The authors changed their plans based on reviewer comments: they decided to focus more closely on the replication and leave extensions for future research. I think this decision is understandable. However, given the wealth of relevant research published over the last 25+ years or so, I would strongly encourage the authors that they also use this replication as an opportunity to clearly outline a future research agenda (based on their findings and this broader literature). I appreciate that they already plan to discuss some of our suggestions.

The authors and reviewers have also identified several limitations in the original study. While one might argue that these limitations weaken the value of a close replication, the authors have revised their study design to better account for them. This enhances the potential for self-correction and significantly improves the replication effort. I look forward to seeing the results. 

I previously reviewed this submission and left several comments. The authors have responded to my comments and adapted the manuscript in multiple ways to address them. In summary, I think my comments have been sufficiently addressed, although I am left with some concerns still.

One of my major comments was that the value of the target article, and why it should be replicated, was not made clear. The author's response is mostly one of arguing for the importance of replications in general and saying that it is beyond the scope of the project to argue for the importance of the target article. I'll defer to the editor whether or not this response is sufficient. I do want to repeat that I do think it is a significant limitation of the proposed study to not more clearly establish the importance of the target article and, in turn, the proposed study. I also think not engaging in this question is a symptom of a more general approach the authors take that I am skeptical of. The authors seem to take a rather mechanical approach to conducting this study. By this I mean they aim to follow a set of rules rather than make decisions regarding the more complex aspects of theory, study design, and data interpretation (as also seen in my comment and their reply on counting successfully replicated hypotheses to determine whether the study as a whole is a successful replication). I think it cannot be avoided that one has to make subjective decisions in doing science and I'd rather see attempts to do so based on argumentation rather than procedure. I can see, however, that this is an approach that might be favored by others and if it falls within the guidelines of PCI-RR, then this seems fine with me. 

Hi all,

The revised manuscript looks much improved, especially in its broader theoretical motivation. In particular I like the paragraph beginning with "Beyond the academic attention it received​..." I also think the summary of the target article's methods is much improved.

I continue to think that defining and motivating the smallest effect size of interest​ would be of interest to readers. Willem made the same point in his initial review. It's up to the editors whether this is a showstopper or not. 

No further comments. 

Dear authors,

Thank you for inviting me to review your planned replication and extension of Rozin et al. (1997), a seminal study on moralization and vegetarianism. I praise the authors for their extensive preparation, and I think they provide a clear and sensible justification for replicating the target article. Given that this article has been published more than 25 years ago, it also makes sense to extend the study. Over the years, the field of vegetarianism has grown substantially, with new studies providing insight on the moralization of animal product comsumption and new measures that have been proposed to measure motivations for meat avoidance. More broadly, there have been interesting developments in the psychology of moralization. 

One study that, to my surprise, was not mentioned in the manuscript is the study of Feinberg et al. (2019), who examined the process of moralization in three longitudinal studies and found that moral emotions (e.g., disgust, guilt) and moral piggybacking (e.g., "“When you think about eating meat, to what extent does it lead you to think about your identity as a moral person?") mediated moralization. I think this study is relevant to mention, as you also examine moral emotions, and the measure of moral piggybacking somewhat overlaps with the "personality reasons" in Table 5; watching animal cruelty videos may cause people to reflect on their moral identity when they think about eating meat. 

