4 Conclusions and Recommendations

This research has explored the factors that can influence levels of trust and considered how these could relate to building trust in official statistics.

The findings of the research point to lots of things that actors across the statistical system and beyond can do to demonstrate the trustworthiness of organisations producing official statistics and of the statistics themselves. Throughout the report, recommendations have been made that, if acted upon, could ultimately build trust in official statistics.

To conclude, these recommendations are reiterated below. They have been grouped by theme, and some overlapping recommendations have been combined. The primary audience for the recommendations – statistical producer, communicator or statistical regulator – is also given. Though a helpful ordering device, recommendations attributed to one group may still be beneficially adopted by others involved in the production or use of official statistics.

Finally, recommendations for future research are provided. This includes avenues that have been prompted by this review, and/or are designed to remedy identified gaps.

4.1 Producers

4.1.1 Reputation and commitments: make a public-facing commitment to behave in a trustworthy manner

Producers should:

• Align themselves with the public’s expectations around statistical production. This includes: public involvement; reflecting real-world needs; clear communication; minimising harm; and best-practice safeguarding. Behaving in line with the public’s expectations is important because predictability and reliability are seen as helpful criteria against which the public can make subjective and rational choices about the trustworthiness of the actor/object. Meeting these criteria can help build trust.
• Make sure the strategies implemented to improve trust are explained, authentically adopted and sustained over time. Evidence to support this should be made publicly available and provided in an easily accessible format. The process of communicating their intentions to act in a trustworthy way can increase the commitment by the producer to abide by them and, as such, contribute towards heightened trust.
• Explore bespoke approaches to building trust among different groups of the public. This is because actions to improve trust may not be universally effective. For instance, producers may consider utilising different communication networks and supporting a wider range of public engagement efforts.
• Ensure a publicly available trustworthy track record to showcase. This is because both positive and negative reputational legacies can shape future trust decisions.

4.1.2 Public engagement: embrace user dialogue and publicise efforts in this space

Producers should:

• Embrace meaningful public engagement: ensure that public engagement is representative and provides opportunities for collaboration from a wide range of voices. Details of the public engagement process, including indications of those invited to participate, should be shared publicly. This provides an opportunity to showcase efforts to capture a representative picture.
• Embrace co-production opportunities: allowing the public to shape the engagement process, and recognising the valuable insights they can provide, is a positive way to demonstrate trust. This enables producers to exhibit their willingness to bare vulnerability and take the first ‘leap of faith’ in trusting relations. This could help build trust because signals of trust are often mutually reciprocated.
• Advertise public engagement opportunities: producers should share information about how users could be involved in future public engagement opportunities. Clearly advertising this would allow those who want to be more involved in the process to have their voice heard and considered. This is constructive from the perspective of trust because being reflected in the story that official statistics tell is reported to reduce suspicion and increase trust.
• Respond to stakeholder views in a thorough manner: the false pretence that simply gathering views ‘will result in enhanced trust’ should be avoided. Performative, superficial and empty efforts to engage public and stakeholder views will not result in increased trust simply by virtue of the process having taken place. In addition to being meaningful, the outcome of any public engagement process should be clearly communicated to stakeholders. Explaining why views have not been taken forward is an important part of public engagement, and bypassing this may have negative ramifications for future trust relations.

4.1.3 Handle challenge with an open mind: look for opportunities to address concerns and exemplify trustworthiness

Producers should:

• Acknowledge and attempt to understand mistrust, avoiding critical or dismissive reactions and statements, which paint mistrust as irrational. Instead, producers should view doubt/challenge from users of statistics as an opportunity to exemplify trustworthiness to the audience, and support them in arriving at a confident, and appropriate, assessment.
• Continue to monitor public levels of trust: producers should be aware of any value shifts and remain at the forefront of any new or developing concerns relating to the use of statistics. This can be achieved through continued investment in their own public engagement and/or social research. Alternatively, or alongside this, engaging with the wider research community, and monitoring upcoming studies, surveys and reports, would support further understanding. This would ensure producers are in a better position to address any concerns as and when they emerge, and hopefully before they become unmanageable.

