Review Reports

Round 1

Reviewer 1 Report

This manuscript provides the results of an analysis of secondary data to assess the impact of victim impact panels on recidivism of DUI offenders. The paper is short and succinct and does a good job of laying out the literature and reasons behind use of VIPs and why their impacts should be further studied. The authors use a simple but appropriate analysis technique and do not overextend the implications of their findings. They are right to be skeptical about the strength of their findings. The findings are important for the field, however, because prior work has provided mixed results. The authors had a unique opportunity to create a comparison group from the apparent (quasi?) random assignment of some offenders to a judge who did not use VIPs. 

On the subject of comparison groups, the authors did note that there were significant differences between groups on two measures. Did the authors consider any kind of matching process (e.g., propensity matching) to better match the treatment and comparison groups? I think the authors should add a statement about why they did not consider matching, or if they did but then decided against using that approach, why they decided not to use it.  

The authors should clarify more explicitly whether there is any system to how people are assigned to judges and to what extent they believe assignments of offenders to judges is random.  

In their literature review, the authors cite a systematic literature review by Miller et al (2015); that paper’s conclusions were that VIP programs did not reduce recidivism. The authors seem to brush off this conclusion and cherry pick specific papers that did find effects and were included in that review to highlight in support of their own work. In the next paragraph, the authors do mention 2 papers that found no effects, so I think this issue is more about how the findings from these studies are presented/discussed and less about having to change the content significantly. I would prefer here that the authors actually discuss why the Miller review concluded that VIPs had no effect on recidivism, where the authors’ claim of a 12% reduction in recidivism came from (if not from the Miller review?) and then mention how their work will improve limitations in the work reviewed by Miller. Not discussing the limitations of prior work and indicating how their paper improves upon those or contributes to the field should also be addressed. Even if the current paper doesn’t improve on anything from prior lit except the length of the follow-up period over which they look for recidivism, it would be useful to know whether there were other factors significantly limiting the prior literature.

In the discussion section, the authors cite and discuss a meta-analysis on restorative justice programs. This seems like something that should be discussed in the literature review. It comes out of the blue in the discussion.

One thing that I thought about while reading was that in the eight-year follow up period, it seems likely that a number of individuals in either the treatment or comparison groups received treatment for alcohol dependency, which would presumably reduce their recidivism rates significantly. This might be part of the explanation of why the eight year improvement in recidivism is higher than the 5-year and 2-year. Just a thought for the authors to consider.

Another thought about differences in offenders – is the location of this program pretty evenly rural or urban? Are there some people who might be able to avoid driving better than others—for example, by having better access to Uber or taxis—creating differences in their outcomes? I am wondering if you have information on the home address of offenders and could use characteristics of that location (like, rural vs. urban or distance to nearest bar, etc.) as a control in the model.

Finally – the observation period is long – were there any significant differences in court services over that time? For example, were remote appearances implemented, or other programs that might also impact DUI recidivism rates?

A few in-text comments:

line 87, sentence starting…”If we exclude…”; this sentence should be rewritten for clarity. It is not clear when reading why you would exclude the study. I suggest switching this sentence with the following sentence, and then clarifying the “If we exclude…” sentence.

Line 98 – “Other studies…” – this sentence could also be written in a clearer way, such as saying that the “…trajectories of individuals with and without criminal records tend to converge over time, so that individuals of the same age have the same risk levels for offending, regardless of prior offending”

Line 105: sentence says ‘greater empathy’ – clarify greater than who or what?

Line 149: Also – did the fact that participants have to pay $50 represent a hardship that discouraged some individuals from participating?

Author Response

Manuscript ID: laws – 1484289

Response to Reviewer 1

• We considered using propensity score matching for this study but the logistic regression results indicated that we would have lost too many cases. I agree with this reviewer that propensity score matching is a superior method for comparison but our power would have been reduced to a point where the findings might not have been interpretable.
• Early on we had chatted with the clerk of court to assess how cases were assigned. The judges rotate from criminal to civil with seven of the nine serving on criminal cases.  She assured us that the cases assigned to the seven judges sitting in on criminal cases were done randomly through a rotation basis. 
• We have modified the Miller et al review to better reflect their summary findings.
• We have also added some narration regarding why this manuscript might differ from the seven articles reviewed by Miller et al, aside from the lengthier follow up period. In our case, the study jurisdiction is more rural and the Victim Impact Panel was delivered in a state which routinely ranks in the top three of drinking and driving behavior.
• A second reviewer requested that we not refer to our study as being characterized by restorative justice practice so we have eliminated the first sentence in the Discussion section.
• We have added some narration to capture the possibility that the VIP group might be more subject to seek chemical dependency treatment over time, relative to the comparison group. Hence, the difference in effect size over time.
• This reviewer inquired whether there were remote appearances during the study period. The data were gathered prior to Covid-19 so all of the appearances in court and with the VIP’s were face to face.
• The court jurisdiction is probably around 70% urban and 30% rural. We do not have records on the residential locations of the VIP and comparison group.  Suffice it to say that like the distribution of these cases, we suspect that there were no systematic differences between the groups in terms of rural vs. urban dwelling.
• We eliminated the sentence referred to in line 87 in which we stated “if we exclude…
• We have modified the sentence referred to in line 98 to reflect more clarity.
• We have clarified the idea that VIP attendees reported increased empathy for victims following their participation in a panel.
• We do not have any information as to whether the payment of $50 posed a hardship for any potential attendees. The Public Health official responsible for the program reported early on that payment was not a problem for these persons.