More theoretical background

I think research on meat-animal dissociation (and compartimentalization) provides additional clues on the moralization process. Most people perceive meat as food, but when people are reminded that meat comes from animals, they experience moral emotions which may motivate meat avoidance (Benningstad & Kunst, 2020). I believe that moral vegetarians moralize meat eating mainly because they view meat not as food, but as remnants of a harmed animal, or more broadly, as a cause of harm. In short, their meat avoidance probably reflects harm avoidance (De Groeve et al., 2022). This aligns with the theory of dyadic morality, according to which moral judgment depends on perceptions of an intentional agent causing damage to a vulnerable patient (Schein & Gray, 2018). When people make the connection between their meat-eating behavior and its harmful consequences, this might cause them to reflect on their self-image as a moral agent (i.e., moral piggybacking) and motivate dietary change. This is also consistent with norm activation theory, in which an awareness of harm and perceived responsibility predict a sense of personal obligation to take action. To better understand the consequences of this moralization process, the paper of Buttlar et al. (2024) seems informative, as they examine meat ambivalence across different stages of behavioral change (also action and maintenance stages). This paragraph is just to share more theoretical background, because I felt this was missing and it might help to interpret findings or further improve the design (e.g., extensions). I will now proceed to more directly comment on the planned study. My main concern is the measurement of reasons for avoiding meat (see below).

Abstract

I know the findings are simulated, but why do the upper boundaries of the confidence intervals for ds* extend to infinity (∞)?  

Authors use the terms "moral/health vegetarian" and "moral/health-origin vegetarian" interchangeably. Personally, I prefer the former terms for readability, but this is a minor issue.

Research questions and hypotheses

Given that the authors plan a replication of Rozin et al. (1997), the research questions and hypotheses are based on this study. The authors want to examine consequences of moralization in vegetarians, by comparing moral and health vegetarians' reasons for avoiding meat and their emotional reactions towards eating meat. The authors found that the t values were not calculated correctly in the target article, and found that hypothesis 3 was no longer supported after recalculation. Based on the note of the authors on p.15, I can confirm that this recalculation is correct. Concerning the note of the authors on p.16, I think equivalence testing makes sense in relation to hypothesis 7, though I do not have experience with this type of analysis. 

Extensions

With regards to the proposed extensions (p.17-18), I'm a bit surprised that the authors want to elaborate on hypothesis 3, given that this hypothesis was actually not supported in the target article. Do they expect that the hypothesis will be supported in their study due to a larger sample size? 

The authors also hypothesize that compared to health-origin vegetarians, moral-origin vegetarians: (1) are more opposed to the use of animals for scientific testing, (2) like a wider range of animals, and (3) reject a wider range of products that directly or indirect involve the use of animals, but the rationale behind these hypotheses is not entirely clear to me. Are these also construed as consequences of moralization? 

If the authors are interested in consequences of moralization, I think some of the following variables could (also/instead) be considered:
- the willingness to engage in advocacy or support policies (Bryant et al., 2024)
- the rejection of motives to eat meat (Hopwood et al., 2021)
- ambivalent attitudes towards meat (Buttlar et al., 2024).  
- perceptions of meat-eaters (e.g., as less moral) or vegetarians (e.g., as more moral, less judgmental) (Aloni et al., 2024) 

Method: Sample size

The authors conducted power analyses and the sample size is considerably larger than the one in the target article. 

Method: Design and procedure

I'm a bit worried about the length of the questionnaire. The authors might consider to shorten it based on pretest results.

Method: Measures

Concerning the note on p.27: I agree with the authors that the quotation marks can be removed.

Meat avoidance:
I think this question is mainly interesting if you would ask it before the fixed measures.

