Review Reports - Marijuana Consumption and Reactivity Are Positively Associated with the Fading Affect Bias for Marijuana Events in Person and Online

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

I enjoyed reading this paper assessing the relation between marijuana consumption and the fading affect bias (FAB) across two experiments conducted in-person and online.

While the manuscript addresses an important and timely question, several issues limit the strength of the conclusions drawn from the data.

First, the interpretation of the findings seems overstated. Many conclusions in this paper overstate the findings. For example, using causal-effect-related terms to refer to their correlational analysis.

And the closing mark of the paper, the conclusion that “getting high can, and is even likely to, lead to emotional highs” extends beyond what the data demonstrate.

The study shows associations between marijuana involvement variables and FAB magnitude; it does not establish that marijuana use produces elevated emotional states.

The manuscript repeatedly interprets larger FAB effects as evidence of general healthy coping, while their data cannot support this.

Second, the manuscript does not report the percentage of participants who were marijuana users versus non-users in either experiment. This weakens the study design and interpretability of the findings. Because marijuana involvement variables are treated as continuous predictors, it remains unclear whether the sample consisted entirely of active users, a mixture of users and abstainers, or included a substantial proportion of non-users scoring at zero on consumption measures. Without reporting the base rate of marijuana use, readers cannot determine (a) whether the moderation effects are driven primarily by variability among active users, (b) whether non-users meaningfully contribute to the analyses, or (c) how representative the sample is of marijuana-using populations. Considering the inconsistencies among findings across 2 experiments in the current paper, the authors should pay more attention to their sample characteristics before draw any other conclusions (though the authors did mention that sample in Experiment 2 has higher marijuana involvement). Moreover, prior research in this area (e.g., comparisons between users and non-users in Pillesdorf and Scoboria (2019) that authors mentioned multiple times in this paper) makes the user/non-user distinction theoretically/empirically relevant.

Third, like I mentioned in my 2nd point, the authors should be more transparent in reporting their sample descriptive data. The cross-experiment inconsistencies suggest that marijuana-specific effects may be sample-dependent or statistically unstable. While the core FAB pattern replicates, the marijuana involvement findings vary in strength, direction, and mechanism across experiments. This variability raises questions about the robustness and generalizability of the claim that marijuana consumption facilitates adaptive emotional processing.

terms such as: “marijuana consumption,” “marijuana involvement,” “marijuana reactivity,” “highness,” are used variably, please be consistent.

duplicated percentages in demographic descriptions (line 654)

Author Response

Reviewer 1 said “I enjoyed reading this paper assessing the relation between marijuana consumption and the fading affect bias (FAB) across two experiments conducted in-person and online.” We greatly appreciate this kind remark.

Reviewer 1 said “While the manuscript addresses an important and timely question, several issues limit the strength of the conclusions drawn from the data.” We agree with this sentiment.

Reviewer 1 said “First, the interpretation of the findings seems overstated. Many conclusions in this paper overstate the findings. For example, using causal-effect-related terms to refer to their correlational analysis.” We worked hard to reduce the amount that we overstated our findings.

Reviewer 1 said “And the closing mark of the paper, the conclusion that “getting high can, and is even likely to, lead to emotional highs” extends beyond what the data demonstrate.” We removed these comments.

Reviewer 1 said “The study shows associations between marijuana involvement variables and FAB magnitude; it does not establish that marijuana use produces elevated emotional states.” Reviewer 1 is correct.

Reviewer 1 said “The manuscript repeatedly interprets larger FAB effects as evidence of general healthy coping, while their data cannot support this.” We typically talk about the FAB as a healthy outcome, rather than a healthy coping mechanism, but we succumbed to allure of the other FAB literature making such statements. We tried to reduce the degree that we made these points and we are open to more suggestions.

Reviewer 1 said “Second, the manuscript does not report the percentage of participants who were marijuana users versus non-users in either experiment. This weakens the study design and interpretability of the findings. Because marijuana involvement variables are treated as continuous predictors, it remains unclear whether the sample consisted entirely of active users, a mixture of users and abstainers, or included a substantial proportion of non-users scoring at zero on consumption measures. Without reporting the base rate of marijuana use, readers cannot determine (a) whether the moderation effects are driven primarily by variability among active users, (b) whether non-users meaningfully contribute to the analyses, or (c) how representative the sample is of marijuana-using populations. Considering the inconsistencies among findings across 2 experiments in the current paper, the authors should pay more attention to their sample characteristics before draw any other conclusions (though the authors did mention that sample in Experiment 2 has higher marijuana involvement). Moreover, prior research in this area (e.g., comparisons between users and non-users in Pillesdorf and Scoboria (2019) that authors mentioned multiple times in this paper) makes the user/non-user distinction theoretically/empirically relevant.” We provided this information in the general discussion and we reduced the number of analyses that we reported to control for Type I error as suggested by another reviewer. Therefore, the results of the two studies are much more similar than different even though the two samples show very different rates of marijuana avoidance.

Reviewer 1 said “Third, like I mentioned in my 2nd point, the authors should be more transparent in reporting their sample descriptive data. The cross-experiment inconsistencies suggest that marijuana-specific effects may be sample-dependent or statistically unstable. While the core FAB pattern replicates, the marijuana involvement findings vary in strength, direction, and mechanism across experiments. This variability raises questions about the robustness and generalizability of the claim that marijuana consumption facilitates adaptive emotional processing.” As I said previously, we provided this information in the General Discussion.

Reviewer 1 said “terms such as: “marijuana consumption,” “marijuana involvement,” “marijuana reactivity,” “highness,” are used variably, please be consistent.” We used marijuana consumption/reactivity because participants reported the amount and time they spend consuming as well as their highness, which is a reaction to the consumption.

Reviewer 1 said “duplicated percentages in demographic descriptions (line 654)”. We corrected this issue.

Reviewer 2 Report

Comments and Suggestions for Authors

In the abstract, it is evident that it is excessively dense and combines multiple objectives into a single argumentative block. The reader has difficulty identifying the central research question: is the study intended to test whether marijuana consumption is an indicator of healthy or unhealthy coping? Or is it primarily intended to examine differences in the fading affect bias (FAB) between events with and without marijuana? The formulation “The current study examined the relation…” accumulates too many elements (consumption, healthy and unhealthy variables, events with and without marijuana, in-person and online study), which undermines clarity. A concrete improvement would be to divide this sentence into two and explicitly state one primary research question.

Furthermore, the concept of “healthy coping” is presented as a conclusion (“which demonstrated specific healthy coping”) without having been clearly operationalized in the abstract itself. It is not explained which variables were considered healthy or unhealthy, nor how this was theoretically defined. Given that the study is correlational, the expression “demonstrated specific healthy coping” is too strong and suggests causal inference. A more cautious alternative would be something like: “consistent with patterns suggestive of event-specific adaptive processing.”

It would also be advisable to include essential information currently missing from the abstract, such as sample sizes and the general type of analyses conducted. In a study with a complex design and multiple interactions, mentioning that ANOVAs and moderation models were conducted would help contextualize the findings.

Turning to the introduction, more significant issues emerge. The text begins with cultural references and colloquial expressions such as “Smoke em if you got em” and uses terms like “stoners” or culturally loaded descriptions. This type of language is inappropriate for an international scientific article and weakens the academic tone. The introduction should begin with a conceptual framing of the FAB or with the scientific gap being addressed, avoiding non-essential cultural references.

