The Impact of Expert Witness Knowledge on Perceived Credibility: Implications for Not Guilty by Reason of Insanity (NGRI) Endorsement

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Behavioral Sciences Review: behavsci-3696538

The Impact of Expert Witness Knowledge on perceived Credibility: Implications for Not Guilty by Reason of Insanity (NGRI) Endorsement – Cain & Smith

 

Summary:

This study examines how different levels of expert witness knowledge affects how credible they seem by lay individuals in the context of a criminal case. Specifically, when an expert witness is called to provide their clinical option about whether the defendant is or is not suffering from a mental illness that absolves them of responsibility for the crime. They operationalized expert knowledge as Master’s vs. Ph.D., Board certification, number of years of experience, evaluations, and cases worked on. These credentials were paired with the same case information and testimony to see if participants’ views about the case and the “Not Guilty by Reason of Insanity” (NGRI) decision could be influenced by the expert’s level of knowledge alone. They did not find a significant effect of expert knowledge on whether participants endorsed the NGRI defense, though.

Comments:

This is a simple study on an important topic. I am always excited to see more work on juror perceptions of evidence and there is no reason to make studies complicated if the question can be answered with a simple study and design.

But I suspect that there are several reasons why they did not find any effects that are due to some features of the materials and design.  These are:

1. They used an unambiguous case example: Generally speaking, when attempting to manipulate decisions via a heuristic-type process, the decision materials shouldn’t be clear cut. So, here, if it was clear from the case materials and testimony that the person was NGRI or not, then the participants would not need to rely on any other information to come to a conclusion—it is clear what the answer is without relying on extraneous information. However, if it had been unclear whether the defense applied in this case, then they might seek to determine how credible the source of the information was—the expert providing their testimony—and they level of experience and knowledge might have had an impact then. They cite that ELM in their paper, which would support this idea. When it is clear what the correct outcome is, they will rely on the Central Route to persuasion and evaluate the message because they feel capable of doing so. If the case was more ambiguous, they would be uncertain and might shift to the Peripheral Route, the route that would result in factors like expert credibility affecting outcomes.
2. Strong pre-existing attitudes: Hudachek & Quigley-McBride (2022) found that initial judgments and pre-existing attitudes about NGRI were more powerful than new information in determining final decisions about NGRI. It took very clear information that contradicted their initial belief to shift even just some participants towards a different decision. In this case, they didn’t provide information that specifically challenged pre-existing beliefs though—just information consistent with relying on the expert’s testimony.
3. Both experts were experienced and qualified: Even if participants had relied upon the expert’s level of knowledge and experience, both experts were qualified and experienced – one was just more experienced than the other. The average person will not know how each expert in this study compared relative to other experts, so it may be that even the low knowledge expert reached the level of qualification required to be deemed credible in this context by a lay person and any additional qualifications really were meaningless without a clear comparison.

Thus, I think these issues with the materials and design meant that they were unlikely to find a significant effect to begin with. I am not sure if this means that they should run another study before publishing this in a journal, or if it is merely sufficient that they discuss these problems with the study in full in the discussion—I’ll leave that to the action editor to decide—but I don’t think these materials and this design were a proper test of their hypotheses as a result given that they were unlikely to find effects with them, in my view. Hopefully, the above comments are helpful to the authors though for any future work that they hope to do on this topic!

