The Relative Role of Knowledge and Empathy in Predicting Pro-Environmental Attitudes and Behavior

Round 1

Reviewer 1 Report

Summary: The authors have surveyed environmental attitudes and behavior along with objective environmental knowledge and self-reported empathy across its cognitive and affective facets. Objective environmental knowledge was the strongest predictor of environmental attitudes and behavior, and a stronger predictor than empathy scores.

Context: I accepted this review as an empathy expert with interests in environmental behavior. I first acknowledge the methodological advancement of the study in measuring environmental knowledge and I appreciate that the division between cognitive and affective empathy has been conducted in all analyses.

Comments: The study is methodologically sound and I have only limited concerns that I am sure the authors will easily address. These mainly concern the measurement and interpretation of empathy: The authors make strong claim about the superiority of knowledge over empathy as a predictor but the methodological rigor applied to measuring knowledge has not been applied for measuring empathy.

• The authors clearly emphasized the need to specifically measure empathy, which they did, but not empathy for the environment – the planet, the animals, the Nature, or else (e.g., https://0-www-sciencedirect-com.brum.beds.ac.uk/science/article/abs/pii/S0272494413000273 ). The empathy scale used only measures general empathic habits and tendencies for other humans. The empathy measured is not the empathy you should have measured. This must be acknowledged as a potential explanation of why empathy was a small predictor.
• The authors clearly emphasized the need to objectively measure knowledge, but the same arguments apply for empathy. Studies, including my own work (e.g., https://0-bmjopen-bmj-com.brum.beds.ac.uk/content/11/9/e048597.abstract) , show that you can obtain very different results when measuring self-reported habits and tendencies versus objectified affective and cognitive empathic competencies. This must be acknowledged as a potential explanation of why empathy was a small predictor.
• The authors focused their empathy literature on manipulation of empathy and perspective-taking (= cognitive empathy), which brings the argument that indeed enhanced empathy is transient and thus insufficient to solve the climate crisis. But there is also studies about measuring empathy and environment/climate (e.g., https://0-www-sciencedirect-com.brum.beds.ac.uk/science/article/abs/pii/S0959378020307470 ), which speak of durable traits and also about what are the predictors of the latent traits. Further, there are also solutions to durably improve empathy. These are known: increase familiarity, attachment, self-relevance, anthropomorphize empathy target, create narrative, are all factors that can durably change how perceive the planet and therefore how we think about it. For instance, the more you anthropomorphize an object the more likely you will empathize for it. These lines of research are absent despite it could give hopeful directions for empathy as a levier for environmental behavior.
• The fact that statistical predictive power of empathy variables is systematically higher in regression analyses than in the correlation matrix strongly suggests that empathy scores are a complementary source of explanatory power to knowledge in predicting environmental attitudes/behavior. Hence, I think the authors underestimate the usefulness of measuring empathy. However, I am already glad they did measure empathy in this study!

Based on these 4 criticisms I have would like the authors to revisit their claim as there is currently a strongly claim of superiority over empathy.

 

At last, there seems to be an interesting coupling: affective empathy rather predicts environmental attitudes whereas cognitive empathy rather predicts environmental behavior. Conceptually it makes sense as attitudes is more about perception and affective empathy measures specifically how you perceive events that happen to others, how you perceive their feelings, and how all these perceptions affect you. A high scorer in affective empathy will be easily and more strongly affected by whatever affective content associated to the person. In contrast, cognitive empathy is more about the tendency/habit to actively consider the other person’s point of view, which demonstrates somehow a proactive interest for the others. This is of course speculative but I want to share this interpretation of your findings.

 

Comments for author File: Comments.pdf

Author Response

We thank the Reviewer for their very positive assessment of our work and a particular appreciation of having measured empathy. We have undertaken major revisions to address the concerns raised and feel that our manuscript now offers a much more balanced evaluation of the findings thanks to this helpful input.

