Introduction

During the 1990s a model for the interviewing of victims, witnesses and suspects was implemented in England and Wales, called the PEACE framework (see Milne and Bull 1999; Shepherd and Griffiths 2013, for a comprehensive description). Since its introduction, there have been several studies examining real-life investigative interviews with suspects that have been conducted in those countries (e.g. Bull and Soukara 2010; Clarke and Milne 2001; Griffiths and Milne 2006; Leahy-Harland 2012; Oxburgh and Ost 2011; Oxburgh et al. 2006; Soukara et al. 2009; Walsh and Bull 2010a, b; 2012a, b; Walsh and Milne 2008).

A common finding is that unethical practices found in earlier studies of police practice in these countries when questioning suspects (such as exaggerating the strength of evidence, manipulating suspects to persuade them to confess, see Irving 1980), have all but disappeared. Many of those studies cited above, however, still find aspects of investigative interviewing that require improvement (such as rapport skills, questioning, summarising and theme development), a research base to identify exemplars of best practice still remains too sparse (Bull 2013). The present study aims to learn from investigators in England and Wales what they consider to be their preferred way of disclosing evidence in interviews.

Evidence Disclosure in Interviews

From research that has been conducted across the world, whether in the laboratory or in the field, it has been found that the disclosure of evidence during interviews with suspects is an important feature in either (i) eliciting confessions (Cassell and Hayman 1996; Gudjonsson and Petursson 1991; Kebbell et al. 2006; Moston et al. 1992; Sellers and Kebbell 2009; Soukara et al. 2009; Walsh and Bull 2012a); (ii) detecting deception (Dando et al. 2015; Granhag et al. 2012; Hartwig et al. 2006; Hartwig et al. 2005; McDougall and Bull 2015; Sorochinski et al. 2014); or (iii) obtaining comprehensive accounts from suspects (Nystedt et al. 2011; van der Sleen 2009; Walsh and Bull 2015).

Prior studies have found that suspects are more likely to admit their wrongdoing when they perceive the incriminating evidence as strong (Baldwin 1993; Moston et al. 1992). Smith et al. (2011), however, found that the public are frequently inaccurate when assessing evidence strength, often overstating its probative worth. Walsh and Bull (2012a) found that evidence weight alone was not a strong indicator of suspects’ willingness to confess. They found that frequent skilled displays of ethical, non-judgemental, and non-aggressive interviewer attitudes were influential in this regard (see also Holmberg & Christianson 2002; Kebbell et al. 2006; Kebbell et al. 2010).

When interviewers believe they hold strong evidence they undertake less preparation (Soukara et al. 2002). Whether investigators’ judgements concerning evidence strength are accurate has been considered by Smith and Bull (2014), who found that a lack of training made police officers’ judgements of evidence strength somewhat tentative, which, in turn, affected how they undertook to reveal that evidence, in terms of its timing during an interview (Bull 2014).

Evidence Disclosure Strategies

Prior research has established that evidence is disclosed either early, gradually, or late in the interview. Bull (2014) provides a more detailed explanation of each of these evidence disclosure mode (EDM) strategies. In brief ‘early’ refers to evidence being disclosed at the start of the interview, before questions have been put to the suspect (see Moston et al. 1992; Sellers and Kebbell 2009; Leo 1996). ‘Gradual’ evidence disclosure involves a phased presentation of information/evidence, provided in a ‘drip-feed’ fashion throughout the interview (Dando and Bull 2011; Granhag et al. 2012). In contrast, ‘late’ disclosure relates to the gathering of the whole story from the suspect, and dealing with all potential alibis and the like, before presenting the evidence (see Hartwig et al. 2005; 2006; Sorochinski et al. 2014). Each of these three approaches will now be reviewed.

In a recent study, conducted in Taiwan, police officers reported that they typically disclosed evidence early in the interview (Tsan-Chang Lin and Chih-Hung Shih 2013). Early disclosure of evidence, however, might lead some suspects (particularly vulnerable ones - see Gudjonsson 2003) to adapt their subsequent accounts to include the information/evidence given earlier, thus providing a distorted portrayal of events. For other suspects, knowledge of what evidence the police possess may allow them to create a false account of events that incorporates this information (Read et al. 2009). Walsh and Bull (2015) found in their field study of PEACE-trained investigators in England that ‘early’ disclosure approaches rarely achieved the PEACE framework’s main objective of gathering a comprehensive and reliable account.