4.2 Communicators

4.2.1 Be transparent: explain how decisions are made and show any working out

Communicators should:

• Caveat all statistical outputs: ensure that the statistics will not be placed in a position where they are taking more weight than they can reasonably bear. Anyone who is publishing statistics should be open and honest about their limitations. This is important as transparency can help build trust.
• Follow the principles of intelligent transparency: open and honest communication about the limitations of statistical products can help shield official statistics from possible misinterpretation. Alongside this, it can also help safeguard against future damage to trust levels, which may occur if hidden inaccuracies or other issues are later uncovered or disclosed.
• Explain decisions about “how” official statistics are made: this willingness to show working out has been signposted as a positive strategy in trust building. This is because it provides a paper trail which the public can then review in order to evaluate the accuracy of the statistical output.
• Provide all methodological details in a format which supports public understanding and avoid technical and specialist jargon: anyone publishing official statistics should be mindful of underlying attitudes and remember that “simply providing more evidence is unlikely to shift attitudes.” To build trust, there should be a qualitative shift, and the complexities of statistical processes should be communicated in a simple and easily interpretable manner.

4.2.2 Highlight user relevance: publish tailored and bespoke statistical products

Communicators should:

• Provide personalised statistics which reflect an individual’s story, wherever possible: This could include regional/localised statistics, and/or statistics which provide a more bespoke reflection of different socioeconomic situations. Developing tailored statistics affords the public opportunities to see how the statistics are relevant to them. Exemplifying user relevance can help build trust as relevance supports wider use, and wider use and familiarity can positively contribute to improved trust levels.
• Consider user reception: when the official statistics feature messages which may contradict the publics established opinions or experiences, they should be presented in ways which are less jarring. This may help prevent individuals feeling neglected in the ‘story’ that the official statistics tell. If this is not possible, anyone communicating official statistics should make a concerted effort to explain to the public why the story the statistics are telling may be odds with their lived experiences. This is crucial because dismissing lived experiences can negatively impact trustworthiness.

4.2.3 Increase awareness: use a variety of publication channels to increase public awareness of official statistics

Producers should:

• Embrace a variety of communications outlets and channels: this should include online intermediaries, social media and more-traditional formats, and within this, direct-to-consumer publication should be encouraged. Frequent exposure to official statistics via media heightens familiarity and improves expectations of trustworthiness. Alongside this, it is important that certain communication channels are not neglected as absence can contribute to distrust.
• Strive to increase public awareness of official statistics. This is because recognition, familiarity and exposure support trust building. As part of this public awareness strategy, the value that official statistics deliver, or could deliver, should be highlighted. This may increase people’s willingness to bear the costs of trusting and the vulnerability this entails.

4.3 Regulators

4.3.1 Emphasise accountability: visibly hold poor behaviour to account

Regulators should:

• Recognise that performance and quality are necessary, but they are not sufficient to build trust. Accountability structures should be emphasised, and poor behaviour (not just poor performance) should be held to account. Regulators should be transparent in this respect, and examples where accountability structures have been actioned should be clearly communicated to the public. This is because it important that accountability is visible and seen to be applied. If not, this can have a negative impact on trustworthiness as it signals to the public that poor behaviour is permissible.

4.3.2 Apply accountability with care: be mindful of individual experiences

Regulators should:

• Adopt a “care-full”, transparent and well-explained approach to accountability. This is because “calling out bad numbers” which are reflective of citizens’ experiences may negatively contribute to low levels of trust, since it may be interpreted as a dismissal of their lived experiences. As such, regulators should take extra care to explain why the statistics are being challenged. This should be explained plainly and without reliance on technical jargon.

4.3.3 Clarify accountability: explain reasoning to the public

Regulators should:

Reassure the public that challenging a statistic does not mean the observations reflect incorrect patterns, nor that lived experience is invalidated. Regulators should make it clear that the issues may be methodological or relate to the way the statistic was communicated. Articulating these reasons to the public may help temper frustration and disenfranchisement, and prevent trustworthiness being undermined

4.4 Recommendations for Further Research

Whilst this review has provided some insight as to the current levels of trust, further research would be welcome. This review has provided a broad indication of the sorts of strategies that organisations and academics recommend for increasing trust levels. This review included exploratory primary research into the types of strategies that members of the public would consider to be effective in building trust in official statistics. In spite of this effort, several gaps and opportunities for further investigation remain.

