The Likelihood of Early Guilty Pleas Following Digitally Recorded Victim Statements for Family Violence

Abstract

This study follows 4715 Family Harm cases for which charges are laid (from around 15,000 events from 2018–2020). Comparisons are made between cases where a digitally recorded victim video statement (VVS) is taken to those who (1) make a written statement, (2) refuse to make any statement and (3) present at the public counter and make a written statement. Findings indicate that VVS increases the rates of an early guilty plea by 95% (OR = 1.95, LCL = 1.34, UCL = 2.7) compared to those who decline a VVS and have a written statement. No difference is observed for those presenting to report an event at a public counter. A more modest effect is observed comparing those who refuse a statement altogether (OR = 1.28, LCL = 1.03, UCL = 1.60). A VVS is nearly twice as likely to lead to an early guilty plea. It is reasoned that there is a poor rate of guilty pleas for written statements, rather than an elevation in rates for VVS. Age and gender are unrelated to the elevated rate of pleading guilty to a VVS. Event seriousness is inversely related to pleading guilty, whereas having many prior convictions or being remanded increases the likelihood of the guilty plea.

Keywords

Family violence Digital Evidence Offender Behaviour Victim Statements Guilty Pleas

The Likelihood of Early Guilty Pleas following Digitally Recorded Victim Statements for Family Violence

New Zealand began trialling the implementation of digitally recorded evidence-in-chief for family harm events in November 2015. Formal implementation started in 2017 following a pilot study (Walton et al., 2018) so that all victims’ statements for family harm offences are now recorded on Police i-Phones and uploaded to a secure site. This effort has led to incremental improvements in procedures, legal challenges and changes to regulation of the Evidence Act (2006). New Zealand has now refined its protocols for the display, storage and sharing of victim video statements (VVS). It is recognised that these practices are quite different from those allowed elsewhere (Backes et al., 2018) and pose a challenge to long-standing judicial practices (Chin, 2020).

Against a background of research that observes the capacity to minimise, exclude or excuse violence within families, even by the victims themselves, video-recorded evidence is considered a means of, ‘limiting the opportunity for perpetrators (and others viewing such evidence) to question, diminish, or deny the reality and impact of such violence on victims’ (McCulloch et al., 2019), p. 3). The early-phase implementation did result in an increase in early guilty pleas for those receiving video statements compared to those who presented written statements (Walton et al., 2018). Similar findings are being observed elsewhere (Backes et al., 2018) and were expected based on earlier studies (Dawson & Dinovitzer, 2001).

This study extends the previous findings by adding detailed comparisons of the different forms of statement, accounting for the seriousness of the offending, and the characteristics of the offender. Three years of data are used to consider whether the original finding is what (McCulloch et al., 2019) describe as the ‘CSI effect’: and elevation in guilty pleas simply because of the novelty of the new technology. The findings have relevance to the spectrum of recorded written statements used to support evidence-in-chief, including those taken by police body worn video cameras (Morrow et al., 2016), as is the practice in across all of Australia (Harris, 2020) and elsewhere (Owens et al., 2014).

In New Zealand’s practice, victims are instructed at the time of making the VVS that it is evidence for the court, that it a criminal offence to falsify the statement and that they will be required to appear in court.¹ Victims are often told that the VVS will make it easier for them to present the case in court, and that it is more likely to lead to an early guilty plea. Legislation requires that the victim must be available for cross-examination. The VVS is a formal statement and can only be recanted if the victim declares that the video is a fabrication of events.²Only in extraordinary circumstances, and then only with significant corroborating evidence (usually an independent witness) can a VVS be presented as evidence-in-chief³ without the victim’s presence in court.

A VVS augments the victim’s opportunity to represent what happened to them. In a typical case, the victim is asked to confirm their identity and confirm they made a VVS at the time of the event. The VVS is then played to the court. The attending officer appears in the video asking questions of the victim who describes the offending in their own words. A VVS can be an emotion-ladened account that brings the court, prosecution, counsel and the victim back to the time of the event. The victim may appear as visibly injured, and the VVS may be taken in any location (for example, in a hospital bed) though it is preferred that the evidence is recorded within their home.

Offenders become aware of the VVS and have an opportunity to view it. Defence counsel negotiates any redactions to the statement with prosecution counsel but in every other respect the VVS presents an ‘at the time of the event’ account of the victim’s experience.

It is well-understood that giving a statement in court is especially difficult for victims who hold a relationship to the offender (Boivin & Leclerc, 2016; Ellison, 2002; Kohn, 2008; Lens et al., 2013; Stephens & Eaton, 2020; Wood et al., 2017). The intended effect of the VVS is to reduce the burden of recalling the event and describing the emotion of the harm. These therefore reduce the chance of being re-traumatised though the court experience (Orth, 2002). The concomitant benefit is a facilitation of the victim’s co-operation in the prosecution because digitally recorded evidence, given as evidence-in-chief, is less intimidating for victims (Kebbell & Westera, 2011)

Two main effects underlie the effect of a VVS inducing early guilty pleas. The first is that the evidence is more accurate (Westera et al., 2013) because the VVS records the questions as they are presented to the witness, along with their response, including their non-verbal responses (Vrij et al., 2001). Related, a VVS is thought to be more credible because it is an account taken immediately after the event (Westera et al., 2011). The second effect is that the VVS makes it more likely that the victim will not recant their statement (Ellison, 2002; Hester, 2006). A VVS assists the victim to co-operate with the prosecution, to remain a credible, capable witness at trial (Ellison & Munro, 2014): a set of prerequisites perceived necessary for a successful prosecution (Dawson & Dinovitzer, 2001; Westera & Powell, 2017).