1. The authors clearly emphasized the need to specifically measure empathy, which they did, but not empathy for the environment – the planet, the animals, the Nature, or else (e.g., https://0-www-sciencedirect-com.brum.beds.ac.uk/science/article/abs/pii/S0272494413000273). The empathy scale used only measures general empathic habits and tendencies for other humans. The empathy measured is not the empathy you should have measured. This must be acknowledged as a potential explanation of why empathy was a small predictor.

• We now justify our choice of this scale at the end of the Introduction section of the manuscript

Given the potential for a dissociation between the effects of cognitive and affective empathy on pro-environmental attitudes and behavior, we used the Questionnaire of Cognitive and Affective Empathy (QCAE) [61] as a scale of trait empathy because it measures both components. Although we did not measure empathy for the planet per se, the QCAE is a well validated and widely used empathy scale, with the advantage of decomposing empathy into its cognitive and affective components. This allowed us to test the prediction that any pattern of correlation between knowledge and pro-environmental behavior would be mirrored more closely by cognitive than affective empathy.

• We also flag this limitation in the abstract and discuss this limitation in the Discussion

Introduction: 

Future research should aim to extend the present findings by testing whether a more exhaustive test of participants’ environmental knowledge and other measures of empathy, including empathic competencies and the recently developed Emotional and Cognitive Scale of the Human-Nature Relationship (ECS-HNR), yield the same dominance of knowledge over empathy.

Discussion:

While scales to measure empathy towards the planet and the connection humans feel towards Nature exist [e.g., 84,14], it is important to note that we found a predominance of knowledge over empathy using a scale that measured human-directed empathy [61]. We selected this empathy scale because it dissociated empathy’s cognitive and affective components [61]. Results revealed a dissociation whereby affective empathy predicted attitudes but not behavior, while cognitive empathy predicted both attitudes and behavior, just like verifiable knowledge. This dissociation was in line with our prediction that the influence of cognitive aspects of empathy (perspective-taking) would mirror that of verifiable knowledge. Why affective empathy only influenced attitudes in our results is not entirely clear, but it is possible that it reflects a dissociation of emotion and action whereby empathic concern is downregulated to protect against burnout [85]. Very recently, a scale encompassing both cognitive and affective dimensions of empathy for nature has been developed [86]. This significant development demonstrates the current interest in disentangling cognitive and affective components of empathy in conservation psychology and provides a much-needed advancement to the field. Future research using this new tool will provide a means of testing whether the pattern of findings found here, that is, the stronger and broader influences of cognition (i.e., knowledge and cognitive empathy) over affect, are also present when empathy towards the planet is measured rather than empathy towards humans.

2. The authors clearly emphasized the need to objectively measure knowledge, but the same arguments apply for empathy. Studies, including my own work (e.g., https://0-bmjopen-bmj-com.brum.beds.ac.uk/content/11/9/e048597.abstract) , show that you can obtain very different results when measuring self-reported habits and tendencies versus objectified affective and cognitive empathic competencies. This must be acknowledged as a potential explanation of why empathy was a small predictor.  

• We thank the Reviewer for pointing us to this line of research. We now refer to it in the discussion

In addition to measuring human-directed empathy, the QCAE is also a self-assessment measure. It has been shown that the use of self-reported empathic habits and tendencies can yield different results to those obtained when using more objectified measures of empathy [87]. Hence, another direction for future work is to test whether empathy becomes a stronger predictor of pro-environmental attitudes and behavior when more objective measures of empathic competencies within the environmental context are used. It should be noted that, to the best of our knowledge, this depends upon future development of a validated measurement scale for this purpose, however.

