tag:blogger.com,1999:blog-4624586630299165335.post8380797156573626835..comments2024-03-14T09:50:44.315+00:00Comments on Psychological comments: Immigrants, scholastic ability, and journalistic abilityAnonymoushttp://www.blogger.com/profile/09320614837348759094noreply@blogger.comBlogger11125tag:blogger.com,1999:blog-4624586630299165335.post-75548517401329831792014-12-08T09:49:45.321+00:002014-12-08T09:49:45.321+00:00Study UK results before London. Look at population...Study UK results before London. Look at population before sub-population. In UK some ethnic groups were getting a higher pass rate than local whites, and likely higher grades. This was already known and predicted from higher intelligence of Chinese. Coverage said "diversity" was good, which obscures the differences in racial groups, some raising and others lowering educational attainment. For London effect, look at Birmingham effect, which is stronger, but conflates the two different meanings of "progress".Anonymoushttps://www.blogger.com/profile/09320614837348759094noreply@blogger.comtag:blogger.com,1999:blog-4624586630299165335.post-45140991583270809142014-12-07T22:07:03.156+00:002014-12-07T22:07:03.156+00:00The presentation of this research in the media imp...The presentation of this research in the media implied that London children were actually getting higher GCSE grades than the benighted white-bread provincials (I paraphrase). <br /><br />Is this not the case ? Laban Tallhttp://ukcommentators.blogspot.co.uk/noreply@blogger.comtag:blogger.com,1999:blog-4624586630299165335.post-40103064087397965632014-11-14T23:44:49.714+00:002014-11-14T23:44:49.714+00:00Thanks for your very detailed explanation of your ...Thanks for your very detailed explanation of your argument, which I appreciate. However, I probably haven't explained sufficiently my strong reservations about "controlling for" variables which are themselves contaminated by differences in ability. I will dig out those posts for you, and then come back later.Anonymoushttps://www.blogger.com/profile/09320614837348759094noreply@blogger.comtag:blogger.com,1999:blog-4624586630299165335.post-90670491076817092692014-11-14T23:33:51.245+00:002014-11-14T23:33:51.245+00:00Second comment to follow on
In this paper Burges...Second comment to follow on <br /><br />In this paper Burgess sets out to see if London’s much higher performance is down to policies (London Challenge) or due to the different nature of the children these schools serve. He starts with a measure of attainment, (the best 8 GCSE, column one table 1 and 2) with no attempt to control for progress. As you can see children in London (and Birmingham) score above average, so their pure scholastic ability is high. However we know that London primary schools are also very good (from an IFS study briefly mentioned in the paper) and London’s performance might just be due to the fact these primary schools are producing bright kids and the secondary schools aren’t doing anything. So in column 2 he controls for prior attainment of pupils and compares just pupils who began secondary school with the same level of attainment at age 11. Here the effect is even larger, and note it won’t penalise bright kids. The analysis isn’t to see which children overall are “making faster progress” relative to all their peers - its to see which children, who began secondary school at the same level, are making faster progress. This looks great for London, and it gets better when he adds in all the things like birth month, socio economic background and gender that are also correlated with school attainment. Here London goes up again, which is where most analysis usually stops and says London is amazing. The value of Burgess’ analysis is he then filters for ethnicity, noting reasonably that London has a very different population from the rest of England and this might have an impact. Sure enough it does, and the London effect basically disappears. After all the fuss made over London it seems that we haven’t solved any problems, London schools do better because, although they are serving a high number of poor children, a lot of these children are drawn from ethnicities such as Chinese and Indian – and, crudely, their ethnicity basically trumps their poverty. Which I again think you wouldn’t find that controversial (although you might disagree with his attributing this to environmental factors). <br /><br />There are three interesting points left to make. Burgess notes that when he focuses just on GCSEs and not “equivalents” (which as Andrew says above, are not equivalent at all) the London effect re-appears. This is actually the table you show in your blog, the reason the Chinese are at the bottom of that is that they have the lowest ratio of taking equivalents to “full” GCSEs. Which is what you’d expect, “full” GCSEs are harder so you’d expect the highest