Round 1

Reviewer 1 Report

The manuscript is interesting, however following points should be addressed carefully. 

The questionnaire was pretested? If yes provide the description and results.

How the selected sample size can reflect the overall trend of brand loyalty? 

What was the criteria tp calculate the acceptable score of each individual or a particular question from all participants??

Was there any correlation between demographic characteristics with other variables

Some Cronbach's α values were lower than 0.70, what justification was provided in discussion.

Data analysis section should be very explicit to provide information of all the analyses carried out, please revisit the data analysis section and add further information. 

Author Response

The manuscript is interesting, however following points should be addressed carefully.

The questionnaire was pretested? If yes provide the description and results.

In this study, the questionnaire reflects the questions posited by Lassoued and Hobbs (2015) in their works. Although the population is different, we decided not to pretest the questionnaire supposing its validity and reliability.

References

Lassoued, R.; Hobbs, J.E. Consumer Confidence in Credence Attributes: The Role of Brand Trust. Food Policy 2015, 52, 99–107, doi:10.1016/j.foodpol.2014.12.003.

How the selected sample size can reflect the overall trend of brand loyalty?

We are conscious that the selected sample size cannot reflect the overall trend of brand loyalty. In light of this aspect, and conscious of this limit, we have declared on page 11, rows 340-342, the need for a probabilistic sample design that allows for more generalisable results referring to the overall trend of brand loyalty.

What was the criteria tp calculate the acceptable score of each individual or a particular question from all participants??

The analysis has taken into consideration the outliers for each question. Finally, 34 answers were withdrawn from the database.

Was there any correlation between demographic characteristics with other variables.

Thank you for this suggestion, which will be interesting for future research on this topic. This study did not aim to find or test relations between sociodemographic aspects and specific constructs or variables nor to verify differences among specific groups of interviewees (e.g. Gen-Z, Gen-Y, Gen-X). This is the reason for not having done this kind of analysis.

Some Cronbach’s α values were lower than 0.70, what justification was provided in discussion.

As discussed in sections 4.3 and 5.1, after checking the hypothesised 4-factor brand trust drivers by performing a confirmatory factor analysis (CFA), an explanatory factor analysis (EFA) was carried out. The EFA produced a two-factor solution (please see Table 2), in which the second component’s low level of internal consistency (α=0.40) led us to drop it as the antecedent of Btrust. Thus, only the first factor named ‘brand performance’, characterised by a Cronbach’s α value of 0.89, was considered a determinant of Btrust. Only the brand loyalty construct has a Cronbach’s α value of 0.68, but this level has generally been accepted in literature.

Data analysis section should be very explicit to provide information of all the analyses carried out, please revisit the data analysis section and add further information.

In section 4.3 we have explicitly reposted in a cronological way all the steps of analysis that we have done. In order: 

• Confirmatory factor analysis (please see on page 7, rows 220-221);
• Explanatory factor analysis (please see on pages 7-8, rows 222-225);
• Test the internal consistency of each extracted factor using Cronbach’s Alpha (please see on pages 8, rows 227-229);
• Examination of the structural part of the model using a Structural Equation Model (please see on pages 8, rows 229-231)

Reviewer 2 Report

The authors present an interesting article on loyalty towards “authentic brand activism” of a sustainable food brand. However, I consider that the work has two major limitations that prevent its acceptance.

The authors propose as an aim to "analysing the extent to which brand activism contributes to developing consumers' brand loyalty through brand trust", however, the model they present (Figure 1) is limited to a classical model of explanation of loyalty. In this sense, the authors state in lines 45 and 46 that “this study 45 adopts the angle of authentic brand activism”. Actually, what they do is apply it to an “Authentic brand”.

Secondly, one of the big problems comes from the model they propose. In it, he establishes competence, reputation, benevolence and credibility as antecedents of trust. Three such antecedents are sometimes considered as components of trust, treating trust as a second-order construct. In fact, the authors themselves state that “In line with prior research on trust, Btrust is considered to be a multidimensional construct encompassing specific attitudes, such as brand competence and reputation”. In their work, they do not consider it multidimensional or second order, but rather treat antecedents as individual elements. Likewise, the justification of the hypotheses is scarce. For example, hypothesis 1 is justified by only two references, and hypotheses 2 and 3 are justified based on three. It would be important to incorporate references linked to the field to which the work is directed. Hardly any analysis is done regarding loyalty for Authentic brands.

The fact of considering trust a second-order construct, or opting for antecedents, requires more reflection on the part of the authors. See, for example, the following works:

Li, F., Kashyap, R., Zhou, N., & Yang, Z. (2008). Brand trust as a second-order factor: An alternative measurement model. International Journal of Market Research, 50(6), 817-839.