Two years prior to the introduction of VVS (2014–2015), 88% of all prosecutions at Counties-Manukau District Court were disposed during the administration or review phases (as the prosecution case was being prepared). In contrast, family harm cases have a much lower rate, at only 62% disposed of during the same phases. Around 38% of family harm cases progress to trial compared to just 12% of other prosecutions. A conviction would be secured in just 28% of family harm, judge-only trials. The average disposal for prosecutions cases was 86 days, but 228 days for family harm prosecutions. (NZ Police, 2017)

There are, at least, four main benefits of increasing early guilty pleas: (1) an efficiency to the justice system, including court time, and police resourcing, reducing the average time to dispose of the case; (2) sparing the victim (and family) the ordeal of a trial; (3) encouraging participation in the process that would not have otherwise been maintained; and (4) under New Zealand legislation, the offender obtains a 25% reduction on their sentence if they plead guilty at first case review.

However, there are concerns for other aspects of the new practice. The emotional content of the VVS may be seen as prejudicial (Estrada-Reynolds et al., 2016), the novelty of the practice might prejudice the decision of the offender (McCulloch et al., 2019), the visibility of injuries is known to be a leading influence in the prosecution of cases (Henning & Feder, 2005) and the proximity of the statement being taken at the event may remove the victim’s agency in navigating an outcome other than criminal prosecution (Moore & Singh, 2018). For example, 40 percent of women in domestic violence shelters report they would not have sought help if they knew a mandatory reporting for domestic violence would occur. Additionally, changes to practice may be especially prejudicial to some sectors of the population (Reeves & Meyer, 2021).

The underlying reasons why an offender is more likely to plead guilty because of a change in the form or evidence are not understood. The offender may recognise evidence-related factors: that the evidence is better, emotionally confronting or more credible. Equivalently, the offender may regard the written statement as more contestable, easier to minimise or reinterpret. Additionally, the offender may recognise that the VVS is easier for the victim to present and is therefore less likely to be recanted. Offenders may have previously relied on the vulnerabilities of the victim, their relationship to them and the difficulty of the judicial process, hoping these factors induce a lack of co-operation with prosecution (Robinson & Cook, 2006). Thus, holding out for trial might be a sensible strategy, knowing that most cases of arrest do not lead to conviction (Garner & Maxwell, 2009). These factors may alter with the nature of the offending, the exposure a person has to the justice system and other offender-related characteristics (Hirschel & Hutchison, 2001).

Moore and Singh (2018) are critical of using digitally recorded evidence-in-chief during a prosecution because they suggest it removes the ability of victims to change or alter their account. Any change to the system that impacts on the ability to alter a statement should not be presumed to be a positive outcome from the victim’s perspective. Their study offers ‘the reliance on visual evidence often serves to facilitate the silencing of domestic violence victims and the further removal of their agency within the prosecution of their assaults’(p. 128). The diminishing of agency may be an important, albeit unintended, consequence of adopting technology for VVS. However, when considering a victim’s agency, it is especially important to know whether early guilty pleas increase because of a VVS if victims are encouraged by, and choose to engage, because of the possibility that they may.

This study addresses the complexities associated with the likelihood of an early guilty plea being entered by the offender following the introduction of VVS. The different forms of statement (VVS compared to written statements) are considered to examine whether the effect of making an early guilty plea is concentrated in key demographics (age, ethnicity, gender and criminal record) or a generalised effect for all offenders and levels of offence seriousness. The analysis also indirectly addresses whether the influence of VVS is a function of the novelty of the practice by addressing 3 years of cases from 2018–2020.

Method

Data Handling and Processing

Data were collated from family harm events in the Counties/Manukau Police District, South Auckland, from 2018 to 2020 where a VVS could have been taken to support charges laid. When combined, there are N = 4975 case files. The data files contained details of each event, charges and outcome, including a detailed account of the progression of the case through the judicial system.

The first data issue was the described detail of the charges laid that lead to the arrest. It is usual in police practice to assign an ANZSOC code, or a NZ Police Code assigned to each of the charges (Australian Bureau of Statistics, 2018). The codes for the charges were not recorded. However, the charges were described with their short form descriptions, such as Male Assaults Female, Injure with Intent and Wilful Damage. To determine the seriousness of the charge, the first 14 letters of the description were matched to the code descriptions in the NZ Crime Harm index (NZCHI) (Curtis-Ham & Walton, 2017).

A stem of a charge description, for example, threatening behaviour, might have several matches in the NZCHI. Where several codes were returned from the matching effort, the average of the returned values was assigned to the charge recorded in the VVS file. This process reduced 1572 individual combinations of charge descriptions to just 377 charges once the most serious case (the first mentioned in the description) was extracted for matching to the NZCHI. Each was assigned its corresponding derived index value.⁴ Table 1 offers the 10 common charge descriptions, their corresponding NZCHI and the values assigned for this study. Note that when a charge with additional minor charges is represented as a unique combination of charges the most serious charge is the value assigned.