3. The authors focused their empathy literature on manipulation of empathy and perspective-taking (= cognitive empathy), which brings the argument that indeed enhanced empathy is transient and thus insufficient to solve the climate crisis. But there is also studies about measuring empathy and environment/climate (e.g., https://0-www-sciencedirect-com.brum.beds.ac.uk/science/article/abs/pii/S0959378020307470), which speak of durable traits and also about what are the predictors of the latent traits. Further, there are also solutions to durably improve empathy. These are known: increase familiarity, attachment, self-relevance, anthropomorphize empathy target, create narrative, are all factors that can durably change how perceive the planet and therefore how we think about it. For instance, the more you anthropomorphize an object the more likely you will empathize for it. These lines of research are absent despite it could give hopeful directions for empathy as a levier for environmental behavior.

• We now refer to the literature examining the predictors of empathy and address the key role of anthropomorphism extensively in the discussion

On the one hand, the narrow focus on fostering empathy and compassion, as proposed by compassionate conservationists, might over-estimate the potential of these specific traits to drive consistent, long-term attitudinal and behavioral change towards the natural world. Indeed, our findings suggest that the power of environmental knowledge should not be overlooked. On the other hand, there are emerging lines of research identifying the predictors of empathy towards nature and suggesting that it might be possible to produce durable changes in empathy, particularly by leveraging people’s capacity to anthropomorphize [39]. For example, empathic concern (i.e., cognitive empathy) appears to be a durable trait which predicts how much one feels affected by environmental crises, with the limitation that it is subject to down-regulation and collapse when it becomes overwhelming [90], as pointed out by Griffin et al. [53]. The more animals are perceived to have human qualities (i.e., suffering), the greater the increase in empathy-related responses towards animals [42]. When one perceives nature as human, the social principles applied when interacting with others (e.g., respect) can be transferred to the environment [43]. Based on these kinds of research findings, the power of anthropomorphizing is gaining political traction [44,45], while also providing hopeful directions for the power of empathy to foster perhaps durable pro-environmental attitudes and behavior.

4. The fact that statistical predictive power of empathy variables is systematically higher in regression analyses than in the correlation matrix strongly suggests that empathy scores are a complementary source of explanatory power to knowledge in predicting environmental attitudes/behavior. Hence, I think the authors underestimate the usefulness of measuring empathy. However, I am already glad they did measure empathy in this study! 

• We now explicitly state the point made by the Reviewer here in the discussion

Our finding that knowledge, cognitive empathy and affective empathy all influence pro-environmental measures to some extent demonstrates that cognition and emotion are joint contributors 

5. Based on these 4 criticisms I have would like the authors to revisit their claim as there is currently a strongly claim of superiority over empathy.

• We have re-written large sections of the discussion to address the Reviewer's four criticisms. Specifically, we now 1. use a more tentative language regarding the relative roles of knowledge and empathy throughout; 2. include the limitation of measuring empathy towards humans rather than towards the planet; 3. include the limitation of measuring self-report empathy rather than empathic competencies; 4. mention the fact that the effect of empathy is stronger in multiple regressions than in the correlation matrix; and 5. call for more research to confirm the relative importance of knowledge and empathy and the internal and external factors that presumably influence their balance.

6. At last, there seems to be an interesting coupling: affective empathy rather predicts environmental attitudes whereas cognitive empathy rather predicts environmental behavior. Conceptually it makes sense as attitudes is more about perception and affective empathy measures specifically how you perceive events that happen to others, how you perceive their feelings, and how all these perceptions affect you. A high scorer in affective empathy will be easily and more strongly affected by whatever affective content associated to the person. In contrast, cognitive empathy is more about the tendency/habit to actively consider the other person’s point of view, which demonstrates somehow a proactive interest for the others. This is of course speculative but I want to share this interpretation of your findings.

• We thank the Reviewer for sharing this interpretation with us. However, the Reviewer has not provided any references for the idea that affective empathy is influenced by perception, and we are not sure we follow this speculative line of reasoning well enough to argue the case in our paper. However, as requested, we have now included an a-priori prediction regarding expected differences between the two components of empathy, as well as our own interpretation of the dissociation found here

Introduction:

This allowed us to test the prediction that any pattern of correlation between knowledge and pro-environmental behavior would be mirrored more closely by cognitive than affective empathy.

