The league tables are out - let's talk Progress 8

Alexander, what's the confidence interval?

So as a parent of e.g. a low prior attainer, do you go for the school with an 'Average' Progress 8 of 0.79 (average because the very lowest end of is confidence interval is 0.01, though the highest is over 1.6) or the 'Well above average' one of 0.69, with a confidence interval of c. 0.2 to 1.1???

There’s a blog somewhere that explains it in detail but:

The system is based on the assumption that children with the same starting point have the same trajectories. But children in the lowest attainment bands at ks1 consist of mainly 2 groups - children with SEN and children with EAL. these groups have very different trajectories but the calculation treats them as being the same. When their ks2 scores are averaged out children with EAL end up with huge positive progress scores and children with SEN huge negative scores (double figures in plenty of cases).

I’d assumed that it might even out by the time p8 was calculated because there would be fewer EAL children in the lower attaining groups, but from your OP that might not be the case.

Wow I'm now so glad that I didn't send Ds to our catchment school. Even the awful private school he went to for 2 years before moving to his current school was a better choice.

Our catchment school is in the worst 100. It actually does very well with kids who enter with less that a SATS level 4 but poorly with average attainers & terribly with high attainers those entering with level 5 & above. (The group Ds would have been in). Out of this group the progress 8 was -1.9 & only 50% of them got a Level 5 in maths & English.

The rather scary thing is, a range of 1.97 isn't even particularly rare.

As the cut off point for not showing the results is 5 pupils, if you can find any subgroup with only 6 -7 in it (disadvantaged pupils in grammar schools are pretty fertile ground for huge confidence intervals, I have found so far), odds are that the confidence interval is well over 1.5, with 1.97 being really quite common.

Funnily enough, the score is thus 'average' because it pretty much can't fail to include 0 within the range....

I've just found a confidence interval of 1.97....

the Problem’s going to be parents thinking they understand it and making judgements about schools, isn’t it? In theory the system looks simple and easy to understand but it’s actually not because two schools with different scores could be similar.

noble are you aware of the cock up with the ks2 progress measure when it comes to children with EAL/SEN?

Just answered my own question - yes, that happens all the time. Ordering the schools in my county for P8 for sub-groups has a school with a small cohort of the specific subgroup listed as 'Average' for a P8 of about 0.8, while one with lots of them and a P8 of around 0.6 is listed as 'Well above average'.

Nowt so odd as statistics...

Oooh, thanks Rafa! That makes sense. So the different figures that fall within the 'average' range for the sub-groups will also be to do with the confidence intervals of those.

Does it also mean that a higher 'absolute' figure can be 'average' (if the cohort size is very small and results scattered, so confidence interval wide) while a lower one with a smaller confidence interval be 'Above average'?

No, thankfully! But they were told there was no way we could keep our Outstanding.

Phew. Was a bit worried it would be RI from your emphatic response!

No Madonna we went to good . So this puts us with every other single school for a bout a 40 mile radius! It is too early to say whether it will affect us. the local feeder school is now an RI so that may have more impact in this area of ridiculous house prices.

The LA is in the bottom something % of some sort of results ... we're a bit meh al round! I went over the border to another county for an interview last year and I couldn't believe how much more modern and thrusting it all seemed!

What was interesting was that the definition of 'average' moved, in a way I wasn't expecting. The reporting at prior attainment level is subtle enough to compensate for the fact that lower prior attainers have higher P8 in general, so the 'numerical P8' which got the school 'average' for lower attainers was significantly higher than the P8 that got them 'above average' for the higher attainers.

If this is the same as the primary figures, then this is to do with confidence intervals. The main figure given is the mean, but there’s also a confidence interval. If that stretches across zero then the result is average. If both boundaries of the confidence interval are above 0 then the result will be above/well above average. And below/well below for a confidence interval where both boundaries are below zero.

And I hope it wasn't a drastic drop down the grading and you aren't facing the pressures that brings.

Do you think it will have a knock on effect on reputation Piggy? I wonder how much weight parents place on P8.

Having seen this thread I've just looked at our local school for updated information

There is none.

Why wouldn't the progress 8 breakdown be published?

Comps in the same town as our grammars(seaside town,in lowest 20% for deprivation)have done well( average,1 higher than average). Both grammar schools boys and girls single sex are above average progress 8.One comp in the same town with higher than average fsm scored 1 point higher in progress 8 than the highest scoring grammar.

Further away in our main city the comp with the higher number of fsm has done really well scoring higher than all the more popular comps with less.

Can you really put much stock in intake as regards progress 8. Life away from London is completely different,surely you can't judge a whole table from the top schools of which most will be in London or the Home Counties. The rest of the country is very different.

IGCSEs in subjects other than Maths and English did count this year. And the more points for moving an A to an A* than an F to an E will not work next year, because the vast majority of GCSEs are now reformed (9-1).

Ah, yes, who is excluded:

Everyone without a KS2 result.

So anyone at a private primary [over 25% at some grammars], anyone moving from abroad......

On a slight schadenfreude note, there are lots of the government's beloved UTCs on the Mirror's hateful list.