Structurally, the review of the marijuana literature is presented in a nearly dichotomous fashion: first studies associating use with negative effects (depression, anxiety, cognitive deficits), then studies showing positive effects (pain reduction, improved mood, sleep). However, there is no critical analysis of the methodological quality of these studies, nor a clear distinction between recreational and medicinal use, frequency of use versus dependence, or control of confounding variables. As a result, the review appears descriptive rather than integrative. It would be advisable to synthesize findings into theoretical categories and discuss conditions under which effects may vary.

Another important point is that the article claims to “correct” the study by Pillersdorf and Scoboria (2019), but it does not sufficiently develop why requesting only non-marijuana-related events constitutes a serious theoretical limitation. A more robust explanation is needed as to why the FAB would be expected to function differently in events involving marijuana use. The hypotheses presented are also vague. For example, the statement that “we expected the FAB to be predicted by marijuana consumption” does not specify the direction of the prediction. It would be preferable to present numbered, clearly directional, and operationalized hypotheses.

The introduction is also rather lengthy and includes a large number of studies, many of which show similar patterns of results and could be condensed. There is also a high number of self-citations, which may create the perception of confirmatory bias, although this is not necessarily problematic if justified by thematic relevance.

At the methodological level, the most critical issues arise. The primary concern relates to the unit of analysis. The article states that the event, rather than the participant, was the unit of analysis, yet events are clearly nested within participants. The use of PROCESS with a nominal “participant” variable as a control does not substitute for an appropriate multilevel model. This methodological decision may compromise statistical validity because it does not properly address the dependency of the data. A clear improvement would be to use hierarchical linear models (mixed models), with events at Level 1 and participants at Level 2.

Another relevant issue is the exclusion of a large number of events (approximately 31% in Experiment 1). Although reasons are provided (missing ratings or incorrect labeling), it is not discussed whether this exclusion may have introduced systematic bias. It would be important to analyze whether heavier users had higher exclusion rates or whether negative events were more frequently removed.

The measures of marijuana consumption also raise concerns. Variables such as “degree of highness” on a 0–100 scale or “mental competency” do not correspond to validated instruments. The absence of psychometric references limits reliability and comparability. An improvement would be to use standardized measures of cannabis frequency, intensity, and dependence, or at least to discuss the limitations of the ad hoc measures employed.

The FAB is calculated based on single-item measures of initial and current affect using a −3 to +3 scale. Although this procedure is common in the FAB literature, it still entails reliability limitations and possible regression-to-the-mean effects. The article tests regression to the mean via initial intensity, but this issue could be discussed in greater depth.

An additional problem concerns the large number of analyses conducted using PROCESS (Models 1, 3, and 11), including multiple interactions and Johnson–Neyman techniques. No correction for multiple comparisons is mentioned, increasing the risk of Type I error. Moreover, many effects present very small ΔR² values (.003, .004, .007). Although statistically significant, these are small-magnitude effects whose practical relevance should be critically discussed.

Causal interpretation is another sensitive issue. Expressions such as “marijuana consumption predicted the FAB” may be read causally, even though the design is cross-sectional and correlational. It would be methodologically more rigorous to use language such as “was associated with” or “was statistically related to.”

There is also a conceptual issue related to the temporal window of the events (past 7 days). The cited literature indicates that the FAB evolves over time and may increase after three months. The article does not theoretically justify the choice of a 7-day window nor discuss how this might affect effect magnitude.

In Experiment 2, some formal weaknesses also appear, such as inconsistencies in the sample description (e.g., repeated percentage errors), suggesting the need for careful editorial revision. Furthermore, the characterization of an MTurk sample aged 18 to 23 would warrant additional explanation, as this age range is not the most typical for that platform.

The results section is extensive, technically detailed, and statistically sophisticated, but it presents problems of organization, interpretation, and methodological grounding.

A first critical aspect concerns structure. The presentation alternates between ANOVAs, PROCESS models (Models 1, 3, and 11), Johnson–Neyman analyses, and multiple two- and three-way interactions. Although this demonstrates analytical complexity, the text becomes excessively technical and insufficiently oriented toward the central hypotheses. At several points, the reader loses track of which hypothesis is being tested. A concrete improvement would be to organize the results according to previously numbered hypotheses (H1, H2, H3…), first summarizing whether each was supported, and then presenting the statistical details.

Second, there is a strong reliance on statistical significance with limited discussion of effect magnitude. Many reported ΔR² values are extremely small (e.g., .003, .004, .007). Although statistically significant, these effects explain less than 1% of additional variance. Nevertheless, the text frequently treats them as substantively meaningful findings. It would be important to critically appraise their practical relevance, for example by explicitly stating that “although statistically significant, the effect size was small and may have limited practical implications.”

Another issue concerns multiplicity of testing. The large number of moderation and interaction analyses substantially increases the risk of Type I error. No correction for multiple comparisons (e.g., Bonferroni or FDR) is mentioned. This weakens confidence in marginal results or those with p values close to .05. A methodological improvement would be to acknowledge this limitation in the results section or, ideally, to reduce the number of exploratory tests.

Regarding the three-way analyses (Initial Event Affect × Event Type × Marijuana Consumption), the authors interpret certain graphical patterns as indicative of “specific healthy coping,” even when statistical effects are not robust or slopes are only “nearly significant.” In quantitative science, interpretation should be based on statistical results, not merely on visual patterns. Expressions such as “the pattern showed specific healthy coping, but statistics did not demonstrate it” reveal a tension between theoretical expectations and empirical evidence. It would be more rigorous to limit interpretation to what was statistically confirmed.

Moreover, the decision to treat the event as the unit of analysis without an appropriate multilevel model continues to affect the validity of inferences in this section. Because events are nested within participants, independence of observations may be compromised, potentially inflating significance levels. This limitation should be explicitly acknowledged.

With respect to mediation analyses using PROCESS Model 11, the authors repeatedly test whether rehearsal mediates complex interactions, but in most cases no significant mediation is found. Despite this, considerable technical detail is devoted to describing the model. A more concise presentation could simply state: “No evidence was found that rehearsal mediated the three-way interactions.” Excessive technical detail hinders readability without enhancing conceptual clarity.

Finally, some results are described using potentially causal language (“marijuana consumption predicted the FAB”), which is not appropriate for a cross-sectional correlational study. The term “predicted” should be replaced with “was associated with” or “was statistically related to.”

The discussion begins by reaffirming the robustness of the FAB, which is consistent with the findings. However, it quickly moves to broad interpretations about healthy coping and emotional regulation, sometimes extrapolating beyond the data.

One of the main problems is the interpretation of the positive association between “marijuana event highness” and FAB as evidence of adaptive functioning. Even if greater “highness” is associated with greater FAB in marijuana-related events, this does not imply that consumption promotes healthy coping. It may reflect memory bias, post-consumption rationalization, or pre-existing individual characteristics. The discussion should consider alternative explanations.

Another point concerns the attempt to reconcile the findings with Pillersdorf and Scoboria (2019). The authors suggest that differences may be due to the type of events recalled (with vs without marijuana). Although plausible, this explanation remains speculative. No direct test comparing methodologies was conducted. The discussion should present this as a hypothesis rather than a conclusion.

