Actually I'm pretty handy with anything from basic probability theory through to bayesian statistics, so nyahhhh.
Yet you use two examples of individuals to try and show that a national data set has no worth in producing target grades for pupils. This suggests you really struggle with probability.
This is a good read for anyone interested in educational statistics: [url= http://physicsfocus.org/lies-damned-lies-ofsteds-pseudostatistics/ ]Lies, damned lies and Ofsted's pseudostatistics[/url]
et you use two examples of individuals to try and show that a national data set has no worth in producing target grades for pupils. This suggests you really struggle with probability.
However, if we apply a modicum of ethics to the argument, my point still stands. No one has shown and nor can I find an evidence basis proving the benefit of using these targets/predictions and from what I gather nor do these targets/predictions seem particularly statistically sound.
ethics are involved in statistics? good deflection though.
The targets are a very effective method of making generalisations about what a pupils progress is likely to look like. Very useful to show a coasting pupils what he/she could be capable of. Very useful in giving a teacher an understanding of what a "good" set of results might look like. (as long as schools remember probabilities multiple up they dont add up, but dont get me started on that)
How have you come to the conclusions these targets are not statistically sound?
ethics are involved in statistics? good deflection though.
The targets are a very effective method of making generalisations about what a pupils progress is likely to look like. Very useful to show a coasting pupils what he/she could be capable of. . (as long as schools remember probabilities multiple up they dont add up, but dont get me started on that)
How have you come to the conclusions these targets are not statistically sound?
The statistical work on the part of the educational community, when you dig a little deeper isn't very good though is it? There are numerous criticisms of it.
The targets are a very effective method of making generalisations about what a pupils progress is likely to look like. Very useful to show a coasting pupils what he/she could be capable of.
That or they are good at telling Bright Key Stage 2/3 coasters that they are going to get D's at GCSE level, when in reality they end up with straight A's. This is where ethics come in, this is especially the case when no one has done controlled pilot studies to show whether this actually helps pupils. I can't find any, can you? If someone in the medical community started using practices that hadn't been proven to help people, they would be vilified - meanwhile the educational community seems to be perfectly happy pushing pseudoscience.
Very useful in giving a teacher an understanding of what a "good" set of results might look like
O RLY?
https://www.tes.co.uk/article.aspx?storycode=6006241
Teachers know the basis of the targets, so they know there will be a variation. Clearly, students will have talents on one or more area and so they will achieve a different result.
The statistical work on the part of the educational community, when you dig a little deeper isn't very good though is it? There are numerous criticisms of it.
any specifics of the question in hand?
That or they are good at telling Bright Key Stage 2/3 coasters that they are going to get D's at GCSE level
you still are not grapsing the idea of a normally distributed set of probabilities are you, nor have you grapsed the idea that a target is not fixed and not a prediction.
I can't find any, can you?
I havent looked
If someone in the medical community started using practices that hadn't been proven to help people, they would be vilified - meanwhile the educational community seems to be perfectly happy pushing pseudoscience.
Mr Gove was hated by the educational community for these very reasons
Different meanings to the word "class" 😉
you still are not grapsing the idea of a normally distributed set of probabilities are you, nor have you grapsed the idea that a target is not fixed and not a prediction.
I get normal distribution. What I'm saying is, is that it's utterly unethical to give students these targets/predictions as they are utterly imprecise, if students...those on the right tail.... with a given grade at key stage 3 score higher in their GCSE's and some kids score on the low side of a normal distribution.....how can you possibly justify using this - as giving students predicted grades or targets based on this could affect the outcome of a students grades for the worse.
Well thats a complete change in your argument. You have a number of times said the stats are wrong, how? They are used to inform teaching and give pupils an idea of what their progress might look like if they were to carry on as they are.
Theres nothing wrong with the statistics they are sound. The question is how you use the information they provide. All the points you make are about miss use of the information as is the frankly idiotic link you posted.