Reasons for avoiding meat: 
- A downside of this measure is that it is not rigorously validated. For example, the "moral" measure is mainly about animal rights/welfare, but includes one double-barreled item referring to religious beliefs and/or movement membership. The moral label is also confusing, because ecological/environmental motives can also be construed as moral motives (see Rosenfeld, 2019). 
- The sum of non-moral, non-health reasons appears to include ecological reasons, but not environmental reasons, which seems inconsistent. Why are environmental and ecological reasons not in the same category?
- To test the hypothesis that moral vegetarians indicate more non-moral(/health) reasons to avoid meat, it seems crucial to ensure that the non-moral reasons are (perceived as) non-moral. For example, the personality reason "Killing and eating animals makes it easier for us to be aggressive and violent" could also be construed as a moral reason (and was indeed highly correlated with "moral/ecological reasons"). Disgust can be a consequence of moral reasons, but not necessarily. In other words, the classification of reasons as "moral" or not by the authors/Rozin et al. (1997) may unintentionally influence whether the hypothesis is supported. Ideally, the authors should pretest reasons for their perceived morality. 
- The authors correctly state that new reasons for vegetarianism have emerged over the years. Recently, Stahlmann et al. (2024) published a new measure of vegetarian motives (VEMI+), including a measure of health, environment, animal rights, disgust, social, pandemic and zoonotic diseases, and farm workers’ rights motives.  
- If you want to adapt the current measurement of reasons or use new measures for reasons to test the hypotheses 1 and 2, you can also find non-moral motives for food consumption in the measure library of PHAIR (p.16-17). Although I understand that a direct replication requires little deviation in the original design to enable comparability, I think deviations can be justified if they allow to better test hypotheses.
- Concerning: "In the original study, participants were instructed to interpret the term meat as beef, unless they do not avoid eating beef."

If the the latter is the case, the participant would not be a (dietary) vegetarian, so this formulation appears a bit strange to me. 

Extensions:

See above. I'm not sure about the theoretical value of these extensions, in particular concerning the 45 animal meats (e.g., gorilla, lion, elephant) and the liking of animals. In my opinion, this takes a lot of space in the questionnaire which could be used for other extensions, but I think it's up to the authors to decide given that they already prepared a lot of work.

Data analysis strategy

I agree that it makes sense to conduct one-sided Welsh's t-tests.
 
With regards to the extensions, do you plan to do conduct t-tests for each type of animal meat separately, or will they be grouped? How is the "range" calculated?

Outliers

It might still be useful to check for the impact of outliers. 

Replication - equivalence test (p.41)

I have no experience with this test, but I'm a bit confused about the null hypothesis of "no effect", 
because I read a null hypothesis in equivalence tests does not assume "no effect"? (For example, see https://en.wikipedia.org/wiki/Equivalence_test).

Supplementary

The authors provide a clear overview of comparisons and deviations between the target article and the planned replication. 

Overall, I appreciate the thoroughness and preparation evident in this replication study, and I am confident it will offer valuable insights into the consequences of moralization and vegetarianism upon revision.

Kind regards

Review of On the moralization of vegetarianism and differences between health and moral vegetarians: Registered Report replicating and extending Rozin et al. (1997) [Stage 1] by Willem Sleegers

The authors propose to replicate Rozin et al. (1997). I think the authors do a great job in handling many aspects of conducting a replication study. Below I offer some of my thoughts, grouped under Major comments, Minor comments, and Writing comments. Naturally, I think my Major comments are the most important and warrant serious consideration. In short, I think there’s much to like about the paper, although I think some of the bigger picture elements of a replication can be improved, such as why the target study needs to be replicated and what the goal is of the replication study. Hopefully my comments below will help in improving the paper.

Major comments

1. I think more could be done to justify why Rozin et al. should be replicated. One of the arguments given is that it has been cited 731 times. This is a decent number of times, but it’s not so much that I think it’s obvious this is a very influential paper. The authors also write that “We believe that Makel et al.’s (2012) note arguing “if a publication is cited 100 times, we think it would be strange if no attempt at replication had been conducted and published” captures the importance of revisiting such impactful findings”, but I’m not sure how much value to place on this argument. It reads more like an opinion that replications should be conducted more (which I agree with), rather than an argument about the value of this specific study. 

The next argument from the authors is that the findings of Rozin et al. (1997) were the foundational empirical evidence for Rozin’s (1999) paper, but it’s not explained how. The process of moralization is not explained, nor how the findings of Rozin et al. (1997) relate to that process. It’s not even clear to me how Rozin et al., (1997) can speak to a process of moralization, since the study appears to be wholly cross-sectional, so how can it reveal something that appears to be temporal (i.e., a process). 

The third argument, that Rozin et al. (1997) sparked interest among other researchers, is also rather weak unless the subsequent work depends on the work by Rozin et al., but there’s no persuasive case being made that the subsequent work really hinges on Rozin et al. 