Some other more minor issues:

1. Introduction - A few things:
   1. I think the topics could be reordered to make the narrative more effective. The ELM could be introduced earlier, for instance, to ground this project in psychological theory. But, I suspect that this is more of an opinion than objective feedback.
   2. They also describe this project as filling a “gap” but there is actually a fair amount of work on perceptions of expert credibility, but somewhat less on how this affects NGRI judgments specifically. Hence, I think it would be better if there was a reason for this study that was a bit more than just “it hasn’t been done very much”. Why is it interesting to know how expert credibility affects NGRI judgments? Why should we care if jurors are relying on the expert’s credentials rather than what the expert actually says? I think that is a better justification for the study. But, again, this is more of an opinion than objective feedback.
   3. They cited Hudachek & Quigley-McBride (2022) as an authority for showing that attitudes measured by scales like the BMI affect how people make NGRI judgments, but actually they may only found an effect with the danger subscale, and it was limited. The better measure of pre-existing attitudes was the first judgment based on the case information only, prior to any expert judgments, rather than the scale they asked them to complete.
   4. There were a few other places where the information provided didn’t quite reflect what is contained in the cited work—double-check your citations and the associated information provided to readers.
2. They used one case vignette only: The did not stimulus sample so, even though they pilot tested this case study, they cannot generalize their findings to a range of cases and cannot determine whether the non-significant findings are just due to this particular case study/facts. If they had found significant results, then they would not have had much justification for generalizing those effects beyond that set of case facts, too. So, I would recommend expanding the set of materials used in future studies just to avoid these issues.

Author Response

Please see attached Word document for responses.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The authors looked at the effect of the level of experience of an expert witness in an mock jury NGRI court case on perceived credibility and endorsement of NGRI.  This is a very interesting topic in a field that is often overlooked (NGRI).   

I appreciate the authors’ efforts in developing this work. My comments are intended to be constructive and are offered with respect for the significance of the research. If any points come across as overly critical, please know this is unintentional and likely a result of time pressure rather than intent — certainly not an attempt to channel my inner Reviewer 2! Thank you again for the opportunity to review this manuscript.

I have some suggestions regarding clarity on certain topics and structure.

Clarity: 

The article is written solely from the perspective of the United States resulting in the fact that certain procedures are not clear for audiences outside of the US. This should be addressed in the introduction giving some more background. also make a sidestep to other countries and in your discussion you should make it clear that this is only relevant in USA context. 

eg: 1-"Research has shown that while expert witness’ credentials and expertise level are relevant in assessing expert witness credibility, they can distract from the evaluation of testimony strength" in other countries a jury is not used to determine NGRI, judges make that decision. how does this work? In practice: do jury members get a bio of the expert and determine credibility based on that? 

2- prolific academic: what is that? 

Especially because the NGRI system is different between countries and even more in de USA. the low number of NGRI cases per year reflect these differences very well. If the authors want to know more about the systems in European countries I recommend this article:  https://doi.org/10.1016/j.ijlp.2022.101775

3- incorrect statements:

a) Line 91 "Forensic psychologists, or expert witnesses, are trained to be unbiased in their evaluations. However, lay people do not receive the same training; consequently, they may be more susceptible to biases." forensic psychologists are not without bias, the opposite is true and they even have a bias blind spot which is a big problem see: https://doi.org/10.1016/j.ijlp.2025.102083

b) line 530. Did you put the IDA-R subscales in the same model as the total scores of the IDA-R as is implied by the table?  This is statistically not a good choice as you add a highly correlated predictor into the model (the subscales are part of the sumscore), You need to redo the analyses with only the sumscore and only the subscales  

Line 579-580: "Furthermore, the expert witness testimony in the current study was approximately 45 seconds, which may not have allowed adequate time for participants to fully evaluate its content." did they only get 45 seconds to read it? why? this is not clear

Line 609 "In the current study, if the participants in the high knowledge condition identified the expert as a hired gun, even subconsciously, it would make sense that they would be more punitive in their verdict decisions by not endorsing NGRI. This bias could have been formed because of the expert’s extensive courtroom experience in the high knowledge condition" this is quite a stretch, there is no reason to believe that the more experienced has a high chance of being regarded as a hired gun

line 629-631 "Research has shown that, in general, those with strong opinions towards a particular sentencing outcome (Skeem et al., 2004) are less likely to be swayed by an expert witness testimony compared to those who may be on the fence or unsure of their position/stance." this is not the same as positive opinions. I do not agree with this section that it aligns with the ELM.