Would it be more ethical to ignore a pupils prior attainment because some people lie at the edges of a distribution?
When I was at school they were called predicted grades and so yes, using something as shitty as a normal distribution curve is awful statistics. 'Target' grades are slightly different and I guess a nice way if getting round the fact that a normal distribution would never be used by an actual statistician in predicting something as complex as a childs educational outcome - complex models would be used to generate individual predictions.
If theres no utilitarian value in predicted/target grades as evidenced by randomised trails of some sort, then yes, it's entirely ethical to ignore a childs previous attainment and treat everyone equally.
For someone who professes to know something about statistics and science, I find you ****ing hilarious.
Some good arguments being made here for grammar schools!
What and make even more wild assumptions about future attainment by sending them to grammar schools at ages 10/11 based on vacuous statistics?
Grammar schools effectively fix it at the same stage, thats very different from using them to inform teaching.
why are normal distributions shitty? Its just a correlation.
Irony, Tom, irony....
WGAF anyway (sorry OP) like most standardised crap, best to simply ignore. At least the link to the GCSE grade. Loads of factors there so any deterministic link is bllx by definition. Keeps a few folk in a non-job though and out of harms way, so look on the bright side!
using something as shitty as a normal distribution curve is awful statistics.Â
On average it will have value though wont it.
Obviously a multi variate model would be far better at giving a more accurate prediction but even then it wouldnt work for all. So getting a rough idea and then adjusting up or down as appropriate is fine as far as I can see.
Which brings us back to the use and interpretation of statistics.
http://trudiiemma.wordpress.com/2013/02/03/predicted-grades-help-or-hinder/
If targets don't actually help students and they aren't actually accurate predictions using complex models (if this is even possible), why are we using them in the first place? Again, I'll reiterate - I can't find the evidence supporting their use.
I had some serious issues potentially caused by being wildly under predicted then subsequently attaining good grades - lack of respect for authority, shit poor motivation as a young adult etc etc.
Got bored quite early on but I dont think that link says what you want it to say. It seemed to be about predicted grades for uni's which someone has already tried to explain the differences to you already.
So quack teachers are changing the way they make predictions for GCSE results and A-level results? Whoooo, that's even better.
If a target is set high, it gives the student something to work towards; although it has been [b]shown to increase the level of stress and anxiety[/b] during an exam, actually lowering the student’s performance level. So these high predicted grades may get them an offer into university (when used at A-level), but they can actually hinder a pupil achieving what they are truly capable of. But looking at things from the opposite end of the scale, students being given lower predicted grades have actually been found to achieve grades a lot higher than expected, so having the opposite effect (Snell, Thorpe, Hoskins, & Chevalier, 2008). Pourgonabadi (2008) [b]suggested that self-perception of abilities and their own expectations were key roles in motivation[/b], and provided the basis of their achievements.
Again, until someone produces some kind of controlled study showing some kind of utilitarian benefit to use of target grades - I'll continue to believe that it's snake oil.
Predicted grades are different to targets grades and are produced in an entirely different manner and used in an entirely different way.
If you refuse to listen and take on board what you are told it makes debate impossible. I'm off to argue about englands useless centre options some more.
You can obviously deduce from this that you are entirely right and the education system in the UK is wrong. Long live Manu Tuilagi.
Predicted grades are different to targets grades and are produced in an entirely different manner and used in an entirely different way.
Explain? Different statistical methodology? Or do teachers now use their own 'judgement' to make predictions in comparison to targets?
Both are shit and the use of both should be eliminated unless proven to have some value.
Predicted grades are based on what the kid needs to get a uni offer on the course he wants. Target grades are based on national data sets gained over a number of years and a pupils attainment in standardised tests. Target grades are used to inform planning and giving parents, teachers and pupils an idea of what expected progress might look like if they continue on the same track. ( also used by ofsted as a stick to beat teachers with)