I suppose I am missing more of a discussion of the moralization theory. It would be nice to discuss this, how the hypotheses derive from it, and how the results support them or not. For example, to me it does not necessarily strike me as necessary that a moral vegetarian has more reasons for being vegetarian than a non-moral vegetarian. In fact the opposite could even be true because having a single powerful moral reason is likely enough to shape behavior while more non-moral (potentially less powerful) reasons might be needed otherwise. A brief discussion of the theory and ways the hypotheses might or might not derive from it could be another argument that justifies replicating the study. 

2. I don’t really follow the reasoning to not focus on Hypothesis 3. According to a re-calculation of the reported statistics, the difference between moral vegetarians and non-moral vegetarians is no longer significant on the outcome measure related to hypothesis 3, and this is the reason for not including it. One of the other hypotheses also didn’t show a statistically significant difference, but that one is not excluded, so I don’t follow the logic here. It seems to me like the theory should dictate what the relevant hypotheses are and whether they are worth testing or not.

3. I am missing a clear discussion of the power and meaning of the effect sizes of the original study. I know that the Cohen d’s of the original study are discussed on p. 16 and p. 17, but it seems valuable to me to discuss these also in the context of what the original study was capable of finding. Was their statistical power high enough to find meaningful effects? Are the effect sizes similar to other effect sizes from other papers on the same topic? Or social psychological in general? What does the size of the confidence intervals of the various effects tell us? Relatedly, I think the sample size of the target study is not even mentioned at this point in the paper, which seems would be good to know. 

4. I am not really convinced that the extensions discussed on p. 17 are necessary. It’s true that “Moral-origin vegetarians and health-origin vegetarians may also differ on other attitudes towards animals”, but I’m not sure what the added value is compared to the measures that are being included. I also went through the survey myself and I found it was quite the list of questions, so I could also see that the study might be improved by making it shorter.

5. Maybe relatedly, I don’t get a clear sense of what the goal of the replication is. For example, is it the goal to see whether the hypotheses are correct but that the effect sizes might be smaller compared to the original? Or is it the goal to show that the statistical evidence of the original is unlikely/not compelling? 

6. The authors include an assumption that 10% of the sample might not be classified as either a moral vegetarian or non-moral vegetarian. Is there any pilot testing of this to see whether this is a plausible number? 

7. The authors write that “The current replication attempt will be considered a successful replication if of the five hypotheses (Hypotheses 1, 2, 4, 5, and 6), four or five were supported, a mixed replication if two or three were supported, and a failed replication if one or none of the hypotheses were supported.” It seems wholly unnecessary to me to classify the success of the replication in this way. It seems quite arbitrary and ignores whether some of the hypotheses might be more important than others, in addition to it not being necessary. The relevant question is whether the hypotheses are replicated or not, not the study as a whole.

8. The authors write that “we determine the consistency of replication and original effects by examining if a given replication effect is in the same direction as its corresponding original effect, instead of examining if the 95% confidence interval of said replication effect include the point estimate of its corresponding original effect.” but I’m an argument justifying this. 

Minor comments

1. Part of the research question is tautological. The question: “Do moral and health vegetarians differ in their reasons for being vegetarian” is by definition true since moral and health vegetarians are defined by having different reasons for being vegetarian.

2. The authors mention on p. 15 that there’s a discrepancy between the degrees of freedom but don’t elaborate on it. Did all of the tests have this discrepancy? If so, what are some of the possible reasons for it? How big were the discrepancies?

3. On p. 16 and p. 17 the Cohen d effect sizes have confidence intervals but the correlations do not. I know the authors write that the original study does not mention the sample size, but you could make some explicit assumptions and include them anyway. Also, I know the authors write that “we were not able to accurately compute CIs for the reported correlations”, but you can still try and give some estimates rather than implying it can't be done.

