Introduction

Private higher education has grown in Brazil and accounts for 88% of Higher Education Institutions (HEI), according to the last 2022 Educational Census of Higher Education. Brazil has 2.595 HEI, of which 2.283 are private. As for the enrollments made in 2022, private HEI received 7.367.080 of the 9.443.597 total enrollments made in Brazil (Inep, 2023).

The significant number of private HEI in Brazil corroborates the studies of Paiva (2011), which addresses the issue of the competitiveness of private HEI and their strategies through the process of commodification of higher education. Among the factors addressed by Paiva (2011), the condition of offering HEI causes the generation of value of institutions, which contributes to their competitive positioning.

The Educational Census of Higher Education, published by Inep (2023), shows the relevance of private HEI in higher education in Brazil, since they include a majority in number of institutions, courses offered, enrolled and graduates. This scenario makes the private HEI seek to differentiate themselves in several ways, being the brand an important strategic asset for competitive advantage of these organizations (Tinto, 1993).

The results of studies on branding in the Scopus database show that 57,496 articles were published between 2000 and 2021, which shows the importance of the concept not only competitively, but also scientifically.

Among the studies that deal with the love of the brand, Carrol and Ahuvia (2006) stand out regarding the number of uses of their model in other studies. His approach presents the constructs Hedonic Product and Self-expressive Brand as antecedents of love for the brand, and the constructs Brand Loyalty and Positive Word of Mouth.

As well as brand love, consumers may develop an opposite relationship, namely, hate. To explain the development of brand hate, Lee et al. (2009) address a study on the antecedents to brand hate, in which the consumer's negative lived experience; incompatibility with the image; and the consumer's moral aversion toward the brand are presented.

The brand is not only a strategic asset, but also the main source of competitive advantage for a company (Aaker, 1998).

The competitiveness of the market leads organizations that work in the educational sector to understand the behavior of customers and create models of evaluation of indicators to obtain effectiveness in the process of brand management, thus inferring competitive advantage.

Therefore, HEI should rethink their strategies to propose contemporary educational management models as a tactic to become increasingly competitive.

From 2009 to 2022, according to data from the National Institute of Educational Studies and Research Anísio Teixeira (Inep, 2023), the number of Higher Education Institutions (HEI), considering universities, university centers and colleges, increased by 25.42%, from 2069 to 2.595, with emphasis on university centers, which, in 2009, were 120, and in 2019, went to 381 HEI in this category, representing an increase of 217.5%.

This growth in market demand reveals an advance in the private higher education sector, which, according to studies by Brown and Carasso (2013), results in a need for brand management in private HEIs, through the commodification of education.

In this sense, by addressing the growth of private higher education and brand love and hate studies, this research is based on the following thesis: The antecedents of brand love and brand hate influence private HEI students in their intention to drop out or stay.

Faced with this context, the research question emerges: How do the antecedents of love and brand hate influence private HEI students in evasion (intention to leave) and permanence (intention to stay)? To answer the question, the following general objective was established with this research: To analyze the effects of the antecedents of love and hate of the brand in the evasion (decision to leave) and in the permanence (decision to stay) of private HEI students.

Brand Hate and the Intetion to Abandon/Stay

According to Gelbrich (2009), when buying a brand, consumers choose for representativeness. To this end, there is an emotional connection in the purchase, which may provoke a positive or negative feeling towards the brand, as pointed out by (Sinha, et al. 2011).

The negative emotions that consumers feel for a brand can vary in intensity (Laros & Steenkamp, 2005). For Bryson et al. (2021), brand hate can be a negative feeling of great intensity, which can go beyond the simple feeling of disliking a brand. In this context, for Bryson et al. (2013), brand hate is an extreme negative emotional outcome towards the brand.

There are four possibilities for the development of brand hate:

1) the level of negative feeling at the apex of the emotion.

2) the negative feeling that remains after a short period.

3) the duration of the negative feeling; and

4) the intensity after the event that generated the hate (Sonnemans & Frijda, 1994).

Lee et al.'s (2009) study demonstrates three antecedents that generate consumer brand hate, namely:

a) Previous experiences

Also called experimental aversion, this antecedent refers to the relationship with the brand that, at a given moment, generates a marked negative experience enough to generate an aversion to the brand (Salvatori, 2007). Examples of negative experiences occur when there is failure in the product/service, dissatisfaction with the service, waiting, among others (Funches et al., 2009).

This situation constitutes an extremely negative experience of the consumer with the brand, causing a negative feeling of high intensity due to the lack of fulfillment of their expectations for the brand (Lee et al., 2009).

b) Symbolic incongruity

It consists in the fact that the brand has an image that conveys something adverse to the consumer's identity, causing in him a barrier to consume the brand, because it comprises values that are against those shared by him (Sinha et al., 2011). With this, the consumer does not want his image to be related to the brand and, in addition to not consuming, also creates a sense of protest against it

Regarding the taxonomy of the various antecedents and associated results, the antecedents that have a stronger influence on Brand Hate, in order of significance, are Ideological Incompatibility, Symbolic Incongruity and Past Negative Experience (Hegner et al., 2017).

c) Ideological incompatibility

Much like incompatibility with image, this principle of it differs in that the hate is generated by issues secondary to the brand, such as the fact that it shares principles that are immoral to the consumer, but does not necessarily represent those principles (Portwood-Stacer, 2012).

