Review of pre-release access in Scotland

Possible options for the future of PRA in Scotland

Both stakeholders and public participants were presented with three potential options and invited to consider the benefits and challenges and potential impact of these, and any practical considerations which should be taken into account. The options presented were to retain the current arrangement of up to 5 working days pre-release access for some statistics, to reduce the pre-release access period to 24 hours, and to remove pre-release access entirely. The options given reflect the current arrangements which are in place in different parts of the UK, and for different types of statistics.

Unlike the stakeholders, there was no broad consensus among the public participants on the different options presented. Public participants’ views on the options tended to align with their opinions on government and their trust in government more generally and they tended not to give detail on the reasoning behind their views.

Retaining 5 working days

From the point of view of stakeholders, the current 5 working days pre-release access period was generally thought to work well and to be beneficial for the reasons previously discussed. Most stakeholders did not identify any concerns about maintaining the current arrangements. Only two stakeholders suggested that 5 working days was potentially too long, with one suggesting that the length may theoretically increase the risk of leaks. On the other hand, two participants questioned why an option to increase the length of PRA beyond 5 working days was not being considered.

I don't want to come across like everything's rosy, but it's one of those things that just works really well. And in this day of in this era of chaos and change, it's quite nice to have something stable, you know. (Comms professional)

Again, the majority of stakeholders stated that maintaining the current arrangements would have no impact on themselves, their workload or their department. For those who did identify potential impact, it was noted that these would be felt at the “statsy corporate level” and would be associated with the fact that Scotland would continue to have a different PRA arrangement to other parts of the UK and against the recommendations made in the Lievesley review. However, this was described as a “minor bone of contention” and not seen as posing a great political risk.

We would remain out of line with England, but I can't say that that causes any issues, you know, operationally or reputationally at the moment. (Comms professional)

Stakeholders put forward a number of suggested improvements to the 5 day PRA period, including:

• Greater transparency of who is on the PRA list, including advanced calls to make sure that PRA lists are up to date for forthcoming publications.
• Centralised publication of PRA lists on the Scottish Government website to aid transparency. Mixed opinions were expressed regarding this, with some participants in favour and others questioning how useful this information would be and whether members of the public would look at this. They also raised concerns that it would increase the workload of their team.
• A more straightforward and streamlined process for swapping people on and off the PRA list in cases of absence or holiday.
• Limits to the number of publications released on any one day.
• To be able to add more people, or a mailbox, to the PRA list, although the need to restrict PRA lists was acknowledged.
• More detail to be given by those requesting PRA as to why people requesting access require PRA to ensure that they are eligible to receive access.
• Greater governance of PRA lists by the ScotStat board or the Office of the Chief Statistician to ensure that these are as limited as possible.

One public participant was in favour of retaining the 5 working days pre-release period as, in their opinion, it allowed for better planning and clearer communication of the statistics.

You know if you tell everybody as it comes out, there could be a state of panic and people are making up their own minds about stuff that they maybe potentially don't even know about…sometimes it's better coming from someone who knows the situation, who's in the power to resolve the situation and deliver the information so that you're at ease. Your mind's at ease. (Public participant)

Reducing PRA

In general, a reduction in the time allowed for pre-release access was seen by stakeholders as less preferable than retaining the current arrangements. Few benefits were identified by participants beyond that 24 hours would be preferable to PRA being removed entirely, and one stakeholder commented that the length of PRA made marginal difference, and it was more important that PRA was retained to some degree. One stakeholder who considered the 5 day period to be too long suggested that 24 hours would be the minimum period for people to be able to use PRA in the same way as they do currently. One participant noted that reducing the PRA period to 24 hours could potentially free up staff capacity as they would no longer need to have the statistics ready for publication 5 working days in advance. Another suggested that decision-making within the Scottish Government would potentially be speeded up as a result of a reduction to 24 hours.

Participants highlighted several potential impacts of reducing the time available for PRA to 24 hours. It was thought that a reduction in the time available to engage with and understand the statistics would lead to a less informed response from the government and from Ministers, leading to the government and Ministers being less well equipped to answer questions on statistics and to be held to account. It was suggested that this would lead to the statistics being less clearly and fully communicated and that this could be detrimental to public trust in both official statistics and in government. It was also suggested that this impact would be particularly felt for complex statistics and large data sets, and on days where a large number of publications were released.

