As the coronavirus (COVID-19) pandemic has had a significant impact on our data, you may need to highlight quality issues to the user. Research shows that while caveats and quality warnings are important to users, their priority is to read analysis and access the latest data. It is important to find the right balance between highlighting the impact on the quality of the data and not overwhelming users with detail.
Use warning boxes to tell users how the coronavirus pandemic is affecting the quality of the statistics.
Experimental Statistics are official statistics that are newly developed or going through development and testing. This is to ensure they meet quality standards. They should always be published primarily as a bulletin; an article using the data can be published alongside the bulletin if more detailed analysis is needed.
Existing official statistics or National Statistics can be labelled as Experimental Statistics if the method or data source changes or goes through a period of development. During this time, the quality of the data may be affected.
It is important to be clear if your analysis contains Experimental Statistics. This is to ensure that users are aware of any quality issues and can interpret and use your data and analysis correctly. You can use a grey warning box at the end of the Main points section saying:
“These are Experimental Statistics. The [method/data source/estimates is/are] currently under [review/development], which [brief detail about how this affects estimates or data quality]. We advise caution when using the data.”
Change the standard wording to describe your release. More information on the methods used and quality limitations of the data can be included in the Measuring the data section for bulletins, or Data sources and quality section for articles.
You can also indicate that your release contains Experimental Statistics in the page summary.
New research into experimental methods and sources
All statistical research into new methods or data sources, where it is not yet considered official statistics, should be published as an article.
It is important to be clear if your analysis is using data gathered through experimental research or methods that are not yet official statistics. You can use a grey warning box at the end of the Main points or Main changes section with this wording:
“These are not official statistics and should not be used for policy- or decision-making. They are published as research into an alternative method for producing [topic] statistics. We advise caution when using the data.”
Change the second sentence to describe your release. More information on the methods used and quality limitations of the data can be included in the Data sources and quality section.
Signposting to existing content
If relevant quality information has already been published in other releases or in ONS statements, hyperlink to this information rather than repeating it. This will help to keep your content short, clear and focused.
Focus on user needs
Choosing the right channel
Consider if a statistical bulletin or article is the correct channel for communicating the information you have.
If you have written a lot of detail on how the data were collected during the coronavirus pandemic, it may be better to publish this as a methodology article and link to it. A high-level look at how the ONS as an organisation is responding to the crisis may be better as a blog. If you wish to pay tribute to a colleague or stakeholder who has passed away due to COVID-19, a Reggie article or social media post would be more suitable.
Anticipating questions from users
Consider what questions users may ask about the data. If you expect there will be a lot of questions about the data quality or differences in data sources, then include a short summary in the Measuring the data section.
More detailed information about the method or quality can be published in a separate article and linked to. This will keep your main content short and avoid duplication. GDS advise against using FAQ articles as these can be difficult for users to navigate. Group the information into topics and structure your article around these using clearly labelled section headings.
If you are not sure how to present any additional coronavirus information, email firstname.lastname@example.org.
Next section: Linking between content