According to Lee et al. (2009), ideological incompatibility occurs when there is a lack of moral identity between the consumer and the brand, due to the fact that consumers' ideological beliefs do not match the brand's moral values. Also called moral aversion, it occurs when consumers associate the brand with moral codes incompatible with their own, causing a moral hate of the brand.

Sternberg (2003) ensures that brand hate manifests itself in different ways. In this perspective, Lee et al. (2009) cite consumer behaviors when feeling brand hate: decreased consumption, evasion, boycott and retaliation.

Gregoire et al. (2009) point out that brand hate can provoke the avoidance and revenge behaviors in consumers. However, they found that over time, revenge behavior decreases because it costs the consumer more energy than simply the behavior of avoiding a product or service.

For Lee et al. (2009), the main behaviors of a consumer who develops a feeling of hate towards a brand are changing or avoiding the brand, rejection of the brand, negative word of mouth, online complaints, revenge.

Brand Love and the Intetion to Abandon/Stay

The topic of brand love has been gaining repercussions over the last ten years, according to bibliometrics carried out on the Scopus database, with 2020 and 2021 having the highest number of publications on the subject (101).

Although data from Scopus show growth in the number of publications, Batra, Ahuvia and Bagozzi (2012, p. 2) state that "(...) the progress in the research of brand love has been hampered by the lack of exploratory studies that guide the subsequent measurement and the development of theories". That is, despite the number of publications, they are impaired in relation to the theoretical construction.

The main questioning of Batra, Ahuvia and Bagozzi (2012) refers to the difference between loving emotion and loving relationship. The authors state that love emotion is instantaneous, and love relationship is something more enduring, citing Richins (1997) and Fournier (1998).

In another perspective of brand love, Christino et al. (2019) address, in their research on brand love of smartphone consumers, brand love as the emotional attachment that satisfied consumers acquire for the product brand name.

The concept addressed by Christino et al. (2019) comprises the model idealized by Carroll and Ahuvia (2006), which presents the constructs Hedonic Product and Self-expressive Brand as antecedents of the love of the brand, and the constructs Brand Loyalty and Positive Word of Mouth as consequent.

For Carroll and Ahuvia (2006, p. 81), brand love is defined as the degree of passionate emotional attachment that a satisfied consumer has for a particular trade name. The authors state that brand love includes passion for the brand, brand attachment, positive brand evaluation, positive emotions in response to the brand and declarations of love for the brand.

The studies by Carroll and Ahuvia (2006) suggested that, in addition to direct effects, brand love mediates the relationships between exogenous variables (hedonic product and self-expressive brand) and the main endogenous behavioral variables (brand loyalty and positive word of mouth).

a) Hedonic characteristics

The hedonic characteristics contemplate the antecedents of love and relate to the fantasy and emotional aspects as a fundamental part of the consumer's decision-making process. According to Holbrook and Hirschman (1982), the symbolic meanings of a product, the hedonic responses and the observed aesthetic criteria summarize the experiential view of consumption. As a result of the hedonic offering, hedonic consumption came to be studied and is defined as "a phenomenon that designates the facets of consumer behavior related to the multisensory and emotional aspects of an individual's experience with products" (Holbook & Hirschman, 1982, p. 92).

For Carroll and Ahuvia (2006), love is a powerful emotional experience. The authors characterize hedonic products as products for which pleasure or fun are major benefits. For the authors, hedonic characteristics tend to generate stronger emotional responses.

Chandon, Wansink and Laurent (2000) point out that the benefits provided by the hedonic characteristics of the offers are of non-instrumental nature, provide experience and arouse affectivity. The same authors counter the utilitarian benefits of offers by stating that they are primarily instrumental, functional, and cognitive. Babin, Darden and Griffin (1994) state that hedonic benefits increase self-esteem and provide entertainment, while utilitarian benefits only help consumers find and buy the best offers. Carroll and Ahuvia (2006) conceptualize the hedonic product in a way that exposes its contraposition to those products of a utilitarian nature, defining it as the consumer's perception of the relative role of hedonic benefits (when compared to utilitarian ones) offered by the product category.

b) Structural characteristics

Carrol and Ahuvia's (2006) model consists in the validation applied to products. In this sense, since it is a brand of educational services, the construct "Structural Characteristics" was inserted, based on the studies of Ambiel et al. (2016), Lobo (2012) and Tontini and Walter (2011), in order to identify the students' perceptions regarding the brand and its relations with the image, reputation, services provided, among other factors observed in the provision of services of an HEI.

c) Self-expressive brand

The last antecedent construct to love is the self-expressive brand. Carroll and Ahuvia (2006) define this as the consumer's perception of the degree to which a particular brand reinforces their social side, and may also reflect their inner self.

