Charles Baird, data architect at the Data Standards Authority, and Phil Richards, chief innovation officer for data analytics at Jisc, discuss how data standards are developing across government, and their impact on higher education.
Making sense of data standards
“Without standards, the internet doesn’t work,”
says Charles Baird, data architect at the Data Standards Authority (DSA).
“Email, messaging, so many things we take for granted couldn't happen without standards to allow them to work with different platforms. And the same goes for data.”
So how do data standards affect education? The Higher Education Statistics Agency (HESA) defines data standards as helping to ‘improve consistency across disparate processes or sources, to reduce the costs of integration between sources, and to manage data quality’.
“It’s relatively early days for standards across the data landscape,”
“We have realised that most organisations have issues with data management and interoperability because there have historically been no external bodies to set standards for data use. Therefore, things like email clients and database CSV files are implemented in lots of different ways, potentially within one organisation.”
Essentially, without standardisation of data, it’s impossible to make it shareable and interoperable.
There are two angles from which to look at data standards as they currently stand in higher education (HE), says Phil Richards, chief innovation officer, data analytics, at Jisc.
“One angle is to consider data standards from an administrative and regulatory point of view, and the other is from an academic or student support perspective. The two angles are complementary, and in future they’re likely to be considered as two sides of the same coin, but for now there is some merit in considering them separately.”
Diverging regulatory approaches
In terms of a regulatory or administrative perspective, says Richards, the UK is in a “really strong position”. This is largely because of the work HESA has done over the years, he says, including developing “a well-established data model that allows meaningful year-on-year comparisons”.
However, there are challenges to maintaining a cohesive approach to data standards across the UK. Richards explains:
“Policy for higher education is devolved to the host nation governments, meaning there are potential emerging variations in approach. For instance, divergence of frequency of regulatory returns between England and Scotland is now a possibility because of the Department for Education (DfE)’s burden review, launched in September 2020.
"This pointed to a potential consolidation of a single, annual collection in England; while in Scotland the intention appears to be to move to multiple in-year collections, which may better fit the more flexible education model we’re increasingly seeing.”
Because the different nations – including Wales and Northern Ireland - have their own emerging positions, it may become increasingly difficult to achieve one UK-wide data standard, and reporting, for HE.
Putting students first
The UK is also in a strong position from the perspective of academic, or student, analytics, says Richards.
“The UK is one of the leading adopters of learning analytics, along with other international leaders such as the US and Australia. We’re also in a unique position because the UK has a national learning analytics service, run by Jisc.”
The learning analytics service is based heavily on standards, says Richards, and approaches two kinds of data.
One set is similar to administrative data, and includes things like assessment data, grades, attendance, etc, focusing on processing information more closely to real-time rather than “keeping all the data in a spreadsheet and uploading them into the system manually at the end of the semester,” explains Richards. The advantage of gathering this data in near real-time “allows insight to be inferred about how students are progressing,” while there is still time to act on that insight, and improve student outcomes before the semester is “done and dusted”.
The other set of data is what Richards describes as “digital exhaust” - the digital footprint that students leave as they navigate through various virtual learning environments, events, video conferences, library systems, and other digital resources.
This data is captured using one of two common data standards: Tin Can, otherwise known as xAPI, and Caliper. The difference between the two systems is becoming academic, says Richards, as they correspond closely and are essentially co-evolving. While for historical reasons the Jisc learning analytics service uses Tin Can, near-real-time mining of patterns in data stored via either standard provides “the maximum opportunity to make interventions, if appropriate, or to advise on improvements to learning styles,” says Richards.
This speaks to the primary point of data standards in this context, as Richards sees it, which is to better support student learning journeys and outcomes.
Driving value and insights
The government’s National Data Strategy, released on 9 September 2020, says that the coronavirus pandemic highlighted ‘massive untapped potential in the way government and public services use and share data’. It proposes, in the third of its five missions, that ‘to succeed, we need a whole-government approach that ensures alignment around the best practice and standards needs to drive value and insights from data’.
But what does this mean for education? An improved government process has far-reaching impact, says Baird.
“If we improve how the DfE works, for example, that will impact HE.”
A streamlined government process for data sharing is the driving force behind the DSA’s new application programming interfaces (API) catalogue.
“APIs aren’t standards per se, but a technology that allows two different applications to talk to each other. They’re the primary way that data is transferred between systems, and we’re trying to streamline their use within central government,”
APIs could be a more granular way of uploading micro-snippets of data returns, rather than having to do a “job lot”, suggests Richards, which might make data returns for universities more flexible. However, it could also make the process more complicated, he warns, as “you have to make sure that there is continuity between the different micro-returns, and it can be easier to check the integrity of a small number of big returns, rather than a large number of small ones”.
At the end of the day, “data sharing shouldn’t be as hard as it is,” says Baird.
“And the reason that it is so hard is that there aren’t always data standards in place, and where there are, they’re not always adhered to”.
If APIs can help streamline government, they should be able to do the same for HE, Baird reasons.
“This situation is not unique to HE, but in a landscape with a number of large, discrete organisations – such as universities – that each have their own individual corporate cultures, but do sometimes collaborate and work as a sector, data standards for sharing are key.”
In this context, the important thing is the benefit to students, concludes Richards.
“Initiatives come and go, and we need to judge them by evidence of their outcomes. So, while this is an appealing concept, with real potential, it remains to be seen what something like the API catalogue could do for education, specifically”.