in addition: 

Line 688-690: 'These results provide support for the Elaboration Likelihood Model (ELM), which suggests that individuals with strong preexisting attitudes are less likely to be influenced by external peripheral cues (e.g., credentials) and are more likely to rely on central cues (e.g., personal beliefs) when making decisions." I do not agree that you can say this based on the information that you provided in the article. IDA-r does not measure the how strongly they believe in something, just if they agree with the statements negative, neutral or positive opinion. you can have a negative opinion about something but about a topic that you do not feel strongly about. You did not provide mean scores on the IDA-r, so it can be possible that you group were mainly positive about the insanity defense or neutral. Also, there is a separate question in the IDA-r that asks: How strongly do you feel about the insanity defense? did you use this in your analyses? Reconsider you statements in the discussion about the ELM.

4) data missing or incorrectly used to support made statements

Please add mean scores per group for you outcome measures (IDA-R scores etc). .

Line 612 "Regarding attitudes toward the insanity defense, results demonstrated that individuals who had more positive attitudes towards NGRI were more likely to endorse it." What were the results on the IDA-r, I would have liked to have seen information mean min and max etc for the two groups. 

Line 621 "Regarding attitudes toward the insanity defense, results demonstrated that individuals who had more positive attitudes towards NGRI were more likely to endorse it, while the opposite was also true." were can I find the data related to this statement? 

Line 632 "The demographics of the current sample seemed to match with what previous  research has found for individuals who are more likely to endorse an NGRI verdict" where is that data?

 

line 507: you find statistical differences, but these differences are really small. You need to address this: are these differences meaningful if it translates to an increase of 1 point on a scale. It is also not clear what range the WCS has (also for line 518), making it difficult for the reading to interpret these differences. In other words, in my opinion, this is not a 'clinical' relevant difference and should be discussed with this in mind. 

Structure

There is a lot of repetition in the article. Do not repeat procedures if the main study was the same as the pilot study. Just mention the differences between the two and why there are differences. This makes it much easier for the reader to understand the procedure and where the differences are. (eg line 425 -426)

Even within the explanation of the pilot study there are repetitions: eg starting from line 338.

On the other hand, information is given for the main study that was not given for the pilot study (line 417-420), I would have liked to have had that information in the explanation of the pilot study. So make the explanation as complete as possible while reporting on the pilot study, when discussion the main study say that the procedure was the same with the exception of... 

line 538. "The converse was also true, that more supportive  attitudes toward the insanity defense was associated with an increased likelihood of endorsing. " you mention this several times. Of course the converse is also true that's what you understand from the direction of the effect. No need to repeat that again. 

Hypothesis 4 is discussed before 3. table name is incorrect? "Table 2. Hypothesis 3 – Statistics for Moderation Model." should be hypo 4? 

Minor comments:

 

Line 447-448: what is DV?  The purpose of using this measure  is to be informed about the extent that participants view the insanity defense as legitimate and how that informs their NGRI endorsement (DV).

Line 571: did Neal and colleagues also use the WCS?

Line 476: give the range of the likert scale so readers do not need to go to the appendix to know. "they were given a Likert scale for the likelihood of rendering an NGRI endorsement."

Line 480: After completion of the demographic portion of the survey, participants received a debriefing." why a debriefing and what was the debriefing?

 

Comments on the Quality of English Language

Line 446-447 use of tense is incorrect: Participants will be given a demographic survey to gather descriptive information about our sample.

 

line 523-525: ' To test Hypothesis 4, which predicted that participants in the high knowledge condition would be more likely to endorse NGRI compared to those in the low knowledge condition, but only if those participants have lower IDA-R scores (i.e., positive or neutral attitudes toward NGRI).' there is somethin missing in this sentence. 