For Carrol and Ahuvia (2006), consumers' love should be greater for brands that play a significant role in shaping their identity.

Aaker (1996) states that the emotional side of a brand, its self-expression and the respective benefits connect with the brand personality. For Clemente (2013), the set of all these factors is responsible for creating a relationship with the consumer, from the perception that he has regarding the brand personality, its emotional side and its self-expression. A product, or a brand, acts as a medium that enables the expression of a consumer's individuality.

Santana (2009) state that if a brand helps a person position themselves better, both internally and socially, their feelings will tend to be positive towards it.

As for studies on brand love in higher education institutions, Montoya-Restrepo et al. (2020), conducted in Colombia, address the measurement of a lovemark and unify a set of elements that form a lovemark in university institutions

The research results of Montoya-Restrepo et al. (2020) validated the positive effects of brand love, brand experience, and brand engagement on brand loyalty as a determinant of college brand love. In this sense, Montoya-Restrepo et al.'s (2020) research suggests that brand love is brand loyalty together.

Bizarrias e Lopes (2017) brings in an analysis of the constructs about the love of the brand and the use of each model.

As exposed the scales of Batra, Ahuvia and Bagozzi (2012) and Carrol and Ahuvia (2006) have the highest number of citations, which demonstrates the importance of these for the study on brand love.

Methodological Procedures

The methodological procedure of this research contemplates the positivist epistemological approach, due to the fact that the proposed objective seeks to elucidate the cause and effect relationship of the antecedents of love and hate in the intention of the students to abandon/stay in private Higher Education Institutions.

The research is characterized as descriptive, because this type of research, as explained by Hair et al. (2014), aims to measure the characteristics represented in a research question. Collins and Hussey (2005) report that descriptive research aims to understand the behavior of phenomena, and is widespread to detect and acquire data about the characteristics of a specific research problem.

In light of the above, the research sought to verify the antecedents to hate and love, and the effects on the intention to stay or drop out; and to identify the perceptions of the students regarding the image of the brand, the brand reputation, and the services provided by HEIs.

The study proposal aimed at a quantitative approach, seeking to identify problems, interpret and understand the opinions, appreciations and perceptions of the researched in relation to the object handled (Creswell, 2010). In the methodological application, the quantitative approach investigated data collected through the application of a survey questionnaire, which sought to clarify the perception of the student public in relation to the brand, identifying, for this purpose, from the perspective of the HEI students surveyed, the antecedents of hate and love in the intention to stay or evade.

The data collection technique comprised a bibliographic research, a documentary research and the application of a survey questionnaire.

The bibliographic research consisted of a secondary data collection, through the collection of published material that approached the research theme, such as loose publications, bulletins, newspapers, magazines, books, articles, dissertations, and theses.

In addition to the search for studies on the aforementioned themes, a bibliometric research was carried out in the Scopus database. In this context, it was identified 2000 publications related to the brand, to the antecedents of brand love and to the antecedents of brand hate, considering, for this purpose, the universe of higher education institutions.

The data collection was structured from a documental research carried out, a priori, in official statistical reports, with the objective of identifying the private HEIs in Brazil.

From the initial survey of private HEIs in Minas Gerais, the sample was defined intentionally, characterizing a non-probabilistic sample.

The criteria of choice for the participation of HEIs were: to be registered with the MEC and to accept participation in the research. The criteria for student participation were: being enrolled in the participating HEIs, and being available to participate in the research.

After choosing the participating HEIs, data were collected by means of semi-structured questionnaires with a 7-point Likert-type interval scale.

The pre-test was conducted contemplating the application of the research instruments in Brazilian private HEI's. Thus, two private HEIs were chosen, whose students were willing to participate in the test phase of the research instruments. The pre-test was applied to 16 (sixteen) students from private HEI's intentionally.

Considering a confidence level of 95% and a margin of error of 5%, the minimum sample size was estimated at 385 respondents. In this sense, foreseeing the possibility of incomplete and/or invalid answers, the sample size was increased by 10%, reaching the number of 424 respondents. At the end of the questionnaire application, 428 answers were obtained, of which 22 were discarded, reaching the final number of 406 respondents, higher than necessary.

Regarding data analysis, the proposed model includes the three main theoretical models focused on brand studies and Educational Institutions. The first consists of studies by Lee et al. (2009), in which the authors explain the antecedents to brand hate: negative experience, incompatibility with the image (symbolic incongruity) and moral aversion (ideological incompatibility). In this sense, the proposed model assumes that there are antecedents to brand hate, considering the parameters suggested by Lee et al. (2009).

As a second model, it contemplates that of Carrol and Ahuvia (2006), which consists of the antecedents to brand love, which are the hedonic characteristics and the Self-expressive brand.

The third model used is the one proposed by Tinto (1993), which considers the academics' experience with HEI, their pre-entry, experiences during the course and the outcome, the last item being observed from the possibilities of intention to stay (permanence) or intention to leave (evasion).