Furthermore, the discussion tends to treat statistically weak interactions as theoretically meaningful. When three-way effects show very small ΔR² values, the discussion should reflect this limitation and question their substantive relevance.

Theoretical integration could also be strengthened. The discussion revisits several prior studies but could better situate the findings within the broader framework of emotional regulation and autobiographical memory. For example, it would be useful to consider whether substance use might influence emotional consolidation processes or whether the observed effect could be explained by motivational selectivity in recall.

The methodological issue of data dependency (events nested within participants) is not sufficiently problematized in the discussion. Given its relevance, it should be explicitly acknowledged.

It would also be important to more thoroughly discuss the limitations of the consumption measures (non-validated scales, subjective self-report of “highness”) and the seven-day temporal window. The cited literature indicates that the FAB evolves over time, but this is not critically explored.

The conclusion summarizes the main findings but tends to emphasize adaptive implications of marijuana consumption more strongly than the data warrant.

In addition, the conclusion could place greater emphasis on the specific contribution of the study: the inclusion of marijuana-related events within the FAB paradigm. This is the primary methodological and conceptual innovation and should be highlighted as such.

It would also be desirable to include clear recommendations for future research, such as:
- Use of multilevel models;
- Longitudinal studies to assess actual changes in the FAB over time;
- Use of standardized measures of cannabis consumption;
- Differentiation between recreational and medicinal use.

Finally, the conclusion could explicitly acknowledge that the observed effects were generally small and that clinical or practical implications should be interpreted with caution.

The article shows a similarity index of approximately 40%, which raises serious concerns about potential plagiarism. This issue should be carefully reviewed and addressed.

Author Response

We would like to thank Reviewer 2 for all the time they put into this review to help us improve our paper.

Reviewer 2 said “In the abstract, it is evident that it is excessively dense and combines multiple objectives into a single argumentative block. The reader has difficulty identifying the central research question: is the study intended to test whether marijuana consumption is an indicator of healthy or unhealthy coping? Or is it primarily intended to examine differences in the fading affect bias (FAB) between events with and without marijuana? The formulation “The current study examined the relation…” accumulates too many elements (consumption, healthy and unhealthy variables, events with and without marijuana, in-person and online study), which undermines clarity. A concrete improvement would be to divide this sentence into two and explicitly state one primary research question.” We reworked the entire paper based on both reviewers’ comments. As a result, the goal of the study is to examine the FAB and its relation to marijuana consumption/reactivity measures across marijuana and non-marijuana events. Another goal is to examine mediation effects, but we fall short there as Reviewer 2 knows.

Reviewer 2 said “Furthermore, the concept of “healthy coping” is presented as a conclusion (“which demonstrated specific healthy coping”) without having been clearly operationalized in the abstract itself. It is not explained which variables were considered healthy or unhealthy, nor how this was theoretically defined. Given that the study is correlational, the expression “demonstrated specific healthy coping” is too strong and suggests causal inference. A more cautious alternative would be something like: “consistent with patterns suggestive of event-specific adaptive processing.” We refer to the FAB as a healthy coping outcome that is positively related to other healthy/adaptive outcomes, such as self-esteem, Grit, and positive PANAS, and it is negatively related to unhealthy/non-adaptive outcomes, such as anxiety, depression, stress, and poor sleep. We made this point in the introduction and we can make it more definitionally explicit if Reviewer 2 lets us know that the change is critical.

Reviewer 2 said “It would also be advisable to include essential information currently missing from the abstract, such as sample sizes and the general type of analyses conducted. In a study with a complex design and multiple interactions, mentioning that ANOVAs and moderation models were conducted would help contextualize the findings.” We added all of this information to the Abstract.

Reviewer 2 said “Turning to the introduction, more significant issues emerge. The text begins with cultural references and colloquial expressions such as “Smoke em if you got em” and uses terms like “stoners” or culturally loaded descriptions. This type of language is inappropriate for an international scientific article and weakens the academic tone. The introduction should begin with a conceptual framing of the FAB or with the scientific gap being addressed, avoiding non-essential cultural references.” We removed the cultural reference. We hope that the introduction is improved but we are open to future suggestions.

Reviewer 2 said “Structurally, the review of the marijuana literature is presented in a nearly dichotomous fashion: first studies associating use with negative effects (depression, anxiety, cognitive deficits), then studies showing positive effects (pain reduction, improved mood, sleep). However, there is no critical analysis of the methodological quality of these studies, nor a clear distinction between recreational and medicinal use, frequency of use versus dependence, or control of confounding variables. As a result, the review appears descriptive rather than integrative. It would be advisable to synthesize findings into theoretical categories and discuss conditions under which effects may vary.” Reviewer 2 is correct. All three of us rewrote a large portion of that section of the paper. We also used headers to guide the reader’s expectations for the sections. The information from each section is very different and it is the literature that one finds when searching. We also provided a brief summary and critique of this literature under its own heading.

Reviewer 2 said “Another important point is that the article claims to “correct” the study by Pillersdorf and Scoboria (2019), but it does not sufficiently develop why requesting only non-marijuana-related events constitutes a serious theoretical limitation. A more robust explanation is needed as to why the FAB would be expected to function differently in events involving marijuana use.” We provided a brief explanation in the abstract and the introduction and a much longer explanation in The Current Study.

Reviewer 1 said “The hypotheses presented are also vague. For example, the statement that “we expected the FAB to be predicted by marijuana consumption” does not specify the direction of the prediction. It would be preferable to present numbered, clearly directional, and operationalized hypotheses.” We minimized the hypotheses and made them specific as directed. Although we did not number the hypotheses, we can make that adjustment if Reviewer 2 thinks it is a critical issue.

Reviewer 2 said “The introduction is also rather lengthy and includes a large number of studies, many of which show similar patterns of results and could be condensed.” Two of us rewrote this section with an emphasis on brevity.

Reviewer 2 said “There is also a high number of self-citations, which may create the perception of confirmatory bias, although this is not necessarily problematic if justified by thematic relevance.” This issue is a weakness of the first author but he knows this literature best and it all helped drive the current study. If Reviewer 2 has specific suggestions about research to add to the introduction from different authors in particular places, we would be happy to add it to the introduction in those places.

Reviewer 2 said “At the methodological level, the most critical issues arise. The primary concern relates to the unit of analysis. The article states that the event, rather than the participant, was the unit of analysis, yet events are clearly nested within participants. The use of PROCESS with a nominal “participant” variable as a control does not substitute for an appropriate multilevel model. This methodological decision may compromise statistical validity because it does not properly address the dependency of the data. A clear improvement would be to use hierarchical linear models (mixed models), with events at Level 1 and participants at Level 2.” Reviewer 2 is correct. We address this issue in the General Discussion.

Reviewer 2 said “Another relevant issue is the exclusion of a large number of events (approximately 31% in Experiment 1). Although reasons are provided (missing ratings or incorrect labeling), it is not discussed whether this exclusion may have introduced systematic bias. It would be important to analyze whether heavier users had higher exclusion rates or whether negative events were more frequently removed.” These studies generally produced equal numbers of pleasant and unpleasant marijuana and non-marijuana events, which is important to address the high level of heterogeneity produced in FAB studies. As for the heavier users having higher exclusion rates, I did not see evidence for that outcome. If anything, users show a higher percentage of missing events than non-users.

