Statistics Group strategic priorities

The following sections give more detail on steers for suggested areas of focus for each of the strategic priorities. 

Users  

The excellent analytical work in the stats group is constrained if we aren’t targeting it appropriately to different audiences, optimising our outputs and proactively promoting it to users and potential users. 

• Undertaking sustained user engagement: OSR recently observed that user engagement with a wider breadth of users of statistics is still too limited across the wider UK statistical system. Developing a deep awareness of what our users value and considering the needs of potential users is vital in the production of our statistics. This requires developing a mindset that is open to different, potential ideas that can be gained from sustained and ongoing interaction with users. For example, this could include sharing information on future improvement plans and encouraging users to get in touch to share their views.  
  
  User input should be core to any design or redesign of data collections and outputs from the outset. In this way user engagement should be embedded into our day-to-day work rather than occasional activity that could be seen as tokenistic.  

• Improving publications so they have more impact: The value pillar in the Code of Practice for Statistics says that the presentation of statistics should be clear, meaningful and meet user needs. It also says that statistics producers should be creative and motivated to improve statistics and data. 
  
  We need to recognise that the appetite for data and how users consume it has changed. OSR noted in their State of the Statistical System report that “The quality of publications varies across government departments and there is an increasing view that they are often too lengthy and wordy”. 
  
  We see four key aspects to good publications to make them more likely to have impact: 
   

1. User engagement: consciously identifying the different users of our statistical products, their level of expertise, what they want from a statistical output and where they are likely to find the information they need. 

2. Generating clearer and more engaging content: thinking about publications in terms of the story the data is telling, and being selective about what to include to avoid distracting focus from the narrative. We recommend writing in plain English in a way that is clearer and easier to understand for all users.  

3. Consistency and trust: implementing a more standard approach to the look, feel, quality and structure of our publications will help establish a more consistent experience for our users. This in turn promotes confidence and trust in our work. 

4. Accessibility: we have a legal requirement to meet accessibility requirements as per the Public Sector Bodies (Websites and Mobile Applications) (No. 2) Accessibility Regulations 2018. 

• Communicating and speaking about our statistics: Our job isn’t finished when our statistics are published. In OSR’s words, “There is a growing need for producers to communicate statistics beyond the traditional statistical bulletin”. We are encouraging statisticians to proactively communicate and highlight the analytical work they do. This can be via our Scotstat mail push system, social media, blogs, conferences and award submissions.  
  
  It also means that statisticians can challenge misuse and misinterpretation of statistics. This provides greater assurances that statistics are properly understood and represented. 

Efficiency  

The ‘efficiency’ priority is about how statisticians do things in an effective, streamlined and less time intensive manner. Central to efficiency is the processes, tasks, tools, services, platforms and infrastructure to produce statistics: 

• Reproducible analytical pipelines (RAP) in open source by default: The Analytical Function RAP Strategy encourages statisticians to adopt RAP by default. Embracing RAP means that the manual aspects of data processing will be minimised. When developed this should free up staff time to do more relevant and interesting analytical work, which should have a benefit on staff morale.  
  
  To support RAP, we’re advocating statistical products to be developed in open-source software, such as R, because we can encourage more collaborative and transparent practices. Within the statistics group we have already signalled our intention to only use open source for the processing of statistical data in Scottish Government by the end of 2026. 
  
  Not everyone needs a full grasp of RAP and open source; managers don’t necessarily need to know how to code but know enough to steer and influence projects and empower assistant statisticians to use and develop their coding skills. 

• Minimising work that adds less value: There are many individual tasks associated with the production of statistics. Teams have good processes to ensure that these tasks are done. But we don’t always consider if these tasks are all necessary or if the amount of time spent on them is proportionate. For example, can publications be discontinued or minimised in frequency if there isn’t a clear user need? Is the amount of time spent on quality assurance of data proportionate to how the data is used? Is all briefing entirely necessary, or could it be condensed to convey the key messages more succinctly?

Data 

The ‘data’ priority is all about how we operate as a data mature profession and maximise the use of our data. Under the data priority, the following are particularly relevant: 

• Enabling coherence of our statistics: OSR describe coherence as “reflecting the degree of similarity between related statistics and the fuller insight achieved by drawing them together”. This means that stats producing teams should think about the suite of data they have available and how to use these more effectively. Statistics producers should think about different sources of data and the pros, cons and applications of each. Examples of coherence include: 

1. Using Census outputs along with our existing data to provide further context and insight for users.  

2. Using administrative data sources alongside other data.  

3. Using complementary statistics based on different measures. 

4. Using statistics relating to different countries in the UK, or different countries across the world to provide a comparison. 

• Considering fitness for purpose of data: The European Statistical System’s (ESS) Dimensions of Quality sets out criteria for assessing the fitness for purpose of data and recognising the trade-offs in using different data sources. A bespoke data collection might provide accurate information but takes time to produce. Whereas management information (MI) may provide timelier information but may not provide as accurate or comparable information with other sources. Both sources can provide value provided their limitations are clearly explained to users. As OSR commented in a recent research paper on quality: “We see value in having more timely data that may be of lower accuracy when there is a clear public interest that can be met… it is a judgement call for producers to make on how to ensure that quality is sufficient for the appropriate use of the statistics”. Using data that is deemed fit for purpose that statisticians may not have previously considered and being transparent with its limitations is something to consider.
  
  Another dimension to fitness for purpose is how far do we need to go with quality assurance, and what is the point where additional time dedicated does not necessarily add value. No data or statistics are perfect, and this is unrealistic to aspire to. 

• Allowing greater granularity of data: We should be striving to produce statistics at a granular level to better support analysis of equality characteristics and richer local level data. For example, the GSS Subnational Strategy encourages statistics producers to “think subnational by default”. Being braver with using data sources that offer greater granularity can open up new insights for users. Key to this is making judgements on the suitability of the quality of the data and risk of identification of individuals. Rather than using blanket rules (i.e. suppressing any figures less than five), we are encouraging statisticians to consider principles-based disclosure control approaches. This means that risk of disclosure is considered on a case-by-case basis. 

People  

Supporting and developing our people is critical, especially as we embark in a change of culture in how we operate as a statistics profession.  

• Enabling intent-based leadership at all levels: This comes from the work of L. David Marquet’s Turn The Ship Around. We advocate empowering all staff to make decisions based on shared purpose and understanding, rather than relying solely on directives from more senior people in the organisation. Those at more junior grades often have a much better understanding of the data they work with and are the best people to make certain decisions. This move to decision making where the information is will enable us to build a confident workforce that engage and can influence the things that affect their day-to-day work.  

• Network building and breaking out of silos: The pandemic and a switch to hybrid working has meant that statisticians have not had the opportunity to network as before in recent years. We are encouraging opportunities for staff to build their networks across the statistical community to help share knowledge and support each other.  

• Using learning and development opportunities: A recent learning and development survey for statisticians issued by the Office of the Chief Statistician (OCS) showed that the main barrier to undertaking continuous professional development is a lack of time. We must allow time for staff to take part in learning and development opportunities and apply the learning in their day job. OSR recognise that the role of statistician is evolving with greater use of data science techniques and complex methodologies. It is important that we keep up to date with skill changes. 
  
  Learning and development can take different forms, for example, structured learning courses, learning from others or self-led development opportunities. It can be on different topics; developing communication, people and influencing skills is equally relevant to statisticians as improving technical and analytical skills. Learning and development should also be applicable and relevant for all grades.