Explanation of anomalous results

[Lysander]

Using the scores posted on this website, I have observed the following:

1. If you sort by Crypt scores, the scores for HSFG, Marling and SHS are in ascending order. I think we can assume from this that these schools ranked pupils by their total score and did not insist on passes for individual sections.
2. If you sort by Pates scores, the scores for Ribston are in ascending order. We can assume from this that Ribston has introduced passmarks for each section.
3. I cannot make full sense of the STRS score. Sorting by Crypt score gives me two outliers: both boys who ranked in the 800s at Pates (MonaB and hmc21). Sorting by Pates score makes one boy an outlier (Glos18). If these scores are all correct then the only hypothesis is that STRS has introduced pass marks for the individual sections, but these allow more of a variation in abilities than Pates, so they are not accepting children with a huge variation in abilities between the different subjects, but are more tolerant than Pates. I would advise these three families to contact the admissions folk to check their scores, explaining the anomaly.
4. I think Pupil Premium is a red herring here: the admissions policies imply that the pupil premium is used to sort applicants once they have been ranked.

Thanks to Jearund for sending spreadsheet details

[Glos18]

I think you have cracked it.

Potential add on theory: I think what pates have done is ranked all the people who passed in 4 areas first, then all those who passed in 3 areas next etc. So presumably someone who otherwise did well but is down in the 800s for pates is probably the best person to pass only 2 sections? Maybe STRS have a system where it is ok if you fail one section but not 2, so you become the best person to fail 2 sections in the 400s. Don't know if that makes any sense???

[muminTewkes]

That’s really interesting, and I would say sits with my DS who I knew one area would be his iffy one where as he was always strong otherwise.

[surfer]

Lysander, do you think Crypt ranked in simple ascending order too (with no pass marks for individual sections)?

[Lysander]

Surfer - yes: this explains why you can have a Crypt ranking in the 100s but a Pates ranking in the 800s: Crypt and a number of other schools simply count the number of marks whereas for Pates you need to get past a certain threshold for each individual section.

[Amber]

I guess someone could just ask STR if they have introduced a 'pass every section' policy this year. I imagine they would be happy to answer.

[cazien]

Surely if STR are using a "pass every section" section they would have to include this in their admissions policy like Pates have?

Looking at the admissions policy on their website, I congratulate them including the last sentence addressing results from cross-county applicants.

"The test consists of multiple choice style questions on verbal ability (e.g. comprehension, vocabulary and verbal reasoning), numerical reasoning and non-verbal reasoning. There are two papers and the standardised marks are added together to give a total score for each applicant. The total score is used to place applicants in a rank order by which places are allocated. This is not a pre-defined pass mark, but reflects a child’s position in the rank order of standardised scores in the Admissions Test.

Any student registering in two authorities (including Gloucestershire) that share the same test, and attempting to sit a late test in Gloucestershire, will have the first sitting taken as their test score".

[stroudydad]

how does that explain my sons score of top 150 for Marling, Pates, Tommies but a lot lower for crypt? Unless i suppose as you said Pates and Tommies have the section requirements, but Crypt dont, and Marling didnt get anywhere near as many shares. So whilst he didnt have the highest aggregate (so lower at crypt), he did well enough on all sections to rank high

[Amber]

It says 'the standardised scores are added together' and 'the total score is used to place applicants in rank order'. I think that allows for them to standardise and then use that score in any way they wish. It isn't as if it materially affects people's decision to share results with the school, whichever way they treat the papers, surely. As in, it doesn't actually materially affect the admissions policy per se.

ETA what if STR and Pates had shared the same algorithm this time? Would that explain it?

[surfer]

And what about these results (these are not my DC, just ones that others have posted):

Crypt: 166
Ribston: 206

Crypt: 172
Ribston: 381

Crypt: 237
Ribston: 74

I would have thought a candidate's ranking for Ribston would generally be higher than their ranking for Crypt as Crypt will include boys too. If both are ranking purely by overall score.