OnPoint by Keith Ng


Because Statistical Rigour

Hey John Hartevelt. Let me get this straight. A lot of people told you not to published this data. Some of them were dicks to you, others gave you very specific reasons:

Anyone who read the National Standards results as a proxy for quality would be quite foolish. We wouldn't do that and we don't suggest you do, either. For starters, they are not moderated, so one school's "well below" may be another's "at" or "above". There is just no way of knowing - yet - exactly how the standards have been applied across schools.

But even if they were moderated, the standards alone could not tell you everything about how a school is doing by its pupils. As many of the experts we canvassed for this project have noted, quality is most evident in what a school does to push its pupils up, not in how well they do at attracting the brainiest, most-privileged kids in the first place.

These are very specific reasons, and not mere technicalities. So, as you asked, why publish the data at all?

Our critics have already suggested this is a "business decision". An official in the Education Minister's office charged that it was "solely aimed at gazumping" the Government's own website. Both accusations reflect the bias of their authors - and both are wrong. Of course we want people to look at what we have published here; to talk about it and to debate it. But that does not mean our decision to publish National Standards data was a "business decision". This project has been led by journalists from the beginning. That has made it subject to our own standards of journalistic rigour. We have not simply dumped all of the new National Standards data online.

"Because journalistic rigour"? That's not a defence, John. You can have the best intentions, but that doesn't change the facts.

Are students Ahuroa School better than those at Abbotsford School because the former has 12.8% below/well-below standard, while the latter has 15%? It is not possible to know because, as you say, the data is not moderated.

Are 12.8% of students at Ahuroa School unable to read adequately? Are 15% at Abbotsford School unable to read adequately? It is not possible to know because, as you say, Ahuroa's "well below" might be Abbotsford's "above".

Is Ahuroa School a better school than Abbotsford School? Are either of them good schools? It is not possible to know because, for the reasons you mentioned, anyone who use these results as a proxy for quality would be quite foolish.

Let me put a question to you. With all these caveats, what *is* the right way to use the data?

You have missed the point of all those people pleading with you not to release the data. It's not that people like me think people like you are unable to draw the right conclusions from the data. It is that, if subjected to standards of statistical rigour, there are aren't many right conclusions that can be drawn by anyone. Maybe longitudinal analysis over time. Maybe looking at demographic differences between ethnicities and gender. But certainly, very few things at a whole-of-system level.

You suggest your readers are smart enough to gain the right insight from this data. They are not. *Nobody* can get the right insight from this data because it's not there.

Also: I think it's pretty disingenuous to say that the data isn't moderated so schools aren't directly comparable, then feature a school comparison tool as the main way of navigating the data. Just sayin'.

Also: To be fair, data side notwithstanding, their reporting is really good.

77 responses to this post

First ←Older Page 1 2 3 4 Newer→ Last