So far in our xAPI Governance series, we’ve looked at what you can do to protect against bad xAPI data and how you can fix it. This week, we explore how using a flexible learning analytics platform can help you work with imperfect data, reducing the impact of xAPI Governance failures.
Now, before I start, let me be clear: You shouldn’t ever rely solely on your learning analytics platform to maintain good xAPI data. You need good governance as well. Keeping your data clean and consistent is always the best approach, but it doesn’t hurt to have a plan B. For instance, a flexible LAP can act as that plan B when:
- data issues slip through the net; or
- you need to work with a data source that can’t conform to your governance rules.
The most important feature of a flexible, xAPI-conformant LAP is its ability to work with any xAPI data, regardless of the identifiers used and the structure of extensions. While an LAP might provide reports designed for specific use cases and types of data, an LAP also needs to be flexible enough to work with the broad mix of learning data that xAPI can support. Otherwise, you might only be able to partially report on data sources—or worse, you might not be able to report on some data sources at all.
If your LAP can't communicate with a variety of xAPI data, you risk incomplete reports.
A flexible LAP will enable you to filter and organize reports by any verb and activity type as well as report on any data type held in an extension. If your LAP requires a particular data structure for reports to work, then it’s unlikely your LAP will be able to support all your requirements—especially as your xAPI implementation grows.
In an ideal world, all your xAPI data will use a single persona identifier that’s unique to each learner and never re-used for a different learner. In practice, organizations might often use systems that assign different identifiers for learners in addition to a central system that maps those various identifiers together.
So where does that leave your xAPI data that uses multiple identifiers for each person? And how can you join that data together?
We'd explain how this image relates to the risks of using multiple persona identifiers for one learner, but we don't talk about it.
A flexible LAP will include functionality to alias all these persona identifiers under a single person so you can report on the data together. Watershed, for example, can import a mapping of identifiers as a CSV file (normally from an HR system) in order to link the data in this way.
Occasionally we see data where multiple activity IDs have been used for what should be a single xAPI activity (e.g. activity IDs that use different capitalization are considered totally separate, even if they’re used for the same activity). We also see cases when multiple verbs with effectively the same meaning are used to describe multiple actions. Using multiple activity IDs to refer to one activity or several verbs to describe one thing can make reporting difficult. That’s because you really want to see these appear as a single thing rather than multiple items.
A flexible LAP will include functionality to treat multiple activity IDs as a single activity as well as multiple verbs as a single verb. The LAP shouldn’t rely on changes to the underlying xAPI data, which is immutable; however, from a reporting perspective it should be as though the underlying data had been changed to merge the identifiers.
A common problem when SCORM data is converted to xAPI is that it only includes identifiers—while leaving out names, descriptions, activity types, or other activity definition metadata. Data from other sources also may include activity definitions with missing or misleading data.
A flexible LAP includes functionality to edit activity definitions. xAPI defines LRS functionality to maintain a canonical definition of each activity ID. The ability to edit that canonical definition, which is what should be used in reports, enables you to fill in any blanks and correct errors in those definitions.
EXAMPLE: Watershed allows you to edit an activity's name, type, and description.
The ability to edit activity type and verb names can also be useful.
WHAT DOES "CANONICAL" MEAN AND WHY IS IT IMPORTANT?
The canonical format requests the LRS to return its own internal definitions of objects, rather than those provided in the statement.
If you trust the LRS, this is normally the most appropriate format when the data will be displayed to the end user. The LRS will build its internal definitions of objects based on statements it receives and other authoritative sources.
That’s all folks!
And this brings us to the end of our series on xAPI Governance. As a reminder, here’s a summary of the series and how to do xAPI Governance well.
- What's xAPI Governance?
- Have processes and rules.
- Document your xAPI statements.
- Test, monitor, and enforce.
- Be ready to fix or remove bad data.
- Use a flexible Learning Analytics Platform.
NOTE: Fight Club is used here only to illustrate the examples in this blog post. Watershed is not associated with, sponsored by, or affiliated with Fox 2000 Pictures or Regency Enterprises.
About the author
As one of the authors of xAPI, Andrew Downes has years of expertise in data-driven learning design. With a background in instructional design and development, he’s well versed in creating learning experiences and platforms in corporate and academic environments.
Subscribe to our blog