Author Response

Please see the attached Word document for responses.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

I appreciate the opportunity to review this submission, which explored the impact of expert knowledge on perceived credibility and determinations regarding a defendant's NGRI plea. In contrast to prior studies, knowledge did not impact credibility, though it was associated with NGRI endorsement - but in opposite direction than anticipated (i.e., lower knowledge associated with NGRI decisions relative to more knowledgeable expert condition). In addition, existing attitudes toward NGRI defense(s) were associated with ultimate decisions (i.e., more favorable attitudes associated with higher endorsement of NGRI and vice versa). 

I appreciated the authors' thorough review of the literature, especially the details of prior studies on expert witness credibility and decision-making. In particular, it was helpful to clearly detail previous findings, along with the associated designs/methodologies, as this helped to contextualize the current study and associated findings (or lack thereof, in some instances). The writing was also very clear, and the paper is well organized and logical in flow/content.

Though a strong submission in terms of writing, clarity, organization, and content, the ultimate contribution to the literature is limited due to aspects of the design/methods. In particular, the testimony to which participants were exposed totaled less than a minute, and there was no exposure to an actual person delivering the testimony, which obviously limits the basis on which to judge the expert's credibility in terms of liking, trustworthiness, etc. I would also note that the case described in the stimulus materials was not particularly strong/persuasive, as there was little to no nexus between the symptoms of the defendant's mental illness and their appreciation of wrongfulness. I imagine this impacted whether participants were willing to entertain/consider an NGRI plea, which could've skewed the results (especially the base rate of NGRI decisions). 

I do have a few comments/suggestions that the authors may wish to consider, which are listed below (in no particular order): 

1. Unless I missed it, I could not find the base rate of NGRI decisions in the sample. This is a crucial piece of information for the reader to know, and it may help to better understand some of the surprising findings. In particular, it would be nice to know whether rate of NGRI endorsement differed from what we might expect (given knowledge of relative success about 25% of time in "real life"), as well as whether this could've contributed to unusual result of low knowledge resulting in greater NGRI findings. 
2. The authors opine that the hired gun effect may explain why the expert with lesser experience/knowledge was seemingly more persuasive (as evidenced by higher NGRI findings), though I would imagine that would have been detected on the WCS, if so. 
   1. It is an interesting idea, however, as it seems to suggest that greater experience and knowledge essentially backfired, resulting in reduced credibility (and therefore persuasiveness). A similar process, termed "the backfire effect" has been found with certain types of mitigating factors during capital sentencing (substance use, in particular), by which jurors seemingly convert mitigating information to aggravating factors. Perhaps this would be relevant to understanding why this might've occurred here too. 
3. I also wonder whether the authors considered general resistance to considerations of NGRI, as this tends to be the prevailing view/perspective of the public (and therefore, mock/real jurors). This would likely be evident in both IDA-R scores and rate of NGRI decisions, which could be explored further. 
4. The authors reference the requirement for death qualification in capital cases, whereby jurors must be willing to impose a death sentence to be eligible for death penalty proceedings, suggesting an analogous process may be worth considering for NGRI cases. This is an interesting idea, though the literature suggests death qualification is problematic, as there is evidence of inherent bias toward capital punishment (i.e., death qualified jurors are more prosecution-oriented, conviction-prone, and authoritarian, raising concerns about reduced fairness for capital defendants). It would be worthwhile for the authors to consider/discuss whether we might expect a similar (albeit opposite) bias if instituting an "NGRI qualification" criterion. 
5. It would be helpful to report the descriptive statistics on the measures included in this study, as it will give reader a sense of baseline attitudes/perceptions, especially toward insanity defense and expert credibility. 
6. In addition, the authors might consider exploring relationship between participant characteristics and variables of interest (e.g., gender, educational level, and political orientation on insanity attitudes - and plea decisions!). 

Author Response

Please see the attached Word document for responses.

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

I found this study to be highly engaging and thought-provoking. Overall, the manuscript is well-organized, and the analyses appear to be conducted with care. Although the results are somewhat mixed, the authors’ interpretations are generally appropriate and reasonable.

