Hard News by Russell Brown

Read Post

Hard News: Media3: Standards Showdown

57 Responses

First ←Older Page 1 2 3 Newer→ Last

  • Jeremy Eade,

    Well done Keith.

    The problem I have is the circular logic of the journalists. A passionate desire to want to raise education standards yet an inability or refusal to the understand the math of statistics. Those books are quite hard to read, well they look harder than they are. They explain how we can take data and compare groups to confirm or deny a hypothesis and this amazing math has bought us many break throughs, but it is a math and you need to understand the principles of statistical math if you write about data. It is a shock that Keith can present his education and get kind of snubbed by the babble-science of badly recorded data.Data that is not fit for the math of statistics. Statistics is a standard.

    auckland • Since Mar 2008 • 1112 posts Report

  • Jeremy Eade,

    Statistical recording of correlations in health studies is where stats rocks, in comparing how we think, they tend to be less certain ,such complex beings we are. Well that's what they said at the University.

    auckland • Since Mar 2008 • 1112 posts Report

  • Ross Mason,

    Keith: "From the stuff you guys have been doing on class sizes, I don't think you have the capacity to do that. The things you were doing with class sizes was not up to scratch". And then Hartevelt had to gall to then tell Keith that "we have analysed our data and there was a weak association but it was an association".

    HELLO, Keith was being polite. Ummmm...he was dissing you.

    IF you ask the right question, IF you collect the right data, IF you analyse it meaningfully, THEN you might - might - have an opportunity to come up with something which may or may not be the answer you are looking for. Fairfax: Fail. Fail, Fail. Fail.

    50% of kids are failures. But it is OK, 1 in 2 kids are above average and that's because half the kids are below average. But it is OK, the average is close to the countrys median. But the shocking revelation is that half of the whole country is below average intelligence. So it makes sense...doesn't it.

    Oh yes, an admission on Media Watch this morning: Well, yes, one of the most important things to do is to make sure we get the story out first. To hell with the facts!

    These guys have probably got BA's in letters but lacking in the numbers. Write but get it wrong.

    Upper Hutt • Since Jun 2007 • 1590 posts Report

  • Sacha, in reply to Ross Mason,

    And then Hartevelt had to gall to then tell Keith that "we have analysed our data and there was a weak association but it was an association".

    Wasn't it the Herald's Jonathan Milne who said that, as Steve noted above?

    Ak • Since May 2008 • 19745 posts Report

  • Kumara Republic,

    I have a theory: lies, damned lies, and statistical classism.

    The southernmost capital … • Since Nov 2006 • 5446 posts Report

  • Ross Mason,

    Sacha: Corrected.

    Reporters. They all look the same.

    Upper Hutt • Since Jun 2007 • 1590 posts Report

  • Rob Stowell,

    There’s a danger with arguments around the ‘dodgy’ data, poor interpretation, and lack of moderation, because they all imply that solid data, well interpreted would be a fine thing- leading to competition between schools, league tables and performance pay.
    So it was timely to hear NZEI present the other story on Morning Report today. (7 mins streaming)

    Whakaraupo • Since Nov 2006 • 2120 posts Report

  • Sacha, in reply to Rob Stowell,

    So it was timely to hear NZEI present the other story on Morning Report today. (7 mins streaming)

    Interesting. Naturally Parata 'disagrees' with the international evidence while refusing to be interviewed.

    Ta for including the audio length too - figure it helps people decide whether they can commit to listening right now, especially with mobile devices or constrained bandwidth.

    Ak • Since May 2008 • 19745 posts Report

  • Jeremy Eade,

    Kids are hard to study.
    The quickest way to find out how well your child is reading and writing is to have them read something to you or write something. Ask the kid.

    auckland • Since Mar 2008 • 1112 posts Report

  • JonathanM,

    Keith is just way too polite :) Pity he didn't say: I'm sorry, but you guys just don't have the statistical skill to know what is appropriate to do with these data and how to do it. You're misleading the public by applying statistical techniques that are not valid to data that is not amenable to the comparisons you are trying to make.

    With that said, I thought he did a great job of retaining his composure, given the body language from the other two (particularly Mr Milne).

    Since Jul 2012 • 64 posts Report

  • mark taslov, in reply to Jeremy Eade,

    The quickest way to find out how well your child is reading and writing is to have them read something to you or write something. Ask the kid

    That is so (too) fucking reasonable Jeremy that it doesn't meet the histrionics quota for significant media/national/blog outrages. You are damned right of course.

    Te Ika-a-Māui • Since Mar 2008 • 2281 posts Report

  • BenWilson, in reply to mark taslov,

    but yeah, you are damned right.

    Sort of. You can find out what your kids are capable of pretty quickly, but knowing whether that is keeping up with their age group is not something you can do in isolation. Whether that's important is debatable, but I'm always personally interested. I don't get anxious/proud about difference from the mean unless it's too far below/above, but I'm certainly not totally indifferent to how my children compare on a number of measures.

    If I found my ten year old (in 4 years time) had a reading age of 6, that would have me wondering if there was something I could do about it. But I do not, myself, know how to assess reading age, just by getting him to read something.

    My first port of call on the what-to-do front would be professional educators, though. His teacher, probably, who would almost certainly know where he was at, and might also know why, from long observation and experience.