I think the public would end up being significantly less informed with a very real risk the public were misinformed. Particularly in relation to some of the more complicated issues that emerge from statistics, with government unable to answer legitimate questions, and that in turn would undermine confidence in government and certainly would not facilitate scrutiny of government. (Special Adviser)

…if we end up in a position where ministers are going on Good Morning Scotland saying “I'm just seeing this”, you know, or is transparency aided by them having some briefing to support what it says and some understanding of them. (Policy professional)

Without the opportunity to fully review and engage with the statistics, participants raised concerns that inconsistencies or errors in the statistics would not be identified or, if they were identified, could not be rectified within the 24 hours, affecting the quality of the statistics. Members of the policy profession also felt that they would lose the opportunity to add value to publications from a policy perspective. One statistics producer with experience of working with statistics which have a 24 hour period highlighted the lost opportunity to act on suggestions which would improve the communication of statistics.

I've definitely found that the discussion within 24 hours has often brought to light things that we could if we had time change in the analysis and actually create more impact with the statistics themselves, but we just wouldn't have time to do it within 24 hours, so that it's like, OK, well, wait till the next release. (Scottish Government statistics producer)

From a communications perspective, it was felt that reducing the time for PRA to 24 hours would result in delays to communications responses to either the statistics which were being published or on other reactive issues. It would also increase the possibility that communications would not be able to engage with the publication or to produce materials such as infographics, and that the opportunity to engage with the public in this way around the statistics would be lost. Workload planning for communications teams would also be affected, with the effect of no longer having specialist communication teams by topic, potentially increasing the risk of communications staff who are less familiar with the topic area making errors and having to ask more questions from those involved in producing the statistics.

The impact on workload and the increased pressure and stress on staff was highlighted across all stakeholder groups, and those who had experience of working with statistics publications which have a 24 hour PRA period spoke about the day before publication as being hectic and high pressure. In some specific situations where staff work with one publication which combines official statistics and management information, they felt that it would no longer be feasible to combine these in a single publication and that these would have to be produced separately, effectively doubling their workload. There was also a suggestion that staff would be put under pressure to release management information to indicate the direction of travel of the statistics.

…on a very personal level rather than a [external stats producer organisation] level, I think my anxiety is going to shoot through the roof. (External statistics producer)

…I'm afraid the pressure and stress it would impress upon people would just be unacceptable. (Special Adviser)

We'd have to do two publications. The statisticians would have to go through the process twice, with the costs involved. (Policy professional)

A number of practical considerations were raised with reducing PRA to 24 hours. Ensuring that the relevant officials would be available in the 24 hour period to be able to attend briefings, develop and approve responses was identified as a challenge, particularly within the current layers of approval that responses have to go through within the Government. A reduction to 24 hours would necessitate the development of new ways of working, such as all relevant parties joining a live call to develop a response rather than a draft response going through successive layers of approval. The 24 hours was seen as particularly challenging for PRA periods which would fall on parliamentary days as these are already very busy for Ministers and their offices. It was highlighted that changes to PRA, and the resulting impact on the extent to which they could be briefed on publications, would have to be clearly communicated to Ministers.

I think ministers in this space would be… they would require to be communicated with, to understand the implications of this in terms of the robustness of analysis that could be provided at on that time period and that would require a change in expectations from Scottish ministers. (Policy professional)

A number of challenges around workload management and flexible working were highlighted, particularly for those members of staff who work part time and may not be able to be involved in some publications if the PRA period did not fall on their working days. Additionally, participants who work compressed hours raised the issue that they would have more or less time to work in the PRA period depending on the day on which it fell. Some participants felt that their departments would require an increase in staff compliment to cover activity in the 24 hours before publication. It was suggested that the number of people on the PRA list would have to be increased as the likelihood of staff members being unavailable due to sickness or holidays was greater.

If it became a 24 hour thing, then, suddenly, there's… I suppose for the management side, there’s fewer people here who can do it because some people don't work on particular days. So yeah, it might affect flexible working a wee bit. (Scottish Government statistics producer)

Participants also raised practical concerns as to whether the statistics would be released to them a full 24 hours prior to publication, as their experience with economic statistics was that these would often be released mid-morning or around midday, further reducing the time available for them to engage with the statistics. Practicalities around how public holidays were accounted for and whether PRA would be given on a Friday for publications on a Monday were raised.

