Age standardising
Moderators: Section Moderators, Forum Moderators
Age standardising
Does anyone know when standardising results are they done by the month rather than the day in the month. ie, if you have a 1st July birthday and 31st July birthday are these both standardised to say 10 years 9 months or does the child with the earlier birthday have the scores adjusted to reflect this extra 30 days. My sister-in-law, a teacher says that when standardising results in school it is just done by the month and not the day in the month
-
- Posts: 9235
- Joined: Wed Jan 11, 2006 8:10 pm
- Location: Buckinghamshire
Hi Sally-Anne
That's not actually correct - NFER do Age Standardisation according to discrete age groups rather than smoothed estimate of percentile points.
Bucks will record the child's actual date of birth and exact day of test but only so that NFER can allocate then to the precise Age Group used for Standardisation.
The NFER Standardisation process is described at
http://apm.sagepub.com/cgi/reprint/14/4/387
So what the teacher says is probably a reasonable interpretation of what goes on in practice
Regards
That's not actually correct - NFER do Age Standardisation according to discrete age groups rather than smoothed estimate of percentile points.
Bucks will record the child's actual date of birth and exact day of test but only so that NFER can allocate then to the precise Age Group used for Standardisation.
The NFER Standardisation process is described at
http://apm.sagepub.com/cgi/reprint/14/4/387
So what the teacher says is probably a reasonable interpretation of what goes on in practice
Regards
KenR,
I'm afraid your link was a little beyond me...
But can you answer a question it probably addresses ?
Last year my son, for various logisitical reasons, took the 11+ one week earlier than everyone else (a few hundred took this particular test). He is also an August birthday so presumably was the youngest to take the test in terms of years and completed months.
How would his score have been standardised if there was no one else exactly the same number of completed years/ months?
Or, if there was only, say, one other child exactly the same age, would they have been disadvantaged if my child scored very highly, as the mean for that particular age group would be higher than the whole group mean?
This would mean an August child could theoretically have points taken off in certain circumstances, not added as most people assume.
Sorry to hijack the thread, but it's something that has intrigued me about this process.(This particular child always showed exceptional ability from toddlerhood, consistently scored 141 at school etc. , unlike my other child I hasten to add who is "normal"!).
I'm afraid your link was a little beyond me...
But can you answer a question it probably addresses ?
Last year my son, for various logisitical reasons, took the 11+ one week earlier than everyone else (a few hundred took this particular test). He is also an August birthday so presumably was the youngest to take the test in terms of years and completed months.
How would his score have been standardised if there was no one else exactly the same number of completed years/ months?
Or, if there was only, say, one other child exactly the same age, would they have been disadvantaged if my child scored very highly, as the mean for that particular age group would be higher than the whole group mean?
This would mean an August child could theoretically have points taken off in certain circumstances, not added as most people assume.
Sorry to hijack the thread, but it's something that has intrigued me about this process.(This particular child always showed exceptional ability from toddlerhood, consistently scored 141 at school etc. , unlike my other child I hasten to add who is "normal"!).
Hi Guest
Your child would have been included in a standardisation column which related to the candicate Age in years and months on the date the candidate took the test. So it's just possible that taking the test a week early may have flipped him into a younger set than the other candidates who took the test a week later.
So it's is possible that a child with exactly the same birthday as your son's might have been slightly disadvantaged - however in practice the difference would be miminal but they would have had an extra week to prepare!
Age Standardisation varies from Test to Test, but if you want an example of what Age Standardisation has on raw scores take a look at the Standardisation tables at the back of the 2007 Key Stage 2 results
http://www.qca.org.uk/libraryAssets/med ... 05_wo1.pdf
Hope this helps
Your child would have been included in a standardisation column which related to the candicate Age in years and months on the date the candidate took the test. So it's just possible that taking the test a week early may have flipped him into a younger set than the other candidates who took the test a week later.
So it's is possible that a child with exactly the same birthday as your son's might have been slightly disadvantaged - however in practice the difference would be miminal but they would have had an extra week to prepare!
Age Standardisation varies from Test to Test, but if you want an example of what Age Standardisation has on raw scores take a look at the Standardisation tables at the back of the 2007 Key Stage 2 results
http://www.qca.org.uk/libraryAssets/med ... 05_wo1.pdf
Hope this helps
Thanks KenR,
I understood that link better!
BTW you're right, I forgot to mention he definitely was "flipped" into the younger month band by taking the test 1 week early, as he had one less "completed month" of life on the test date.
He sat the test at 10 yrs 1 complete month and virtually all other August birthday children would have been 10 yrs 2 complete months on the test date and of course some children would have been as old as 11 yrs 1 month.
I can see how this system works well when large numbers of children are sitting the test (e.g. Bucks), but surely it must be less valid for smaller cohorts.
I understood that link better!
BTW you're right, I forgot to mention he definitely was "flipped" into the younger month band by taking the test 1 week early, as he had one less "completed month" of life on the test date.
He sat the test at 10 yrs 1 complete month and virtually all other August birthday children would have been 10 yrs 2 complete months on the test date and of course some children would have been as old as 11 yrs 1 month.
I can see how this system works well when large numbers of children are sitting the test (e.g. Bucks), but surely it must be less valid for smaller cohorts.