In Bull and Soukara’s (2010) study of real-life interviews conducted by police officers in England, information/evidence was often revealed gradually. Walsh and Bull (2014) found that a particular form of ‘gradual’ evidence disclosure (labelled ‘deferred gradual’) was optimal in terms of gathering a fulsome account. Interviewers released evidence in stages, but only after each sub-topic within interviews had been thoroughly covered. Another form of ‘gradual’ evidence disclosure, described as ‘reactive gradual’, was less likely to obtain a comprehensive account compared to ‘deferred gradual’ (or, indeed, ‘late’) evidence disclosure. ‘Reactive gradual’ evidence disclosure typically occurred when challenges were immediate and triggered by apparent contradictions with the held evidence, before any, or much, attempt to elicit a fulsome account had been made .

‘Deferred gradual’ disclosure, on the other hand, may (i) lessen the chances of suspects providing no responses to questions (Bull and Soukara 2010); (ii) enhance lie detection (Dando et al. 2015); and (when used skilfully) (iii) enhance the chances of obtaining a reliable account (Walsh and Bull 2010b; 2012a; 2015). Indeed, the PEACE framework prescribes just such an approach. On the other hand, Granhag et al. (2015) argue that gradual disclosure strategies may over time be thwarted by those resistant suspects who repeatedly experience police interviews. Further, Sorochinski et al. (2014) raise the possibility that incremental disclosure could lead to passive compliance from suspects and, in turn, (in worst case scenarios) even prompt false confessions. However, these concerns with ‘gradual’ evidence disclosure strategies did not discriminate between ‘reactive’ or ‘deferred’ gradual approaches (See also Luke et al. 2014).

With regard to ‘late disclosure strategies, alongside Fahsing (2014), Walsh and Bull (2015) expressed concern at investigators’ frequent failure to cover all the alibis and the like in their preceding questioning strategy before putting evidence to suspects, being a fundamental requirement of the ‘late’ evidence disclosure model. Such concerns raise the question of whether a ‘late’ approach is even feasible in interviews with many or several pieces of complex information/evidence. The pioneering experimental studies which involved ‘late’ evidence disclosure, conducted by Hartwig and her associates (2005; 2006), were typified by less complex crime scenarios (but see Sorochinski et al. 2014, who found ’late’ disclosure to be effective in uncovering liars when using a more detailed paradigm). On the other hand, Jordan et al. (2012) found no significant difference in the rate of detecting deception between interviews where evidence was presented either ‘early’ or ‘late’ (likewise for confession rates in their study too).

The use of simpler crime scenarios is also a concern when examining experimental research concerning other evidence disclosure strategies. For example, Granhag et al.’s study (2012) involved an uncomplicated theft scenario when examining the effectiveness of ‘gradual’ disclosure (which these researchers termed ‘incremental’). Interestingly, however, they also utilised as part of their study an ‘Evidence Framing Matrix’, finding that when information relating to the evidence was gradually released in order of its strength (beginning with the least strong first), deception was more frequently identified.

The foregoing reflects the controversy regarding which disclosure strategy is most effective. Walsh and Bull (2010a) found no single strategy apparent in their British field study, with early, ‘gradual’ and ‘late’ strategies occurring across their sample of 142 interviews. Nor was evidence framing used consistently, when evidence was revealed gradually.

In light of the competing views in the literature the present study involved a survey of investigation professionals in England and Wales, to seek their beliefs as to which methods they undertake (and their reasons for their choice). Considering the training these investigators are likely to have received (which emphasises gradual disclosure) it was hypothesised that respondents would opt for such a phased approach and that their reasoning for their choice would focus upon the gathering of fulsome and reliable accounts. However, it was also hypothesised that no single viewpoint would emerge concerning the framing of various items of evidence.