The third construct also considers the antecedents to love, structural characteristics, which was incorporated based on the studies of Ambiel et al. (2016), Lobo (2012), and Tontini and Walter (2011), aiming to identify students' perceptions regarding the brand (image, reputation, services provided) of HEI.

In this investigative effort, Structural Equations Modeling, also known by its acronym in English SEM (Structural Equations Model), was used. The Lavaan (Yves Rosseel, 2012), semTools (Jorgensen, et al., 2021) and semPlot (Sacha Epskamp, 2019) packages of the R software (R Core Team, 2021) were used to build the structural equation model.

Thus, the observed variables were grouped into their respective constructs, being 4 (four) items related to Negative Experience (EN), 5 (five) items related to Symbolic Incongruence (IS), 4 (four) items related to Ideological Incompatibility (II), 6 (six) items related to Brand Hate (OM), 5 (five) items related to Hedonic Characteristics (CH), 4 (four) items related to Structural Characteristics (CE), 8 (eight) items related to Self-Expressive Brand (ME), 5 (five) items related to Brand Love (AM), and finally, 4 (four) items related to Intention to Stay (IP).

Next, regressions were run, indicating: a) the relationship between Brand Hate and its predecessors Negative Experience, Symbolic Incongruence, and Ideological Incompatibility; b) the relationship between Brand Love and its predecessors Hedonic Characteristics, Structural Characteristics, and Self-Expressive Brand; and c) the relationship between Intention to Stay and Brand Hate and Self-Expressive Brand.

Data Presentation and Analysis

The research results begin by presenting the univariate analysis of the questionnaire items, followed by the validity and reliability analysis.

a) Unvariate Analysis of the Questionnaire Items

A univariate analysis of the questionnaire items was performed, where the following constructs are highlighted here:

onsidering the averages, Table 1 shows that OM1 (I am disgusted with my HEI) presented the highest value (2.63), which represents a higher agreement of respondents regarding being disgusted with their HEI. On the other hand, item OM6 (The world would be a better place without HEI.) showed the lowest value (1.29),

Considering the averages, Table 1 shows that OM1 (I am disgusted with my HEI) presented the highest value (2.63), which represents a higher agreement of respondents regarding being disgusted with their HEI. On the other hand, item OM6 (The world would be a better place without HEI.) showed the lowest value (1.29), which represents a lower agreement of respondents regarding their belief that the world would be a better place without their HEI. However, the averages presented in this construct are low, which indicates that the respondents showed low rates in relation to brand hate with their HEI.

Regarding standard deviation, item OM6 (The world would be a better place without HEI.) showed the lowest value, indicating low variability of respondents regarding this item. On the other hand, item OM1 (I am disgusted with my HEI.) presented the highest value (2.63), which indicates a greater dispersion of the responses in relation to it.

This result follows Lee et al.'s (2009) finding that prior experience (Salvatori, 2007), symbolic incongruity (Sinha et al., 2011) and ideological incompatibility (Portwood-Stacer, 2012) are determining factors for brand hatred. The fact that the students have established a long-term relationship with HEI means that these factors are better identified.

onsidering the averages, table 2 shows AM2 (My HEI's brand makes me feel good.) presented the highest value (5.11), which represents a higher agreement of respondents regarding how they feel about their HEI's

Considering the averages, table 2 shows AM2 (My HEI's brand makes me feel good.) presented the highest value (5.11), which represents a higher agreement of respondents regarding how they feel about their HEI's brand. On the other hand, item AM5 (I love my HEI's brand.) showed the lowest value (4.52), which represents a lower agreement of the respondents regarding the love they feel for their HEI. However, the averages presented in this construct are high, indicating that the respondents showed high indexes in relation to the love for their HEI's brand.

As for the standard deviation, item AM1 (The brand of my HEI is wonderful.) showed the lowest value, indicating low variability of respondents in relation to it. On the other hand, item AM5 (I love my HEI's brand.) showed the highest value, indicating a greater dispersion of responses in relation to it.

As discussed in brand hate, better indicators can be observed in brand love, which is motivated by a long relationship with the brand, something typical for a higher education service. This type of love is linked to relationship love, presented by Batra, Ahuvia and Bagozzi (2012). Relationship love then helps to reinforce hedonic characteristics (Carroll and Ahuvia, 2006), structural characteristics (Ambiel et al., 2016 & Lobo, 2012) and the self-expressive brand (Montoya-Restrepo et al., 2020).

Table 3 explains that, considering the averages, IP1 (I intend to graduate in the HEI I currently study.) presented the highest value (6.62), which represents a higher agreement of respondents regarding the intention to graduate in their HEI. On the other hand, item IP2 (I intend to stay in this course in this HEI.) presented the lowest value (6.53), which represents a lower agreement of the respondents regarding the intention to stay in the course in their HEI. However, the averages presented in this construct are high, indicating that the respondents showed high indexes in relation to the intention to leave/stay.

Regarding standard deviation, item IP1 (I intend to graduate from the HEI where I currently study.) showed the lowest value, indicating low variability of respondents in relation to it. On the other hand, IP2 (I intend to stay in this course in this HEI.) presented the highest value, which indicates a greater dispersion of responses in relation to it.

