cognitive IAT procedures.

• Affective Explicit Procedure and Cognitive Explicit Procedures.

• Affective IAT Procedures and Cognitive Explicit Procedures.

• Cognitive IAT and Affective Explicit Procedures.

The results are presented in Table 1 and Figure 1.

Primary Ethnic Group. No relationships were found.

Ethnic Minority Group. Three relationships at a significance level of α = .001 between the results of the following measurements were revealed:

• Affective and Cognitive IAT, the effect size is large,

• Affective and Cognitive Explicit Procedures, the effect size is large,

• Cognitive IAT and Affective Explicit Procedure, the effect size is almost large.

The results indicate that in the Primary Ethnic Group, the affective bases of attitudes towards brand and the cognitive bases of attitudes towards brand are independent psychological constructs.

Affect and cognition have long been regarded as independent entities. There are three classical views of the independence of affect ([60], part I). These are:

• Affective independence hypothesis, that emotion is processed independently from cognition ([60], pp. 1213–1217),

• Affective primacy hypothesis, that evaluative processing precedes semantic processing ([60], pp. 1213, 1217–1221),

• Affective automaticity hypothesis, that affectively potent stimuli commandeer attention and evaluation is automatic ([60], pp. 1213, 1221–1225).

Research in the fields of Cognitive Science and Neuroscience suggests that these two concepts are closely related. Therefore, the absence of affective-cognitive correlations in the Primary Ethnic Group should be considered as a peculiarity inherent in this group, the meaning of which will be clear in the future.

4.1.4 Self-concept IAT and affective IAT, cognitive IAT relationships

The results of measurements using Self-Concept IAT correlate with the results of measurements using both Affective and Cognitive IAT. All results are statistically significant, with the above-average effect sizes (Table 1, Figure 1).

4.2 The second research question: “overall” attitudes

In each of the ethnic groups, with the help of factor analysis, two “overall” attitudes, which are independent constructs, were identified and stored. For them, “cut points” were calculated.

In the Ethnic Minority Group, these were “Explicit ‘Overall’ Brand Attitude”, and “Implicit ‘Overall’ Brand Attitude”, for which the property of implicit-explicit independence was satisfied. In the Primary Ethnic Group, these were “Affective Base of ‘Overall’ Brand Attitude” and “Cognitive Base of ‘Overall’ Brand Attitude”, for which the property of affective-cognitive independence was satisfied.

The large values of “Overall” Brand Attitudes and their affective and cognitive bases correspond to the preference for the domestic brands, small values—to the preference for foreign brands.

The structure of all “overall” attitudes is shown in Table 4.

Measurements and Results	Implicit “Overall” Attitude	Explicit “Overall” Attitude	Affective Base of “Overall” Attitude	Cognitive Base of “Overall” Attitude
	Ethnic Minority Group		Primary Ethnic Group
Affective Explicit		46%	49%	1%
Cognitive Explicit		47%		18%
Affective IAT	40%	2%	42%	4%
Cognitive IAT	35%	5%	2%	40%
Self-Concept IAT	25%		7%	37%
Associations	100%	7%	51%	81%
Propositions	0%	93%	49%	19%
Affective Associations	40%	48%	91%	5%
Cognitive Associations	35%	52%	2%	58%
Self-Others Associations	25%	0%	7%	37%

Table 4.

Percentages of variability in the components of ‘overall’ attitudes.

Notes. The results of measurements implicit (automatic) associations or explicit preferences of domestic or foreign food brands:

Affective Explicit. The explicit preferences, which are based on the assessment of bipolar adjectives characterising the affective properties of perception of brands.

Cognitive Explicit. The explicit preferences based on the assessment of bipolar adjectives characterising the perception of the price and quality of the brands of these foods.

Self-Concept IAT. The automatic associations between categories: “Self”—“Others” and attributes characterising food brands.

Affective IAT. The automatic affective associations with attributes having positive or negative valence.

Cognitive IAT. The automatic semantic (cognitive) associations with attributes that characterise the price and quality of foods.

For a long time, consumer attitudes have been viewed as one-dimensional constructs reflecting the overall evaluation of a product, brand, or ad, formed by evaluating specific attributes or characteristics of the attitude object (s. [61]). At present, the affective and cognitive (utilitarian and hedonistic) foundations on which the attitude is formed and changed are considered. The affective basis refers to the sensations, feelings, and emotions that a person experiences in response to the object of the attitude. The cognitive basis of attitude contains positive and negative characteristics and beliefs about an object (e.g. nutritional value, health effects, convenience). For example, the affective and cognitive bases of attitudes towards foods in [62, 63, 64] have been identified.

The study [65] also examined “overall” implicit attitude towards food, which had two independent bases: cognitive and affective. All three constructs were defined as measured by three IAT procedures having the same categories of “chocolate” and “apple”. The attributes were different. IAT attributes for “overall” attitude were affectively coloured words. Attributes of affective IAT were affectively coloured words reflecting taste. Attributes of cognitive IAT were words reflecting the choice of healthy and unhealthy foods. A regression analysis was carried out with the dependent variable “overall” attitude. Predictors: the results of measurements of the cognitive and affective bases of attitude that were independent of each other. At high cognitive load [65], choice was unambiguously predicted by Affective IAT, and at low cognitive load and for participants with low impulsivity, choice was unambiguously predicted by cognitive IAT. The affective and cognitive bases of implicit attitudes have been shown to influence choice in a variety of settings.