Table 1.

The Most Common Charges Laid in Cases of Family Harm for the 3 Years 2017–2019 in Counties/Manukau New Zealand.

Code Description	N	Assigned CHI-Value
Assault on a person in family relationship	512	10.0
Male assaults female	454	10.0
Contravenes protection order	426	10.0
Wilful damage	326	5.0
Assault with intent to injure	187	69.3
Common assault (domestic)	106	5.7
Behave threateningly	51	5.0
Assault on a person in family relationship and wilful damage	50	10.0

An additional file was sought from the national database with a query of identification details to match offender demographics into the file. Around nine percent could not be matched based on a simple query, and these were manually identified to extract the details of age, gender, ethnicity and number of prior convictions. After data manipulation the removal of duplicates leaves a remaining core of serviceable data of N = 4715.

A thousand and fifteen files are separated from consideration of the main effect (whether a guilty plea is entered) because these cases involve no further action being taken by Police, or where the file recorded is assessed as inadmissible through procedural failure. This leaves a core of 3699 files that constitute the main analysis. Where no further action is recorded, Police may have issued a Police Safety Order or a warning, or initiated some other course of action, other than a decision to lay charges. Demographic details include all cases in the files (N = 4715) unless otherwise specified.

Method of Analysis

The main model is generalised linear model with whether a guilty plea is entered as the main dependent measure. The approach to building the model follows the recommendations of Hosmer and Lemeshow (2000). The first step is to identify factors holding a univariate relationship to the rate of guilty pleas. A broad set of factors are considered including person characteristics such as age, ethnicity, gender and number of previous convictions. Two factors related to the event are also included: (1) the seriousness of the charge and (2) whether the offender was remanded in custody. These factors were found to be important in the pilot study (Walton et al., 2018).

The second step placed the significant factors into a combined model to adjust the combination of factors that interact and impact on the comparison of interest: whether VVS increases the likelihood of an early guilty plea compared to a written statement. This is a general approach to ensure any bias towards guilty pleas in the constituents of the subgroups representing VVS is considered. The sample is reduced because the analysis recognises a potential ‘survivor bias’; named after Wald (1943), it is a type of sample filtering that occurs in large natural experiments (see Mangel Samaniego, 1984).

An additional analysis represents the relationship between those who wholly refuse any statement. This subgroup of the data has a special interest because they have at first sight a higher rate of early guilty pleas compared to those who offer a written statement. This is a counter-intuitive result and cannot remain without proper consideration. The analysis interpretation of why offenders might plead guilty more to a VVS.

Event Seriousness

Family Harm events constitute approximately 40% of frontline Police activity and represent over 300 different offence types (Walton, 2020). To reduce the associated complexity, the NZCHI (Curtis-Ham & Walton, 2017) is introduced to rank-order the offence codes on a dimension of ‘harm caused’ as inferred from calculation based on the actual sentences given for each offence-type. It was previously determined that event seriousness independently influences the likelihood of an early guilty plea (Walton et al., 2018)

Characteristics of the Offenders

Most offenders were males (89%). The mean age was 34 years though this span of ages is from 18 through to 81 years old. It is previously recognised that most family harm events do not result in a charge being laid (Walton & Brooks, 2019). Here, around 15% of the samples have no prior convictions prior to being charged with the family harm event (see Table 2). There is an observed relationship between the seriousness of harm derived from the charge, and the prior convictions of the offender H (5) = 13.22, p. <. 001. A post-hoc pairwise test established that those with no prior convictions are more likely to have committed a lower-ranked offence compared to those with 10 or more prior convictions.

Table 2.

The Number of Prior Convictions for all Main Aggressors in 4715 Events of Family Harm in Counties/Manukau, 2017–2019.

Number of Prior Convictions	N	Mean CHI-value	SD
0	707	29.97	68.82
1	228	28.62	54.55
2–5	716	35.54	80.96
6–10	632	32.75	67.00
11–20	856	36.31	78.20
20+	1576	30.34	66.99
Total	4715

Offender Ethnicity

The high proportion of Māori represented in Table 3 is an overrepresentation compared to the 2018 National Census which estimates 15.9% Māori in the Counties/Manukau Police District. However, compared to Walton & Brooks, 2019 the portion is only slightly less than the 48% Māori observed for all family harm events from 2015 to 2018 in the District. Pacific Peoples represent around 21% of the population in Counties/Manukau but also have an observed higher portion appearing in the cases represented in the sample.

Table 3.

The Ethnicity of the Main Aggressor for 4715 Events of Family Harm in Counties/Manukau, 2017–2019.

Ethnicity	N	Percent
Māori	2068	44
Pacific Island	1484	31
European	569	12
Indian/Asian	445	9
Unknown	93	2
Native African (or cultural group of African origin)/Middle Eastern/Latin American/Hispanic/Phillipino	56	1

Results

Early Guilty Pleas

The univariate relationship between key characteristics of the individual case and the rates of pleading guilty are presented in Table 4. This sample is reduced to those that face charges (n = 3699) because, of course, those who do not face charges (through being diverted by formal warnings, for example) will never need to enter a plea.