performing group to take more of them. It also re-appears if you focus only on the % of children getting the very highest grades at GCSE. This was my point on twitter – even controlling for ethnicity, gender, month of birth, economic background and the level children started at – London schools still do a better job of getting more pupils to the highest grades than we would expect. And finally there appears to be potential peer effect where white British pupils (the lowest performing group) are doing better in London but their performance cannot be explained by anything London schools are doing and they seem to do worse in areas with less ethnic integration (which is might be why Manchester doesn’t do as well as Birmingham or London. All three cities have quite large percentages of ethnically diverse children, but they are much better integrated in Birmingham and London). Although as Burgess says, this might actually be because the white British people in London are inherently different to the white British people in the rest of the country. So maybe there is something good going on in the capital (or maybe it’s just attracted all the smart people). <br /><br />Sorry for such a long post – I hope it’s useful and reasonably clear. Please just let me know if not, or if you feel I’ve made an error anywhere!<br />nph500https://www.blogger.com/profile/09961175258291178403noreply@blogger.comtag:blogger.com,1999:blog-4624586630299165335.post-91327349372414898652014-11-14T23:33:19.400+00:002014-11-14T23:33:19.400+00:00Hi James, re twitter will try to explain what I me...Hi James, re twitter will try to explain what I mean. <br /><br />Progress in education measures a child's absolute attainment conditional upon their previous score. In the example you gave of the child being in the top 80% at 11 and again at 16 making "no progress" isn't quite right. In the current system children are assessed at 11 on a series of standardised test and given a score, most children get a 4, some do worse and get a 3, some do better and get a 5 and some do even better and get a 6. The progress measure comes in by tracking these same children to GCSE and seeing what they got. Given previous historical examples we basically expect children who get a 4 to continue to demonstrate reasonable academic ability and "progress" through the education system to get around 5 C's at GCSE (they would therefore show "expected progress"). Equally we would expect children who get a level 3 at age 11 to struggle more (although still to "progress" and learn new things), and to probably get Cs and Ds at GCSE. And then we'd expect the level 5 children to do very well, and the level 6 children to do really well. The point is that we fully expect previously low attaining children to struggle at school and high attaining children to learn more. Which I think is relatively uncontroversial, as you say - lots of research demonstrates kids who do well at 11 tend to do well at 16. <br /><br />The reason expected progress is used here (and in education) rather than pure attainment is to try and isolate as much as possible the impact a particular school (or area) is having on the attainment of its pupils. So we fully expect a school serving pupils who enter with high attainment at age 11 to get better absolute GCSE scores than a school serving children who enter with low attainment at age 11. But that tells me about the pupils in the intake, and I’m interested in the school and it’s teaching. In particular what we are interested in are schools (or areas) where for some reason children are making much more or much less progress than we would expect. London is one of those areas, poor children in London do way way better than children of similar ethnic and economic backgrounds in other parts of the country – both in terms of absolute attainment and progress. And only 10 years ago London was one of the worst performing areas, so we were seeing significant improvement. Interestingly they also do better if they move into London and worse if they move out. (I really recommend the work of Chris Cook on this) The question for people like me who are trying to figure out if there’s anything to learn from London is why?