Methods

Materials

In the present study a questionnaire was designed (see Appendix 1) that initially asked for demographic information relating to (i) length of service; (ii) whether or not they had received either basic or advanced training in the PEACE framework; and (iii) the type of investigations they conducted. Subsequent questions concerned the preferred disclosure tactic (‘early’, ‘late’ or ‘gradual’). Respondents were also able to describe any alternative strategy.

Participants were also asked about the order in which they disclosed items of evidence, by relative evidence weight, incident chronology, or a strategy deemed (in their view) to be ‘logical’. Open ended questions also sought explanations from participants for their given answers to explain their reasoning for (i) their chosen strategy; (ii) any lack of adherence to their planned strategy; (iii) any revised plan; (iv) their rejection of other strategies; and (v) any presumptions of guilt.

Participants and Procedure

Ethical approval was obtained from the first author’s University. Participants were recruited during a series of face to face presentation sessions given by the first author to a range of serving professionals between September 2012 and May 2013. These professionals were employed in various parts of England and Wales either as police officers or as civilian investigators (the latter group being employed either by government departments, such as income tax/social security benefit fraud investigators, or in the private sector (e.g. insurance fraud).

Before undertaking these sessions the first author asked the participants to complete the questionnaire. They were advised not to look at their neighbours’ questionnaire. Additionally, the first author used the assistance of a serving police officer, who distributed the questionnaire to fifteen of his colleagues in Wales. Again, those involved were advised that these questionnaires should be completed autonomously by respondents. Once completed, these questionnaires were returned by the conduit to the first author. This method was replicated for eight further questionnaires, where an investigations manager recruited participants, using his contacts in the fraud /financial crime investigation arena. None of the participants were either financially incentivised or rewarded for their participation.

General Details of the Sample.

A total of 266 questionnaires were issued, from which 224 were returned (a response rate of 84 %). Table 1 shows that 58 % of respondents (n = 130) were from investigators with over 10 years’ professional investigative experience, while another 26 % (n = 59) had undertaken their investigative role for between 5 and 10 years. It was also found that 88 % of respondents had undergone training in the PEACE framework, while 60 % of respondents stated that they had undergone further or advanced interview training. Thirty-three percent (n = 74) of the respondents were benefit fraud investigators, 30 % (n = 67) were volume crime police officers, and 29 % (n = 65) were either internal fraud investigators or tax, insurance, or trade fraud crimes investigators. The residual 8 % of respondents (n = 18) were police officers who investigated most serious crimes (e.g., murder, sex offences, or terrorism).

Table 1 Questionnaire response rates: investigative experience by function of job role
Full size table

Table 1 shows that over half of those least experienced respondents (i.e., those with 3 years or less service) were fraud financial crime investigators, while a similar proportion of benefit fraud investigators were among those most experienced of those completing the questionnaire. Ninety-two percent of serious crime investigators (n = 17) had more than 5 years’ experience, the largest proportion of experience by function of job role.

Inter -Rater Reliability of Open Ended Questions.

To avoid potential subjectivity of the data that was yielded through open ended questions inter-rater coding was undertaken of the descriptive data that emerged from the open-ended responses. Two raters (i.e. the first author and a trained Research Assistant, very familiar with this area of study) undertook independently to code the responses thematically. The data were then subjected to Cohen’s Kappa analyses (See Table 2). Following Landis and Koch (1977), the inter-rater values ranged from good to perfect, providing reassurance concerning the qualitative data results.

Table 2 Inter rater reliability on thematic coding to open ended question
Full size table

Results

Preferred Evidence Disclosure Strategies

As can be seen from Table 3, the majority of the 221 respondents, who answered this question (three not doing so), stated that they planned to undertake a gradual evidence disclosure strategy. Using open-ended questions to invite explanations for their beliefs, 70 % (n = 107) of respondents, who said they planned to disclose evidence gradually, stated their planning involved allowing suspects to provide a detailed first account before revealing the information/evidence held (i.e. a ‘deferred gradual’ approach). The remainder said that they disclosed evidence immediately (i.e. a reactive gradual strategy) either when a conflict arose between any given account and the evidence at hand (n = 26), or when denials were received (n = 19).