However, I would like to raise one methodological concern. If I understand correctly, the moderation analyses include both the total score of the IDA (Sum) and its two subscales (Strict Liability and Injustice and Danger), along with their respective interaction terms, within the same model. If this is the case, it is highly likely that the Sum score and the subscales are strongly correlated, which may lead to multicollinearity problems—such as unstable parameter estimates and inflated standard errors.

Moreover, from a theoretical perspective, including both the Sum and its subscales in the same model introduces conceptual redundancy, which may obscure the interpretation of the results. Since the Sum score is a direct aggregation of the subscales, simultaneously entering them into a single model may result in statistical overcontrol—partially adjusting for the variance that constitutes the construct itself—thereby complicating the interpretability of the predictor effects.

In light of these concerns, I recommend analyzing separate models: one using only the Sum score and another using only the subscales. Alternatively, presenting such models as supplementary analyses could enhance the robustness and interpretability of the findings.

Author Response

Please see the attached Word document for our response.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Review of behavsci-3696538: The Impact of Expert Witness Knowledge on Perceived Credibility: Implications for Not Guilty by Reason of Insanity (NGRI) Endorsement.

 

I have now completed my second review of “The Impact of Expert Witness Knowledge on Perceived Credibility: Implications for Not Guilty by Reason of Insanity (NGRI) Endorsement”. The authors have made some changes to the manuscript and, while some aspects have been improved, I do not believe the manuscript is yet ready for publication.

 

The authors have addressed a number of my earlier suggestions and concerns, incorporating some of my feedback into the manuscript. However, I find that many of the responses still require further attention. Below are the key areas that I believe warrant re-consideration:

 

Hypothesis 2 Results:

• The authors note that they found an effect in the opposite direction than expected. While this is acknowledged, the discussion does not fully explain why this result deviates from the predictions based on psychological theories relevant to the study. It would be helpful if the authors could offer a more detailed interpretation of these unexpected findings in the discussion.

 

Ambiguity in Materials:

• In response to my comment regarding the lack of ambiguity in the case example, the authors indicated that they aimed to streamline the study and maintain control. While I understand the desire for control, it seems that if only one case example is being used, it would be strategic to select one that is more likely to reveal the effects of interest. In line with the Elaboration Likelihood Model (ELM) and heuristics, ambiguous cases are typically the ones where the influence of expert knowledge and credibility would be most pronounced. The authors' choice of a non-ambiguous case may limit the generalizability of their findings.
• The authors also mention that the lack of ambiguity was intentional, as they wanted to focus on expert credibility in NGRI cases, not conduct a mock trial. However, I believe that using an ambiguous case would still allow the authors to assess the impact of expert credibility in a more nuanced context. Furthermore, if the goal is to examine public perceptions of expert credibility in media (e.g., news reports), the materials and the task would look quite different, with individuals not tasked with evaluating evidence or determining a verdict.

 

Influence of Manipulations Interacting with Existing Beliefs about NGRI:

• The authors argue that existing beliefs about NGRI were a secondary hypothesis and may be additive, but this runs counter to the way the hypothesis was presented (Hypothesis 4). They indicated that they expected a moderating effect of pre-existing attitudes, but this suggests a more complex relationship than simply an additive effect. I encourage the authors to clarify this inconsistency and reconsider how this issue is framed in the paper.

 

Altering Qualifications and Credibility:

• Upon revisiting the methods, I believe I understand why altering the expert’s qualifications did not significantly affect credibility perceptions. It appears that the credentials were presented after the participants read the case information, which may have limited the influence of the qualifications on their perception of the testimony. For the qualifications to have a stronger effect, they would ideally be introduced before the critical case information, as would be the case in a courtroom or media context. If this is indeed the case, the authors should address this limitation in the discussion.
• The knowledge scale reflected the factual information that the participants received about the experts qualifications, but see changes on the other credibility subscales they would probably need to have the credentials color their participants interpretation of the testimony. So, the credentials should have been read before the critical case information. This is also ecologically valid in court (expert is first qualified, then testifies), and probably in the media (Dr. X has a PhD in Y and 20 years of experience. They examined the defendant and concluded [insert]).

