TAG Meeting 2020-09-14 - Subgroup on Certification Improvements 2
Participants
Doug Q
Rosh D
John R
Jean-Francois G
David Clements, Ed-Fi Alliance
Eric J
Notes
The discussion picked up from the previous meeting: TAG Meeting 2020-08-10 - Subgroup on Certification Improvements
It was noted that Ed-Fi support by SIS systems and demand for data has a strong state flavor to it, as state SIS editions often build in specific fields needed for state compliance, and LEA requests for data (via Ed-Fi) often reflect those state-specific fields.
However, LEA requests for data do not always reflect those fields: sometimes districts are looking for data from the "core" product tables, and not the state-compliance tables.
This problem is further complicated by the fact that the SIS may have data "duplicated" in the state compliance tables. This "duplication" may not be exact - the semantics and values may differ for a variety of reasons.
The group also acknowledged that there is still the problem of LEAs asking for data not defined in Ed-Fi's specifications - i.e. seeing Ed-Fi as a way to get "all my data" or "any data I want"
The Alliance further noted that it has had some success in driving state data specifications to more closely conform to Ed-Fi's specs - the Alliance annually reviews state data specifications, and makes specific, itemized requests to SEAs to conform more closely in their specifications. These are mostly but not always honored.
Despite these problems, the group saw value in seeing certification more closely allied with state specifications and SIS support for state editions, and in certification surfacing that support publicly.
State-based focus is still strongly defining SIS Ed-Fi development priorities
Surfacing state-based supports and state-based info creates regional and state-based opportunities for the state and LEA to have collective data governance
How would this be operationalized? The method discussed was to surface state-conduced certifications. in the Ed-Fi certification. The Alliance would continue to run a certification, but would also publish state-certifications. The action taken was to create a visual sample for further discussion.
There was also some discussion of the current Ed-Fi certification, and if it's format or purpose should change? The general direction seemed to be that it should not - the current practice of testing vendor API interaction/support, enforcing best practice in vendor implementations, and testing for a default support for domain-based data representation by SIS vendors was seen as valuable.
The group briefly discussed 2 other topics:
Possibly breaking certification down into series of narrower certifications. It was noted that this has been a problem for previous standards and was also rejected by the Ed-Fi GAT for that reason.
Harvesting "best practices" that are tested from state certifications.