Consistent with the findings of cognitive neuroscience studies of implicit learning and memory systems (e.g. [27]), overall implicit attitude towards food is determined not only by automatically perceived appetency (i.e. a measurement of the affective basis of implicit attitude) but also by automatically perceived health (i.e. measuring the cognitive basis of the implicit attitude).

The study [66] argues that cognitive implicit attitudes and affective implicit attitudes towards “green” products are different constructs. The cognitive component can be distinguished from the affective one, which is consistent with the results obtained in [65].

4.3 The third research question: consumer preference distributions

In both ethnic groups, preferences of domestic food brands for Implicit “Overall” Brand Attitude, Affective Base of “Overall” Brand Attitude, and Cognitive Base of “Overall” Brand Attitude dominate: 58.6% to 78.3% of preferences. However, the process of brand preference formation proceeds in different ways.

The result of our research, in our opinion, is interesting in that a phenomenon is mathematically built which is quite plausible and very similar to the construct of consumer ethnocentrism [1] even though it was not part of our goal.

In the study [1], which aimed at finding out whether consumer ethnocentrism is a purely conscious mechanism based on ideology [67] or whether it is an automatic, unconscious process, the concept of Implicit Consumer Ethnocentrism (ICE) was introduced. The latter is defined as a strong and automatic preference for domestic products over foreign ones. ICE was measured using the IAT [1]. Along with conscious, deliberate judgements, consumer ethnocentrism may also be based on automatic “in-group brand favouritism” rather than real knowledge of the brand image or experience with brands. Regardless of conscious evaluation, domestic products can be perceived on an automatic level more favourably than foreign ones. As a result, in some situations, consumers will choose foreign, and in other—domestic products, regardless of their conscious preferences. At an automatic level, consumers prefer domestic products [1].

4.3.1 Ethnic minority group

In the Ethnic Minority Group, the Implicit “Overall” Brand Attitude maintains a pattern found in separate implicit dimensions using Affective, Cognitive, and Self-Concept IAT. The structure of its variability also testifies to this: 100% associations, 0% propositions. As a result, it was found that 78% of consumers prefer domestic food brands. Consumer preferences based on implicit overall attitude are much stronger for domestic food brands than for foreign ones. This is very similar to the phenomenon of implicit consumer ethnocentrism [1].

The variability of the Explicit “Overall” Brand Attitude, consisting of 7% of associations and 93% of propositions, leads to a different pattern: most consumers (57%) do not have any revealed preferences. This is a statistically significant dominance with an average effect size. Consumer preferences based on overall explicit brand attitude are mostly uncertain: either ambivalent or unpronounced. This fact might indicate the presence of some process of changing explicit consumer preferences. It is most likely that preferences formed on propositions will lead to the fact that the established stereotypes will be questioned.

4.3.2 Primary ethnic group

In the Primary Ethnic Group, the variability of Cognitive Base of “Overall” Attitude contains 81% associations and 19% propositions (Figure 3) and is essentially an implicit attitude. It reveals the fact that 66% of the participants prefer domestic brands. A representation based on explicit measurements using Affective and Cognitive Explicit Procedures separately (Figure 4) no longer corresponds to reality. Consumer preferences based on the cognitive basis of “overall” brand attitude are much stronger for domestic food brands than for foreign ones. This is also very similar to the phenomenon of implicit consumer ethnocentrism [1].

Variability of Affective Base of “Overall” Brand Attitude contains 51% associations and 49% affective propositions (Figure 3). About 98% of the components have an affective “origin”. It reveals the fact that 59% of the participants prefer domestic brands, and most of them are based on emotions. Consumer preferences formed on the affective basis of “overall” brand attitude are much stronger for domestic food brands than for foreign ones. This is reminiscent of the phenomenon of implicit-explicit consumer ethnocentrism based on emotions.

5. Conclusions

The aim of the research is achieved, the answers to the research questions are obtained.

Subject to the methodological consistency of measurement procedures, both implicit-explicit independence and affective-cognitive independence of attitudes towards domestic and foreign food brands can manifest themselves.

In the Primary Ethnic Group, affective-cognitive independence was revealed, which made it possible to develop two constructs using factor analysis: the Affective and Cognitive Bases of “Overall” Brand Attitude. It was demonstrated that in the structure of the variability of these constructs there are both associations and propositions. Consumer preferences formed on affective bases were much stronger for domestic food brands than for foreign ones. This has some analogies with the implicit-explicit consumer ethnocentrism, as well as the explicit consumer ethnocentrism with an affective focus.

Consumer preferences formed on cognitive bases were much stronger for domestic food brands than for foreign ones. This has some analogies with consumer ethnocentrism, mainly, with implicit consumer ethnocentrism.

In the Ethnic Minority Group, implicit-explicit independence was revealed, which made it possible, using factor analysis, to build two “overall” attitudes towards domestic and foreign food brands: Implicit “Overall” Brand Attitude and Explicit “Overall” Brand Attitude. It was demonstrated that in the structure of the variability of “overall” attitudes there are both affective and cognitive components. It was found that consumer preferences based on Implicit “Overall” Brand Attitude are much stronger towards domestic food brands than towards foreign ones. This has some analogies with implicit consumer ethnocentrism.

Consumer preferences based on Explicit “Overall” Brand Attitude are generally uncertain, that is either ambivalent, or unpronounced.

Indirectly, the results of the research are consistent with the concept of multiple systems of separate information storage developed within the framework of Multiple Memory Systems Model of Attitudes [26].