Discussion:

Results revealed a dissociation whereby affective empathy predicted attitudes but not behavior, while cognitive empathy predicted both attitudes and behavior, just like verifiable knowledge. This dissociation was in line with our prediction that the influence of cognitive aspects of empathy (perspective-taking) would mirror that of verifiable knowledge. Why affective empathy only influenced attitudes in our results is not entirely clear, but it is possible that it reflects a dissociation of emotion and action whereby empathic concern is downregulated to protect against burnout [85]. 

Reviewer 2 Report

 In the manuscript, the authors conducted one correlational study and found that knowlege about the enviroment is a robust predictor on pro-enviromental attitudes above and beyond empathy. Here are my concerns.

1. I found the study was not strongly theoretically-based. You argued that regarding pro-enviromental behaviors as a mixture of prosocial and self-interested is more reasonable. However, you studied the factors that could influence pro-enviromental behaviors from the perspective of moral norm.
2. A moral norm is activated by an interplay of cognitive, emotional, and social factors. However, you did not study social factors. It seems the manuscript was not theoretical-driven.
3.  I do not think empathy is an emotion. As you said, it has cogntive and affective dimensions.
4. I do not think a19-item self-report questionnaire is sufficient to measure accurate knowledge of enviroment.
5. It's great to distinguish cognitive and affective empathy. However, how do they differ, what do you expect about the role of cognitive and affective empathy in pro-enviromental attitudes remains unclear to me. Also, you should also explain the inconsistent results on cognitive and affective empathy.
6. I do not think you should study the two items as pro-enviromental behaviors. Acutually, you did not measure acutal behaviors but behavioral tendency. It seems reasonable to study them as pro-enviromental attitudes.
7. Why did you have age, education, participants' source as covariates? I am wondering what the results would be like if you drop these covarates. I would like to suggest to drop these covariates if you do not have enough evidence suggesting that they would impact the relationship between predictors and outcome.
8. I think one correlational study could not provide enough solid evidence.
9. Nature benefits does not necessarily mean people would be motivated to preserve nature. It is likely that people would utilize the natural resources to benefit themselves.

 

Author Response

We thank the Reviewer for their thoughtful input. We have undertaken major revisions to address the concerns raised.

1. I found the study was not strongly theoretically-based. You argued that regarding pro-enviromental behaviors as a mixture of prosocial and self-interested is more reasonable. However, you studied the factors that could influence pro-enviromental behaviors from the perspective of moral norm.

• We have removed the sections on pro-social and self-interested modelling from the introduction to avoid setting the reader up with the expectation that we would test these theories.

2. A moral norm is activated by an interplay of cognitive, emotional, and social factors. However, you did not study social factors. It seems the manuscript was not theoretical-driven.  

• We have removed the section on moral norms from the manuscript, immediately focusing instead on the factors of knowledge and empathy and the compassionate conservation debate. We feel that this avoids building the expectation that we will test these theories.

3. I do not think empathy is an emotion. As you said, it has cognitive and affective dimensions.

• We thank the Reviewer for drawing our attention to this slip. We have edited the manuscript text to make sure we avoid implying that the dichotomy between cognition and affect more generally is mirrored in the dichotomy between knowledge and empathy. This was not our intention, and we appreciate the reviewer drew our attention to this inaccuracy. For example, we noticed that in the introduction, we moved from introducing cognition and emotion straight to knowledge and empathy, implying a parallel. This section has been removed.

4. I do not think a19-item self-report questionnaire is sufficient to measure accurate knowledge of enviroment.

• We were interested in investigating participants’ objective, verifiable environmental knowledge, including knowledge of the Australian environment, climate change, and the role of biodiversity more generally. To the best of our knowledge, there is no existing questionnaire designed for an Australian audience to test verifiable environmental knowledge. Therefore, we designed our own for for the purpose of this research. Our study revealed the predicted significant correlation suggesting that our questionnaire, despite being small was adequate to address our questions of interest. We thank the Reviewer for having pushed us to think about that and have added this reflection to the methods of the manuscript.