Reviewer 2 Report

The paper aims to explore the impact of victim impact panels (VIP) on recidivism at the 5 and 8 year post-VIP engagement points, building on a previous study which looked at outcomes at the 2 year point. The researchers employ a quasi-experimental design, comparing those who participated in VIP with those who did not (because the judge in their cases did not refer any defendants to the program). This study’s findings show that participation in VIP contributes to reduced recidivism 5 and 8 years post-participation, which is fascinating.

The methodology is appropriate for the questions under study. The article is accessible and logically and clearly written, including in the reporting out of their analytic findings.

The authors offer a thorough review of previously conducted studies on VIP outcomes, laying the groundwork for their particular focus. Their discussion of the findings suggest future directions for research – e.g., studying the mechanism by which VIP impact future behavior (see lines 367-368) and how those who did not participate in VIP rate their agreement on drunk driving being a risk, etc. (see lines 369-374). This study was not set up to answers these questions but a new study could.

I have one critical concern with the article related to how the authors situate VIP in relation to restorative justice and victim offender dialogue (VOD), especially around conflating VIP with VOD. They are two very different practices. While some restorative justice theorists/practitioners situate VIPs within a restorative justice framework, it is in a different placement within that framework than VOD. Some theorists/practitioners would not even consider VIP as a form of restorative justice. A great process that should be offered? Yes. But restorative justice? No.

I agree with the authors that VIPs are “compatible” (line 33) with reintegrative shaming efforts (which are often discussed alongside restorative justice and VOD). However, there are important differences between VIP and VOD which are not discussed – doing so would clarify how VIP, RJ, and VOD are compatible but different. From line 33 on, however, VIP, RJ and VOD seem to be used interchangeably. For instance,  

• In lines 34-35, the authors write that “VIPS have been employed to combat….homicide,” citing Umbreit et al. (2005). This article by Umbreit and colleagues is specifically about VOD, not VIPs, thus not a demonstration of VIPs being used in response to homicide.
• Lines 48-58 starts with “Restorative justice scholars contend that this process empowers….” Which process are they talking about? VIP? Based on the rest of the paragraph, one would think RJ and VOD. For instance, line 54 states that “negotiation and agreement transpires when offenders consent to model….” This is a feature of VOD (e.g., the perpetrator and their victim actually making agreements together based on the victim’s specific needs) but I am unclear of how this comes up in VIP. The previous paragraph to this one doesn’t suggest that such occurs in a VIP and my own experience with them doesn’t suggest it either.
• The discussion starts by contextualizing the study and its results within Latimer et al. who conducted a meta-analysis of VOD programs, not VIP programs (lines 319-339). The study should be contextualized within literature pertaining to VIP. If restorative justice or VOD is brought up at all, it should be clear that it is being done so due to compatibility, not because they are the same.

This conflation suggests that the authors are not familiar enough with RJ and VOD to properly use it to contextualize the study or the results. It also makes me wonder where else in the article there are similar conflations or problematic uses of previous studies, thus reducing my confidence in the authors. The use of RJ and VOD either needs to be eliminated from the framing (which would not harm the final article as the VIP literature is strong) or its inclusion needs to be more carefully framed and discussed, noting compatibilities as well as distinctions and making more tentative linkages across findings. I also encourage the authors to update their RJ citations as they are quite dated.

 

 

 

 

Author Response

Manuscript ID: laws – 1484289

Response to Reviewer 2:

1. We agree with this reviewer that the process and structure of Victim Impact Panels is not entirely compatible with restorative justice, such as encouraging both victims and offenders to have a voice and negotiate disputes. We have eliminated any reference to restorative justice other than mentioning on p. 1, paragraph 2 that this process is only somewhat related to both restorative justice and reintegrative shaming.  We also wish to retain the study by Latimer et al. mentioned in the second paragraph of the Discussion section.    We felt it was important to note that the effect sizes of a large volume of restorative justice studies largely match the effect sizes reported in our study.  All of these studies assess recidivism effects so we felt it was important to note that the effect sizes are similar, even for less compatible victim processes.  We have done this throughout the manuscript.
2. We have eliminated reference in the Introduction to restorative justice processes that have previously been employed such as studies that have found reductions in homicide (Unbreit et al. 2005) and others. We agree that the process of victim-offender dialogue was different for these studies when compared to ours.
3. We have eliminated again any reference to victim voices as they pertain to restorative justice process as we agree that the process is not entirely compatible with VIP’s.

Round 2

Reviewer 2 Report

Thank you for addressing the concerns about the use of restorative justice.