Research and analysis

Summary: Fraud and error deterrence/prevention message testing

Published 28 March 2019

By Michelle Lloyd and Philip Wilson (Solutions Strategy Research Facilitation Ltd).

Background and aims

The Department for Work and Pensions (DWP) is looking to improve the way it communicates with benefit claimants, supporting them to manage their claims more effectively and establish new behaviour change techniques to discourage fraud and reduce error. DWP analysts looked at the literature and evidence around:

• penalties and deterrence
• barriers to reporting changes of circumstances
• lessons learnt from previous benefit fraud communications campaigns

The analysts then highlighted a number of possible approaches, using new messages across the communications product range aimed at claimants.

This qualitative research was commissioned to provide DWP with:

• evidence indicating which approaches may be most effective in discouraging undesirable behaviour and/or supporting claimants’ understanding of benefit rules and processes
• findings which help to refine the themes covered in the messages
• an indication of how any identifiable claimant groups may respond differently to messaging to allow tailored messaging as needed to specific groups

Methodology and sample

The research used qualitative methods with 64 individuals, including 10 small group discussions (3 to 5 respondents each) and 18 depth interviews (3 paired and 15 one-to-one interviews). Thirty-nine respondents were current or previous claimants of Universal Credit including a mix of age and life stages, but biased to those who were currently unemployed. Twenty-five respondents were current or previous claimants of Pension Credit.

The research sample was drawn primarily from DWP records of claimants who had previously experienced problems with their claim. The sample was therefore not representative of the wider Pension Credit or Universal Credit claimant population. Additional free-found recruitment was used to boost numbers for the study, creating a separate sub-sample of respondents who fitted the attitudinal profile of claimants who may be prone to making mistakes or not being fully compliant with the benefit rules.

The research took a staged approach to allow for the test messages to be amended between stages 1 and 2, and for the recruitment of stage 2 to focus on the audiences of greatest interest emerging from stage 1. Follow up telephone calls were conducted with a selection of respondents who were happy to be re-contacted to discuss the messages that respondents felt had made an impact.

Main findings

Knowledge and attitudes

The research highlighted many similarities in terms of attitudes to DWP and knowledge and awareness of the rules around benefit claims across both Universal Credit and Pension Credit audiences. However an important difference identified was that although mistakes could be made by both groups, mistakes were more likely amongst the Pension Credit audience who did not necessarily understand how Pension Credit worked, and did not see it as a classic ‘benefit’, more as a top up to their State Pension. This suggests that clarity is needed about what changes of circumstance affect Pension Credit and how claimants come to receive it. The research also showed that communications directed at Pension Credit claimants needed to be clearly labelled as such, as they will not necessarily engage with communications that call out directly to ‘benefit claimants’ more generically.

The research highlighted the importance of considering claimants’ personal circumstances when predicting reactions to DWP messaging and communications, including:

• reasons for not updating changes of circumstances
• attitudes to DWP
• attitudes to their own financial situation and benefit fraud

Across the sample, the phrase ‘changes of circumstances’ was a commonly understood term, however respondents, particularly Universal Credit claimants, had a variable understanding of specific changes of circumstances they were required to report. Alongside lack of knowledge and understanding, other factors also influenced non-reporting, including poor understanding of the significance of reporting certain changes and the short-term nature of certain changes, such as someone moving in temporarily.

Reporting changes of circumstances was also linked to a perception of loss with assumptions that reporting would most likely lead to:

• a reduction of benefits
• a suspension of benefits
• a new claim or investigation

In addition, there was a concern that retrospective reporting would lead to:

• sanctions
• having to pay back overpayments
• the temporary halting of a claim

As cash flow was described as a constant problem, interruptions to benefits or the need to make repayments caused significant anxiety and problems.

There was a perception amongst respondents that serious benefit fraud does exist but that making small mistakes, or at times just bending the rules slightly, was not benefit fraud. Given this mindset, when the word ‘fraud’ was used in messaging it signalled large-scale fraudulent behaviour, not the type of behaviour that they themselves engaged in. Statements that highlighted ‘fraud’ were therefore easily dismissed.

Responses to messages

The research identified useful themes in terms of how to foster a mutual respect between claimants and DWP and how helpful reminders of changes of circumstances might be combined with messages about the potential penalties for fraud. The findings also offered fresh insights into why some messages may provoke a negative emotional response that leads to them being poorly received or ignored by claimants.