Table 4.

The Unadjusted Univariate Relationships between Demographic and Other Factors Related to the Likelihood of Entering an Early Plea for a Family Harm Event.

	Plea Type (N = 3699			Unadjusted Univariate Relationships to Early Guilty Pleas
	Not Guilty	Guilty	% EGPs	Crude Odds Ratio	LCL 2.5%	UCL 97.5%	p-Value
ALL	1549	2150	0.58	—	—	—	—
VVS	1136	1620	0.59	1.11	0.96	1.30	0.17
Charged but no VVS	413	530	0.56	—	—	—	—
VVS versus refused statement	170	189	0.53	1.28	1.03	1.60	< .05
VVS versus written statement	81	61	0.43	1.89	1.35	2.67	< .001
Public counter versus VVS	33	44	0.57	1.07	0.67	1.69	0.77
Remanded versus bailed	310	649	0.68	1.73	1.48	2.02	< .001
On bail	1239	1501	0.55	—	—	—	—
Māori	680	977	0.59	1.06	0.93	1.2	0.36
Pacific peoples	452	695	0.61	1.15	1.01	1.37	< .05
Priors 20+	455	851	0.65	1.57	1.37	1.81	< .001
Fewer than 20 priors	1094	1299	0.54	—	—	—	—
Higher severity charges	428	443	0.51	0.67	0.58	0.79	< .001
Lower severity charges	1121	1707	0.60	—	—	—	—
Male offenders	1404	1989	0.59	1.27	1.01	1.61	< .05
Female offenders	145	161	0.53	—	—	—	—
Under 25 years versus over 25 Years	297	362	0.55	.85	.72	1.01	0.68
Over 25 years	1252	1787	0.59	—	—	—	—

VVS = victim video statement.

Through the period of observations, the rate of early guilty pleas is approximately 58% of all cases (n = 2150). A person can plead guilty at different stages throughout the judicial process. In cases of family harm, a person is unlikely to receive a ‘police bail’ due to policy that prioritises protection of the victim. An offender is usually presented to a court for a judge to decide their initial bail and conditions. This first hearing occurs a day or so after the event (depending on the day of the week in which the event took place). A second hearing should occur within 2 weeks where a judge will hear the plea and reassess bail and impose any conditions of bail. What follows is an administration and review phase, where a person who has not pleaded guilty at either of the first two opportunities has a case prepared for trial and in which they prepare their defence.

Table 4 outlines the univariate relationship of the collated variables associated with the offender and the likelihood of an early guilty plea being entered. A simple assessment that does not account for a proper comparison would not find any advantage in VVS over every other way of proceeding. In Table 4, analysis shows that the odds of pleading guilty are about the same between those that have a VVS (59%) and all others (58%).

Comparing Victim Video Statement to Written Statements

However, it matters who the others are in any such natural experiment. Analysis is indicated by the comparison to those who provided a written statement. Here it is found, at first sight, that an early guilty plea is significantly more likely for VVS (59%) compared with a written statement (43%), and between a third to one and a half times more likely (OR = 1.89, LCL = 1.35, UCL = 2.67). However, complicating a simple relationship is the large positive effect on guilty pleas of an offender being remanded or having previously committed more than 20 offences. Against this, it is much more likely that a person facing a higher CHI-valued charge pleads not guilty. Small effects are observed for gender, with males and those identifying as Pacific Peoples, more likely to plead guilty. No difference is observed for those presenting to report an event at a public counter.

Refusing a statement altogether also increases guilty pleas

A paradoxical effect is found showing a modest effect for those who refuse a statement altogether (OR = 1.28, LCL = 1.03, UCL = 1.60) as they lead to higher rates of guilty pleas than those offering a VVS or a written statement. The effect of the high rate of guilty pleas when the victim refuses to make a statement is because the sample is limited to those prosecutions that proceed to do so with a witness statement, where the witness is often an attending police officer. Consequently, ‘refused statement’ really represents only ‘independently witnessed family harm events’.

The Influence of Charge Severity

An obvious consideration is that those who face higher charges, with a larger number of previous convictions are more likely to be remanded. To account for this complexity, the multivariate relationships are modelled and presented in Table 5.

Table 5.

Adjusted Estimates Showing Early Guilty Pleas Considered by Type of Statement Taken When Controlling for Differences in Offender Demographics, Prior Convictions, Charge Severity and Remand Status.