Although such a short survey does not provide an exhaustive test of a person’s environmental knowledge, we aimed first and foremost to keep participation time short to maximize engagement. Our findings suggest that it was adequate to address our questions of interest. 

5. It's great to distinguish cognitive and affective empathy. However, how do they differ, what do you expect about the role of cognitive and affective empathy in pro-enviromental attitudes remains unclear to me. Also, you should also explain the inconsistent results on cognitive and affective empathy. 

• We have added the prediction to the Introduction that any pattern of correlation between knowledge and our pro-environmental measures would be mirrored more closely by cognitive than affective empathy. 

Given the potential for a dissociation between the effects of cognitive and affective empathy on pro-environmental attitudes and behavior, we used the Questionnaire of Cognitive and Affective Empathy (QCAE) [61] as a scale of trait empathy because it measures both components. Although we did not measure empathy for the planet per se, the QCAE is a well validated and widely used empathy scale, with the advantage of decomposing empathy into its cognitive and affective components. This allowed us to test the prediction that any pattern of correlation between knowledge and pro-environmental behavior would be mirrored more closely by cognitive than affective empathy.

6. I do not think you should study the two items as pro-enviromental behaviors. Acutually, you did not measure acutal behaviors but behavioral tendency. It seems reasonable to study them as pro-enviromental attitudes.

• Although the reviewer may disagree, we employed the official categorization presented by this validated scale. We consider our approach therefore justified. We appreciate that these two items are not direct measures of environmental behaviours, but rather self-reported pro-environmental behavioural tendencies. However, we consider these to be distinct from the other items because of their focus on action rather than belief. Further, because the pattern of significant and nonsignificant predictors was different between preservation attitudes and preservation behavioural tendencies, we consider it appropriate to keep these outcome variables separate instead of collapsing them into a single attitudinal variable, which may obscure these distinctions.

7. Why did you have age, education, participants' source as covariates? I am wondering what the results would be like if you drop these covarates. I would like to suggest to drop these covariates if you do not have enough evidence suggesting that they would impact the relationship between predictors and outcome.

• Age and education were included as covariates because of their associations with the outcome variables, influences which need to be accounted for. Higher levels of education are associated with higher levels of environmental concern and older age is associated with more pro-environmental behaviours and less environmental concern (e.g., Gifford & Nilsson, 2014). Participant source was included to ensure that the overall conclusions were not driven by one or the other of the two different samples who participated in the study. Each of these three factors was significant in at least one of the regression models, further demonstrating their relevance in explaining variance in the outcome variables. By incorporating these predictors into our model alongside our key variables of interest, we determined whether knowledge and empathy were significant predictors even when variation due to these variables was accounted for. We therefore consider it appropriate for these covariates to remain in the analyses.

8. I think one correlational study could not provide enough solid evidence.

• Our study included a battery of analyses on our outcome variables, and we included the covariates to check whether a similar trend was seen to that expected regarding demographics (e.g., education level). We have reworded the abstract, the discussion and the conclusion to make our take-home message more tentative, calling for more research on this important question.

Abstract:

Future research should aim to extend the present findings by testing whether a more exhaustive test of participants’ environmental knowledge and other measures of empathy, including empathic competencies and the recently developed Emotional and Cognitive Scale of the Human-Nature Relationship (ECS-HNR), yield the same dominance of knowledge over empathy.

Conclusion:

Overall, our study in an Australian sample suggested that verifiable environmental knowledge, and to a lesser extent empathy, are related to people’s attitude towards the environment and their tendency to act sustainably. When considering the relative influence of both psychological dimensions, environmental knowledge was the stronger and more consistent predictor. At the very least, this indicates that focusing solely on fostering empathy and compassion as suggested by proponents of compassionate conservation is unlikely to achieve the far-reaching outcomes for which the movement argues [46-48]. More research is needed to extend the present findings by using a larger participant pool and testing whether other measures of empathy, including empathic competencies and empathy towards nature, yield the same dominance of knowledge over empathy.