For instance, when asking the public “do you trust official statistics?”, it is not clear what they interpret “trust” to mean. Cognitive interviews to help get a better understanding of the responses provided in the PCOS survey have been conducted. This reveals that there is variation in the ways that people interpret “statistics” (OSR, 2025). However, similar cognitive interviews to understand how the public interpret “trust” were not conducted. Further research into this would be helpful to get a fuller picture of what aspects of trust the public consider to be sufficient, and which are missing when responses of low or no trust are given. The different definitions and variants of trust given in this review could provide a useful basis from which to design a survey which would facilitate investigation of which variant(s) of trust the public are referring to when they signal their levels of trust.

Further research to help establish what characteristics the public regard as valuable for statistical producers would also be beneficial. As mentioned in the review, the professional standards and characteristic attributes which are seen as valuable vary across different professions – with honesty and care considered important for politicians, and competence and expertise respected for those in the scientific profession (Seyd et al., 2022). However, it is not clear what configuration of characteristics the public would consider to denote a trustworthy producer of official statistics. Research dedicated to this question would ensure positive qualities can be emphasised and prevent professionals positioning themselves in ways that may damage trustworthiness.

Studies have suggested that the area of science (hard or soft) has an impact on whether people trust its findings. However, this could be expanded further within the remit of official statistics. Specifically, future research to understand if there is a difference in the type of official statistics could help develop a fuller understanding of trust dynamics and if they function differently across different areas, policies and/or status of the official statistic. This could help establish a better picture of whether the “type of official statistic” influences levels of trust. In other words, is the hard/soft distinction maintained?

Evidence in favour of network dynamics has been proposed, showing that trust in official statistics improves the likelihood that people trust other parts of government apparatus (STATEC, 2023). Further research to investigate whether the diffusion of trust also operates in the reverse (i.e., does increasing trust in producers and/or others across government increase trust in official statistics) would provide fuller detail of trust diffusion across government networks.

Due to the wide variety of ways official statistics are accessed, the evidence to be able to confidently assert the network via which people are accessing them is not available. Further research to help fill this gap and get a better picture of where people access official statistics would be valuable. This would make the discussion surrounding network dynamics more practically applicable. For instance, for questions relating to the communication of official statistics, it may be possible that the chain is broken for one outlet or mode of delivery (say, reports on the national news), yet it can remain intact for another (i.e., the official bulletin). Getting a more complete picture of this will help tailor the recommendations.

In addition to this, getting a better picture of the network via which the public are accessing official statistics may also provide insight as to what mode of delivery (or what platform) future research would be best positioned to focus on. The current challenges associated with accessing the data required to be able to identify which platform the public are using to access official statistics (directly via the statistical bulletin, traditional news, online intermediaries, social media, word of mouth, etc.) makes this a difficult recommendation to action. This is acknowledged. However, it is included as a recommendation, as if the capability arises, this acts as a signpost to point to possible future avenues for valuable research. This recommendation, specifically involving exploration into the way official statistics are communicated via intermediaries, and the possibility of collaboration as an approach to measure access, was also suggested in OSR’s research project, Statistics in Personal Decision-Making (2025).

Finally, commentators are vocally proclaiming a crisis of trust in science and evidence-based decision making. This has inspired further research on this theme, with the fifth wave of the Public Attitudes to Science (PAS) survey adding new questions on the trustworthiness of science and refreshing the section dedicated to science in the media.

The survey, which is delivered by Ipsos and the British Science Association, is scheduled to be completed in spring 2025. Familiarising oneself with the results would be advantageous for anyone interested in the topic of trustworthiness in science and scientific evidence. In addition to this, reviewing the PAS data highlighted in this review and comparing the new figures (2025) to those published in 2019 (which are used in this review) would help identify if the patterns are up-to-date and enable any necessary adjustments to be signposted, particularly as the 2025 survey is the first PAS survey to be carried out in a post-COVID-19 context.