	Plea Type			Unadjusted Univariate Relationships to Early Guilty Pleas (N = 2898)				Adjusted Relationships to Early Guilty Pleas (N = 2898)
	Not Guilty	Guilty	% EGPs	Crude Odds Ratio	LCL 2.5%	UCL 97.5%	p-Value	Adjusted Odds Ratio	LCL 2.5%	UCL 97.5%	p-Value
ALL	1217	1681	0.58	—	—	—	—
VVS	1136	1620	0.59	1.11	0.96	1.30	0.17
VSS versus written statement	81	61	0.43	1.89	1.35	2.67	< .001	1.95	1.38	2.77	< .001
Remanded versus bailed	250	522	0.69	1.74	1.47	2.07	< .001	1.58	1.31	1.90	< .001
On bail	1239	1501	0.55	—	—	—	—	—	—	—	—
Pacific Peoples versus all others	350	552	0.61	1.21	1.03	1.42	<.05	1.26	1.07	1.49	<.05
Priors 20+	355	361	0.65	1.57	1.34	1.84	<.001	.92	0.76	1.13	< .001
Fewer than 20 priors	1094	1299	0.54	—	—	—	—	—	—	—	—
Higher-severity charges	428	443	0.51	0.67	0.58	0.79	< .001	0.68	0.57	0.81	< .001
Lower severity charges	1121	1707	0.60	—	—	—	—	—	—	—	—
Male offenders	1116	1577	0.59	1.37	1.03	1.82	< .05	1.21	0.90	1.62	n.s
Female offenders	101	104	0.51	—	—	—	—	—	—	—	—
Under 25 years	236	277	0.54	.82	.68	0.99	< .05	0.92	.76	1.13	n.s
Over 25 years	981	1403	0.59	—	—	—	—	—	—	—	—

VVS = victim video statement.

Table 5 again offers the univariate relationship and the combined model that controls for the observed influences presented in Table 1. However, the sample is further reduced to n = 2898 as the focus is on comparing those who progress with VVS and those who refused this and instead have a written statement. This removes those cases that did not have VVS but progress when the victim refused to make any statement, public file statements taken some time after an event, those with technical errors, or when no reason for not progressing is given.

When comparing VVS to those who refused VVS but took a written statement it is observed that the influences on the rate of early guilty pleas are increased by a mean of 95% (OR = 1.95, LCL = 1.38, UCL = 277, p < .001) when other factors that influence the rate of guilty pleas are accounted for in the adjusted model. Importantly, the two previous demographic factors drop out of consideration so there is no effect for age or gender. However, there is an effect for those with a self-reported ethnicity aligned to Pacific peoples. A check on the interaction effects shows no significant interaction between VVS and Pacific Peoples, indicating a general response to pleading guilty early is not one specific to receiving VVS.

Discussion

The programme developed from the pilot implementation of VVS offers a much larger set of cases in which to examine the likelihood that a person charged with a family harm-related offence will plead guilty at some stage prior to the completion of a trial. The expansion of the number of cases allows for consideration of demographic factors concerning both the likelihood of pleading guilty prior to trial and the characteristics of those who refuse to cooperate as victims. The reason for the rate of early guilty pleas being relatively high for VVS is revealed by comparing the rates of early guilty pleas for other types of prosecutorial processing. At first sight, VVS does not fare much better than ‘any other statement’, ‘those who make no statement’ or ‘those that make a statement at the public counter’. However, these comparisons mask the effect due to a type of selection bias in the sample frame. Proper comparisons to those who make a written statement reveal a large increase in the relative rate of early guilty pleas for those cases with a VVS.

The research here follows the initial findings of Walton et al., 2018 that determined VVS increased the likelihood of an early guilty plea by 77%. Here, using over three thousand files, over a 3-year period, a similar result is found. When VVS is compared to written statements, VVS is found to increase the rate of early guilty pleas by around 95%. Equivalently, the odds of an early guilty plea are nearly double for VVS compared with those who give a written statement. That the finding is replicated and endures suggests that the heightened rate of early guilty pleas is not a response to the novelty of the technology. Some aspect of having a veridical account of the victims’ statements via a digital record induces a higher rate of guilty pleas, as was anticipated with the introduction of the technology and new practice (Backes et al., 2018; Dawson & Dinovitzer, 2001; Westera & Powell, 2017)

The effect is considered against a set of confounds that are controlled for in a more sophisticated modelling, though when these factors are controlled, they elevate the odds slightly (albeit with a wider confidence interval). Minor effects are observed for gender, with males and those identifying as Pacific Peoples, more likely to plead guilty. No difference is observed for those presenting to report an event at a public counter to volunteer a statement sometime after the event.

It is thought that offenders perceive an opportunity to overturn a written statement’s veracity based on the interpretation of the wording and the recalled circumstances of the event. This may well remain a motivation for some, however, we find large positive effects on the rate of guilty pleas if an offender is remanded or has previously committed more than 20 offences (including offences unrelated to Family Harm). Those with more experience of the justice system are less likely to defend the charge and plead guilty early. At least some of this effect is accounted for by the delay in reaching trial with the time in remand accumulating to the point where there is little to benefit (to the offender) of contesting the charge as the sentence is likely to be considered as ‘time-served’ (on remand).

There is some evidence in our study that those facing more compelling evidence are likely to plead guilty, all other things held equal. The reasoning derives from the seemingly paradoxical result that those whose victim refuses to give any statement plead guilty at much higher rates than those who offer a written statement. The finding occurs when witnesses give evidence, with Police officers themselves often being the witness to family harm events. In these cases, the offender is more likely to plead guilty than when confronted, even with a written statement, from a witness.