Table 3 Questionnaire response rates: preferred planned strategy of sample (n = 221)
Full size table

Ninety-six percent of respondents (n = 215) stated that they adhered to their planned strategy either most or all of the time with serious crime investigators stating they always followed their prepared evidence disclosure strategy. For other respondents, changes to planned strategy were attributed to various causes, such as suspects (i) making admissions or denials (n = 75); (ii) providing new/unforeseen revelations or evidence (n = 44); (iii) being willing/unwilling to co-operate (n = 42); or (iv) having their lawyer present (n = 3).

Twenty-five respondents, however, stated that their planned strategy was dependent upon certain factors, citing as reasons for their decision-making (i) the strength of the evidence available; (ii) offence type or complexity; (iii) the expected presence of the suspect’s lawyer; or (iv) suspects’ willingness or refusal to co-operate. However, given that this question pertained to planning ahead of interviews, it remained unclear from their answers as to how they could know at the planning stage whether suspects would either be subsequently compliant or obdurate when questioned.

Investigators’ Reasoning for their Chosen Strategy

Respondents were also encouraged, via open-ended questions, to provide justifications for their chosen strategy. Those who stated that they would rather undertake to disclose evidence ‘early’ commonly justified their choice by saying that this approach would (i) show suspects the strength of the case against them (n = 8); (ii) prompt an account (n = 6); and (iii) encourage admissions (n = 8). Further comments also included that an ‘early’ evidence disclosure strategy would also expedite the interview process (n = 5), and avoid the possibility of suspects remaining silent (n = 4). Early evidence disclosure was also reckoned to reduce the opportunity for suspects to think (n = 2). The presence of a lawyer was also stated as a reason for ‘early’ disclosure (n = 2).

Those who reported that they opted for a ‘gradual’ strategy, offered a rationale that such an incremental approach would assist them in controlling the interview more effectively and be more likely to obtain a fulsome account from suspects (n = 98), particularly when the case involved (i) complexity; (ii) suspected offences that may have occurred over a period of time; or (iii) many pieces of evidence (58 respondents). Further reasons for preferring a ‘gradual’ strategy included the opportunity to expose inconsistencies and lies (n = 43). As such, it was also stated by 17 respondents that in these circumstances challenges could then be more effectively made, while building up the pressure on the suspect so that he/she becomes increasingly aware of the evidence held (n = 35). In this way the gradual production of evidence was stated by 55 respondents as one that avoided the possibility of suspects providing a false version of events (that might occur if all the evidence was disclosed early in full).

Other viewpoints for opting to reveal evidence in a phased manner included allowing time and opportunity for suspects to think and (perhaps) be better placed to provide a more comprehensive account (n = 71). Three serious crime investigators stated that because of the often urgent need to make both an arrest and then undertake initial interviewing of suspects (i.e. before all the evidence had been gathered), gradual disclosure was somewhat inevitable as further evidence frequently came to hand post-arrest.

Those who favoured ‘late’ disclosure argued their case by stating that this approach enabled lies to be better uncovered by such evidence-withholding strategies (n = 13). They reasoned that by asking questions beforehand that would close off potential alibis and the like suspects would have no opportunity to provide further ones that could reasonably excuse them from continuing suspicion (n = 9). As such, it was suggested that a guilty suspect would in these circumstances be unable to account for any conflicts between his/her provided version of events and the evidence once revealed latterly in the interview (n = 11).

Examining Differences Between Investigative Roles

We also examined differences in disclosure choice between the civilian investigators (whose role would be likely more homogeneously defined, reflective of the more narrow nature of their investigative task) and police officers (whose investigations were likely more varied). A series of 2 × 4 Chi square analyses showed a significant association between job role and the choice of evidence disclosure strategy; χ2(4, N = 220) = 16.87, p < .01, phi = .28. While only 6 % of civilian investigators said they used an ‘early strategy’, 24 % of police officers preferred this approach.