9. Nature benefits does not necessarily mean people would be motivated to preserve nature. It is likely that people would utilize the natural resources to benefit themselves.

• We are unsure what part of the manuscript this comment refers to or whether the Reviewer is requesting a change. We have included a section in the Introduction where we explain that nature benefits to one’s personal wellbeing should incentivize individuals to act sustainably, but that we also know from research in health psychology that motivating change in human behavior is surprisingly challenging even when people are aware of these personal benefits. We think that by investigating both preservation and utilization attitudes, our study makes some progress on examining both altruistic (preservation oriented) and more selfish/personal gain-based (utilization oriented) motivating factors.

Experience with the natural world provides numerous benefits to wellbeing [12], including stress reduction [13], reduced blood pressure and improved mental health, and increased life satisfaction [14]. These benefits should incentivize individual actions that preserve nature. Yet decades of research in health psychology have demonstrated that changing human behavior even when it benefits the individual remains surprisingly challenging [15,16].

Reviewer 3 Report

The authors presented an interesting manuscript about the power of empathy and knowledge in predicting pro-environmental attitudes and behavior, producing some important findings.

It is very well written, analyzed and discussed and worthy of publication.

Comments

---Lines 150-162 – The sample is not representative of the study population. This must be emphasized as limitation in the Discussion.

---Lines 180-234 – The questionnaires used as measures should be given in supplementary files, along with item means and standard deviation. This information is important for the reader, allowing for a deeper understanding. Especially so for the Environmental Knowledge Survey which was firstly presented in this manuscript.

---Lines 210-223, 255 – It is not clear how were answers to the 19-item Environmental Knowledge Survey rated. Please clarify.

Author Response

1. Lines 150-162 – The sample is not representative of the study population. This must be emphasized as limitation in the Discussion.

• We now explicitly state that research with a larger sample size is needed

Conclusion:

More research is needed to extend the present findings by using a larger participant pool and testing whether other measures of empathy, including empathic competencies and empathy towards nature, yield the same dominance of knowledge over empathy.

2. Lines 180-234 – The questionnaires used as measures should be given in supplementary files, along with item means and standard deviation. This information is important for the reader, allowing for a deeper understanding. Especially so for the Environmental Knowledge Survey which was firstly presented in this manuscript.

• Questionnaires used are presented in the Appendix section of the manuscript. Item means and standard deviations are presented in table format in the document titled “Item Means and Standard Deviations for Measures” in Supplementary Materials.

3. Lines 210-223, 255 – It is not clear how were answers to the 19-item Environmental Knowledge Survey rated. Please clarify.

• We have now incorporated more detail into the methods:

Participants’ total scores ranged from 0-38 with correct answers scored as +1 and incorrect answers scored as 0. Scores between 27-39, denoting a high score, were indicative of an extensive environmental knowledge base. Scores between 14-26 were indicative of an intermediate knowledge of environmental and biodiversity issues, and scores between 0-13 indicated participants were uninformed on issues and information about the environment and biodiversity.

Reviewer 4 Report

In this paper it is not clear to what theory authors referred and what type of the behavior was analysed. Because the separate type of behavior is determined by different factors. Furthermore authors analysed the environmental knowledge impact on behavior, but for me it is also not clear what type of knowledge was analysed whereas in the literature review authors provided that different knowledge differently influence behavior. To analyse the knowledge impact in general it is not logical becuase it shows anything maybe only the education level, because speaking about species it is difficult to answer to these questions correctly even for people who attribute himself as environmentalist. Therefore, in this case I suggest to revise or to clarify this factor.