4. I don’t know what the ‘safeguard power approach proposed by Perugini et al. (2014).’ is. It would be better to briefly describe the method (which doesn’t seem that difficult to describe based on my brief reading of Perugini et al.)

5. Removing the quotation marks from ‘meat’ in the instructions seems fine to me.

6. The authors write that “We felt that this list was too limited and that there was a need to distinguish between attitudes towards the consumption of animal meats and that towards animal products/by-products.”. I think the phrasing is a bit too informal (‘we felt’) and unclear (why is it needed?).
Is there a benefit of using the cocor package over simply running a model that includes the interaction test, in order to show whether the correlations differ by group?

7. The authors write on p. 37 that “note that because they are all “individual tests of individual null hypotheses” and not tests of an intersection null hypothesis (Rubin, 2024, p.3) we did not adjust alpha for multiple comparisons”. I’m not familiar with Rubin (2024) but I can easily imagine this claim could be controversial. It’s not at all clear to me why multiple comparisons are not a potential problem here. In fact, if you say that they are all individual tests, then to me that seems like a reason why multiple comparisons IS a problem.

Writing comments

I found the structure of the text to make it a bit difficult to understand everything in the right order. Several times things were discussed or mentioned without the necessary details being discussed first. For example, in the abstract Rozin et al. is mentioned but is not briefly summarized to help readers unfamiliar with the study understand what the core measures and findings were. Another example is that the reasons for replicating the study are explained before Rozin et al. is actually summarized. It would be useful to first describe the target article is more detail and afterwards discuss its limitations, strengths, and influence, followed by the details related to replicating it.

Seems like a nice project and I look forward to seeing the results.

You've basically convinced me that the paper is important (though I don't agree with the claim that every paper with 100+ citations should be replicated -- that does not seem like a good use of our collective time 😃). I was looking forward to hearing why the results are important to you -- is this your subfield? Were its ideas influential on your work in particular? I am someone who accepts the idea that most published research findings are false, so my threshold for replicating a past study is a bit higher. Ideally it would be one that you, or people around you, had special interest in. I think you get part of the way there when establishing its influence, but I didn't really understand why I should care because I didn't get a sense of why *you* care.

There are some ways in which the writing can be tightened up a bit, but it can probably wait until publication (it's definitely good enough for a manuscript at this stage). For example, when I look at this sentence:

> One such attitude is disgust towards meat; several studies have claimed to have found evidence supporting the notion that moral vegetarians are more disgusted by meat than health vegetarians

I would go with something like 

For example, several studies have found evidence that moral vegetarians are more disgusted by meat than health vegetarians are.

or "has had much impact on the literature" can become "has been influential." I also see that you use the word effectually on p. 10 where the appropriate word is effectively, delineated on p. 8 where I think distinguish would work better,  & In essence, this serves as a test and an extension of the target article’ is missing an s at the end of article.

(BTW the phrase 'target article' feels a litlte warlike to me but perhaps that is a convention I just don't know 😃)

A few technical notes. First, you distinguish between reproducibility and replicability, but these aren't universally established terms, so I'd either define them or stick to replicability. Second, you distinguish between two types of cohen's d, please define them. Third, you spend a lot of time on the minium detectable effect and power, but I am more interested in: what's a meaningful effect size here? What does it mean for someone to have more reasons for being vegetarian than someone else -- twice as many? five times as many? I really have no idea. what are your beliefs on this question? And here I'm not that interested in a cohne's d metric -- literally can you tell me how many reason count as a litlte or a lot?

My final conceptual note: you spend a fair bit of time trying to reproduce their original results. Again, as a reader, I don't really care about whether a t-test on an article from three decades ago is precisely reproducible. That doesn't change my beliefs about the world at all because I assume errors in this category are common and I have no idea what this article's publication process looked like. For my tastes I'd just stick to: we find these questions interesting because...we think a meaningful effect size would be...this previous paper found [whatever] and we found [whatever].

But all of thtis can be adfdressed at the publication stage. I look forward to seeing the results!