 

Stimulus Sampling:

• The authors mention stimulus sampling as a means to enhance generalizability, but it is more critical for establishing construct validity. It is important to ensure that the study materials reflect the construct of interest—in this case, NGRI endorsement—rather than being confounded with other factors. The authors mention stimulus sampling as a means to enhance generalizability, but it is more critical for establishing construct validity. It is important to ensure that the study materials reflect the construct of interest—in this case, NGRI endorsement—rather than being confounded with other factors. I raised this point not due to concerns about generalizability, but to ensure that the materials are valid indicators of the construct being measured. It may be useful for the authors to elaborate on how they ensured the materials accurately reflected the intended construct. It may be useful for the authors to elaborate on how they ensured the materials accurately reflected the intended construct.

 

Some other minor issues:

• While the additions to the introduction and discussion are relevant, the paragraphs have become quite lengthy, and some sentences are overly complex. I recommend breaking up these sections to improve readability and clarity.
  • The description of the Koehler et al. (2016) study is an example of how the text could be more concise (see lines 176 to 180). They say that “One study found that even if evidence was scientifically validated, jurors inconsistently recognized or weighted scientific validity when judging evidence strength (Koehler et al., 2016). In other words, jurors were not translating scientific validity into evidential strength. Rather, it was the experience level of the expert witness (high/low) that served as a proxy for scientific validity (Koehler et al., 2016).”
  • Instead, they could say something like “One study found that, even when the method used by an expert to evaluate evidence had a strong scientific foundation, people tended to rely on the experience and background of the expert to determine how much weight to give to that evidence (Koehler et al., 2016).”
• There are a few typos in the new additions, such as “Given the ambiguity of a NGRI verdicts…” (line 156). A quick proofread would help clean this up.
• Hypothesis 1 and 2 specify that these are “main effects” but then the pattern of specifying the type of effect does not continue for Hypothesis 3 and 4.
• It would be helpful for the reader to add the knowledge subscale results to Table 1 so that it is easy to look at the results in one place.
• Hypothesis 3 is addressed after Hypothesis 4 in the results – it would flow more logically if they were switched.
• The authors suggest using a more direct measure of attitudes in future research. It might be helpful if they could propose specific paradigms or methodologies that could address this challenge, as direct measures of attitudes can be difficult to create in lab or online settings.

 

Summary:

The authors have made notable efforts to address some of the concerns raised in the first round of reviews. However, there are still several important areas that require further revision and clarification. I believe the manuscript has the potential to make a meaningful contribution to the literature, but additional work is needed to address the remaining issues.

Author Response

Please see attached response to reviewer comments.

Author Response File: Author Response.pdf

Round 3

Reviewer 1 Report

Comments and Suggestions for Authors

Thank you for engaging with all of my comments - I am satisfied that the authors have responded sufficiently to my suggestions and concerns. I still believe that there are parts of the paper that could be more concise or that could be phrased differently, but I believe that these are mostly a matter of personal preference rather than anything wrong with the paper or study itself. They have adequately stated the limitations of their design and approach now, and corrected any errors I pointed out. I have no further suggestions.

Author Response

Thank you for engaging with all of my comments - I am satisfied that the authors have responded sufficiently to my suggestions and concerns. I still believe that there are parts of the paper that could be more concise or that could be phrased differently, but I believe that these are mostly a matter of personal preference rather than anything wrong with the paper or study itself. They have adequately stated the limitations of their design and approach now, and corrected any errors I pointed out. I have no further suggestions.

1. We appreciate your willingness to review the manuscript. We believe it is much improved due to your contributions.