In the section of methodology and results authors should provide the main assumptions of regression analysis. Absolutely it would be more interesting if the knowledge would be analysed as mediator because in the vast of literature authors analysed this point.

The representativeness of the survey also should be reasoned.

The conclusion section and policy implication sections should be provided as well.

Author Response

We thank the Reviewer for this helpful input, which we think has helped improve the clarity of the manuscript.

1. In this paper it is not clear to what theory authors referred and what type of the behavior was analysed. Because the separate type of behavior is determined by different factors. Furthermore authors analysed the environmental knowledge impact on behavior, but for me it is also not clear what type of knowledge was analysed whereas in the literature review authors provided that different knowledge differently influence behavior. To analyse the knowledge impact in general it is not logical becuase it shows anything maybe only the education level, because speaking about species it is difficult to answer to these questions correctly even for people who attribute himself as environmentalist. Therefore, in this case I suggest to revise or to clarify this factor.

• We have added an operational definition of knowledge to the Introduction

As per Geiger et al.’s [33] argument about the importance of how knowledge is measured, we quantified environmental knowledge using a questionnaire designed to test objective, verifiable knowledge of environmental issues. The questionnaire was designed for an Australian sample. Consequently, in the context of the current study, environmental knowledge was operationalized as specific verifiable knowledge of the natural world beyond that of general scientific literacy. 

2. In the section of methodology and results authors should provide the main assumptions of regression analysis. Absolutely it would be more interesting if the knowledge would be analysed as mediator because in the vast of literature authors analysed this point. 

• An outline of the regression assumption results has been added to the Results section.

Regression assumptions were met for all three tested models. Variance inflation factors for each variable ranged between 1.05 and 1.84, indicating no evidence of multi-collinearity issues. The skewness and kurtosis values of the residuals were well within ± 1 (skewness between 0.34 and -0.74 [SE = 0.10] and kurtosis between 0.15 and 0.89 [SE = 0.89]), indicating a reasonably normal distribution. Durbin-Watson values were close to 2 (1.53-1.85 across the three models), indicating sufficient independence of errors.

• As for the mediation analyses suggestion above, we initially decided not to take this approach because there are better ways to test mediation effects, such as experimental manipulation or time-lagged (short longitudinal) survey designs. Being cross-sectional, our data are less appropriate for making a mediation argument, so we have not conducted mediation analyses.
• We did conduct sensitivity analyses in which we re-ran the regressions excluding outliers. The pattern of results remains the same except for the regression on preservation attitudes

All regression analyses were reconducted excluding outliers (i.e., values with a Cook’s Distance value greater than three times the mean Cook’s Distance). The results of these sensitivity analyses are reported only when the pattern of significant and non-significant predictors or the relative magnitudes of the coefficients differed from the main analyses.

• In the main analyses, cognitive empathy was not a significant predictor of preservation attitudes, but in the sensitivity analyses, it became significant. A note of this has been made to the results section “The pattern of results was the same for the sensitivity analyses, except that cognitive empathy was also a significant predictor of preservation attitudes (β = 0.07, p = 0.014).”

3. The representativeness of the survey also should be reasoned.

• We were interested in investigating participants’ objective, verifiable environmental knowledge, including knowledge of the Australian environment, climate change, and the role of biodiversity more generally. To the best of our understanding, there is no current species-specific knowledge questionnaire designed for an Australian audience. Therefore, the Environmental Knowledge Survey was designed by our broader research team for the purpose of this research. We have added a reflection on this point to the methods

Although such a short survey does not provide an exhaustive test of a person’s environmental knowledge, we aimed first and foremost to keep participation time short to maximize engagement. Our findings suggest that it was adequate to address our questions of interest. 

4. The conclusion section and policy implication sections should be provided as well.

• We now include additional conclusion and policy implication sections at the end of the discussion.