The improved quality of the evidence is given as the main motivation for supplying VVS considered in the District Court by R v Brown (2019). Here, the judge considered all evidence ought to be admissible unless it is explicitly inadmissible (which VVS is not) and concludes that:

The use of the VVS does not of itself give the imprimatur of truth and accuracy to the evidence. It is merely an attempt to get the best evidence that will have the greatest relevance. If there are issues that were unknown at the time, such as not being on medication, which will have an effect on reliability and accuracy then they can be dealt with at the trial. That is when and where findings of fact are determined. (Judge N Sainsbury, R v Brown, [34], 2019)

Judge Sainsbury’s ruling should also not be interpreted as giving licence to VVS. However, it goes to the concern that the defendant might also acknowledge the same: that VVS is better evidence than a written statement. The analysis of VVS outcomes here suggests that a defendant facing high-quality evidence has a reasonable motivation for an early guilty plea. This is a likely mechanism for the pattern of results we observe.

The effect that Pacific Peoples plead guilty is a concern for the broader judicial system. Further examination may find the result generalises across a wider range of offending and is therefore likely a cultural response. This may represent an opportunity for considering alternatives resolution and restorative justice techniques (Cook et al., 2004; Robinson & Cook, 2006).

Pleading guilty early is also less likely when a person faces a higher CHI-valued charge. Thus, the more serious the offending the more likely the person is to defend their position. Conversely, for those who have time-served the likelihood of pleading guilty is elevated. Kingsnorth and Macintosh (2004) found that felony charges rather than lower-level misdemeanours or violations of probation were negatively associated with levels of victim support for prosecution. This points to an important dynamic relationship between being charged with higher-level criminal offences and the likelihood that the victim will be less co-operative.

The results are consistent with the theory that a digitally recorded statement is harder to contest and that a written statement is perceived as an opportunity to re-interpret. There seems to be nothing inherent in the VVS that drives it to have a higher rate of guilty pleas compared to other categories, such as when witnesses attest the events, or those statements made with purpose at the public counter. It is more likely that digital records remove the opportunity to reinterpret what was meant by a statement and in so doing make it harder for both the victim and the offender to exempt or diminish the behaviour, sometime after the event. It is this core idea that should be further considered in studies that explore reasons for retractions, or refusals to give statements.

Moore & Singh’s (2018) idea that a victim’s agency is diminished by digital evidence remains a challenge. It is always in difficult, confusing and strained circumstances that a victim is asked to consent to a process supporting a prosecution. It is known that this is not always what they want (Hoyle & Sanders, 2000). However, in this study, the same is largely true whether the victim makes a written or digitally recorded statement. The victim is warned that the statement is to be used in court, and their permission is obtained for this to occur. Even the potentially prejudicial component of having visible injuries is somewhat offset by the traditional practice of supplying photographs to the court. The important unresolved issue is whether the option of a VVS, by being better evidence, somehow induces an outcome that is not what the victim wants. It remains essential to the victim’s agency that researchers confirm that a VVS increases the likelihood of an early guilty plea if being told this influences a victim’s decision to participate.

Limitations

This study has not manipulated the opportunity for a victim to receive VVS, nor would such be appropriate given the accumulated evidence that shows a strong positive effect for VVS on promoting early guilty pleas. As a natural experiment it simply compares the rate of VVS to the baseline of guilty pleas from the traditional practice of recording a written statement.

Similarly, we have not asked offenders why they are more likely to plead guilty when faced with VVS, though it is a presumption that they can fully account for their reasoning in such complex circumstances. The influence of content on the rate of early guilty pleas might be better indicated by, for example, comparing those VVS containing visible injuries and comparing their outcomes with those where no injury is visible (Schmidt & Steury, 1989). All the VVSs go through a moderation process so they could be coded for content characteristics. Obtaining the perspective from the offender is important to resolve concerns that the process might unfairly induce the plea.

There is a need to understand victim demographics and compare these to those of the offender. The full spectrum of family harm dyads are represented in the data presented here. For example, it might be expected that child/parent and parent/child relationships have different rates of early guilty pleas to those represented overall. The importance bears upon the opportunity for diversion and other schemes (such as restorative justice practices, Hoyle & Sanders, 2000; Walton et al., 2019). Equally, understanding the victim’s perspective of a VVS is important, especially whether the intent of the practice aligns with their actual experiences.

Finally, there is no baseline rate of pleas against the practice prior to the introduction of VVS and the opportunity for such is eliminated by the nationwide implementation of VVS. However, it is known that the rate of convictions following family violence events is low (Garner & Maxwell, 2009) and comparing prior practice to current, especially following the points of attrition from arrest to prosecution, may help resolve whether VVS does facilitate co-operation with a successful prosecution.

Conclusions

The odds of an ‘early guilty plea’ are nearly doubled (195%) for those having a VVS compared to those who refuse this but take a written statement. The adjusted model controls for ethnicity, gender, age and other factors like whether a person was on remand, whether they face higher-severity charges and the number of prior convictions they hold. VVS has the clear benefit of producing a higher rate of early guilty pleas, an effect that is retained through 3 years of observations. It is reasoned that there is a poor rate of guilty pleas for written statements, rather than an elevation in rates for VVS. Age and gender are unrelated to the elevated rate of pleading guilty with a VVS. Event seriousness is inversely related to pleading guilty, whereas having many prior convictions and being remanded increase the likelihood of the guilty plea. The introduction of VVS has had a significant impact on the processing of family harm cases but this impact appears to be applied independent of other factors.