Clear objective reminders to report changes of circumstances were welcomed. The research identified a role for more frequent nudges about the changes of circumstances that need to be reported so claimants do not fall behind on reporting and then fear coming forward due to penalties or repayments. The research also suggests that using the positive framing of a useful reminder allows reference to avoiding penalties, making the message feel beneficial to the claimant. General messages of support and help were also broadly welcomed but would need to be followed through effectively and be credibly backed up by DWP’s behaviour.

Many of the test messages, particularly stronger warnings about the negative consequences of failing to comply, triggered spontaneous negative and fearful responses. Although some claimants responded to these out of fear and a desire not to be dishonest, others did not feel encouraged to take a measured and thoughtful approach to the information or instructions being supplied. The research therefore suggests that these types of warnings need to be used and framed carefully.

The research also showed that certain tonal factors encouraged a more negative emotional mindset which appeared to stand in the way of objective reasoned reflection. These included messages that:

• used personalised language around cheating or fraud
• used short staccato sentence structure which felt aggressive
• used emotive words such as ‘fraud’ or ‘fraudster’
• linked penalties to making genuine mistakes

Alternative tonal approaches appeared to trigger a less negative emotionally charged mindset. For example using polite, de-personalised language encouraged acceptance and using a formal, informative, authoritative but serious tone showed DWP was serious in its intent.

Messages that personalised fraudulent behaviour also risked being read as assuming fraudulent behaviour before proof, which provided the individual claimant with an excuse to dismiss them. There were those in the sample who wanted to avoid the stigma of being referred to as a benefits cheat or fraudster and wanted to distance themselves from this behaviour by ensuring they were compliant, however using this provocative language also provided more negative responses. The findings suggest therefore that such language could be used carefully within more rational descriptions of what fraud is to avoid being seen as too personalised.

Many research participants responded negatively to messages which directly highlighted that the responsibility rested solely on them to maintain accurate benefit claims because they felt that DWP was in control, and they also argued that DWP sometimes made mistakes. In this respect, the Universal Credit online portal seemed to give the small number of its users in the sample a valuable sense of control over their claim, which engendered empowerment and a greater sense of responsibility, which could theoretically deliver better compliance.

Messages about the social cost of benefit fraud and error encouraged thoughtful discussion, but as tested in the research, were not felt likely to encourage changes in behaviour and attitude. In addition, respondents often dismissed official statistics, or found them hard to understand or relate to their personal circumstances.

The research suggests that it might be helpful to state that avoidance of penalties is also DWP’s goal. Where DWP’s responsibilities and commitment to helping people in financial need were also highlighted, a more reciprocal relationship of ‘mutual trust’ could develop. This helped claimants to accept DWP communications more positively, leading potentially to more pro-active and honest management of their claims. Framing the claimant’s obligations to stand by the terms and conditions set out when they signed up to benefits with DWP’s commitment to provide them with ongoing support, made messages about escalating penalties more acceptable, setting up the framework for the benefit claim in partnership terms, and not appearing to assume that claimants would attempt to defraud the system.

Further considerations

A number of specific perceptions held by those who were involved in this research should be further considered within the overall dialogue with claimants.

Claimants’ fear of coming forward and facing retrospective penalties or repayments was identified as an obstacle to full disclosure and honest management of their benefit claims. DWP should therefore consider how to help claimants overcome their fears. Building a sense of mutual respect between DWP and claimants might help. The research suggested that claimants might be more willing to report all types of changes if they felt that they could trust DWP to be open about what qualifies a person for benefits and assured that DWP was equally concerned about changes which might be to claimants’ financial gain. This would need to be underpinned by improvements to the customer experience to have some credibility.

There can also be a belief that claimants will not get caught for minor infringements which in themselves do not feel significant compared to much larger scale benefit fraud. There could therefore be value in fostering an understanding that fraud or lower level fiddling the system is not acceptable, nor the norm, and that DWP is increasingly effective at detection. Since the official fraud and error statistics shown to claimants during the research often left claimants unconvinced, DWP should consider exploring presenting the data in more emotively connecting ways that highlight how low value infringements also matter.

Finally, the research highlighted that if claimants think they are acting properly then they will not reconsider their behaviour. DWP should therefore explore new ways of more regularly reminding and/or obliging people to verify their current claim details and helping them understand the range of changes they have to report and why. There could therefore be value in encouraging claimants to recognise it is in their own interests to check promptly that any updates they make have been correctly recorded by DWP. Claimants who are comfortable with and prefer web-based means of communication should be informed that managing your claims to benefit through an online portal can make this much easier.

Read the full report Fraud and error deterrence/prevention message testing.