Hegner, S. M., & Jevons, C. (2016). Brand trust: a cross-national validation in Germany, India, and South Africa. Journal of Product & Brand Management.

Regarding the field work, this is based on a survey carried out four years ago, which already supposes a hindrance, likewise data was collected exclusively in one city, has the bias that such an element may have implied been taken into account? However, the most striking element is that, since hypotheses 2, 3, 4 and 5 are rejected, the authors decide to unite the antecedents in a single construct. In this way, they generate a construct that practically measures trust as an antecedent of trust. If a hypothesis is rejected, there is no problem, but the approach of the constructs is not usually modified and a new hypothesis is incorporated.  The results of the discriminant validity deserve to be reviewed and commented on by the authors, some results of said analysis do not seem to adjust to the canons.

Author Response

The authors present an interesting article on loyalty towards “authentic brand activism” of a sustainable food brand. However, I consider that the work has two major limitations that prevent its acceptance.

The authors propose as an aim to “analysing the extent to which brand activism contributes to developing consumers’ brand loyalty through brand trust”, however, the model they present (Figure 1) is limited to a classical model of explanation of loyalty. In this sense, the authors state in lines 45 and 46 that “this study 45 adopts the angle of authentic brand activism”. Actually, what they do is apply it to an “Authentic brand”.

Thank you for this reflection and suggestion. We have better aligned the aim of the study as follows:

Page 2, withdrew the  rows 49-54

Page 2, rows 52-57

“This study aims to understand the roles of brand reputation, benevolence, credibility, and competence in forming brand trust and, indirectly, brand loyalty towards a brand charged with political meaning and devoted to promoting a sustainable social change”.

Secondly, one of the big problems comes from the model they propose. In it, he establishes competence, reputation, benevolence and credibility as antecedents of trust. Three such antecedents are sometimes considered as components of trust, treating trust as a second-order construct. In fact, the authors themselves state that “In line with prior research on trust, Btrust is considered to be a multidimensional construct encompassing specific attitudes, such as brand competence and reputation”. In their work, they do not consider it multidimensional or second order, but rather treat antecedents as individual elements. Likewise, the justification of the hypotheses is scarce. For example, hypothesis 1 is justified by only two references, and hypotheses 2 and 3 are justified based on three. It would be important to incorporate references linked to the field to which the work is directed. Hardly any analysis is done regarding loyalty for Authentic brands.

The fact of considering trust a second-order construct, or opting for antecedents, requires more reflection on the part of the authors. See, for example, the following works:

Li, F., Kashyap, R., Zhou, N., & Yang, Z. (2008). Brand trust as a second-order factor: An alternative measurement model. International Journal of Market Research, 50(6), 817-839.

Hegner, S. M., & Jevons, C. (2016). Brand trust: a cross-national validation in Germany, India, and South Africa. Journal of Product & Brand Management.

Thank you for this reflection. To reinforce the hypotheses justification and better specify that we have initially considered the trust as a second-order construct (as illustrated in figure 1), we have reported the following sentence on page 3, rows 49-54, and inserted the suggested references:

“Building on previous studies that established brand trust as a second-order construct [18,23], the conceptual framework adopted in this study to verify the direct relationship between “Libera Terra” brand trust (hereafter called Btrust) and brand loyalty (hereafter called Bloy) was gauged considering five constructs. Btrust and Bloy represent the model’s latent endogenous variables, whereas brand competence (Bcomp), brand reputation (Brep), brand benevolence (Bben), and brand credibility (Bcred) represent the latent exogenous variables (Figure 1)”

References:

Li, F.; Kashyap, R.; Zhou, N.; Yang, Z. Brand Trust as a Second-Order Factor: An Alternative Measurement Model. Int. J. Mark. Res. 2008, 50, 817–839, doi:10.2501/s1470785308200225.

Hegner, S.M.; Jevons, C. Brand Trust: A Cross-National Validation in Germany, India, and South Africa. J. Prod. Brand Manag. 2016, 25, 58–68, doi:10.1108/JPBM-02-2015-0814

Furthermore, to reinforce H1, we have inserted the following sentence on page 3, rows 130-131

“Furthermore, previous studies have shown that loyalty directly consequences a trust triggered by the brand [15]

References:

Chaudhuri, A.; Holbrook, M.B. The Chain of Effects from Brand Trust and Brand Affect to Brand Performance: The Role of Brand Loyalty. J. Mark. 2001, 65, 81–93, doi:10.1509/jmkg.65.2.81.18255

In their work, they do not consider it multidimensional or second order, but rather treat antecedents as individual elements.

As mentioned before, we have initially considered a second-order construct of brand trust, but unlike what we had expected, the analysis confirmed that the antecedents of brand trust are not distinct constructs. Thus, all the items referred to in each construct were merged and loaded into a single factor named brand performance (see Table 2 and Fig. 2).