Reviewer 2 said “The measures of marijuana consumption also raise concerns. Variables such as “degree of highness” on a 0–100 scale or “mental competency” do not correspond to validated instruments. The absence of psychometric references limits reliability and comparability. An improvement would be to use standardized measures of cannabis frequency, intensity, and dependence, or at least to discuss the limitations of the ad hoc measures employed.” We addressed this issue in the General Discussion.

Reviewer 2 said “The FAB is calculated based on single-item measures of initial and current affect using a −3 to +3 scale. Although this procedure is common in the FAB literature, it still entails reliability limitations and possible regression-to-the-mean effects. The article tests regression to the mean via initial intensity, but this issue could be discussed in greater depth.” We addressed this issue briefly in the General Discussion. Reviewer 2 is correct that regression to the mean is a potential issue and it should be examined and we did and it was not an issue, which has been true for most of the FAB literature.

Reviewer 2 said “An additional problem concerns the large number of analyses conducted using PROCESS (Models 1, 3, and 11), including multiple interactions and Johnson–Neyman techniques. No correction for multiple comparisons is mentioned, increasing the risk of Type I error. Moreover, many effects present very small ΔR² values (.003, .004, .007). Although statistically significant, these are small-magnitude effects whose practical relevance should be critically discussed.” First, we used a Bonferroni correction for the continuous two-way interactions because they involved 14 measures. All of the effects for the significant measures remained significant when testing at the .003571 level in Experiment 1 and all the effects were significant for all the measures except for one in Experiment 2. As for the three-way interactions, we simply focused on the marijuana measures for the continuous measures, so we tested at .01 because there are five continuous marijuana consumption/reactivity measures. We found that two of the three-way interactions were significant in Experiment 1 and only one three-way interaction was significant in Experiment 2, but one interaction replicated across the experiments. This suggestion and changed helped streamline the paper. Thank you.

Reviewer 2 said “Causal interpretation is another sensitive issue. Expressions such as “marijuana consumption predicted the FAB” may be read causally, even though the design is cross-sectional and correlational. It would be methodologically more rigorous to use language such as “was associated with” or “was statistically related to.” We agree and we apologize for indicating causation. We tried our best to limit such statements, so know that we are open to further suggestions.

Reviewer 2 said “There is also a conceptual issue related to the temporal window of the events (past 7 days). The cited literature indicates that the FAB evolves over time and may increase after three months. The article does not theoretically justify the choice of a 7-day window nor discuss how this might affect effect magnitude.” Based on the study in 2011, the FAB fades within 12 to 24 hours, and remains stable for 3 months. The Walker article in 1997 suggested that events begin to fade again around 3 months, so the 7-day window is fairly inconsequential. Some researchers use a 2-to-6-month window and I think a 2-month window is pushing it.

Reviewer 2 said “In Experiment 2, some formal weaknesses also appear, such as inconsistencies in the sample description (e.g., repeated percentage errors), suggesting the need for careful editorial revision. Furthermore, the characterization of an MTurk sample aged 18 to 23 would warrant additional explanation, as this age range is not the most typical for that platform.” We corrected the repeated percentage error. We asked MTurk for 18 to 23 year olds and their reported ages fit in that age range. We learned a great deal in the pandemic with MTurk and people using robots or being farmers to try to get money. We revamped our process in that we asked the eggplant question and we told people in the directions that we would contact them if we saw events that looked like a machine made or copied them and research assistants contacted many individuals. Can I guarantee that every age is correct? Could participants have lied? Yes. Participants in any study could lie but I cannot let that thought keep me up at night.

Reviewer 2 said “The results section is extensive, technically detailed, and statistically sophisticated, but it presents problems of organization, interpretation, and methodological grounding.” We thank Reviewer 2 for their kind comment.

Reviewer 2 said “A first critical aspect concerns structure. The presentation alternates between ANOVAs, PROCESS models (Models 1, 3, and 11), Johnson–Neyman analyses, and multiple two- and three-way interactions. Although this demonstrates analytical complexity, the text becomes excessively technical and insufficiently oriented toward the central hypotheses. At several points, the reader loses track of which hypothesis is being tested. A concrete improvement would be to organize the results according to previously numbered hypotheses (H1, H2, H3…), first summarizing whether each was supported, and then presenting the statistical details.” I have never numbered hypotheses in a paper. I think it could make sense to say something in the heading about which hypotheses are being tested. The ANOVAs examining fading affect across initial event affect H1, all the continuous 2-way interactions and the fading affect across initial event affect and event type test H2, the three-way interactions test H3, and the mediation analyses test H4. Would that change help?

Reviewer 2 said “Second, there is a strong reliance on statistical significance with limited discussion of effect magnitude. Many reported ΔR² values are extremely small (e.g., .003, .004, .007). Although statistically significant, these effects explain less than 1% of additional variance. Nevertheless, the text frequently treats them as substantively meaningful findings. It would be important to critically appraise their practical relevance, for example by explicitly stating that “although statistically significant, the effect size was small and may have limited practical implications.”” We understand and we addressed this issue in the General Discussion.

Reviewer 2 said “Another issue concerns multiplicity of testing. The large number of moderation and interaction analyses substantially increases the risk of Type I error. No correction for multiple comparisons (e.g., Bonferroni or FDR) is mentioned. This weakens confidence in marginal results or those with p values close to .05. A methodological improvement would be to acknowledge this limitation in the results section or, ideally, to reduce the number of exploratory tests.” We addressed this issue in the Analytic Strategy and the Results for the continuous 2-way interactions and 3-way interactions.

Reviewer 2 said “Regarding the three-way analyses (Initial Event Affect × Event Type × Marijuana Consumption), the authors interpret certain graphical patterns as indicative of “specific healthy coping,” even when statistical effects are not robust or slopes are only “nearly significant.” In quantitative science, interpretation should be based on statistical results, not merely on visual patterns. Expressions such as “the pattern showed specific healthy coping, but statistics did not demonstrate it” reveal a tension between theoretical expectations and empirical evidence. It would be more rigorous to limit interpretation to what was statistically confirmed.” We understand and agree and we removed all reference to non-significant patterns.

Reviewer 2 said “Moreover, the decision to treat the event as the unit of analysis without an appropriate multilevel model continues to affect the validity of inferences in this section. Because events are nested within participants, independence of observations may be compromised, potentially inflating significance levels. This limitation should be explicitly acknowledged.” We explicitly acknowledged this limitation in the General Discussion.

Reviewer 2 said “With respect to mediation analyses using PROCESS Model 11, the authors repeatedly test whether rehearsal mediates complex interactions, but in most cases no significant mediation is found. Despite this, considerable technical detail is devoted to describing the model. A more concise presentation could simply state: “No evidence was found that rehearsal mediated the three-way interactions.” Excessive technical detail hinders readability without enhancing conceptual clarity.” We followed the direction for Experiment 1 but we put the description in Experiment 2 because we had partial mediation.

Reviewer 2 said “Finally, some results are described using potentially causal language (“marijuana consumption predicted the FAB”), which is not appropriate for a cross-sectional correlational study. The term “predicted” should be replaced with “was associated with” or “was statistically related to.”” As we said previously, we apologize for that language. We tried to remove it all but let us know if and where we failed and we will change it.