A further 2 × 4 Chi square test was conducted, relating to the disclosure choices of police officer respondents (i.e. between those 68 volume crime officers and those 18 who investigate serious crime), finding a significant effect; χ2(3, N = 86) = 6.52, p < .01, phi = .28. That is, only 6 % of police officers who investigate serious crime preferred an ‘early’ strategy, while 29 % of their volume crime counterparts opted for this approach. No significant differences were found between job roles and either ‘gradual’ or ‘late’ disclosure strategies (similarly, no significant differences were found between any of the EDMs and investigative experience).

Evidence Framing

In connection with evidence framing (regardless of whether respondents preferred an ‘early’, ‘gradual’, or late’ disclosure strategy) Table 4 shows that the most favoured approach among the 199 respondents (twenty-five abstaining) was found to be that of initially presenting least strong evidence first. However, one in five respondents admitted to having no plan. A further 25 respondents stated that their choice of order was contextually dependent (as with the preferred EDM) with twelve citing reasons such as offence sophistication or suspect type. Others were less specific (e.g. Respondent #36 stated “if the opportunity is right”).

Table 4 Questionnaire response rates: planned framing of pieces of evidence
Full size table

Responses were also given that suggested that the order of evidence disclosure was more a reactive measure to what the suspect said. For example, if the suspect willingly made admissions then respondents felt that it was less important to consider the order of disclosure (n = 5), whereas resistance would require a more structured approach, with any “conclusive” piece of evidence, generally withheld till the last. In contrast, a benefit fraud investigator opined, that strong evidence might be introduced first to “more quickly weaken the suspect’s resolve”.

With regard to those respondents who preferred to disclose evidence ‘gradually’ (where it might be argued the importance of evidence framing’ is more likely important), responses closely followed a similar distribution as the overall sample (see Table 4).

Presumption of Guilt and Evidence Disclosure

As Table 5 shows, 56 % (n = 120) of the 216 respondents who answered this question, declared that they believed the suspect was guilty either often or very often. Seventy-five respondents stated that they presumed the suspect was guilty due to the information/evidence to hand (including suspects ‘being caught in the act’). Other views included guilt being presumed due to the suspect’s criminal antecedents or personality (n = 91), and pre-interview utterances made by the suspect that indicated his/her guilt (n = 14). Additionally, 70 % (n = 46) of volume crime investigators, 61 % (n = 38) of financial crime/other fraud investigators, and 46 % (n = 33) of benefit fraud investigators stated that they regularly believed that the suspect was guilty prior to the interview. In contrast, just three of the 17 responding serious crime investigators (18 %) professed to assume guilt from the outset.

Table 5 Questionnaire response rates: frequency of guilt presumption
Full size table

Of the 199 respondents who admitted that they believed suspects were guilty at least some of the time before the interview, 43 % (n = 86) acknowledged that it would affect their chosen strategy at least some of time. However, after undertaking a 2 × 2 chi- square test, no significant difference was found between the tendencies to change/retain EDM strategy (always/sometimes or never), regardless of whether guilt presumption or open-mindedness existed; χ2(1, N = 216) = 0.28, p = .60, phi = .04.

What Variables Might Influence the Choice of Evidence Disclosure Strategy?

Standard multiple regression analyses were conducted to determine whether any of (i) training in the PEACE model; (ii) advanced training; (iii) investigative experience; (iv) job role; or (v) guilt presumption were more influential regarding preferred evidence disclosure strategy. Table 6 shows that little relationship was found between any of the predicted variables and the stated evidence disclosure choice. Investigator role, was the only variable significantly associated with the chosen EDM, although (as with all the other variables) only with a modest correlation, F(5, 216) = 2.07, p < .01, R 2 = .05.

Table 6 Results of multiple regression analysis: predictor variable of the chosen EDM
Full size table

Discussion

The PEACE framework, pioneered in England and Wales over 20 years ago, is argued to have contributed (along with the mandatory tape recording of interviews with suspects and the implementation of legislation in England and Wales) to the decline of unethical tactics, previously found in several field studies. What remains less certain is how particular tactics and strategies help advance the central aim of gaining a reliable and comprehensive account from interviewees. The present study set out to examine evidence disclosure strategies, consistently found in the extant literature to be influential in achieving such aims, hypothesising that a ‘gradual’ approach would be dominant in professionals’ stated choice.