Footnotes

Acknowledgments

The authors acknowledge the work undertaken by police officers and employees who implemented this trial and collected data and information.

Declaration of Conflicting Interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article:The authors are independent researchers working across Government, contracted to New Zealand Police.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Darren Walton

Notes

Author Biographies

Darren Walton is an adjunct Associate Professor of Psychology at the University of Canterbury. His work is multi-disciplinary and focuses on data science and analytics. He has published widely in applied psychology working within public health, road safety and criminal justice.

Ross Ellwood is an Inspector in the New Zealand Police. He is the National Manager for Victim Video Statements and has maintained a role within the development of VVS since the 2017 pilot. Ross has been a Police officer for 28 years, with 23 years as a Criminal Investigation Branch Investigator. He has worked extensively with victims and families, and survivors of family harm.

Samara Martin (TeRarawa and Ngāpuhi) is a Māori criminologist with special interest in victims and victim advocacy with previous roles as a victims’ support worker and with the Salvation Army.

References

Australian Bureau of Statistics . (2018). ANZSOCcodes. http://www.abs.gov.au/ausstats/abs@.nsf/mf/1234.0

Backes

B. L.

Wasim

LaMotte

Kammer-Kerwick

Wood

McDonald

(2018). Recording victim video statements as evidence to advance legal outcomes in family violence cases (ReVEAL) 2.0: Case characteristics and initial impact. Domestic Violence Research Reports.

Boivin

Leclerc

(2016).Domestic violence reported to the police: Correlates of victims’ reporting behavior and support to legal proceedings. Violence and Victims, 31(3), 402–415. https://doi.org/10.1891/0886-6708.VV-D-14-00076

Chin

J. M.

(2020). Pre-recorded expert evidence in intimate partner violence cases. Current Issues in Criminal Justice, 32(4), 458–474. https://doi.org/10.1080/10345329.2020.1830505

Cook

Burton

Robinson

Vallely

(2004). Evaluation of specialist domestic violence courts/fast track systems. London: Crown Prosecution Service (CPS) and Department for Constitutional Affairs (DCA).

Curtis-Ham

Walton

(2017). The New Zealand crime harm index: Quantifying harm using sentencing data. Policing: A Journal of Policy and Practice, 12(4), 455-467. https://doi.org/10.1093/police/pax050.

Dawson

Dinovitzer

(2001).Victim cooperation and the prosecution of domestic violence in a specialized court. Justice Quarterly, 18(3), 593–622. https://doi.org/10.1080/07418820100095031

Ellison

(2002). Prosecuting domestic violence without victim participation. The Modern Law Review, 65(6), 834–858. https://doi.org/10.1111/1468-2230.00412

Ellison

Munro

V. E.

(2014).A ‘special’ delivery? Exploring the impact of screens, live-links and video-recorded evidence on mock juror deliberation in rape trials. Social & Legal Studies, 23(1), 3–29. https://doi.org/10.1177/0964663913496676

10.

Estrada-Reynolds

Schweitzer

K. A.

Nuñez

(2016).Emotions in the courtroom: How sadness, fear, anger, and disgust affect jurors’ decisions. Special Section: Psychology Of Persuasion Symposium, 16(2), 17.

11.

Garner

J. H.

Maxwell

C. D.

(2009). Prosecution and conviction rates for intimate partner violence. Criminal Justice Review, 34(1), 44–79. https://doi.org/10.1177/0734016808324231

12.

Harris

B. A.

(2020). Visualising violence? Capturing and critiquing body-worn video camera evidence of domestic and family violence. Current Issues in Criminal Justice, 32(4), 382–402. https://doi.org/10.1080/10345329.2020.1831730

13.

Henning

Feder

(2005).Criminal prosecution of domestic violence offenses. Criminal Justice and Behavior, 32(6), 612–642. https://doi.org/10.1177/0093854805279945

14.

Hester

(2006).Making it through the criminal justice system: Attrition and domestic violence. Social Policy and Society, 5(1), 79–90. https://doi.org/10.1017/S1474746405002769

15.

Hirschel

Hutchison

I. W.

(2001). The relative effects of offense, offender, and victim variables on the decision to prosecute domestic violence cases. Violence Against Women, 7(1), 46–59. https://doi.org/10.1177/10778010122182299

16.

Hosmer

D. W.

Lemeshow

(2000). Applied logistic regression. John Wiley & Sons.

17.

Hoyle

Sanders

(2000). Police response to domestic violence. British Journal of Criminology, 40(1), 14–36. http://doi.org/10.1093/bjc/40.1.14

18.

Kebbell

Westera

(2011).Promoting pre-recorded complainant evidence in rape trials: Psychological and practice perspectives. Criminal Law Journal, 35(6), 376–385.

19.

Kingsnorth

R. F.

Macintosh

R. C.

(2004).Domestic violence: Predictors of victim support for official action. Justice Quarterly, 21(2), 301–328. https://doi.org/10.1080/07418820400095821

20.