<br />nph500https://www.blogger.com/profile/09961175258291178403noreply@blogger.comtag:blogger.com,1999:blog-4624586630299165335.post-59272807899231646582014-11-14T20:27:47.784+00:002014-11-14T20:27:47.784+00:00Dear Andrew, thank for your observations, particul...Dear Andrew, thank for your observations, particularly those regarding GCSE equivalents, which I sidestepped, feeling the post was getting too long. The better approach would be to try to define a core of key GCSEs and then we would have a better grip on real achievement. By the way, I want to get back to the "streaming" debate, and have a half-written post in progress.Anonymoushttps://www.blogger.com/profile/09320614837348759094noreply@blogger.comtag:blogger.com,1999:blog-4624586630299165335.post-46135351925098413462014-11-14T19:08:03.852+00:002014-11-14T19:08:03.852+00:00Thanks. I don't know exactly how the research ...Thanks. I don't know exactly how the research has been done of course, but to take an extreme example, if the grading system were confined to awards of either A or B at both ages, and all ethnic minorities were scored B initially, everyone else an A, then when the children are regraded at an older age, the progress of the former group is certain to be as good or better than the latter. That reflects entirely the nature of the examination system, of course, and has nothing to do with the changing abilities of the different groups of children. Any examination system will have these characteristics. Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-4624586630299165335.post-61195146394966488552014-11-14T18:50:57.895+00:002014-11-14T18:50:57.895+00:00one correction; Chinese pupils have apparently BOT...one correction; Chinese pupils have apparently BOTH above-average "progress" and attainment. They manage this by doing slightly better than average (whites) at younger ages, and then doing a little bit better again at KS2, and then better again at GCSE (i.e the Chinese-remainder gap is present to start with but gets bigger over time). This could reflect any number of things; acculturation, the increasing heritability of IQ over development, both, etc.<br /><br />Another problem is that the achievement of London schools is not strictly comparable to that of schools elsewhere, because up until recently schools were allowed to enter pupils for GCSE 'equivalents' that were equivalent to GCSEs in name only - in reality they were much easier. Many more non-London schools played this game than London schools.<br /><br />So yes - this data is a huge mess and the analyses are frequently farcical. The various measures seem designed to confuse rather than illuminate. Thanks for restoring some order to the chaos. Andrew Sabiskyhttps://www.blogger.com/profile/13017718231961934541noreply@blogger.comtag:blogger.com,1999:blog-4624586630299165335.post-53399952923848865912014-11-14T18:10:13.939+00:002014-11-14T18:10:13.939+00:00Yes, there are apparent regression to the mean eff...Yes, there are apparent regression to the mean effects, but those are not the main issue here, I think. The assessments done at 11 and at 16 are spread out in terms of time and exam content, so are very probably a fair approximation to achievement. it is simply that the immigrant effect appears to coincide with an acculturation effect which does not boost school achievement of itself. Some ethnic groups boost the overall average, some don't. I have just published work on this, using international data, and will describe it later.Anonymoushttps://www.blogger.com/profile/09320614837348759094noreply@blogger.comtag:blogger.com,1999:blog-4624586630299165335.post-39487171930193121412014-11-14T14:32:48.216+00:002014-11-14T14:32:48.216+00:00PS: Upper and lower bounds come into play with eve...PS: Upper and lower bounds come into play with even greater force with a grading system, which greatly narrows the variation in the data. Is it purely a coincidence that the Professor uses grades rather than percentage marks?Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-4624586630299165335.post-32786196267180615362014-11-14T13:59:10.176+00:002014-11-14T13:59:10.176+00:00Many thanks James for tackling this well-publicise...Many thanks James for tackling this well-publicised piece of "research". It's very clear from Professor Burgess's comment "There is nothing inherently different in the educational performance of pupils from different ethnic backgrounds" that this work is motivated by political considerations. He therefore tries to obfuscate by not publishing the raw data, as you have found. <br /><br />I have not read this paper and hence can't claim to provide a thorough critique. But surely one point (the main point?) of relevance is given that scholastic tests always have an upper and lower bound of test scores, even a random score generator at both the younger and older ages will show the (randomly generated) average initial low-score "pupils" making much more "progress" than the (randomly generated) average initial high-score "pupils", some of whom can only go backwards relatively speaking. Given this principle, a researcher looking to play political tricks will seek to find a population of pupils (eg ethnic minorities) that is likely to score lowly on tests at a young age, and compare them with another group of pupils (eg the indigenous population, or even better, Chinese youngsters) that will heavily out-score them initially. The subsequent conclusion of the "research" is almost guaranteed I would imagine. Anonymousnoreply@blogger.com