To the extent that the ‘gradual’ strategy was found to be preferred by over 2/3 of the respondents the hypothesis was satisfied. This finding, perhaps, should not be wholly surprising since the training investigators receive in England and Wales explicitly advocates this approach. Prior to the introduction of the PEACE framework, Moston et al. (1992) found ‘early’ disclosure to be commonplace, and (more recently) Tsan-Chang Lin and Chih-Hung Shih (2013) found that Taiwanese police officers (most likely to be untrained in the PEACE model) believed that they undertook an early strategy. In the present study PEACE training also appears to influence reasoning. That is, when justifying their choice of EDM, respondents argued that a ‘gradual’ approach would allow opportunity for investigators to gather, and for suspects to give, a reliably comprehensive account (being the aims of the PEACE model). Nevertheless, a quarter of respondents, whose choice was a ‘gradual strategy’, indicated that before carrying out this EDM they would consider such matters as (i) the complexity of the case; (ii) the offences under investigation; (iii) the amount of evidence available, (iv) the suspect; or (v) the presence/absence of a lawyer.

In the present study a ‘deferred gradual’ evidence disclosure strategy was more likely to be one that investigators said they undertook. It was argued that such an approach, with its integral ‘drip-feed’ characteristic, may well make suspects incrementally aware, as the interview develops, of the amount of information held by investigators (and, where they have not told the truth, make them also aware that their account was inconsistent with evidence being gradually revealed). Such an approach, it was speculated, may prompt or encourage suspects to make further disclosures that explain, refute or concede the disclosed information. Such phased responses by suspects, it was claimed, would then be analysed by interviewers for their plausibility, either against suspects’ earlier revelations (not unlike the ‘late’ disclosure strategy) or against information still being withheld, for ‘deferred gradual’ disclosure at later stages of the interview.

Twenty-five percent of respondents, however, said that they favoured an ‘early’ disclosure strategy (arguing that this approach would efficiently reveal to suspects the evidence against them. prompting an early confession). Walsh and Bull (2015) have, however, found that interviews tended to end without obtaining either a confession or a comprehensive account, when ‘early’ disclosure was conducted. Read et al. (2009) caution that some vulnerable suspects may well integrate the revealed evidence into their accounts, thus compromising what should be the aim of interviews with suspects, i.e., to search for the truth.

Respondents in their first 3 years of experience more often stated that they preferred an ‘early’ strategy, a finding which might be explained by their being yet to be fully comfortable with the more demanding ‘gradual’ or ‘late’ strategies. Preference for ‘early’ disclosure (for some investigators) may also be a product of their main duties, which due to their inexperience might be expected to generally involve less complex, volume crimes. The present study found that those opting for early disclosure tended to be those police officers investigating volume crime, where the matters that they investigate, due to their regularity, might lead to a rather expeditious (as opposed to efficient) approach to investigation.

Other survey respondents, albeit fewer in number, stated they would choose to undertake ‘late’ disclosure strategies. ‘Late’ disclosure, it was argued would likely detect any lies through within-interview inconsistencies. ‘Late’ disclosure may also be viewed as congruent with the principles of the PEACE framework, particularly because the ‘late’ approach aims to exhaust all potential alibis and the like that could possibly occur before revelation of the evidence held. However, Walsh and Bull’s field study (2015) found that such endeavours are unlikely to meet their goals as investigators do not tend to cover all alibis and the like in practice, particularly when the case possesses more complex characteristics.

A further concern with the ‘late’ approach could be that prolonged withholding of information/evidence, while interviewers proceed to conduct questioning (that is aimed to cover all possible excuses that a suspect might give) could lead to a lack of co-operation from suspects, through fear or mistrust (Kebbell et al. 2006). Indeed, Granhag and Hartwig concede that the ‘late’ approach is ineffective if the suspect remains silent when being questioned. On the other hand, ‘late’ (and, indeed, ‘gradual’) approaches might enable the maintaining of rapport - a key element in interviews; see Walsh and Bull 2012a, b). ‘Gradual’ and ‘late’ approaches continually invite suspects to provide explanations in response to what the interviewer either asks or (in the case of ‘gradual’ EDMs) reveals, having first requested him/her to freely give an account. Both approaches include a requirement to probe for micro-details, where inevitable turn-taking will occur between the interviewer’s questions and the suspect’s responses. Since previous studies have not examined which of the EDM approaches are associated with prompting suspects to talk, we encourage research in this area.