    Auckland • Since Nov 2006 • 10657 posts Report

  • mark taslov, in reply to BenWilson,

    My first port of call on the what-to-do front would be professional educators, though. His teacher, probably, who would almost certainly know where he was at, and might also know why, from long observation and experience.

    Certainly Ben. I recall being given extensive standardised reading comprehension tests on a number of occasions, results passed on to parents. The means to make similar assessments (reading age etc) are also readily available as they're required by parents who home school.

    Te Ika-a-Māui • Since Mar 2008 • 2281 posts Report

  • BenWilson, in reply to mark taslov,

    The means to make similar assessments (reading age etc) are also readily available as they're required by parents who home school.

    Classic, there you go. One of the bands is 6-10 year old. Which means statistically their progress is lost in the noise for several years at a time, but if your child hasn't had any improvement in 4 years of primary schooling, it's not like you wouldn't notice! Unless you weren't looking.

    Auckland • Since Nov 2006 • 10657 posts Report

  • mark taslov, in reply to BenWilson,

    One of the bands is 6-10 year old

    Looking at that test Ben it's very basic vocab recognition exercise and not in anyway a comprehensive indicator, in the test pdf itself the results are perhaps too precise (to the month). Fortunately cursory googling reveals a range of tests. I imagine a financial transaction may be in order for something of quality (more e.g:)

    http://www.gl-assessment.co.uk/products/suffolk-reading-scale
    http://www.hoddertests.co.uk/tfsearch/reading/AccessReading.htm

    Having said that, the international tests we were given back in the eighties seemed more than adequate. and yes:

    Unless you weren’t looking

    Te Ika-a-Māui • Since Mar 2008 • 2281 posts Report

  • mark taslov,

    I thought Keith presented a strong case on the show, well done. Transcription query - Roger Shepherd describing Ben Howe as a vital 'cog' or 'clog' 15:12?

    Te Ika-a-Māui • Since Mar 2008 • 2281 posts Report

  • Jarno van der Linden, in reply to JonathanM,

    Pity he didn't say: I'm sorry, but you guys just don't have the statistical skill to know what is appropriate to do with these data and how to do it.

    But the Herald sent their reporters to an intro to statistics course! So they must know what they're talking about.

    Nelson • Since Oct 2007 • 82 posts Report

  • Russell Brown,

    And here -- a bit late, unfortunately -- we see confirmation that the design and/or interpretation of the standards is a shambles:

    High error rate in National Standards marking

    Sheeting the blame home to teachers making "errors" is, of course, rather missing the point. This doesn't happen with properly moderated assessments.

    Auckland • Since Nov 2006 • 22850 posts Report

  • Kyle Matthews,

    I read that story this morning and listening to John Hartevelt on the mediawatch podcast on the way to work.

    It really is looking like a system designed badly enough to be an argument for national standards testing. Oh the reason this data is no good is because it's based on individual teacher assessments so it's not very useful. We should have a national test that all children sit each year...

    Since Nov 2006 • 6243 posts Report

  • linger,

    Milne’s response to having statistical errors pointed out was a classic example of the Dunning-Kruger effect in action. The original paper (Kruger & Dunning 1999) notes that those with insufficient competence in some field ipso facto cannot recognise and respond to corrections in that field, but instead may interpret any such corrections as irrelevant personal attacks. Kruger & Dunning give the example of students who “explain” their low test scores as indicating that their teacher hates them.

    Tokyo • Since Apr 2007 • 1944 posts Report

  • Russell Brown, in reply to Kyle Matthews,

    It really is looking like a system designed badly enough to be an argument for national standards testing. Oh the reason this data is no good is because it’s based on individual teacher assessments so it’s not very useful. We should have a national test that all children sit each year…

    I'm still on the side of cock-up over conspiracy, but ... yeah.

    Auckland • Since Nov 2006 • 22850 posts Report

  • Martin Lindberg, in reply to Kyle Matthews,

    Oh the reason this data is no good is because it's based on individual teacher assessments so it's not very useful. We should have a national test that all children sit each year...

    If we use Finland as a shining example, then no. Teachers set grades based on their own assessments. But that's part of their training, of course.

    Stockholm • Since Jul 2009 • 802 posts Report

  • Ross Mason,

    Checkpoint monday night: (18 minutes)

    Chat to a Finnish " Expert on Education Reform" who spoke at Teachers conference in Rotorua. Finland: No Nat standards. Education system based on equity not achievement. Controlled at the municipality level......

    Finishes with:

    "Finland lacks the extreme excellencies and the extreme poor schools."

    Katherine: "Everyone is kind of average."

    " That's what makes the Finnish taxpayer very happy where the education and schooling of your children is never a topic at the dinner table."

    Upper Hutt • Since Jun 2007 • 1590 posts Report

  • Matthew Poole, in reply to Kyle Matthews,

    We should have a national test that all children sit each year…

    There used to be something similar. I have strong memories of taking written language and maths aptitude tests at the beginning of at least two years of my primary schooling, using standardised-format answer sheets. Someone out there will know the ones I mean. Are those no longer used?

    Auckland • Since Mar 2007 • 4097 posts Report

  • Sacha, in reply to Matthew Poole,

    PAT tests?

    Ak • Since May 2008 • 19745 posts Report

First ←Older Page 1 2 3 Newer→ Last

Post your response…

This topic is closed.