Reviewer 2 said “The discussion begins by reaffirming the robustness of the FAB, which is consistent with the findings. However, it quickly moves to broad interpretations about healthy coping and emotional regulation, sometimes extrapolating beyond the data.

The methodological issue of data dependency (events nested within participants) is not sufficiently problematized in the discussion. Given its relevance, it should be explicitly acknowledged.

The conclusion summarizes the main findings but tends to emphasize adaptive implications of marijuana consumption more strongly than the data warrant.

Statements suggesting that consumption may “enhance healthy coping” or that certain patterns “demonstrated adaptive processing” are too strong for a correlational study with small effects. The conclusion should adopt more cautious language, for example: “The findings suggest complex associations between marijuana-related experiences and emotional memory processes.”” We changed the Discussion a great deal and we strongly changed the Summary and conclusion as well. We tried to make sure we were being informative and providing factual information without going beyond the data.

Reviewer 2 said “In addition, the conclusion could place greater emphasis on the specific contribution of the study: the inclusion of marijuana-related events within the FAB paradigm. This is the primary methodological and conceptual innovation and should be highlighted as such.” We did our best to listen and provide our contributions and nothing else.

Reviewer 2 said “It would also be desirable to include clear recommendations for future research, such as:
- Use of multilevel models;
- Longitudinal studies to assess actual changes in the FAB over time;
- Use of standardized measures of cannabis consumption;
- Differentiation between recreational and medicinal use.” We provided all of this information in the Discussion.

Reviewer 2 “Finally, the conclusion could explicitly acknowledge that the observed effects were generally small and that clinical or practical implications should be interpreted with caution.” We emphasized the small effects in the General Discussion and we did not mention clinical or practical implications. If Reviewer 2 thinks we need to add additional statements, we are glad to comply.

Reviewer 2 said “The article shows a similarity index of approximately 40%, which raises serious concerns about potential plagiarism. This issue should be carefully reviewed and addressed.” The first author does not know how to use Chat.GBT, so AI was not used for this paper. The first author did write 80 to 90% of the first draft and has a particular writing style that could have copied the writing style and phrasing in previous papers written by the first author, but that similarly was not done on purpose. I hope we reduced the similarity greatly. If I may, how did you check that match? Can anyone do that or do editors only have access to that technology?

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript presents several relevant limitations that should be considered and discussed before any potential publication. One of the most evident issues concerns the excessive reliance on the work of the main author. Throughout the article, a high number of self-citations associated with the author Gibbons can be observed, totaling approximately 13 references. Among these are studies such as Gibbons et al. (2011), Gibbons et al. (2013), Gibbons et al. (2015), Gibbons et al. (2016), Gibbons et al. (2017), as well as more recent works such as Gibbons & Lee (2019), Gibbons & Bouldin (2019), Gibbons et al. (2021), Gibbons et al. (2022), and Gibbons et al. (2024). These references are frequently used to support central aspects of the theoretical framework concerning the phenomenon of the fading affect bias (FAB). While it is natural for authors to cite previous work from their own line of research, in this case the frequency and positioning of these self-citations may convey the impression that the relevant literature is dominated almost exclusively by the same research group. Such a situation may limit the perception of theoretical and empirical diversity in the field and increase the risk of confirmation bias. It would therefore be advisable for the authors to incorporate a broader set of independent studies on autobiographical memory, emotional regulation, and psychological effects associated with cannabis use, in order to present a more balanced theoretical framework that better reflects the current state of the literature.

Another point that deserves attention concerns the conceptual fragility of the central hypothesis. The article starts from the premise that the fading affect bias can be interpreted as an indicator of adaptive emotional regulation or psychological coping processes. It is then suggested that marijuana use may be associated with positive or negative emotional experiences and that, for this reason, it would be relevant to investigate whether FAB also occurs in events related to the consumption of this substance. However, the logical connection linking these two elements is not entirely convincing. The manuscript does not present a clear theoretical model explaining why cannabis use should influence the differential fading of emotions associated with autobiographical memories. It would therefore be important for the authors to further develop this conceptual link, for example by discussing specific cognitive or affective mechanisms that might explain such an association.

From a methodological perspective, the study is essentially based on correlational data obtained through self-report measures. Participants are asked to recall events related to marijuana use and to evaluate both the original emotional intensity and the current emotional intensity associated with those events. This type of procedure allows the exploration of associations between variables but does not allow causal relationships to be established. Nevertheless, in some parts of the manuscript the discussion of the results appears to suggest causal interpretations, such as the idea that marijuana consumption could directly influence the fading affect bias or certain emotional regulation processes. It would therefore be important for the authors to clarify that the study design does not allow causal inferences and that the conclusions should be interpreted only in terms of associations.

Another aspect that may limit the robustness of the conclusions concerns the way some variables were measured. Certain indicators appear to rely on single items or rather general estimates provided by participants, such as the subjective level of “highness,” the approximate number of hours associated with consumption, or the initial emotional intensity of the recalled event. Measures based on very few items tend to present lower psychometric reliability and greater susceptibility to measurement error. One possible improvement would be to use multi-item scales or previously validated instruments designed to assess subjective experiences related to cannabis consumption.

Furthermore, the design of the recall task itself raises some concerns. Participants are asked to recall past events, evaluate the emotions they experienced at the time, and compare them with the emotions currently associated with those memories. This process relies heavily on retrospective memory and on participants’ ability to accurately estimate past emotional states. However, it is well known that autobiographical memories are often reconstructed and may be influenced by current beliefs, attitudes, or personal narratives. For this reason, it would be important for the authors to discuss more thoroughly the possible distortions associated with this type of data collection.

Another relevant methodological issue concerns the unit of analysis used in the statistical analyses. Although the study was conducted with several hundred participants, the analyses are based on a much larger number of observations because each participant reports multiple events. For example, a sample of approximately 328 participants generates more than 1,800 analyzed events. When individual events are treated as independent observations, there is a risk of artificially inflating statistical power and introducing problems of pseudo-replication, since several events originate from the same participant. Even if statistical techniques are applied to control for this dependency, this analytical strategy remains methodologically debated and should therefore be more clearly justified in the manuscript.

The statistical strategy adopted in the study also appears relatively complex when compared with the research questions being addressed. The article employs several different analyses, including ANOVA, multiple moderation models using the PROCESS macro (for example, Models 1, 3, and 11), analyses of regions of significance using the Johnson–Neyman technique, and multiple Bonferroni corrections. While these tools may be appropriate in certain contexts, the accumulation of numerous analyses increases the risk of statistically significant findings that may reflect very small or unstable effects. In fact, some of the reported effects present very small ΔR² values, around 0.004, which indicates that the proportion of explained variance is quite limited. In such cases, it would be important to discuss more clearly the practical or theoretical relevance of these effects.

Related to this issue, the interpretation of the results sometimes appears stronger than what the data actually support. In some parts of the manuscript it is suggested that the fading affect bias represents an indicator of healthy psychological coping and that marijuana consumption may be associated with this process. However, the results reported are relatively modest and, in some cases, not entirely consistent across the two studies conducted. A more balanced discussion should acknowledge these inconsistencies and avoid overly broad generalizations.

In summary, although the article addresses a potentially interesting question and presents a considerable amount of empirical data, several aspects could be substantially improved, particularly the diversity of the literature review, the theoretical grounding of the hypotheses, the discussion of methodological limitations, the clarity of the statistical analyses, and the quality of the scientific writing. A thorough revision of these aspects would contribute to making the manuscript more robust and convincing.