Relationship love, as presented by Batra, Ahuvia and Bagozzi (2012), shows a greater intention to stay than to leave the brand. Higher education courses generate a sense of belonging in their students, which reinforces this intention. The brand then comes to be seen in an emotional way, according to other research presented by Gelbrich (2009) and Sinha, et al. (2011).

b) Validity and Reliability Analysis

Convergent validity and reliability analysis was used. For convergent validity, first, they analyzed standardized factor loads. Items that are indicators of a specific construct must converge or share a high proportion of common variance. In this sense, Hair et al. (2009) propose that standardized factor loads should be greater than 0.5, ideally greater than 0.7. All standardized factor loads are greater than 0.5, and only four are less than 0.7, considering what was proposed.

Remaining in the convergent validity analysis, the authors suggest that the Extracted Mean Variance (AVE) greater than 0.5 is a good rule to determine convergence. For ordinal data, polychoric correlations are recommended to the detriment of Pearson correlations (Zumbo, Gadermann & Zeisser, 2007; Zumbo & Kroc, 2019). For the reliability analysis, Hair et al. (2009) consider that there is an extensive debate about the best option to estimate reliability. The authors consider that the Alpha coefficient is often applied, despite underestimating reliability.

For the use of categorical variables, Zumbo, Gadermann and Zeisser (2007) recommend the use of the Ordinal Alpha coefficient, which produces more robust results. In this sense, we used the Ordinal Alpha coefficient combined with the Construct Reliability (CR) so that the results are able to produce good indicators of reliability, as long as they exceed the value of 0.7 (Hair et al., 2009).

The results presented show good validity and reliability indicators, since for all constructs the estimated values are above the recommended ones.

It is worth noting that the construct "to brand" has the highest AVE and highest Alpha value, while the construct with the highest CR value is "self-expressive brand".

Once the model was validated, we set out to analyze the correlation between the constructs. Since these are categorical variables, the estimates of the polychoric correlations were used, as recommended by Zumbo, Gadermann, and Zeisser (2007) and Zumbo and Kroc (2019).

Table 5 shows that in the first group, highlighted in blue, there are the correlations of the construct hate to the brand and its antecedents. It is verified that all correlations are positive, which indicates a behavior of the set of these constructs. The highest correlation coefficient (0.877) is between Negative Experience and Brand Hate. On the other hand, the lowest correlation (0.795) is between Symbolic Incongruity and Ideological Incompatibility.

The relationship between Negative Experience and Brand Hate is in line with Salvatori (2007) when he states that this antecedent refers to the relationship with the brand that, at a given moment, generates a negative experience marked enough to generate brand aversion. On the other hand, the lowest correlation (0.795) is between Symbolic Incongruence and Ideological Incompatibility.

Although they have similarities, Symbolic Incongruence and Ideological Incompatibility are different.

The former refers to a brand that has an image that conveys something adverse to the consumer's identity (Sinhq et al., 2011); and the latter, occurs when there is a lack of moral identity between the consumer and the brand, due to the fact that consumers' ideological beliefs do not match the brand's moral values (Lee et al., 2009).

The second group, highlighted in green, shows the correlations of the brand love construct and its antecedents. It can be seen that all correlations are positive, which indicates a behavior of these constructs as a whole. The highest correlation coefficient (0.836) is between Hedonic Characteristics and Brand Love. On the other hand, the lowest correlation (0.523) is between Self-Expressive Brand and Structural Characteristics.

When analyzing the inter-group correlations, it is clear that all of them are negative. This shows that love and hate and their antecedents have antagonistic behaviors. In modulus, the highest correlation coefficient (-0.853) was between Hedonic Characteristic and Negative Experiences. The correlation coefficient between Structural Characteristic and Symbolic Incongruence (-0.452) presented the lowest absolute value. Finally, when analyzing the correlation coefficients between the constructs of Love and Hate and their antecedents with the intention of remaining, it is verified that in the group Love to the brand, the highest correlation coefficient is386, between Hedonic Characteristics and Intention to Remain; and the lowest coefficient is 0.256, between Structural Characteristics and Intention to Remain. In the Brand Hate group, the highest value is -0.352, between Symbolic Incongruity and Intention to Remain; and the lowest value is -0.413 between Brand Hate and Intention to Remain.

These results are directly related to the statements of Tinto (1993) when it comes to the influence between the positive experiences and the intention of the students to remain in the HEI.

Table 6 shows the regression coefficient and p values of each construct (hypothesis). From the analysis of these values, the significance of the hypotheses is considered.

The Hypothesis H1 - The antecedent "Previous experiences" positively influences brand hate was significant as it showed p-value less than 0.05 (0.00). This hypothesis states the fact that the negative experiences experienced by students in HEIs influence their HEI's brand hate feeling.

The hypothesis H2 - The antecedent "Symbolic incongruence" positively influences brand hate was significant, since it presented a p-value less than 0.05 (0.019). This hypothesis consists of the fact that when consumers do not want their image to be related to the brand, they create a feeling of protest against it.

