Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

AVETMISS address changes in version 7

ariari
edited March 2013 in AVETMISS
It has been just announced that AVETMISS 7 starting in 2014 will have the student address changed. The following fields will be replaced:

- Address first line
- Address second line

with:

- Building/property name
- Unit/flat details
- Street number
- Street name

At first glance, this seems like a trivial change, but as we submitted in our response to NCVER earlier this year, the change is really quite absurd.

Firstly let's look at WHY this change is being introduced. NCVER and their government bosses want to be able to uniquely identify every student and to geolocate that student. This makes it easier for them to correlate the data against census data and databases like MOSAIC. This allows them to determine household income, ethnic background, religion and much more. This can be correlated against student study choices and all the information you provide in AVETMISS.

Why is this so absurd?

Privacy

Completely gone is the pretence that the NAT85 file was optional and that student details were somehow scrambled before being part of the national collection. Every student is uniquely identified in the results and students should be informed of that fact. Not everyone feels comfortable with the fact that every bit of study you do should be recorded and cross-referenced against your social and political choices.

USI

You would have thought that the introduction of yet another layer of red-tape, the Unique Student Identifier, would eliminate the need for increasing the complexity of AVETMISS. NCVER's response to me on this question was that they have no access to the USI data and so they need to effectively collect the same data separately. So in 2014 we move to having two separate mechanisms to uniquely identify individual students.

The RTO as a data analysis expert

Is splitting up these fields really so hard? Surely "4/30 Wilson St" is simple. Put the "4" into the unit field. The "30" into the street number. And the "Wilson St" into the street name. We could even write a computer algorithm to do this for us. But if it were that simple, the people at NCVER could do it themselves. Let's take the address that I put on all our business correspondence: "level one, 30 Wilson St". That is the address that causes the postoffice to put mail in the right mailbox. But is "level one" a building name? Do I put it in the unit field?

Take a look at this US guide for properly encoding address details. This is really complicated stuff.

NCVER are asking the RTO (or the student!) to develop enough expertise in data analysis in order to put the right thing in the right place. There will undoubtedly be a set of arbitrary validation rules which we'll all struggle to comply with. Can a street number have letters? How about a unit? Can you have a building name as well as a street number?

If we consult the QAS software from Experian, I find that "level one, 30 Wilson St" should actually be entered as "Unit 4, 30-34 Wilson St". Does that get my post delivered correctly? Will NCVER require that every RTO pass all their address details through a $5000 piece of software in order to pass validation?

In short, NCVER are pushing their inability to properly analyse the data back onto students and RTOs. Just students are not allowed to be born in East Germany or Yugoslavia according to NCVER rules, they will only be allowed to live at "proper" street addresses regardless of whether that is where the student believes they live and whether Australia Post deliver mail there. Street address is being used as a surrogate for "student identifier" and all the extra work is being pushed onto RTOs.

What onCourse will do

For now we are writing an algorithm which will split up the address details into separate pieces for the AVETMISS export. It isn't simple, but we believe our loyalties must lie with the RTO. Our approach means zero extra work for you or for the student, since every hurdle you put in front of online signup reduces your enrolments. It isn't going to magically fix "level one, 30 Wilson St" but it will pass the validation and everything after that is NCVER's problem, not ours and not yours.

Comments

  • One would have thought the Australia Post Addressing Standards document (http://auspost.com.au/media/documents/australia-post-addressing-standards-1999.pdf) might have been a good starting point, but it seems even that has been ignored. At least ish are taking the initiative to help split existing addresses, albeit to very limiting standards imposed by NCVER. Sigh.
  • You are quite right about the lack of clarity in the change - both the justification for the change and the transition/conversion arrangements.
    They may need to back out this change as I can see nothing but a data quality disaster ahead. We would be interested in collaborating on a conversion routine (if it comes to that).
  • Some time ago I thought about writing a little open source application which all software providers in the sector could contribute to. It would specify a clearly documented XML schema (the way AVETMISS should have been written) and convert it to 15 flavours of pseudo-AVETMISS nonsense.

    If nothing else it would be satisfying to show NCVER how standards are supposed to be written, but my free time is spent these days on anything but AVTEMISS. It would be more fun to poke my eyes out with pointy sticks...

    onCourse 4.1 now supports a full AVETMISS 7 export (and the South Australian variant... the only state to have their documentation together so far). This is very tedious work when there are so many more exciting things to be building into our software. And imagine if they gave the $21m it is costing to implement just the USI over the next four years toward, oh, training.

    I recently found the explanatory memorandum for the legislation. It is interesting in its analysis of the relation to privacy and right to education provisions in our international treaties. But also interesting is this fabrication:

    "An individual’s identity and contact details will be held by the Agency, separate from their VET activity records held by the NCVER in its National VET Provider collection. By storing these records separately, it will not be possible to identify an individual by their student identifier alone and will therefore mean that there is greater protection of the individual’s privacy."

    NCVER (and every state reporting body) will hold records for all training with full identifying data for each student. And the USI Agency will hold both identifying information and VET records. So where is this division of data which ensures our privacy? There would have been many ways to structure the data so that true privacy was created while still meeting the objective of reducing government funding to students who study too much (really, that is the whole point here). Zero knowledge proofs are a way to store data securely and separately but still be able to answer questions like "have I already completed a certificate IV"?

Sign In or Register to comment.