Go easy on the local kids, y'hear?

As is well known, Australian students in grades 3, 5, 7, and 9 will be sitting their first nationwide test on 13-15 May 2008.

But something I haven’t seen mentioned in the press is that the tests will be marked at a state/territory level. Given that one purpose of national testing is to compare across states and territories, does this strike anyone else as sub-optimal?

Advertisements
This entry was posted in Australian Politics, Economics of Education. Bookmark the permalink.

8 Responses to Go easy on the local kids, y'hear?

  1. LuckyPhil says:

    I know somebody who was involved in marking the tests last year, when it was found that some questions were getting a high fail rate they changed the marking criteria.

  2. conrad says:

    Thats what they should have done LuckyPhil. If you have questions almost everyone fails, then they don’t distinguish between people well, and are therefore not useful (aside from knowing what the upper bound of performance is).

  3. derrida derider says:

    Andrew, marking has to be done by qualified, experienced people – that is, teachers. It seems to have escaped your notice that the Commonwealth does not employ significant numbers of primary or high school teachers.

    Of course the marking has to be done by the states – who else can do it?

  4. Damien Eldridge says:

    I doubt very much that bias is likely to be a big problem here. However, there are ways in which it could be avoided without creating additional resource burdens for any state or territory. Perhaps each test paper should be allocated a unique ID and then randomly distributed to students. The completed papers could then be combined into one group and randomly allocated to states for grading. That way, so long as there are no state identifiers on the paper, none of the graders would know which state the student came from. If each state recieved the same number of papers for grading as they had students sitting the exam, the resource burden would be the same. Of course, there would be some additional resources required to implement the randomisation processes and to compile the final results and distribute them to the states and territories.

  5. Damien Eldridge says:

    Given the final sentence of my previous comment, I guess the second sentence of that comment should have said: “However, there are ways in which it could be avoided, hopefully without creating too much of an additional resource burden for any state or territory.”.

  6. conrad says:

    I have an alternative way to save time Damien — just use multiple choice questions that can be marked automatically. If the results don’t need to be given back and you really must give non-multi choice, then you could also use automatic marking. The reliability is as high as human markers. The main problem is that students don’t like it, which isn’t a problem if they don’t need the results.

  7. Andrew Leigh says:

    DD, we already have nationwide tests that are marked centrally. They’re the LSAY tests, marked by ACER. The only difference is that less than 100% of the cohort participates (I think it’s 5-10%). A similar approach could apply here – NSW marks grade 3 tests, Vic marks grade 5, Qld marks grade 7, and we move money around to pay for it. Assuming the markers can’t distinguish their own state’s papers, that would solve the problem.

  8. Pingback: National Student Testing « David Mathews

Comments are closed.