Author Response

We address every comment made by Reviewer 2 below and in the paper. We would like to thank Reviewer 2 for their guidance and improving our paper.

Reviewer 2 said “The manuscript presents several relevant limitations that should be considered and discussed before any potential publication. One of the most evident issues concerns the excessive reliance on the work of the main author. Throughout the article, a high number of self-citations associated with the author Gibbons can be observed, totaling approximately 13 references. Among these are studies such as Gibbons et al. (2011), Gibbons et al. (2013), Gibbons et al. (2015), Gibbons et al. (2016), Gibbons et al. (2017), as well as more recent works such as Gibbons & Lee (2019), Gibbons & Bouldin (2019), Gibbons et al. (2021), Gibbons et al. (2022), and Gibbons et al. (2024). These references are frequently used to support central aspects of the theoretical framework concerning the phenomenon of the fading affect bias (FAB). While it is natural for authors to cite previous work from their own line of research, in this case the frequency and positioning of these self-citations may convey the impression that the relevant literature is dominated almost exclusively by the same research group. Such a situation may limit the perception of theoretical and empirical diversity in the field and increase the risk of confirmation bias. It would therefore be advisable for the authors to incorporate a broader set of independent studies on autobiographical memory, emotional regulation, and psychological effects associated with cannabis use, in order to present a more balanced theoretical framework that better reflects the current state of the literature.” Reviewer 2’s point was well made and taken. We reduced the number of citations for the first author in the paper. It would be extremely difficult to completely eradicate the first author citations, but we are open to specific suggestions for removal.

Reviewer 2 said “Another point that deserves attention concerns the conceptual fragility of the central hypothesis. The article starts from the premise that the fading affect bias can be interpreted as an indicator of adaptive emotional regulation or psychological coping processes.” Many FAB papers suggest that the FAB is a healthy coping mechanism and the first author has resisted using that phrasing in many papers until recently. We took Reviewer 2’s point to heart and we talked about FAB as a healthy coping outcome throughout the paper, not a healthy coping mechanism.

Reviewer 2 said “It is then suggested that marijuana use may be associated with positive or negative emotional experiences and that, for this reason, it would be relevant to investigate whether FAB also occurs in events related to the consumption of this substance. However, the logical connection linking these two elements is not entirely convincing. The manuscript does not present a clear theoretical model explaining why cannabis use should influence the differential fading of emotions associated with autobiographical memories. It would therefore be important for the authors to further develop this conceptual link, for example by discussing specific cognitive or affective mechanisms that might explain such an association.” We thank Reviewer 2 for pointing out this oversight. We corrected for it in the introduction. As the mobilization-minimization hypothesis suggested that biological, cognitive, and emotional resources are mobilized to minimize the harmful effects of unpleasant events and the literature on marijuana showed that marijuana consumption inhibited biological processes, cognitions, and mood, we combined the two points to explain the reason marijuana consumption reduces the FAB, as demonstrated by Pillesdorf and Scoboria (2019).

Reviewer 2 said “From a methodological perspective, the study is essentially based on correlational data obtained through self-report measures. Participants are asked to recall events related to marijuana use and to evaluate both the original emotional intensity and the current emotional intensity associated with those events. This type of procedure allows the exploration of associations between variables but does not allow causal relationships to be established. Nevertheless, in some parts of the manuscript the discussion of the results appears to suggest causal interpretations, such as the idea that marijuana consumption could directly influence the fading affect bias or certain emotional regulation processes. It would therefore be important for the authors to clarify that the study design does not allow causal inferences and that the conclusions should be interpreted only in terms of associations.” We thank Reviewer 2 for making this point again, as they did in the last review, but we addressed this issue in our last revision, and we double-checked the paper to make sure we did not make causal interpretations.

Reviewer 2 said “Another aspect that may limit the robustness of the conclusions concerns the way some variables were measured. Certain indicators appear to rely on single items or rather general estimates provided by participants, such as the subjective level of “highness,” the approximate number of hours associated with consumption, or the initial emotional intensity of the recalled event. Measures based on very few items tend to present lower psychometric reliability and greater susceptibility to measurement error. One possible improvement would be to use multi-item scales or previously validated instruments designed to assess subjective experiences related to cannabis consumption.” We thank Reviewer 2 for making this point again as they did in the previous review, but we addressed this point in the previous review and we checked the paper again to make sure we address it sufficiently.

Reviewer 2 said “Furthermore, the design of the recall task itself raises some concerns. Participants are asked to recall past events, evaluate the emotions they experienced at the time, and compare them with the emotions currently associated with those memories. This process relies heavily on retrospective memory and on participants’ ability to accurately estimate past emotional states. However, it is well known that autobiographical memories are often reconstructed and may be influenced by current beliefs, attitudes, or personal narratives. For this reason, it would be important for the authors to discuss more thoroughly the possible distortions associated with this type of data collection.” We thank Reviewer 2 for making this important point. Every retrospective study should address this limitation, and we had not done so previously. Therefore, we made the point that retrospective procedures bring in inherent reconstructive memory processes. We also cited Ritchie et al.’s (2009) article, which found in a combined diary and retrospective memory procedure that participants in their accurately estimated initial intensity of pleasant events and underestimated the intensity of unpleasant events.

Reviewer 2 said “Another relevant methodological issue concerns the unit of analysis used in the statistical analyses. Although the study was conducted with several hundred participants, the analyses are based on a much larger number of observations because each participant reports multiple events. For example, a sample of approximately 328 participants generates more than 1,800 analyzed events. When individual events are treated as independent observations, there is a risk of artificially inflating statistical power and introducing problems of pseudo-replication, since several events originate from the same participant. Even if statistical techniques are applied to control for this dependency, this analytical strategy remains methodologically debated and should therefore be more clearly justified in the manuscript.” We thank Reviewer 2 for making this point again. We addressed this point in the previous version of the paper, and we added to this discussion in the current General Discussion.

Reviewer 2 said “The statistical strategy adopted in the study also appears relatively complex when compared with the research questions being addressed. The article employs several different analyses, including ANOVA, multiple moderation models using the PROCESS macro (for example, Models 1, 3, and 11), analyses of regions of significance using the Johnson–Neyman technique, and multiple Bonferroni corrections. While these tools may be appropriate in certain contexts, the accumulation of numerous analyses increases the risk of statistically significant findings that may reflect very small or unstable effects.” We addressed this issue in the Analytic Strategy of Experiment 1 and the General Discussion. Timothy Ritchie is a fairly savvy analyst as he makes his living that way, and he is the one who created this analytic strategy as the main analyst on the alcohol study in 2013 by Gibbons et al. Although Tim did not use Bonferroni corrections in that study, we have used them in previous studies, and we were glad to add them after Reviewer 2 asked us to address Type I error in the current study. As Reviewer 2 knows, Bonferroni corrections are very conservative, and they guided us to delete one 2-way interaction and several 3-way interactions in Experiment 2. We made this change according to both Reviewers’ suggestions in the previous review.

Reviewer 2 said “In fact, some of the reported effects present very small ΔR² values, around 0.004, which indicates that the proportion of explained variance is quite limited. In such cases, it would be important to discuss more clearly the practical or theoretical relevance of these effects.” We thank Reviewer 2 for making this point and we discuss this issue in the General Discussion.