Conclusions

In the summer of 2019, the east coast of Australia was ravaged by unprecedented fires, estimated to have killed or displaced over 3 billion animals [93]. Unprecedented financial donations flowed to environmental charities during the fires and in the months that followed [93]. Images of burnt koalas hanging to the top of one tree amongst a landscape or burnt trees were perhaps powerful triggers of empathic suffering, a negative emotion potentially alleviated by the action of donating [89,88]. Yet Australia also has one of the highest prevalence of climate change deniers in the world and advocating for action on environmental issues has caused systematic political suicide [94]. To design interventions that effectively lead to long-term consistent individual tendencies and habits to behave sustainably [16], we need to achieve a better understanding of how objective knowledge and empathy interact to influence pro-environmental attitudes and behavior.

Overall, our study in an Australian sample suggested that verifiable environmental knowledge, and to a lesser extent empathy, are related to people’s attitude towards the environment and their tendency to act sustainably. When considering the relative influence of both psychological dimensions, environmental knowledge was the stronger and more consistent predictor. At the very least, this indicates that focusing solely on fostering empathy and compassion as suggested by proponents of compassionate conservation is unlikely to achieve the far-reaching outcomes for which the movement argues [46-48]. More research is needed to extend the present findings by using a larger participant pool and testing whether other measures of empathy, including empathic competencies and empathy towards nature, yield the same dominance of knowledge over empathy.

Policy implications

Our results suggest that environmental knowledge was a stronger and more consistent predictor of pro-environmental attitudes and behavior. The finding suggests that advocates of environmental protection and environmental policy makers need to ensure that people’s environmental knowledge is increased and not rely solely on eliciting empathic suffering. More generally, developing policies that encourage and support pro-environmental engagement will benefit strongly from an in depth understanding of the relative importance of affect and cognition in driving change, as well as of how this balance is moderated by internal (e.g., personality) and external factors.

Reviewer 5 Report

Thanks for giving the opportunity to review the paper titled as “The relative role of knowledge and empathy in predicting pro-environmental attitudes and behavior”. This paper is well developed, conducted and well written. It addresses a significant topic likely to be of interest to individual pro-environmental behavior and sustainability. Despite all of this, there are some minor revisions as the following:

1. The authors should better to present the specific items of scales for readers in the part of Appendix.

2. The participants information should be merged into one table.

3. There latest references about environmental knowledge, environmental literacy, pro-environmental behavior should be incorporated into the paper to strengthen the literature basis.

“Factors Influencing Public-Sphere Pro-Environmental Behavior among Mongolian College Students: A Test of Value–Belief–Norm Theory”, Sustainability, 2018.

“Application of the Modified College Impact Model to Understand Chinese Engineering Undergraduates’ Sustainability Consciousness”, Sustainability, 2020.

Author Response

We thank the Reviewer for helping us determine the right level of detail and provide the reader with all the required information and new references.

1. The authors should better to present the specific items of scales for readers in the part of Appendix. 

• Questionnaires used are presented in the Appendix section of the manuscript.

2. The participants information should be merged into one table. 

• Table 1 in the Materials and Methods section of manuscript has been updated to include participant demographic information.

3. There latest references about environmental knowledge, environmental literacy, pro-environmental behavior should be incorporated into the paper to strengthen the literature basis. 

• We have added the following references

Kukkonen et al. Factors Influencing Public-Sphere Pro-Environmental Behavior among Mongolian College Students: A Test of Value–Belief–Norm Theory”, Sustainability, 2018.

Zhao et al. Application of the Modified College Impact Model to Understand Chinese Engineering Undergraduates’ Sustainability Consciousness”, Sustainability, 2020.

Nelson, S.-M., Ira, G., & Merenlender, A.M. Adult Climate Change Education Advances Learning, Self-Efficacy, and Agency for Community-Scale Stewardship. Sustainability 2022, 14, 1804. https://0-doi-org.brum.beds.ac.uk/10.3390/su14031804

Round 2

Reviewer 4 Report

Accept

Author Response

Authors thank you for your kind comment.