Challenges with the format in which statistics produced by external organisations are shared were also highlighted. Some teams are provided high level statistics from external statistics providers and then request more detailed statistics in areas of interest, and do much of the analysis of these statistics themselves. A reduction in PRA to 24 hours would mean that the external statistics producers would need to do this analysis themselves in advance, and a participant noted that this would cause difficulties as additional analysis depended on what the minister wants to see and in how much detail. Another participant mentioned that another external statistics provider currently shared statistics with key findings written in a report format but had recently been trialling providing statistics in a dashboard which required much more work to understand and to extract findings, potentially leading to misinterpretation of the statistics if those receiving them did not have sufficient time to understand and interrogate the dashboard.

So I think right now with a shorter PRA period, we'd cope, we’d make it work - if we ever though got into a position where we didn't have pre-release access and statistics are just published on dashboards, there'd be an awful difficult adjustment period where I think we'd misuse a lot of statistics. (Scottish Government statistics producer)

In terms of implementing a reduction in the PRA period, it was suggested that this would need to be piloted in particular areas or with particular types of statistics. It was noted that UK Government departments had moved to 24 hours PRA, and that the stakeholders in the Scottish Government could engage with their counterparts in the UK Government to learn from their experiences.

One participant in the public engagement group felt that PRA should be reduced to 24 hours, arguing that 5 working days is too long and that Scotland should be aligned with other parts of the UK.

I think after having a better think about it after this discussion, I think I probably wouldn't keep the five days, I think it's too long, but I would say up to 24 hours because I think just what I touched on earlier on, you know they’re human and need time to process it a little bit, but I think 5 days is potentially too long, and that's when maybe when people could start sort of not trusting things because they've had a bit too long, but up to 24 hours, I think is fair. (Public participant)

Removing PRA

Participants did not identify any practical benefits associated with the removal of PRA entirely. It was acknowledged that the argument could be made that removing pre-release access would have benefits in terms of increased transparency and therefore public trust. However there was broad disagreement with this. Participants countered that, without any kind of pre-release access, the risk of public misinformation would be greatly increased, and that this would be harmful to public trust in statistics.

The potential impacts of removing PRA which were identified by participants were broadly similar to those for reducing PRA, but participants felt that these would be compounded by the removal of any access to statistics prior to publication. Participants were of the opinion that losing the opportunity to engage with and understand the statistics would lead to a delayed response from government regarding statistics, ministers being unable to answer questions on the statistics or to counter inaccurate reporting of statistics in a timely manner, misquoting statistics or giving responses which are not based in evidence. This was seen as injurious to public trust in both statistics and government. It was suggested that this would have the impact of reducing the extent to which the government can be directly questioned held to account by the media and members of the Scottish Parliament, and one Special Adviser stated that it may no longer be practicable for ministers to undertake media interviews regarding statistics publications.

I think you get the situation where government can't answer the question, which I think is an absurd position if we're meant to be people who are furnishing people with information and in a time where we have an awful lot of disinformation online, if that's allowed to go in gallop before a factual response can be provided, I think we actually harm public discourse. (Special Adviser)

I think it would actually reduce the ability of people to hold government to account because government would be literally unable to answer legitimate and significant questions in a meaningful time frame and in the modern world in the way that the modern media operates, once the story has moved on or once a false narrative which is perhaps a greater risk here has taken hold, it's incredibly difficult to shake it. (Special Adviser)

…and I think to take away PRA like, it'll be not to the benefit of the public. We’ll be putting ministers out there and they won't know what they're talking about. And so, yeah, I think it would be extremely bad. (Communications professional)

From a communications perspective, participants felt that no pre-release access would negatively impact the speed and accuracy of the communications response. The speed of the media cycle was frequently highlighted and it was suggested that the delay in the Government’s response would mean that the news cycle would have moved on by the time a response could be put out, meaning that the opportunity to counter misrepresentation of the statistics would be lost and news coverage would simply say “The Scottish Government refused to comment”. It was also a concern that, depending on workload and priorities, the government may lose the opportunity to respond to some media stories at all. While it was again noted that skeleton communications responses could be drafted in advance, it was stated that this would not be possible for large and complex publications. One participant shared their experience of engaging with communications regarding the publication of statistics which currently have no PRA and felt that there was little communications engagement with this release. It was their opinion that replication of this approach for all official statistics would result in a reduction in the quality of information being communicated by the Scottish Government. It was also noted that it would not be possible to plan ministerial visits around the publication of statistics, meaning that the opportunity to raise the profile of publications and bring attention to them would be lost.