Reviewer 2 said “Related to this issue, the interpretation of the results sometimes appears stronger than what the data actually support. In some parts of the manuscript it is suggested that the fading affect bias represents an indicator of healthy psychological coping and that marijuana consumption may be associated with this process. However, the results reported are relatively modest and, in some cases, not entirely consistent across the two studies conducted. A more balanced discussion should acknowledge these inconsistencies and avoid overly broad generalizations.” After the initial review, we softened the language about coping. As stated previously, we changed all language about the FAB being a healthy coping mechanism to “healthy coping outcome”. The FAB is related to marijuana consumption as shown by Pillesdorf and Scoboria (2019) and in the current study. However, we find very different relations between marijuana consumption/reactivity and the FAB than the seminal study and it is important to tell others that this relation seems to have changed. We are open to suggestions for softening other language if Reviewer 2 can provide that guidance for specific parts of the manuscript; we are merely asking for the problem and the place in the manuscript, not the wording. We have already taken enough of Reviewer 2’s time and attention.

Reviewer 2 said “In summary, although the article addresses a potentially interesting question and presents a considerable amount of empirical data, several aspects could be substantially improved, particularly the diversity of the literature review, the theoretical grounding of the hypotheses, the discussion of methodological limitations, the clarity of the statistical analyses, and the quality of the scientific writing. A thorough revision of these aspects would contribute to making the manuscript more robust and convincing.” We addressed each of the issues raised by Reviewer 2 in the paper and this response letter, and they improved the quality of the paper. We feel that the current version of the paper is much improved and ready for publication, and we hope Reviewer 2 and the editors agree with us.

Round 3

Reviewer 2 Report

Comments and Suggestions for Authors

One of the most critical aspects is the quality of the writing. The text contains numerous grammatical errors, editing lapses, and inconsistencies that hinder readability and undermine its scientific credibility. For example, constructions such as “reported found that,” word duplications (“itthe researchers”), and errors in author citations (“Scorboria” vs. “Scoboria”) indicate a lack of careful proofreading. These issues are not merely formal: they interfere with the comprehension of the argument and may lead reviewers to question the overall rigor of the work. A professional language revision is strongly recommended, preferably by a native speaker or a specialized academic English editing service.

At the structural level, the article tends to be redundant and at times lacks focus. The introduction, although rich in references, repeats ideas particularly in the description of the FAB and its implications and would benefit from greater conciseness. Rather than extensively listing studies with positive and negative findings on marijuana consumption, it would be more effective to organize the review around clear theoretical tensions (e.g., “adaptive vs. maladaptive effects”) and to more explicitly show how these tensions lead to the study’s hypotheses. In addition, some sections mix the presentation of results with interpretation, which undermines the conventional separation between results and discussion. The manuscript should be reorganized to ensure greater linearity: introduction (with clear hypotheses), method, results (objective and without excessive interpretation), and discussion (critical interpretation).

From a methodological perspective, although the sample sizes are adequate, there are weaknesses that require further justification. The use of events as the unit of analysis, rather than participants, raises concerns about the independence of the data. Although the authors mention controlling for “cluster,” the explanation is unclear and insufficient to reassure the reader about the validity of the analyses. It would be important to more clearly specify the statistical model used (e.g., multilevel models) and to justify the decision to treat events as independent units. Furthermore, the measures of marijuana consumption, such as “degree of highness” on a 0–100 scale, are highly subjective and lack validation. The article should more explicitly acknowledge these limitations and discuss their potential impact on the findings.

Another problematic aspect is the complexity and volume of statistical analyses. The use of multiple models (ANOVA, Process Model 1, 3, and 11) with several two- and three-way interactions makes the paper difficult to follow and raises the possibility of excessive data exploration (a “fishing expedition”). Although the authors apply corrections such as Bonferroni, the presentation of results remains dense and not very intuitive. One possible improvement would be to focus on analyses directly related to the main hypotheses, relegating exploratory analyses to supplementary material. Additionally, it would be helpful to include a conceptual framework or diagram to guide the reader in understanding the relationships being tested.

Regarding the interpretation of results, the article at times goes beyond what the data can support. For instance, when suggesting that marijuana consumption may be associated with adaptive outcomes in certain contexts, the authors rely on correlational relationships that do not allow causal inferences. A more cautious formulation would acknowledge that the findings indicate specific associations depending on event type, without implying direct beneficial effects of the substance. The discussion would benefit from a more critical and balanced tone, better integrating methodological limitations and avoiding overgeneralizations.

The article contains an excessive number of self-citations. For example, the author Gibbons includes 11 self-citations.

Author Response

1.2 Marijuana Literature Shows Bifurcated Findings (pages 2-3)

The literature surrounding marijuana contains many contradicting findings pertaining to the adaptivity of marijuana consumption. For instance, in a 4-week diary study, Gruber et at. (2012) found that marijuana use lowered daily mood ratings across time in healthy participants. Similarly, Lex et al. (1989) asked a sample of 30 participants to complete daily diary questionnaires that monitored their use of marijuana and other drugs and found that heavy marijuana users reported higher negative moods (e.g., confusion, fatigue, and anger) and lower positive moods (e.g., friendliness, elation, and vigor) than light marijuana users. In a cross-sectional study, Hines et al. (2020) found that depression, generalized anxiety disorder and psychotic experiences were positively related to high-potency marijuana consumption in a sample of 1087 participants. Petrucci (2020) examined a sample of 1,168 men and women and reported that cannabis use was positively related to apathy. Similarly, Looby and Earleywine (2007) found that a large sample (N = 2500) of cannabis users meeting DSM-IV-TR criteria for dependence reported low levels of motivation, and Lac and Luk (2017) reported that consumption was associated with low initiative and persistence that worsened over time in a sample of 505 college students.

While those studies found marijuana consumption to be positively associated with unhealthy outcomes, providing evidence of marijuana consumption being a maladaptive outcome, other studies have found marijuana to be an adaptive outcome. For example, Roitman et al. (2014) found that the main psychoactive element in marijuana, THC, may be effective in reducing nightmares and improving sleep quality in patients with PTSD. Sznitman et al. (2022) conducted a 14-day diary study and found that intended cannabis use and recent cannabis use were related to enhanced positive affect and decreased negative affect. In addition, participants with bipolar disorder, PTSD, and depression who consumed cannabis reported improved moods. Gruber et al. (2012) found similar results in a 4-week diary study where bipolar participants who consumed marijuana reported improved mood symptoms when compared to the bipolar participants who did not consume marijuana. Lake et al. (2019) uncovered similar effects in the context of PTSD in a sample of 225,113 participants where PTSD was associated with the likelihood of experiencing a major depressive episode and suicidal ideation in non-cannabis users, but the relation was not found in cannabis users. Similarly, Li et al. (2020) examined the effectiveness of cannabis to immediately relieve symptoms of depression in a sample of 1,819 participants who self-administered cannabis.