From the analysis of these values, it is found that the hypothesis H3 - The antecedent "Ideological incompatibility" positively influence the hate to the brand was significant, since it presented p-value below 0.05 (0.005). This result shows that when consumers associate the brand with moral codes incompatible with their own, a moral hate of the brand develops.

The analysis of these values shows that the hypothesis H4 - The antecedent "Hedonic Characteristics" positively influences the love of the brand was significant, since it presented a p-value lower than 0.05 (0.000), in line with Chandon, Wansink and Laurent (2000) which affirm that hedonic characteristics provide experience and arouse affection, attachment, loyalty, leading to brand love.

Only the hypothesis H5 - The antecedent "Structural characteristics" positively influences the love of the brand was not significant for the construct love of the brand, since it presented a p value greater than 0.05. However, in the period in which the survey was conducted, HEI were conducting classes remotely (online) due to the distancing restrictions imposed by health agencies for the prevention of COVID-19.

Analyzing these values, we have that the hypothesis H6 - The antecedent "Self-expressive Brand" positively influences the love of the brand was significant, since it presented p-value below 0.05 (0.000). Carrol and Ahuvia (2006) define self-expressive brand as the consumer’s perception of the degree to which the specific brand enhances the social self and/or reflects the inner self.

The hypothesis H7 - Brand hate positively influences the intention to leave and negatively influences the intention to stay was significant, since it presented a p value of less than 0.05 (0.003), in line with Lee et al. (2009), when they state that the main behaviors of a consumer who develops brand hate are brand switching or avoidance, brand rejection, negative "word of mouth", online complaint, revenge and market aggression.

Finally, the hypothesis H8 - Brand love negatively influences the intention to leave and positively influences the intention to stay was significant, since it presented a p-value less than 0.05 (0.022). The positive feeling generated by brand love generates students' intention to stay in their HEI. As per research conducted by Montoya-Restrepo et al. (2020), the positive effects of brand love, brand experience and brand engagement on brand loyalty as a determinant of educational institution brand love were validated.

Conclusions and Recommendations

The findings of this research underscore a critical relationship between students' emotional responses toward a brand—specifically love and hate—and their decisions to remain enrolled or leave a private Higher Education Institution (HEI). This study highlights that these emotional responses are significant predictors of student retention or attrition. The analysis of the model's constructs revealed that all but "structural characteristics" were statistically significant. This suggests that while the physical infrastructure of an institution may not heavily influence student emotions during the pandemic-driven shift to online learning, other factors play a crucial role.

Specifically, the antecedents to brand hate—such as negative experiences, symbolic incongruence, and ideological incompatibility—were found to negatively impact students' intentions to remain at or leave an HEI. The data showed that the respondents generally did not report these negative experiences, indicating a lack of substantial discontentment with their institutions. However, the existence of brand hate, even in minimal forms, was significantly correlated with a higher likelihood of students considering leaving their HEIs.

Conversely, antecedents to brand love—encompassing positive emotional connections and perceptions—were associated with favorable student attitudes towards their HEI. Despite the positive overall sentiment, "structural characteristics" did not show a significant impact on brand love. This anomaly could be attributed to the remote learning circumstances during the COVID-19 pandemic, which limited students' interactions with the physical aspects of the campuses.

The hypotheses testing provided deeper insights:

H1 posited that "Previous Experiences" positively influence brand hate, and this was confirmed. Negative past interactions with the institution can linger in students' perceptions, fueling a dislike for the brand.

H2 suggested that "Symbolic Incongruity" leads to brand hate, which was also supported. When students perceive a mismatch between the institution's image and their own values, it can result in negative emotional responses.

H3 confirmed that "Ideological Incompatibility" contributes to brand hate, illustrating that when an institution's values or practices conflict with students' beliefs, it fosters dislike.

H4 established that "Hedonic Characteristics," such as enjoyable experiences and emotional gratification, positively influence brand love.

H6 demonstrated that a "Self-expressive Brand," which allows students to express their identity through association, also fosters brand love.

H7 and H8 revealed that brand hate increases the likelihood of leaving and decreases the intention to stay, whereas brand love decreases the likelihood of leaving and increases the intention to stay.

However, H5, which suggested that "Structural Features" influence brand love, was not supported by the significance analysis. This could be due to the lack of direct interaction with campus facilities during the pandemic, leading to a neutral impact on students' emotional connections with the institution's physical attributes.

These findings contribute significantly to the marketing literature, especially in the educational services sector. They demonstrate the importance of understanding the emotional drivers behind brand perceptions and their subsequent impact on student retention. This study emphasizes that emotional connections—both positive and negative—are not merely superficial but are deeply ingrained in the students' decision-making processes regarding their educational journey.

For practical implications, this research suggests that HEI managers and marketers should not only focus on enhancing the tangible aspects of the institution but also work on building and maintaining a positive emotional connection with students. This can be achieved by ensuring high-quality academic experiences, fostering a supportive and inclusive community, and aligning the institution's values with those of the students. Furthermore, regular assessments and adaptations of marketing strategies to reflect the evolving emotional needs and preferences of students are crucial. By doing so, institutions can improve brand love and mitigate brand hate, thus positively influencing student retention and satisfaction.