Another potential impact of removing PRA was that individuals who currently have access to the statistics prior to publication may try to find other ways to access this information, potentially in a less controlled and transparent manner such as release under the purposes of quality assurance or by statistics being leaked. Pre-release access was thought to give “a legitimate route where there's a legitimate need” and it was suggested that that, without this, statistics producers may be put under pressure to release management information which would give an indication of the overall statistics in a way which was not regulated.

I think it's important in that there's a real transparency, it feels, with the current process in that we know who has seen what and when. And so that's very, very clear. Without pre-release access, there's that sort of concern that it would almost become a bit unregulated with the management information where people would potentially have a very good idea of what statistics we're going to show, but there just wouldn't be that same record of knowing this person saw this. (Scottish Government statistics producer)

The potential impacts on workload of reducing PRA were also relevant to the removal of PRA. It was suggested that the need to understand and respond to statistics on the morning that they were published would put “stress levels through the roof”. Those who currently work with statistics for which there is no PRA reflected on the expectation for staff to be working early in the morning ready to respond to the publications. Again, it was noted that this would not be accessible for those who work part time or have flexible working arrangements, and it was suggested that this may make roles working with these publications inaccessible to some. It was also suggested that, instead of engaging with the statistics themselves to distil meaning and key messages, those who currently receive pre-release access may expect those who had been involved in the production of statistics to provide these, increasing workload for statistics producers.

If we had to do that for every stat that gets published across government, that's a massive increase in cost and it's a massive impact on the well-being of a lot of people because all of a sudden I'm going to need people whose job starts at 5 o’clock in the morning to be ready to turn things round. (Special Adviser)

…people who would normally see the pre-release content itself would surely come to us asking for “OK, what are the headlines then?” or “what's the big news?” as opposed to getting that information themselves. It was… just it feels like a lot more effort on our part here. (Scottish Government statistics producer)

It was suggested that moving to no PRA would necessitate a practical change in the way that stakeholders within the Scottish Government approach the development of responses to publications. It would no longer be practicable for responses to be approved using the current model of the “five hurdles” consisting of “various layers of internal governance” of statisticians, directors, policy leads, special advisers and Ministers. To allow the Government to respond in a timely manner, all parties would need to be available to join a live call, and it was suggested that this approach may require greater involvement of staff at more senior levels.

We’d need to change our processes, so essentially everybody's available to live-write a statement. Otherwise it'll be middle of the afternoon before we get something out, and that's too late. (Communications professional)

I think you'd also probably find more senior officials having to get roped in to where they may be dealing with other things. But to provide that sort of assurance to cabinet secretaries and ministers, I think it would almost escalate to do that in such a short time. (Special Adviser)

It was also suggested that changes would be needed in the way that the Scottish Government produces and publishes reports if these contain statistics which are produced by external statistics producers as these statistics would no longer be available in advance.

One participant in the public engagement group believed that PRA should be removed entirely. This participant argued that the statistics should be released to representatives of all parties simultaneously and that these representatives should convene at a later date to discuss the statistics. However, this participant did acknowledge that the media was unlikely to wait for this to happen.

Give everybody the information at the same time, let everybody digest it, investigate it and see what they want, and then come back, at a later point. (Public participant)

…that's never going to happen, the media is always hounding a story. So you would like to think they would have the grace to say they will get a couple of days, but they're looking for the knee jerk reaction and who's going to say the silly thing and get that out there first… (Public participant)

Other public participants disagreed with this view and felt that the lack of preparation time would lead to those in power making incorrect statements which were not supported by the statistics, to the detriment of public trust.

I know they'd be thinking on their feet, so they'd just be… they might say something, and then they didn't mean that …and then people wouldn't trust what they're saying. I just think it would be fumbling… (Public participant)