The existing literature suggests that marijuana consumption is associated with positive outcomes, but many studies reveal that cannabis use is associated with unwanted, unhealthy, and maladaptive mental and physical outcomes. In contrast, marijuana consumption seems to aid individuals suffering from mental health challenges, such as PTSD and bipolar disorder. Nevertheless, several limitations within the cited research should be noted. Much of the current research relies heavily on self-reported measures of marijuana use, which may produce bias in recall as well as other measurement areas. In addition, many of the studies employed cross-sectional designs, which limit the ability to determine the exact causal relation between marijuana use and psychological/health outcomes. Differences in dosage, potency, frequency of use and the context on consumption (medical versus recreational, flower versus vaping) may also contribute to the mixed findings observed across studies. Taken together, these considerations suggest that the effects of marijuana are complex and may depend on contextual factors and individual differences.

REVIEWER COMMENTS AND RESPONSES:

Reviewer 2 said “One of the most critical aspects is the quality of the writing. The text contains numerous grammatical errors, editing lapses, and inconsistencies that hinder readability and undermine its scientific credibility. For example, constructions such as “reported found that,” word duplications (“itthe researchers”), and errors in author citations (“Scorboria” vs. “Scoboria”) indicate a lack of careful proofreading. These issues are not merely formal: they interfere with the comprehension of the argument and may lead reviewers to question the overall rigor of the work. A professional language revision is strongly recommended, preferably by a native speaker or a specialized academic English editing service.”

We are very thankful for the reviewer’s comments that highlighted some of the egregious errors we had accidentally made in the writing of our manuscript. Such spelling and grammatical errors have been meticulously corrected (pages 2 and 4), especially regarding the spelling of the names of Pillersdorf and Scoboria, a mistake which we regret deeply (pages 1, 2, 12, 13, 22, and 24).

Reviewer 2 said “At the structural level, the article tends to be redundant and at times lacks focus. The introduction, although rich in references, repeats ideas particularly in the description of the FAB and its implications and would benefit from greater conciseness. Rather than extensively listing studies with positive and negative findings on marijuana consumption, it would be more effective to organize the review around clear theoretical tensions (e.g., “adaptive vs. maladaptive effects”) and to more explicitly show how these tensions lead to the study’s hypotheses.”

We thank the reviewer for this insight, and we have modified the introduction accordingly. Importantly, many citations were removed (page 2). In addition, the section related to marijuana research and theory was truncated; it became one section with three paragraphs. This section directly addressed the adaptive and maladaptive outcomes found in marijuana consumption literature. We concluded the section by mentioning the limitations of the research, as Reviewer 2 suggested, and we emphasized that the effects of marijuana are complex and may depend on contextual factors and individual differences (pages 2-3).

Reviewer 2 said “In addition, some sections mix the presentation of results with interpretation, which undermines the conventional separation between results and discussion. The manuscript should be reorganized to ensure greater linearity: introduction (with clear hypotheses), method, results (objective and without excessive interpretation), and discussion (critical interpretation).”

While we thank reviewer 2 for their suggestion, we believe that the current layout of the manuscript provides the most truthful and accurate representation of the study as it was conducted. Formatting the manuscript as the intro, then experiment one, then experiment two, with a concluding general discussion supports transparency in the two experiments being conducted via separate modalities and ensures the reader will not become confused as to which experiment served as the origin for any given finding, as the results were different across the experiments. However, we have processed the comments made about objectivity of the results section and we have attempted to ensure bias was removed from the results sections of each experiment (pages 1-25).

Reviewer 2 said “From a methodological perspective, although the sample sizes are adequate, there are weaknesses that require further justification. The use of events as the unit of analysis, rather than participants, raises concerns about the independence of the data. Although the authors mention controlling for “cluster,” the explanation is unclear and insufficient to reassure the reader about the validity of the analyses. It would be important to more clearly specify the statistical model used (e.g., multilevel models) and to justify the decision to treat events as independent units.”

We thank the reviewer for continuing to press on this matter. We had initially included several statements regarding the reason the previous analytic approach was utilized and explaining potential shortcomings of the methodology. Now, we have recompleted the analyses using the newest version of the process macro, utilizing its cluster-robustSE feature, which allows for the accurate analysis of clustered data such as those in the current study. The new analytic approach changed a significant number of the findings, but allows for a less liberal correction measure than a pure Bonferroni correction, as we are analyzing exclusively the relations we initially set to analyze a priori (pages 7-13 and pages 14-19).

Reviewer 2 said “Furthermore, the measures of marijuana consumption, such as “degree of highness” on a 0–100 scale, are highly subjective and lack validation. The article should more explicitly acknowledge these limitations and discuss their potential impact on the findings.”

We thank the reviewer for reiterating the importance of these methodological concerns. We have added a section to the description of the measure in the methods section acknowledging the potential limitations of the measure and stating plainly that it is not a validated measure. We hope that these additional statements will meet the reviewer’s expectations for precautioning readers to the potential limitations of the related findings (page 5).

Reviewer 2 said “Another problematic aspect is the complexity and volume of statistical analyses. The use of multiple models (ANOVA, Process Model 1, 3, and 11) with several two- and three-way interactions makes the paper difficult to follow and raises the possibility of excessive data exploration (a “fishing expedition”). Although the authors apply corrections such as Bonferroni, the presentation of results remains dense and not very intuitive. One possible improvement would be to focus on analyses directly related to the main hypotheses, relegating exploratory analyses to supplementary material. Additionally, it would be helpful to include a conceptual framework or diagram to guide the reader in understanding the relationships being tested.”

We appreciate the reviewer’s comments on this matter and have come to agree as to the density of the findings. As such, we have removed the mention of mediation analyses entirely from the paper, as they did not contribute to the central hypotheses of the manuscript. Further, we have removed the Neyman-Johnson analyses from our results, as they were especially detrimental to the readability of the section without providing significantly to our findings (pages 7-13 and pages 14-19).

Reviewer 2 said “Regarding the interpretation of results, the article at times goes beyond what the data can support. For instance, when suggesting that marijuana consumption may be associated with adaptive outcomes in certain contexts, the authors rely on correlational relationships that do not allow causal inferences. A more cautious formulation would acknowledge that the findings indicate specific associations depending on event type, without implying direct beneficial effects of the substance. The discussion would benefit from a more critical and balanced tone, better integrating methodological limitations and avoiding overgeneralizations.”

We thank the reviewer for their continued advocation for the removal of causal statements from the manuscript. We have added qualifiers and limitations to references to the correlation between consumption measures and the FAB, such as “may be considered” and “positively associated with” to ensure that the statements cannot be mistaken for claims of causality (pages 20, 21, and 22).

Reviewer 2 said “The article contains an excessive number of self-citations. For example, the author Gibbons includes 11 self-citations.”

We once again thank the reviewer for this comment. While we fundamentally believe that the inclusion of those citations provided a strong foundation for the current research, we understand that the optics of such a high number of self-citations may cause distrust in our methods and findings. We are sincerely appreciative of the reviewer’s foresight in saving us potential trouble in this capacity. In accordance with the suggestion, we have removed 6 first author citations from Gibbons in the paper. The remaining citations were believed to be too important to remove and not theoretically pertinent enough to the primary hypothesis as to suggest undue bias in our methodologies, rather, they served to display general facts about the FAB or to support our choices for analytical approaches, which were also then further modified to meet the requests of the reviewer stated above (pages 2, 7, and 12).

The new analyses did not provide numbers for the positive continuous predictors that would have been used to create Figure 3. Therefore, we removed Figures 3 and 4 from the paper. The paper now includes eight figures instead of 10 figures.