Kohn

(2008). The justice system and domestic violence: Engaging the case but divorcing the victim. New York University Review of Law & Social Change, 32(2), 2012–2134.

21.

Lens

K. M. E.

Pemberton

Bogaerts

(2013). Heterogeneity in victim participation: A new perspective on delivering a Victim Impact Statement. European Journal of Criminology, 10(4), 479–495. https://doi.org/10.1177/1477370812469859

22.

Mangel

Samaniego

F. J.

(1984).Abraham Wald’s work on aircraft survivability. Journal of the American Statistical Association, 79(386), 259–267. http://doi.org/10.2307/2288258

23.

McCulloch

Pfitzner

Maher

Fitz-Gibbon

Segrave

(2019). Digitally Recorded Evidence-in-Chief Family Violence Trial: Final Report (pp. 1-98). Melbourne: Monash Gender and Family Violence Prevention Centre, Faculty of Arts, Monash University.

24.

Moore

Singh

(2018).Seeing crime, feeling crime: Visual evidence, emotions, and the prosecution of domestic violence. Theoretical Criminology, 22(1), 116–132. https://doi.org/10.1177/1362480616684194

25.

Morrow

W. J.

Katz

C. M.

Choate

D. E.

(2016). Assessing the impact of police body-worn cameras on arresting, prosecuting, and convicting suspects of intimate partner violence. Police Quarterly, 19(3), 303–325. https://doi.org/10.1177/1098611116652850

26.

New Zealand Police (2017). Victim video statements training manual. New Zealand police.

27.

Orth

(2002).Secondary victimization of crime victims by criminal proceedings. Social Justice Research, 15(4), 313–325. https://doi.org/10.1023/A:1021210323461

28.

Owens

Mann

Mckenna

(2014).The Essex body worn video trial. The impact of body worn video on criminal justice outcomes of domestic abuse incidents. College of Policing.

29.

Reeves

Meyer

(2021). Marginalized women, domestic and family violence reforms and their unintended consequences. In Oxford research encyclopedia of criminology and criminal justice. Oxford: Oxford University Press.

30.

Robinson

Cook

(2006).Understanding victim retraction in cases of domestic violence: Specialist courts, government policy, and victim-centred justice. Contemporary Justice Review, 9(2), 189–213. https://doi.org/10.1080/10282580600785017

31.

Schmidt

Steury

E. H.

(1989).Prosecutorial discretion in filing charges in domestic violence cases. Criminology, 27(3), 487–510. https://doi.org/10.1111/J.1745-9125.1989.TB01043.X

32.

Stephens

Eaton

(2020).Cultural factors influencing young adult indian women’s beliefs about disclosing domestic violence victimization. Journal of Social Issues, 76(2), 416–446. https://doi.org/10.1111/josi.12385

33.

Vrij

Edward

Bull

(2001).Stereotypical verbal and nonverbal responses while deceiving others. Personality and Social Psychology Bulletin, 27(7), 899–909. https://doi.org/10.1177/0146167201277012

34.

Wald

(1943). Tests of statistical hypotheses concerning several parameters when the number of observations is large. Transactions of the American Mathematical Society, 54(3), 426–482. https://doi.org/10.2307/1990256

35.

Walton

(2020). Te Manawa Titi - Whāngaia Ngā Pā Harakeke: Auckland District: 2019-2020 (pp. 1-23). Wellington: Crow’s Nest Research.

36.

Walton

Brooks

(2018). The use of victim video statements in family violence cases increases the rate of early guilty pleas. Journal of Interpersonal Violence, 36(13–14), 6098-6116.0886260518817065. https://doi.org/10.1177/0886260518817065.

37.

Walton

Brooks

(2019). Whāngaia Ngā Pā Harakeke Pilot: Counties Manukau District Outcomes Evaluation (pp. 1-37). Wellington: New Zealand Police.

38.

Walton

Brooks

(2019). Iwi community justice panels reduce harm from re-offending. Kōtuitui: New Zealand Journal of Social Sciences Online, 15(1), 75–92. https://doi.org/10.1080/1177083X.2019.1642921

39.

Westera

N. J.

Kebbell

M. R.

Milne

(2011). Interviewing rape complainants: Police officers’ perceptions of interview format and quality of evidence. Applied Cognitive Psychology, 25(6), 917-926. https://doi.org/10.1002/acp.1770.

40.

Westera

Kebbell

Milne

(2013).Losing two thirds of the story: A comparison of the video-recorded police interview and live evidence of rape complainants. Criminal Law Review, 19(4), 290–308.

41.

Westera

N. J.

Powell

M. B.

(2017).Prosecutors’ perceptions of how to improve the quality of evidence in domestic violence cases. Policing and Society, 27(2), 157–172. https://doi.org/10.1080/10439463.2015.1039002

42.

Wood

Wasim

Busch-Armendariz

Kammer-Kerwick

Bassett

Kellison

(2017). Methods report reveal: Recording victim video statements as evidence to advance legal outcomes in family violence cases: Determining efficacy, cost efficiency, & impact on outcomes. Domestic Violence Research Reports, 1-47. https://sites.utexas.edu/idvsa/files/2019/03/ReVEAL-Methods-Report-formatting.pdf.