In summary, this study provides a comprehensive understanding of the factors contributing to brand love and brand hate in the context of private HEIs. It offers valuable insights for academic administrators and marketers seeking to enhance student retention by addressing both the affective and cognitive dimensions of student experiences.

References

Aaker, D. (1998). Marcas: brand equity: gerenciando o valor da marca (2a ed.). Editora Negócio.

Aaker, D. (1996) Building strong brands New York: Free Press.

Ambiel, R. A. M., Lamas, K. C. A., & Melo-Silva, L. L. (2016). Avaliação dos interesses profissionais no Brasil: revisão da produção científica. Avaliação Psicológica, 15(esp.), 1-9. https://doi.org/10.18222/eae.v33.9022

Babin, B. J., Darden, W. R., & Griffin, M. (1994). Work and/or Fun: Measuring Hedonic and Utilitarian Shopping Value. Journal of Consumer Research, 20, 644-656. https://doi.org/10.1086/209376

Batra, R., Ahuvia, A. C., & Bagozzi, R. P. (2012). Brand love. Journal of Marketing, 76, 1–16. https://doi.org/10.1509/jm.09.0339

Bizzarias, F. S. & Lopes, E. L. (2017). Escalas Concorrentes para Mensuração do Amor à Marca: Um Estudo no Contexto Nacional. Revista de Administração IMED, 7(1), 180-207. http://dx.doi.org/10.18256/2237-7956/raimed.v7n1p180-207

Brown, R., & Carasso, H. (2013). Everything for sale? the marketisation of UK higher education. Liverpool Hope University, United Kingdom.

Bryson, D., Atwal, G., & Hultén, P. (2013). Rumo à conceituação dos antecedentes do afeto negativo extremo em relação às marcas de luxo. Pesquisa de Mercado Qualitativa, 16(4), 393-405. https://doi.org/10.1108/QMR-06-2013-0043

Bryson, D., Atwal, G., Hultén, P., & Heine, K. (2021). Antecedents of luxury brand hate: A quantitative study. Strategic Change, 16(3), 326-280. https://doi.org/10.1002/jsc.2387

Carrol, B.A., & Ahuvia, A.C. (2006). Some antecedents and outcomes of brand love. Marketing Letter, 17 (1), 79–89. https://doi.org/10.1007/s11002-006-4219-2

Chandon, P., Wansink, B., & Laurent, G. (2000). A Benefit Congruency Framework of Sales Promotion Effectiveness.. Journal of Marketing, 64(4), 65–81. https://doi.org/10.1509/jmkg.64.4.65.18071

Christino, J., Silva, T., Moura, L. R., & Fonseca. L. H. (2019). Antecedentes and Consequents of Brand love in the Smartphone Market: and extendend Study of the impact of witching cost.Journal of Promotion Management, 26:3, 301-321, 10.1080/10496491.2019.1699630.

Clemente, C. (2013) O Amor à Marca e seus determinantes: um estudo comparativo entre as marcas Apple e Samsung. [Dissertação de Mestrado]. Faculdade de Economia, Universidade de Coimbra.

Collis, J., & Hussey, R. (2005). Pesquisa em administração: um guia prático para alunos de graduação e pós-graduação (2ª ed.). Bookman.

Creswell, J. W. (2010). Projeto de Pesquisa: método qualitativo, quantitativo e misto (3ª ed.). Artmed.

Fournier, S. (1998). Consumers and their brands: developing relationship theory in consumer research. Journal of Consumer Research, 24(4), 343-373, https://doi.org/10.1086/209515

Funches, V., Markley, M., & Davis, L. (2009). Reprisal, retribution and requital: Investigating customer retaliation. Journal of Business Research, 62(2), 231-238, https://doi.org/10.1016/j.jbusres.2008.01.030

Gelbrich, K. (2009). Beyond just being dissatisfied: how angry and helpless customers react to failures when using self-service technologies. Schmalenbach Business Review, 61(2), 40–59. https://doi.org/10.1007/BF03396779

Grégoire, Y., Tripp, T. M., & Legoux, R. (2009). When customer love turns into lasting hate: the effects of relationship strength and time on customer revenge and avoidance. Journal of Marketing, 73(6), 18-32. https://doi.org/10.1509/jmkg.73.6.1

Hair, J. F., Black, B., Babin, B. J. & Anderson, R. E. (2014). Multivariate data analysis (7ª ed.). Pearson Education Limited.