Regarding the field work, this is based on a survey carried out four years ago, which already supposes a hindrance, likewise data was collected exclusively in one city, has the bias that such an element may have implied been taken into account?

As mentioned in section 7 (page 11, rows 344-346”), “…this research was carried out based on a non-probabilistic sample; thus, results cannot be generalised. Therefore, future research should be carried out through probabilistic sample designs that would allow for more generalisable results.”

However, the most striking element is that, since hypotheses 2, 3, 4 and 5 are rejected, the authors decide to unite the antecedents in a single construct. In this way, they generate a construct that practically measures trust as an antecedent of trust. If a hypothesis is rejected, there is no problem, but the approach of the constructs is not usually modified and a new hypothesis is incorporated. The results of the discriminant validity deserve to be reviewed and commented on by the authors, some results of said analysis do not seem to adjust to the canons.

As reported on page 8, rows 244-258, the considered antecedents of brand trust are not distinct constructs, and the items measuring competence, credibility and benevolence are part of a unique “umbrella” that was named “brand competence”. The meaning is that different dimensions of brand trust towards products charged with political meaning collectively contribute to the formation of brand trust, not to measure brand trust.

As posited on page 11, rows 319-325, “This result could be since the focus of this research was a brand characterised by authentic brand activism, in which the practical actions taken by the management and the marketing communications carried out were perceived to be consistent and in support of one another. As a result, we may expect that the different constructs making up ‘brand performance’ contributed to brand trust in the broader case of authentic brand activism.”

Reviewer 3 Report

The manuscript „sustainability-1745814” presents the results of a study related to „authentic brand activism”. The authors have chosen the Italian brand „Libera Terra” for the survey. „Libera Terra” is an Italian organization that uses lands confiscated from the mafias to grow agri-food products that are environmentally and socially sustainable. The interviewees were selected from the consumers they met in the food aisles of supermarkets in Bologna and confirmed that they knew about „Libera Terra” brand.

The manuscript is organised in several sections. The introduction is comprehensive and provides a correct statement of the problem and aims of the paper. It also sets the manuscript in an international context and contains pertinent citations. The second section is a review of literature on authentic brand activism, brand loyalty, and brand trust” followed by the conceptual framework in which the developed hypotheses are presented. Materials and methods section contains the sample description, questionnaire structure, and data analysis. Brand competence, credibility, benevolence, and reputation were measured and analysed to find the influence on brand trust, then how brand trust determines brand loyalty. The results showed that the Exploratory Factor Analysis produced a factor named brand performance as a determinant of brand trust. This factor led to the rejection of several hypotheses, reduction of conceptual model and formulate a new hypothesis: „Libera Terra” brand performance leads to increased brand trust. The results were further discussed and compared to literature, theoretical contributions and managerial implications being also presented. The research showed that the authentic brand activism is effective.

The manuscript is well written. However, a few observations and suggestions. Are presented below.

Observation and suggestions

L62 Add a comma after the law number „109/96”.

L69 Avoid the repetition of „study”. Suggestion: replace the second „study” (after „empirical”) with „research”.

L145 Avoid redundancy by removing „own”.

L168 It appears that „outbreak” may be unnecessary in the sentence. Consider removing it.

L170 Replace the verb „was” with „were” to obtain „Data were collected”. („Data” is in the plural.)

L172 – 173 Consider changing „were passing” with „passed”.

L181–183 Reword the sentences to avoid starting any sentence with digits.

L191 It appears that „of” may be unnecessary in this sentence. Consider removing it.

L208 The second „to” may be unnecessary in the sentence. Consider removing it.

L202 Remove „was” before „carried”.

L343 Consider replacing „was made up” with „consisted”.

Author Response

The manuscript “sustainability-1745814” presents the results of a study related to “authentic brand activism”. The authors have chosen the Italian brand “Libera Terra” for the survey. “Libera Terra” is an Italian organisation that uses lands confiscated from the mafias to grow agri-food products that are environmentally and socially sustainable. The interviewees were selected from the consumers they met in the food aisles of supermarkets in Bologna and confirmed that they knew about “Libera Terra” brand.

The manuscript is organised in several sections. The introduction is comprehensive and provides a correct statement of the problem and aims of the paper. It also sets the manuscript in an international context and contains pertinent citations. The second section is a review of literature on authentic brand activism, brand loyalty, and brand trust” followed by the conceptual framework in which the developed hypotheses are presented. Materials and methods section contains the sample description, questionnaire structure, and data analysis. Brand competence, credibility, benevolence, and reputation were measured and analysed to find the influence on brand trust, then how brand trust determines brand loyalty. The results showed that the Exploratory Factor Analysis produced a factor named brand performance as a determinant of brand trust. This factor led to the rejection of several hypotheses, reduction of conceptual model and formulate a new hypothesis: “Libera Terra” brand performance leads to increased brand trust. The results were further discussed and compared to literature, theoretical contributions and managerial implications being also presented. The research showed that the authentic brand activism is effective.

