1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.
  2. Hi Guest, welcome to the TES Community!

    Connect with like-minded education professionals and have your say on the issues that matter to you.

    Don't forget to look at the how to guide.

    Dismiss Notice

How do you boost the accuracy of GCSE predictions?

Discussion in 'Senior Leadership Team' started by thearns, Aug 17, 2015.

  1. Greetings everyone,

    The link below takes you to some recent analysis carried out on GCSE predictions in 2014 across OCR exams (680,000 grades). Here's what I found interesting - 43.6% of grades were spot on, and teachers find it easier to predict higher grades (53% accuracy for A*-C) compared with lower grades (20% accuracy below C).


    I'm starting a new job in September: Raising Standards Leader (KS4 focus). I'd be grateful if any school leaders out there could share strategies you have implemented in your schools that improved the accuracy of forecast grades. Any schools out there that regularly achieve over 50% accuracy?

    Many thanks

  2. Hi

    I'm also interested in this area and am concerned at the lack of research available. I always would doubt the forecast sent to the boards as very optimistic. When staff are asked for realistic predictions I have always expected 80-90% accuracy from a dept as rule of thumb As I look through results over the last couple of years in several school and across several subjects this figure has rarely been reached and I'm seeing closer to 50% as the norm. The need to get this right has never been greater yet it seems to be getting worse-does anyone have any more detailed, realistic research?
  3. ValentinoRossi

    ValentinoRossi Star commenter

    "Research" generalises and is, more often than not, biased to the expectations of the researcher.

    1. Employ leaders who know what they are doing.

    2. Employ teachers with excellent subject knowledge and good teaching skills.

    3. Put in place a reliable, consistent and workable Tracking Progress System.
  4. Skeoch

    Skeoch Star commenter

    Depends on what you mean by predicted grades....

    The grades submitted to the Awarding Bodies are likely to be a bit political. There's a remote chance that the grades might be used to support a Special Circs application, so there is some motivation to err on the side of generosity.

    Grades submitted internally and not released to candidates might be expected to be more accurate - though again coloured by what the teacher thinks SLT/HoD will use them for.

    Grades given to candidates might again be coloured by what the teacher thinks the response might be!

    We have used two predicted grades, a "most likely" and a "best you are likely to get," internally and this has been well understood by parents and candidates; the "best" is a motivator for many.

    There's no question in my mind that a good tracking system (whether formalised or just done well by an individual teacher) does a lot for achieving accurate predictions. At the other end of the process, HoDs and individual teachers need to be prepared to explain why their predictions were wrong, if they were adrift.

    And despite all that, and VR's wise comments above, I suspect we've all wished for a "Your guess is as good as mine" tick-box for some wayward candidates!
  5. Scintillant

    Scintillant Star commenter

    Can you point out where the bias is in the research in question please?
  6. Scintillant

    Scintillant Star commenter

    If someone could make these "reliable" or "good" tracking systems available nationwide, then the accuracy of predicted grades might jump from "worse than guessing" to around 95%.

    Maybe. Just a thought.
  7. MrMedia

    MrMedia Star commenter

    If you predict grades below the department's targets then you invite micromanagment by SLT. If you hedge your predictions so that collectively they look like they will be ok then you stave off the micromanagement until the real grades. Then, if you pull out the grades needed you can start the cycle of avoiding bullying all over again.
    There is no incentive to accurately predict GCSE grades.
  8. briggs1209

    briggs1209 New commenter

    This is something of a pet project of mine and I think you might have to be happy with a success rate of 40 to 50%.

    I would echo your findings that more able pupils are easier to predict - they work harder in the build up to the mocks and are generally more consistent in their studies over a two year period.

    Equally as the poster above says there are clear motivations (in some schools) to tell SLT what they want to hear to avoid micromanagement and constant scrutiny.

    However, even when there was no pressure the departments I have ran only achieved a success rate of around 40 to 45%. I found it most useful to focus on eliminating the shock results where the predictions were out by two grades.

    Interestingly I found that where multiple staff taught towards a single certificate (e.g. three science teachers contributing towards double-award science) there combined predicted grade was no better than the efforts on single teachers.

    Obviously in subjects where coursework played a greater role there was more accuracy.

Share This Page