Most investigators said that their planned strategy was one they actually carried out. Of course, what investigators say they do, and what they actually do has been found to be different (a limitation also of this study) (O’Neill and Milne 2014). Whether the level of planning is thorough remains open to conjecture. Several studies have shown that officers admit that they do not always plan, or (when their interviewing practice is examined) they do not always demonstrate preparedness (Clarke and Milne 2001; Walsh and Bull 2010a; Walsh and Milne 2007; 2008). Indeed, the present study found that one in five respondents acknowledged that they had not thought that much, if at all, about their evidence framing strategy, in contrast to every serious crime investigator who said they always planned, consistent with Griffiths’ (2008) study of their practice.

Respondents, regardless of job role, tended to suggest that their evidence framing strategy was influenced by their own perceptions of the strength of the evidence (stating that they presented the strongest/weakest evidence either first or last). While this suggests that some forethought has been undertaken, such an approach depends on the assumption that the evidence is reliable. Smith and Bull (2014) found investigators had rarely received any training to assist them in making such judgments. Further, the public (some of whom may either have been or may become suspects) were regularly imprecise in estimating evidence weight (Smith et al. 2011). As such, respondents’ views in the present study (that they gradually disclose evidence either increasingly or decreasingly in regard to its strength) is dependent on both interviewers and suspects possessing similar views as to its weight. It might well be the case that such escalation or de-escalation strategies might not be as effective as respondents suggest, due to suspects’ possible misperceptions concerning the probative value of the disclosed evidence.

Limitations of the Study and Future Implications

The present study breaks new ground in eliciting from investigative practitioners their beliefs about their preferred interview strategies concerning what has been found in the literature to be a critical aspect of investigative interviews; evidence disclosure. Nevertheless, there are possible shortfalls in the methodology that may inhibit how generalizable the results are. Firstly, while the sample size is reasonably large, convenience sampling was employed some of the time from those with whom we secured co-operation. However, respondents (i) were recruited through various avenues over several months, and (ii) reflected a wider range of professionals than just police officers (the sample involved several employers, and many different job roles), suggesting that a cross-section of investigator opinion may well have been achieved. Further, as already noted, one of the frailties of self-report studies is that beliefs do not always match realities. That said, the tendency of respondents in the survey to favour gradual evidence disclosure has been corroborated in other studies that involve actual practice (Soukara et al 2009; Walsh and Bull 2015). With regard to concerns over social desirability, our participants were assured anonymity and that their responses would be kept confidential.

The present study is the first to examine in detail investigative practitioners’ preference for evidence disclosure strategies and their reasoning behind such choices. There is, among investigative practitioners, a division of opinion as to what is an effective information/evidence disclosure strategy. Of those who did exercise a choice, a gradual strategy was invariably preferred, with the majority of this particular group opting for, what Walsh and Bull (2015) have coined, ‘deferred gradual’ evidence disclosure. For some participants, however, any strategy would need to be adapted case by case to take into account various factors (whether in regard to either timing or framing of evidence disclosure). If this is indeed the case in practice, policy that argues for one universal approach may need to be re-visited. Further, research can help us better understand in what circumstances certain approaches are more effective than others in gaining comprehensive accounts from suspects.

While the PEACE training may have influenced belief in a gradual EDM, it remains unknown whether this strategy is actually more efficacious than the other approaches in either obtaining reliable confessions or gaining comprehensive accounts. Laboratory studies examining EDMs have thus far provided no conclusive decision. Published field studies are rare, although Walsh and Bull (2015) found a ‘deferred gradual’ approach was both more proficient and more likely to obtain a fulsome account. Nevertheless, more studies (including field ones) are required that involve all the approaches before any solidarity of view as to the most effective EDM is likely to emerge.