The manuscript is well written. However, a few observations and suggestions. Are presented below.

Observation and suggestions

L62 Add a comma after the law number “109/96”

L69 Avoid the repetition of “study”. Suggestion: replace the second "study" (after "empirical") with "research".

L145 Avoid redundancy by removing “own”.

L168 It appears that “outbreak” may be unnecessary in the sentence. Consider removing it.

L170 Replace the verb “was” with “were” to obtain “Data were collected”. (“Data” is in the plural.)

L172 – 173 Consider changing “were passing” with “passed”.

L181–183 Reword the sentences to avoid starting any sentence with digits.

L191 It appears that “of” may be unnecessary in this sentence. Consider removing it.

L208 The second “to” may be unnecessary in the sentence. Consider removing it.

L202 Remove “was” before “carried”.

L343 Consider replacing “was made up” with “consisted”.

Thank you for these recommended suggestions, all of them were implemented in the manuscript.

Round 2

Reviewer 2 Report

I believe that the authors have tried to respond to the elements highlighted in the previous review. However, certain elements still need to be clarified.

In the first place, as already mentioned, it must be specified whether trust is considered a second-order construct or not. At the beginning of section 3, the authors indicate that they have considered trust a second-order construct. However, such an element is not taken into account in the writing of the hypotheses that present them as background.

Likewise, it is necessary to be more precise in the support in the references. In this way, it is stated that trust encompasses “specific attitudes, such as brand competition and reputation [23, 24]. Does such work raise only such attitudes?

Also, the analysis should be clarified. In this way, the authors rely on the work of Lassoued & Hobbs (2015). However, in the work they cite, the justification for using the construct "Per" has nothing to do with the justification presented in the work under analysis.

Therefore, authors are requested to:

- Clearly present the antecedents of trust as antecedents or as a second-order construct. In the case of using them in both ways, such extreme must be clearly justified.

- Justify in detail, in the analysis, the step from considering trust from a second-order construct to a single construct.

Author Response

Reviewer(s)’ Comments to Author:

Reviewer #2

I believe that the authors have tried to respond to the elements highlighted in the previous review. However, certain elements still need to be clarified.

In the first place, as already mentioned, it must be specified whether trust is considered a second-order construct or not. At the beginning of section 3, the authors indicate that they have considered trust a second-order construct. However, such an element is not taken into account in the writing of the hypotheses that present them as background.

Appreciating the reviewer’s comment, the following statements were inserted:

Page 3, rows 144-145: “Considering Btrust as a second-order construct, it can be hypothesised that it is a function of consumers’ trust in Bcomp and Brep”

Page 4, rows 165-167: “This trust-based conceptual model hypothesises that Btrust is a second-order construct affected by the individual’s perception of competence, reputation, benevolence, and credibility of the “Libera Terra” brand.”

Likewise, it is necessary to be more precise in the support in the references. In this way, it is stated that trust encompasses “specific attitudes, such as brand competition and reputation [23, 24]. Does such work raise only such attitudes?

As the reviewer has seen, this work didn’t encompass only brand competence and brand reputation. The sentence mentioned by the reviewer it is merely used to introduce the section. For this reason, “specific attitudes” is followed by “such as”.

Also, the analysis should be clarified. In this way, the authors rely on the work of Lassoued & Hobbs (2015). However, in the work they cite, the justification for using the construct “Per” has nothing to do with the justification presented in the work under analysis.

Technically the justification for considering only one construct can be found on page 8, rows 244-256. Specifically, after conducting a CFA that did not confirm the four hypothesised trust antecedents, an EFA produced two factors, of which only one has an acceptable level of internal consistency, and it encompassed the items of the four previous constructs. Unlike Lassoued & Hobbs (2015), which found good composite reliability of all the constructs but didn’t find a significant relationship between trust and its antecedents, in our research, as previously mentioned, it was not possible to discriminate the four factors. For this reason, we conducted an EFA suggesting only one factor’s presence, named brand performance.

Therefore, authors are requested to:

- Clearly present the antecedents of trust as antecedents or as a second-order construct. In the case of using them in both ways, such extreme must be clearly justified.

Please, see the abovementioned answers.

- Justify in detail, in the analysis, the step from considering trust from a second-order construct to a single construct.

Please, see the abovementioned answers.

Round 3

Reviewer 2 Report

I think the authors have explained the steps developed in their work.

Also, the other reviewers agree the acceptance.

So, I accept the publication