Hegner, S. M., Fetscherin, M., & Delzen, M. van. (2017). Determinants and outcomes of brand hate. Journal of Product and Brand Management, 26(1), 1–29. https://doi.org/10.1108/JPBM-01-2016-1070

Holbrook, M. B., & Hirschman, E. C. (1982). The Experiential Aspects of Consumption: Consumer Fantasies, Feelings, and Fun, Journal of Consumer Research, 9 (2), 132–140, https://doi.org/10.1086/208906

Instituto Nacional de Estudos e Pesquisas Educacionais Anísio Teixeira [INEP]. (2023). Censo da Educação Superior 2022 - Notas Estatísticas INEP. https://download.inep.gov.br/publicacoes/institucionais/estatisticas_e_indicadores/notas_estatisticas_censo_escolar_2022.pdf

Jorgensen, T. D., Pornprasertmanit, S., Schoemann, A. M., & Rosseel, Y. (2021). SemTools: Useful tools for structural equation modeling. R package version 0.5-5. https://CRAN.R-project.org/package=semTools

Laros, F. J. M., & Steenkamp, J. B. E. M. (2005). Emotions in consumer behavior: a hierarchical approach. Journal of Business Research, 58(10), 1437–1445. https://doi.org/10.1016/j.jbusres.2003.09.013

Lee, M. S., Conroy, D., & Motion, J. (2009). Brand avoidance: a negative promises perspective. Advances in Consumer Research, 36 (1), 421-429. https://www.acrwebsite.org/volumes/14318/volumes/v36/NA-36

Lobo, M. B. C. M. (2012). Panorama da evasão no ensino superior brasileiro: aspectos gerais das causas e soluções. In: Instituto Lobo para desenvolvimento da educação, da ciência e da tecnologia (Eds.). Instituto Lobo. São Paulo. http://www.institutolobo.org.br/imagens/pdf/artigos/art_087.pdf.

Montoya-Restrepo, I. A., Sánchez-Torres, J. A., Rojas- Berrio, S. P., & Montoya-Restrepo, A. (2020). Lovemark Effect: Analysis of the Differences Between Students and Graduates in a Love Brand Study at a Public University. Innovar, 30(75), 43-56. https://doi.org/10.15446/innovar.v30n75.83256

Paiva, R. C. V. (2011). Competitividade, estratégia e desempenho financeiro: um estudo das instituições privadas de educação superior brasileiras [Tese de Doutorado]. Faculdade de Ciências Sociais Aplicadas, Universidade de Minas Gerais.

Portwood-Stacer, L. (2012). Anti-consumption as tactical resistance: anarchists, subculture, and activist strategy. Journal of Consumer Culture, 12(1), 87-105. https://doi.org/10.1177/146954051244202

R Core Team (2021). R: A language and environment for statistical computing. R Foundation for Statistical Computing. https://www.r-project.org/

Richins, M. L. (1997). Measuring Emotions in the Consumption Experience. Journal of Consumer Research, 24(2), 127–146. https://doi.org/10.1086/209499

Sacha Epskamp (2019). SemPlot: path diagrams and visual analysis of various SEM packages' output. R package version 1.1.2. https://CRAN.R-project.org/package=semPlot

Salvatori, E. (2007). Brand hate: the dark side of consumer attitudes towards a brand. [Dissertação], Università Commerciale Luigi Bocconi di Milano.

Santana, E. E. P. (2009). O amor à marca e seu relacionamento com algumas variáveis que o antecedem e o sucedem: um estudo sob a ótica de torcedores-consumidores de times de futebol brasileiros [Tese de Doutorado]. Faculdade de Administração, Universidade Federal do Paraná.

Sinha, N., Ahuja, V., & Medury, Y. (2011). Corporate blogs and internet marketing – using consumer knowledge and emotion as strategic variables to develop consumer engagement. Journal of Database Marketing and Customer Strategy Management, 18(3), 185-199. https://doi.org/10.1057/dbm.2011.24

Sonnemans, J, & Frijda, N. H. (1994) The structure of subjective emotional intensity, Cognition and Emotion, 8 (4), 329-350. https://doi.org/10.1080/02699939408408945

Sternberg, R. J. (2003). A duplex theory of hate: development and application to terror-ism, massacres, and genocide. Review of General Psychology, 7(3), 299-328. https://doi.org/10.1037/1089-2680.7.3.29

Tontini, G. & Walter, S. A. (2011). Análise da satisfação do aluno para melhoria de um curso de administração. FACES Adm, 5 (2), 57-70. https://doi.org/10.21714/1984-6975FACES2006V5N2ART73

Tinto, V. (1993). Leaving college: rethinking the causes and cures of student attrition. University of Chicago Press.

Yves Rosseel (2012). lavaan: An R Package for structural equation modeling. Journal of Statistical Software, 48(2), 1-36. https://doi.org/10.18637/jss.v048.i02

Zumbo, B. D., Gadermann, A. M., & Zeisser, C. (2007). Ordinal versions of coefficients alpha and theta for Likert rating scales. Journal of Modern Applied Statistical Methods, 6(1), 21–29. https://doi.org/10.22237/jmasm/1177992180

Zumbo, B. D., & Kroc, E. (2019). A measurement is a choice and Stevens’ scales of measurement do not help make it: A response to Chalmers. Educational and Psychological Measurement, 79(6), 1184–1197. https://doi.org/10.1177/0013164419844305

IR